Advanced
Advanced
Advanced
Edited by
Keith Green
Medical College of Georgia
Augusta, Georgia
Henry F. Edelhauser
Emory University
Atlanta, Georgia
Robert B. Hackett
Alcon Laboratories, Inc.
Fort Worth, Texas
David S. Hull
Medical College oj"Georgia
Augusta, Georgia
David E. Potter
Morehouse School of Medicine
Atlanta, Georgia
and
Ramesh C. Tripathi
University of South Carolina School of Medicine
Columbia, South Carolina
http://www.plenum.com
AII rights reserved
\0987654321
No part of this book may be reproduced, slored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written
permission from the Publisher
To Myra
for all the years of love, care,
encouragement and support,
and without whom the Congress
and this resulting text would
not have been possible
-Keith
PREFACE
This volume represents the proceedings of the Fifth Congress of the International
Society of Ocular Toxicology (ISOT), which was held at the Grove Park Inn and Resort in
Asheville, North Carolina, October 13-17, 1996. We are delighted to present this volume
to the ophthalmic community, especially those with a significant interest in ocular toxicol-
ogy. The Fifth Congress was developed around themes relating to ocular drug metabolism,
the ocular pathophysiological effects of nitric oxide, government issues relating to the use
of alternative methods for toxicity testing, and a workshop that encompassed comparisons
of both in vitro versus in vivo models as well as different animal models.
The outcome of this congress, embodied in this volume, is a contribution to the
methodologies currently employed or under development and to various drug or physical
effects on different ocular tissues. While the focus of this proceedings is on ocular effects
of drugs or other materials, many of the contributions deal with topics that have a much
broader interest. The workshop concerning the use of different model systems and the
choice of the best animal model for drug testing covers a wide range of interests that ex-
tends far beyond specific ocular effects. This is especially true in the area of alternative
methods and in the choice of the best animal model for examination of different disease
entities. This information will prove of value in the performance and assessment of studies
designed to determine drug, or other chemical, effects on the eye that will be used to pro-
vide data pertinent to the protection of human health.
The editors of this volume, who are members of the Congress Organizing Commit-
tee, thank the presenters, who gave their time, experience, and knowledge that made this
gathering a success. The discussions in this volume contain the essence of the presenta-
tions and provide the material given orally or in poster format in an excellent manner.
Our thanks go to all participants, who offer a broad range of backgrounds, varying
from regulatory, industrial, and academic interests, for their attention to detail and their
contemporary contributions. This is one of the unique aspects of our society, where the in-
termingling of individuals with such varied backgrounds can exchange information, ideas,
and concepts. These exchanges enable each of us from different environments to remain
fully conversant with the advances in other areas and to maintain an overview of the field
of ocular toxicology. We are particularly grateful to all of the workshop presenters and
discussants for their frank and open interactions that added so much to the outcome of this
session. All of the contributing authors require a special expression of gratitude for meet-
ing deadlines imposed by the volume's editors so that the text could be as contemporary
as conceivably possible. Their rapid responses to requests for any component related to
the chapters not only kept the courier mail services in business, but also facilitated pro-
gress in finalizing the present text.
vii
viii Preface
Once the workshop was recorded and transcribed and editorial changes were made
(many through the process of one of us [Keith Green] listening to all of the tapes while
reading the transcript), further editorial revision and rewriting of certain portions for clar-
ity followed. Each presenter and/or discussant then performed his or her own readjust-
ments of the already twice-editorialized remarks. These changes were again retyped and
reread by the workshop editors to arrive at the present chapters. We hope that this contri-
bution, which contains the information in a relatively informal style, will be of interest
and value to many readers. We retained the style of oral presentation rather than that of
formal manuscripts because this reflected the tone of the workshop and the interaction be-
tween the participants. Each participant, if contacted, will gladly provide references to any
quotes or citations of results.
A glance at the List of Participants will show the almost equal geographic distribu-
tion between Europe and North America, as well as a similar distribution pattern between
academic and industrial representation. This same ratio applies to the membership of the
society, indicating a broad range of interests. Our meetings, held every two years, offer an
excellent forum for informational exchange on various aspects of ocular toxicology. This
year was no exception, as there was almost universal participation of the attendees in pa-
per and poster discussions.
We also thank the generous corporate sponsors whose contributions helped make the
congress run successfully. Without their assistance and support, the congress would have
not had many of the scientific contributions that made the meeting such a setting for the
exchange of information. These funds are an essential part of the meeting, since they al-
low the organizers latitude in providing an atmosphere where the interchange of ideas is
foremost.
Finally, many thanks are due to Ms. Brenda Sheppard, who was responsible for tran-
scribing the entire workshop (over 100 pages) retyping the participants' amended remarks.
She was also responsible for typing many other parts of this text, including at least two
chapters, the index, and other components. Her dedication to this enormous task, as well
as her efforts before, during, and after the meeting, is appreciated by us all.
The next congress of the society will be held during the summer of 1998 (sometime
in the period of mid-September to mid-October) in Riom, France. Dr. Pierre Duprat, ISOT
president, will act as the host for this meeting. Announcements will be made as decisions
concerning the specific dates for the congress precipitate.
K. Green
H. F. Edelhauser
R. B. Hackett
D. S. Hull
D. E. Potter
R. C. Tripathi
BOARD OF DIRECTORS, INTERNATIONAL
SOCIETY OF TOXICOLOGY, 1995/1996
ix
x Board of Directors
CORPORATE SPONSORS
Alcon Laboratories, Inc., Fort Worth, Texas
Allergan, Irvine, California
Boehringer Ingleheim GmbH, Ingelheim, Germany
Ciba Vision Ophthalmics, Duluth, Georgia
Gillette Medical Evaluation Labs, Gaithersburg, Maryland
Laboratoire Chauvin, Montpellier, France
Merck and Co., Inc., West Point, Pennsylvania
Pfizer, Centre de Recherche, Amboise, France
Procter & Gamble Co., Cincinnati, Ohio
Santen Pharmaceutical Co. Ltd, Osaka, Japan
Senju Pharmaceutical Co. Ltd, Osaka, Japan
CONTENTS
xi
xii Contents
12. Testing of Ocular Viscoelastics by Injection into the Rabbit Vitreous: Historical
Control Data Resulting from Various Dosing Techniques .............. 103
P. J. Upman, K. A. Herkowski, and M. J. Shepherd
16. Role of Nitric Oxide in Vascular Dysfunction Associated with Ocular Diseases. . 133
Ronald G. Tilton
17. Effects of the Inhibition of Nitric Oxide Synthase and Lipoxygenase on the
Development of Endotoxin-Induced Uveitis ........................ 151
Juan L. Bellot, Nuria A1coriza, Mercedes Palmero, Alfonso Blanco,
Rafael Espi, Claude Hariton, and Alfredo Orts
22. Computer-Assisted Evaluation ofIris Color Changes in Primate Toxicity Studies 203
W. H. Bee, F. Vogel, and R. Korte
23. Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing. . . . . .. 207
Henry F. Edelhauser and Keith Green
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 273
Advances in Ocular Toxicology
WELCOME AND OPENING OF CONGRESS
Keith Green
President of ISOT
Department of Ophthalmology
Medical College of Georgia
Augusta, Georgia 30912-3400
On behalf of the Board of Directors and the Congress Organizing Committee, I wish
to extend a hearty welcome to everyone attending the 5th Congress of the International
Society of Ocular Toxicology. Unique locations are becoming a tradition for our Congress
and the current site in Asheville, North Carolina, is no exception. The magnificent setting,
especially at this time of year when the autumn leaves are at their most prevalent and col-
orful, offers the opportunity for nurture of the physical and mental being when away from
the scientific sessions.
The primary themes of our 5th Congress are "Ocular Metabolism" organized by
Robert B. Hackett; "Nitric Oxide: Actions and Pathological Roles in the Eye" organized
by David E. Potter; "Workshop on in vitro vs. in vivo Methods and Different Animal
Models" organized by Henry F. Edelhauser; and "Government Issues Pertinent to the
Adoption of Alternative Testing Models". These sessions are in addition to those relating
to drug or physical effects on ocular tissues, whether from local or systemic treatments,
which constitute a regular component of our Congresses. Our invited speakers represent
the forefront of these special research interests and will update our knowledge in these ar-
eas. They are:
Dr. M. L. Schwartzman, New York Medical College, Valhalla, New York
Professor H. F. Edelhauser, Emory University, Atlanta, Georgia
Professor D. A. Fox, University of Houston, College of Optometry, Houston, Texas
Dr. M. F. Lou, University of Nebraska, Lincoln, Nebraska
Dr. J. B. Allen, North Carolina State University, Raleigh, North Carolina
Dr. R. Tilton, Texas Biotechnology Corp., Houston, Texas
Dr. J. Avalos, Food and Drug Administration, Rockville, Maryland
Our thanks are also extended to all speakers and presenters of posters who provided
their skills to bring us the latest in their areas of expertise. Please take advantage of this
Congress to meet with all other attendees and benefit from the unique opportunity to learn
from our fellow toxicologists. Each individual has unique knowledge, skills and expertise
and this gathering offers the chance to discuss topics with renowned authorities in their
field. These interactions will enhance our ability to perform our studies using the best of
technology and with the maximum of understanding of our field. In a very practical sense,
much of what we hear and learn at this Congress can be taken back to our own setting and
immediately incorporated into our efforts. Furthermore, the friendships developed from
these contacts may prove beneficial in establishing new and fruitful professional relation-
ships that hopefully will be of long-lasting duration.
1
Department of Pharmacology
New York Medical College
Valhalla, New York 10595
Injury to the cornea provokes an inflammatory response with the degree of inflam-
mation correlating to the severity of the injury.' The in situ appearance of the inflamed
cornea usually consists of conjunctival vasodilation, edema and neovascularization. 2
While edema and vasodilation are reversible, neovascularization mayor may not be de-
pending upon the degree of vascular invasion into this normally avascular tissue. There-
fore, a vascularized cornea presents a serious clinical problem since visual acuity is
impaired by the physical presence of the vessels. If the vessels do not regress fully, perma-
nent loss of visual acuity follows. Despite the uniform manifestations of inflammation, the
mediators may be tissue and injury specific. '.2
Numerous mediators are involved in the development and progression of corneal in-
flammation. 3 The source(s) of these mediators is the anterior surface tissues or inflamma-
tory cells (PMNs and mononucleocytes) which have invaded the avascular cornea. 4 While
the continuation of an inflammatory response is usually dependent on the inflammatory
infiltrate, its initiation is usually the result of inflammatory mediators released from the
damaged tissue. 5 Corneal inflammatory mediators released from the epithelium include
arachidonic acid-derived,6 platelet activating factor (PAF)/ cytokines (ILl and IL8)8 and
serine proteases (plasmin and plasminogen activators).9 These autocoids are believed to
function together in the orchestration of a corneal inflammatory response. However, only
the eicosanoids have been implicated in the initiation, development and progression of an
inflammatory response,1O while the others are mostly involved in its progression and con-
tinuation.' I.'2 Although eicosanoic release from infiltrating inflammatory cells can aid in
the promotion of the inflammatory response,IO·13.14 their ability to be produced endo-
genously from the injured tissue renders them able to initiate the inflammatory response.'5
This is relevant since endogenously released eicosanoids are mostly responsible for initiat-
ing events such as alterations in vascular permeability and leukocyte chemotaxis. 16,'7 Fur-
ther, the effectiveness of corticosteroids which inhibit the synthesis of these
pro-inflammatory mediators by manipulating substrate (arachidonic acid) availability;8
supports the involvement of eicosanoids in corneal inflammatory processes.
The arachidonic acid-derived eicosanoids can be synthesized in mammalian tissues
through three enzymatic pathways '9,2o (Figure 1): (i) cyclooxygenase to form prostagland-
ins (PGs), prostacyclin (PGI2) and thromboxane (TxA2); (ii) lipoxygenase to form hy-
droxyeicosatrienoic acids (HETEs) and leukotrienes (LTs); and (iii) cytochrome P450
(CYP) monooxygenases to form epoxyeicosatrienoic acids (EETs) and HETEs, The three
pathways can be differentiated via their subcellular localizations, inhibitor sensitivity and
product stereoselectivity, Cyclooxygenase and CYP monooxygenases aremicroso-
mal/membrane-associated enzymes: 9,21 Classically, lipoxygenases have been thought to
be associated with the cytosolic fraction of the cell and require an increase in intracellular
Ca2+ for their translocation to the membrane in order to gain access to the substrate,
arachidonic acid,z2 However, there is evidence to support the existence of microsomal
lipoxygenases in certain tissues. 23-25 The mammalian lipoxygenase enzymes produce only
the'S' enantiomers of the various HETEs/6 while CYP enzymes have been shown to pro-
duce both the 'R' and'S' enantiomers. 27- 29 Therefore, one must consider the subcellular
distribution of the enzymes involved in HETE production, the stereochemical differences
of the products and the selectivity of metabolic inhibitors when striving to characterize
HETE biosynthetic enzymes,
INJURY
ETYA
1
PHOSPHOLIPIDS
~1
.
Indomethacin
BW755C
. -_ _ _ _ _ _ _ _ _ ~COO-H------------.
Arachidonic Acid
o ETYA
CLOTRIMAZOLE
o
t
Prostaglandins
~
EETs
~
Leukotrienes
DHETs HETEs
Thromboxane
HETEs HETrEs
Prostacyclin
HETrEs
Figure 1. The three enzymatic pathways of arachidonic acid metabolism: Major metabolites and known enzy-
matic inhibitors.
Cytochrome P450 and Arachidonic Acid Metabolism in' the Corneal Epithelium 5
genases and have a minimal cofactor requirement and confined substrate specificity, CYP
proteins are monooxygenases with an absolute requirement for NADPHINADH and a fla-
voprotein, NADPH-CYP (c) reductase and/or cytochrome b5, for substrate activation.
This enzyme system, in the presence of NADPH and molecular oxygen, can oxidize
arachidonic acid via three reaction types (Figure 2): I) olefin epoxidation which involve
the reductive cleavage of molecular oxygen with the formation of water and an activated
form of atomic oxygen, an oxene, that is added in a regio- and stereoselective manner to
the double bonds of arachidonic acid to form the four regioisomeric epoxides (5,6; 8,9;
11,12; 14,15 EETs); 2) allylic oxidation, a lipoxygenase-like activity, in which the acti-
vated oxygen intermediate attacks an electron rich carbon atom resulting in hydrogen ab-
straction and oxygen insertion to form the six regioisomeric cis-trans conjugated
mono-HETEs; and 3) m-hydroxylations that forward the hemoprotein-bound reactive oxy-
gen to the terminal and adjacent sp3 carbon atoms (C-16 through C-20) of the arachidonic
acid molecule to form the 16-, 17-, 18-, 19-, and 20-HETEs.20 Several purified CYP
isozymes have been shown to metabolize arachidonic acid in a reconstituted system to
EETs and HETEs (mainly hydroxylation at C16 -C20); among them are CYPIA1, IA2,
2Bl, 2B4, 2C2, 2Cll, 2C23, 2El, 2Gl and 4A1. 5{}-54
ARACHIDONIC ACID
,
/
Cis-trans conjugated
/
Allylic Oxidation
........_.-_..
~
Olefin Oxidation
n
EETs
(5,6; 8,9; 11,12; 14,15)
wlw-1Iw-2
, Oxidation
......f - - - - -....
18·HETE
19-HETE
mono-HETEs
(5,8,11,12,15) 20-HETE
I
Epoxide Hydrolase •
20-COOH
~
DiHETrE
12(R)-HETrE 12(R)-HETE
Vasodilator; Increases Barrier Na-K-ATPase Inhibitor
Permeability; Potent PMN
Vasoconstrictor
Chemotactic factor (pM); Angiogenic
Factor; Endothelial Cell Mitogen; Stimulates PMN Chemotactic Factor
(loWIlM)
antigen-induced hypersensitivity skin
reaction.
Figure 3. Structure and biological activity of the two major corneal epithelial cytochrome P450-arachidonic acid
metabolites, 12(R)-HETE and 12(R)-HETrE.
8 M. L. Schwartzman
These metabolites are structurally similar except for the absolute stereochemistry at
C-12 and the presence of a saturated bond at C-lO,ll. Designation of the stereochemistry
of the hydroxyl group at C-12 for both 12(R)-HETE and 12(R)-HETrE is a consequence
of the nomenclature priority rules and, in fact, these metabolites have opposite absolute
stereochemistries. Little is known concerning the mechanistic details of the conversion of
arachidonic acid into such monohydroxylated metabolites. The mechanism of CYP reac-
tions with olefins typically involves the formation of an epoxide, and several CYP
isozymes have been identified as arachidonic acid epoxygenases. Thus, the initial step of
oxygenation would be the formation of an 1l,12-EET intermediate formed from molecular
oxygen and arachidonic acid catalyzed by a specific CYP isozyme(s). Alternatively, CYP
enzymes can directly oxidize arachidonic acid to 12(R)-HETE via a lipoxygenase-like re-
action without an intermediate epoxide. 27 The formation of 12(R)-HETrE would require
additional rearrangement of any purported 11,12-EET, either through direct epoxide rear-
rangement to an isomeric 12-keto-eicosatrienoic acid or through oxidation of 12(R)-HETE
to 12-keto-HETE (Figure 4). To this end, we have shown that 12(R)-HETrE can be
formed not only from arachidonic acid but also from CYP-derived 12(R)-HETE and
lip oxygenase-derived 12(S)-HETE via an oxidation/keto reduction activity.71.76 The cor-
neal epithelial synthesis of both metabolites is greatly enhanced following injury. How-
ever, while 12(R)-HETE synthesis is readily detected in the normal cornea, 12(R)-HETrE
formation is only observed following injury.73,74,77.78
To date, no specific CYP isozyme(s) for the formation of 12(R)-HETE has been
identified. However, in searching for one, the most likely candidates would be those that
either insert an hydroxyl at carbon 12 via allylic oxidation, or catalyze specific epoxida-
tion at the 11,12 double bond to yield II(S),12(R)-EET (Figure 4). The latter step was
postulated to yield, via direct hydrogen abstraction at CI0 or epoxide hydrolase, 12(R)-
ARACHIDONIC ACID
~~
12-LIPOXYGENASE CYTOCHROME P450
MONOOXYGENASES
+
'" ,
12(S)-HpETE
12(S)-HETE
•
[12-oxo-ETE]
12-oxo-ETrE
12(S)-HETrE
/,~ .--'-112(R)-HETrEI
Figure 4. Metabolic pathways for the formation of 12(R)-HETE and 12(R)-HETrE in the corneal epithelium. The
solid arrows indicate established steps, whereas the dotted arrows indicate postulated steps based on literature evi-
dence.
Cytochrome P450 and Arachidonic Acid Metabolism in the Corneal Epithelium 9
HETE.79 The rabbit CYP2C2, and another kidney CYP2C isoform termed 2CAA, have
been shown to oxidize arachidonic acid regioselectively to 11,12-EET, specifically to the
11(S),12(R)-EET enantiomer. 53 .80 In that study, the HPLC profile showed a significant
amount of 12-HETE.53 The same pattern of metabolism has been observed for the rabbit
liver CYPIA2. II, 12-EET is also produced in incubation of arachidonic acid with purified
CYP 2B2, 2C23 and 2B4.52.53 Incubation of purified rabbit CYP2Gl and IA 1 with arachi-
donic acid also yielded a metabolite with an HPLC retention profile of 12-HETE.81 All to-
gether, these studies clearly point to the involvement of CYP proteins of gene families I
and 2 in olefin epoxidation and allylic oxidation of arachidonic acid.
Little is known about CYP isozymes in the corneal epithelium and factors that may
regulate their expression. Matsumoto et al. 62 have demonstrated the presence of a pheno-
barbital-inducible CYP in the rabbit corneal epithelium. Lin et al. 66 showed that copper
ion induced CYP-derived 12(R)-HETE production in porcine corneal epithelium. Asakura
and Shichi 67 ,68 demonstrated induced levels of CYP activity in porcine ciliary and corneal
epithelial microsomes capable of synthesizing 12(R)-HETE after treatment with 3-MC or
clofibrate. Furthermore, the 3-MC-inducible 12(R)-HETE formation was inhibited by anti-
bodies directed against purified CYPIAl. 67 ,68 We obtained preliminary results (Figure 5)
demonstrating the presence ofCYPIAIIlA2, 2EI, 2BI, and 2GI immunoreactive proteins
in the rabbit corneal epithelium; however, CYPIAl-immunoreactive protein(s) was the
dominant. Furthermore, treatment of rabbit corneas with 3-MC (10 11M) in corneal organ
culture increased epithelial CYP-arachidonic acid activity by 2_fold. 75
1 2 3
1 1
2Bl
2Gl
Figure 5. Western blot analysis of rabbit corneal epithelial microsomes. Immunoblotting was performed with anti-
bodies against CYPI A I , 2B 1 and 2G I. Lane I, rabbit liver; lanes 2 and 3, 10 and 50 Ilg corneal epithelial mi-
crosomes.
10 M. L. Schwartzman
crease urine volume and electrolyte excretion in the isolated perfused rat kidney;86 effects
consistent with inhibition ofNa,K-ATPase activity. Evidence for a link between pump in-
hibition by 12(R)HETE and the function of specialized transport epithelia has also been
found in the cornea. Infusion of 12(R)HETE onto the endothelial side of the isolated per-
fused rabbit cornea produced a dose-dependent increase in corneal thickness, an effect
characteristic of ouabain and other inhibitors of corneal endothelial Na,K-ATPase activ-
ity.87
An inhibitor of Na,K-ATPase may have fundamental importance in ocular-transport-
ing epithelia that relay on this pump activity. Deturgescence of the cornea mediated by the
corneal endothelium and secretion of aqueous humor by ciliary body are examples of ba-
sic physiological functions dependent upon metabolic pumps that are readily inhibited by
digitalis glycosides such as ouabain. A product(s) of the P450-dependent AA metabolism
such as 12(R)HETE, may modulate these and other important functions in the eye (e.g.,
maintenance of lenticular transparency). Perturbations of this pathway may playa role in
the pathogenesis of the development of spontaneous cataract formation. Kinoshita 88 has
shown that in mice of the Nakano cataract strain that develop a spontaneous cataract after
birth, there is a failure of lens Na,K-ATPase due to the presence of an unidentified inhibi-
tor (or inactivator) of the pump activity. If indeed the pathway functions in the modulation
of aqueous humor secretion, then manipulation of the pathway might be useful in the man-
agement of ocular hypertension (glaucoma). We measured the effect of topically applied
12(R)HETE on lOP of rabbit eyes. Our studies clearly demonstrated that a low dose
(0.5 J.lg) of 12(R)HETE lowers lOP by as much as 7 mmHg and that this effect is main-
tained for several days, without causing ocular side effects such as conjunctival hypere-
mia, aqueous flare and miosis. 89 This investigation did not determine the mechanism by
which the 12(R)HETE reduces lOP, but the fact that 12(R)HETE is a potent inhibitor of
Na,K-ATPase activity in several tissues including the corneal epithelium, suggests a de-
crease in the rate of aqueous humor secretion at the ciliary body via a possible inhibition
of the ciliary epithelial metabolic pump. Socci et al. 90 confirmed the lOP lowering effect
of 12(R)HETE in the rabbit eye and further demonstrated the inhibition of the ciliary epi-
thelial Na,K-ATPase by l2(R)HETE. In addition, cardiac glycosides decrease the rate of
aqueous humor secretion in rabbits, cats and humans. 9l ,92 These agents in order to be ef-
fective must be administered intravitreally or systemically and in potentially toxic higher
doses. Another Na,K-ATPase inhibitor that lowers lOP by topical application is vanadate,
but its lOP lowering effect can also be related to changes in adenylate cyclase activity or
with prevention of cyclic AMP effects. 93 ,94 Another possibility for the lOP lowering effect
of 12(R)HETE is that it may enhance aqueous outflow facility or uveosc1eral outflow as it
has been shown for PGF 2u ' 95.96
These findings, including enantiomer specific biologic and enzymatic effects, sug-
gest that endogenous levels of 12(R)HETE may serve a variety of physiological and/or
pathophysiological functions in vivo. 12(R)HETE from one or more tissues in the eye may
be a potent endogenous modulator of lOP. Further studies are needed to elucidate the
mechanism of 12(R)HETE effects on Na,K-ATPase and lOP as well as to determine the
therapeutic potential of this novel compound.
the conjunctival blood vessel with a potency 4-8 times greater than acetylcholine.?7 In
doses of as little as 5 ng, it increases the aqueous humor protein concentration by 7 fold
indicating that it is at least 20-fold more potent than PGE2 in producing the breakdown of
the blood-aqueous barrier. J9 The vasodilatory and permeability-increasing properties of
this compound indicating that it is a pro-inflammatory agent. Several studies in a variety
of different model systems have indicated that corneal vascularization, in which blood
vessels invade the normally avascular cornea, is usually a manifestation of a chronic in-
flammatory response, and that leukocytic infiltration of the corneal stroma precedes and
accompanies the corneal vascularization. 12(R)-HETrE at the amount of 0.5 Jlg induced an
extensive neovascular response within 4 days (Figure 6). As the implant was depleted of
this compound, blood flow through the new vessels ceased, suggesting that the neovascu-
lar response and the presence of functioning new vessels in the cornea may be dependent
on the presence of 12(R)_HETrE. 97 .98 The direct chemotactic property of this compound
was tested in human polymorphonuclear leukocytes and found to be more potent but with
less efficacy than LTB4, one of the most potent chemotactic agents so far reported. 99 In
fact, we recently obtained evidence to suggest that 12(R)-HETrE is a powerful agonist
(Kd= 1O-lOM) for the LTB4 receptor in human leukocytes. Thus, 12(R)-HETrE may be an
indirect angiogenic factor, i.e. a molecule which induces the release of a primary angio-
genic factor from other cells present in the biological system. IOO The neovascular effect of
12(R)-HETrE could be due to its ability to attract neutrophils, lymphocytes or macro-
phages into the cornea. These blood elements may themselves, or in combination with
12(R)-HETrE, elicit neovascularization. 1ol However, we can not exclude the possibility
that 12(R)-HETrE is a direct angiogenic factor as well, i.e., a molecule which acts primar-
ily on endothelial cells. Indeed, in recent studies we showed that 12(R)-HETrE is a potent
mitogen in microvessel endothelial cells including cells isolated from the limbal microves-
sels. 102,lo3 Furthermore, 12(R)-HETrE, but not 12(S)-HETrE, induces cultured limbal mi-
crovessel endothelial cells to organize themselves as a network of branching cords
reminiscent of capillaries. This effect was evident within 48 hours, maximum by 5 days of
culture, and paralleled the effect observed with bFGF,103 thus, further strengthens the
angiogenic properties of 12(R)-HETrE by demonstrating a direct effect on limbal mi-
crovessel endothelial cells.
12(R)-HETrE stimulated the growth of quiescent endothelial cells in a time- and
concentration-dependent manner with a maximal effect at 0.1 nM.lo2 This effect was
highly stereospecific since its enantiomer, 12(S)-HETrE, had no effect within the same
concentration range. 12(R)-HETrE may exert its mitogenic effect on the microvessel en-
dothelial cells by activating transcription factors and altering the expression of immediate
early response genes such as the protooncogenes cjos. c-jun and c-myc. We demonstrated
by Northern blot analysis and transient transfection with CAT constructs of the protoonco-
gene promoter regions that 12(R)-HETrE (0.1 nM) causes a rapid, marked and stereospe-
cific increase in c-fos. c-myc and c-jun mRNAs and expression. 102 Electrophoretic
mobility shift assay demonstrated a stereospecific and rapid (within IS min) increase in
the binding activity of NFkB (20-fold over control) in 12(R)-HETrE (0.1 nM) but not in
12(S)-HETrE treated cells. \02.104 NFkB is a transcription factor that is activated in endothe-
lial cells following injury and has been implicated as an immediate early mediator of im-
mune and inflammatory responses. The capacity of the injured corneal epithelium to reach
and exceed the effective concentration (0.1 nM) of 12(R)-HETrE is readily observed,73,75
Hence, activation of NFkB by 12(R)-HETrE under such circumstances infers that it may
playa role in mediating the mitogenic and angiogenic signals of this eicosanoic in mi-
crovessel endothelial cells. Indeed, a recent study demonstrated that inhibition of NFkB
activation markedly attenuated 12(R)-HETrE-induced in vitro capillary formation of mi-
crovessel endothelial cells. 104
Eliason and ElliottlOS,lo6 demonstrated that corneal epithelial homogenates contain
heat stable factor(s) which stimulate the proliferation of endothelial cells in vitro and vas-
cularization in the cornea in vivo. These authors concluded that the corneal epithelium is
the source of a secreted stimulant(s) for the growth of vascular endothelial cells. 12(R)-
HETrE, which is an endogenous metabolite of arachidonic acid in the corneal epithelium,
has the characteristic of such a factor; it not only possesses pro-inflammatory properties,
but also has very potent angiogenic capabilities. Thus, 12(R)-HETrE may qualify as that
intrinsic corneal angiogenic factor and may, in association with other inflammatory
mechanisms,lo7.108 account for the growth of new vessels in the cornea which frequently
appears in chronic inflammation or in the reparative stages of an acute process. Vasodila-
tation, neovascularization and changes in blood vessel permeability are events common to
many pathological processes in the eye beyond injury to the corneal epithelium. It will be
of interest in the future to investigate whether 12(R)-HETrE is produced in other transport
Cytochrome P450 and Arachidonic Acid Metabolism in the Corneal Epithelium 13
epithelia such as the RPE and CBE, especially in relation to conditions such as uveitis and
diabetic retinopathy.
The formation of these metabolites is also seen in other tissues and their topical ap-
plication elicits biological activities typical of inflammatory factors. Woollard and co-
workers 28 ,109,IIO demonstrated the production of 12(R)-HETE in human psoriatic lesions
and suggested that it was generated by a CYP monooxygenase. The presence of an ei-
cosanoic with an HPLC retention time similar to that of 12-HETrE was also documented
by these investigators,I09 Human and rat epidermis has been shown to produce 12-HETE
enantiomers by both lipoxygenase and CYP monooxygenase pathways, and rat epidermal
microsomes have been shown to produce 12-HETrE enantiomers by an NADPH-depend-
ent mechanism. 29,111 Production of 12-HETrE from 12-HETE was also documented in por-
cine neutrophils. 1I2 12(R)-HETE is a chemoattractant for human neutrophils and mixed
peripheral blood lymphocytes at high nM concentrations, I13,1 14 an activity thought to be
mediated via binding to the LTB4 receptor. 115 It also causes erythema following topical
administration of nmol doses to human skin. 116 In guinea pig skin, 12(R)-HETrE enhances
delayed-type hypersensitivity inflammatory reactions, with a maximal 50% enhancement
at 1 fmol doses, and significant enhancement at doses up to 1 pmo!. This effect is specific
for 12(R)-HETrE; no significant enhancement is observed with its enantiomer nor with
other 12-hydroxyeicosanoids tested.1I7 The implication is clear: this pathway and its prod-
ucts may be involved in other inflammatory responses besides the cornea. Skin and cor-
neal epithelium are derived from surface ectoderm and are well known to be clinically
related. It is possible that these two tissues share a similar response to injury as well as
similar mechanism for regulation of the inflammatory response.
CYP activity by depleting the levels of microsomal heme, an important component for
CYP activity. Several studies have demonstrated that certain heavy metals, at relatively
low, non-toxic doses, could induce the synthesis and activity of the ubiquitous heme meta-
bolizing enzyme, heme oxygenase (HO-I). This microsomal enzyme normally functions
to convert cellular heme to biliverdin via an NADPH-dependent reaction which is sub-
sequently converted to bilirubin in the presence of cytosolic biliverdin reductase. 12 1l--123
With induction, various heme containing enzymes, particularly those with a more rapid
turnover, are eventually depleted due to a lack of available heme within the cell. Further
studies showed a selective depletion in CYP hemoproteins while other more stable hemo-
proteins such as those of the respiratory chain, remained unaltered. '24 Tin (Sn) is such a
metal. It is non-toxic at effective doses, and yet quite efficacious as a HO-I inducer, lead-
ing to depletion of CYP levels and subsequent inhibition of enzyme activity. 124-127 Using
this metal, we have demonstrated that treatment of the contact lenses with SnCl2 (100
J,1g/ml) resulted in the significant inhibition of CYP-arachidonic acid metabolism to 12-
HETE and 12-HETrE. This was associated with an attenuation of the inflammatory re-
sponse, determined by subjective scoring, in the same time frame (Figure 6). Thus,
selective manipulation of the CYP enzyme system via HO-1 induction in the corneal epi-
thelium with subsequent inhibition of 12-HETE and 12-HETrE synthesis, resulted in a sig-
nificant attenuation of the inflammatory response to closed eye-contact lens wear74 and
provided strong pharmacological evidence for the role of CYP-arachidonic acid metabo-
lites in this model of ocular surface inflammation.
In summary, although the biological activities of these arachidonic acid metabolites
and their enhanced production with injury to the cornea are well established, our knowl-
edge of the scope of the CYP-catalyzed reactions is still incomplete as the isoform(s) re-
sponsible for their production has yet to be fully characterized or even identified. Our
findings and those of Asakura and Shichi 67 ,68 suggest the possibility that a CYPIAI or
-----
70 Vehicle
! 60 ---Q- Stannous Chloride
0
u
1/1 50
~ 40
0
ai 30
E
E 20
.!!!
:s 10
0
DO 01 03 05 07
B
50 ___ Vehicle
--0-- Stannous chloride
40
30
Figure 7. Effect of SnCl, on the inflammatory parame-
w ters of closed eye-hydrogel contact lens wear at 6 days.
t=
w
20
Lenses were treated with either vehicle or SnCl, (100
:r: J.lgfml) for 30 min prior to lens placement and tarsor-
...
N 10
rhaphy. At 6 days, sutures were removed and eyes exam-
ined. A) Inflammatory score, B)12-HETE+12-HETrE
DO 01 03 05 07 synthesis, Results are the mean±SEM, n=3--8. *=p<0,05
Days of closed eye-hydrogel lens wear versus the vehicle treatment (taken from ref. 74).
Cytochrome P450 and Arachidonic Acid Metabolism in the Corneal Epithelium 15
IAI-like isoform is present in the corneal epithelium and may be involved in the metabo-
lism of arachidonic acid to 12(R)-HETEI12(R)-HETrE. However, at this point further
characterization of the CYP isoform composition in the corneal epithelium is needed be-
fore concluding on a specific isoform. The relevance of these two metabolites to ocular
pathophysiology rests on the premises that their synthesis is induced following injury,
their production rate correlates with the inflammatory response, and that both compounds
are potent pro-inflammatory agents capable of affecting the cellular and vascular re-
sponses of inflammation which can ultimately culminate in the evolution of corneal ne-
ovascularization. Due to their wide spectrum of biological activities and potency at low
picomolar concentrations which are readily obtained in the injured cornea, 12(R)-HETE
and 12(R)-HETrE may indeed be major inflammatory eicosanoic mediators under condi-
tions of corneal and conjunctival injury where NSAIDs are ineffective and yet corticos-
teroids are quite efficacious. Knowledge of their synthesis, actions and regulation may
lead to novel insights into the treatment of ocular surface inflammation, a major problem
in clinical ophthalmology.
7. ACKNOWLEDGMENT
This study was supported in part by grant EY06513 from the National Eye Institute.
8. REFERENCES
I. Thoft RA, Friend J, Kenyon KR. Ocular surface response to trauma. Int Ophthalmol Clin 1979;
19:111-131.
2. Waring GO, Rodrigues MM. Patterns of pathologic response in the cornea. Surv Ophthalmol 1987;
31 :262-266.
3. Leopold IH, Gaster RN. Ocular inflammation and anti-inflammatory drugs. In: Kaufman HE, Barron BA,
McDonald MB, Waltman SR, eds. The Cornea. New York: Churchill Livingston Inc., 1988:67--84.
4. Smolin G. Cellular response to inflammation at the limbus. Eye 1989; 3: 167-171.
5. Cotran RS, Kumar V, Robbins SL. Robbins Pathologic Basis of Disease. 4th ed. Philadelphia: WB Saun-
ders Co., 1989:39-86.
6. Bazan HEP, Birkle DL, Beuerman RW, Bazan NG. Inflammation-induced stimulation of the synthesis of
prostaglandins and lipoxygenase-reactions products in rabbit cornea. Curr Eye Res 1985; 4: 175-179.
7. Bazan HEP, Reddy STK, Woodland JM, Bazan NG. The accumulation of platelet activating factor in the
injured cornea may be interrelated with synthesis of lipoxygenase products. Biochem Biophys Res Com-
mun 1987; 149:915-920.
8. Elner VM, Strieter RM, Pavilack MA. Human corneal interleukin 8. IL I and TNF-induced gene expression
and secretion. Am J Patho11991; 139:977-988.
9. Lantz E, Andersson A. Release of fibrinolytic activators from the cornea and conjunctiva. Graefes Arch
Clin Exp Ophthalmol1982; 219:263--267.
10. Williams KI, Higgs GA. Eicosanoids and inflammation. J Path 1988; 156:101-110.
II. Berman M, Winthrop S, Ausprunk D, Rose J, Langer R, Gage J. Plasminogen activator (urokinase) causes
vascularization of the cornea. Invest Ophthalmol Vis Sci 1982; 22: 191-199.
12. Strieter RM, Kunkel SL, Elner VM. Martony CL, Koch AE. Polverini PJ. Elner SG. Interleukin-8. A cor-
neal factor that induces neovascularization. Am J PathoI1992; 141:1279-1284.
13. Higgins AJ. The biology, pathophysiology and control of eicosanoids in inflammation. J Vet Pharmacol
Ther 1985; 8:1-18.
14. Bazan HEP, Birkle DL. Beuerrnan RW, Bazan NG. Cryogenic lesion alters the metabolism of arachidonic
acid in rabbit cornea layers. Invest Ophthalmol Vis Sci 1985; 26:474--480.
IS. Srinivasan BD, Kulkarni PS. The role of arachidonic acid metabolites in the mediation of the polymor-
phonuclear leukocyte response following corneal injury. Invest Ophthalmol Vis Sci 1980; 19: 1087-1093.
16 M. L. Schwartzman
16. Stjernschantz J, Nilsson SFE, Astin M. Vasodynamic and angiogenic effects ofeicosanoids in the eye. In:
Bito LZ, Stjernschantz J, eds. The Ocular Effects of Prostaglandins and Other Eicosanoids. New York:
Alan R. Liss, 1989:155-170.
17. Srinivasan BD, Kulkarni PS. Polymorphonuclear leukocytes response. Inhibition following corneal epi-
thelial denudation by steroidal and anti-inflammatory agents. Arch Ophthalmol1981; 99:1085-1089.
18. Dennis EA. Regulation of eicosanoic production: role of phospholipases and inhibitors. Biotechnology
1987; 5:1294-1300.
19. Smith WL. The eicosanoids and their biochemical mechanisms of action. Biochem J 1989; 259:315-324.
20. CapdeviJa JH, Falck JR, Estabrook RW. Cytochrome P450 and arachidonate cascade. FASEB 1992;
6:731-736.(abstract)
21. Lu AYH, West SB. Multiplicity of mammalian microsomal cytochrome P450. Pharmacol Rev 1980;
31 :277-295.
22. Baba A, Sakuma H, Inoue T, Iwata H. Calcium induces membrane translocation of 12-lipoxygenase in rat
platelets. J BioI Chern 1989; 264: 15790-15795.
23. Shornick LP, Holtzman MJ. A cryptic, microsomal-type arachidonate I2-lipoxygenase is tonically inacti-
vated by oxidation-reduction conditions in cultured epithelial cells. J BioI Chern 1993; 268:371-376.
24. Chang W, Ning e, Lin MT, Huang J. Epidermal growth factor enhances a microsomal 12-lipoxygenase ac-
tivity in A431 cells. J BioI Chern 1992; 267:3657-3666.
25. Nakadate T, Aizu E, Yamamoto S, Kato R. Some properties of lipoxygenase activities in cytosol and mi-
crosomal fractions of mouse epidermal homogenate. Prostagland Leukot Med 1986; 21:305-319.
26. Yamamoto S. "Enzymatic" lipid peroxidation: reactions of mammalian lipoxygenase. Free Rad Bio Med
1991; 10:149-159.
27. Capdevila J, Yadagiri P, Manna S, Falck JR. Absolute configuration of the hydroxyeicosatrienoic acids
(HETEs) formed during catalytic oxygenation of arachidonic acid by microsomal cytochrome P450. Bio-
chern Biophys Res Commun 1986; 141:1007-10011.
28. Woollard PM. Stereochemical difference between 12-hydroxy-5,8.10,14-eicosatrienoic acid in platelets
and psoriatic lesions. Biochem Biophys Res Commun 1986; 136: 169-172.
29. Holtzman MJ, Turk J, Pentland A.. A regiospecific monooxygenase with a novel stereo preference is the
major pathway for arachidonic acid oxygenation in isolated epidermal cells. J Clin Invest 1989;
84: 1446-1453.
30. Kulkarni PS, Fleisher L, Srinivanan BD. The synthesis of cyclooxygenase products in ocular tissues of
various species. eurr Eye Res 1984; 3:447-452.
31. Williams RN, Bhattacherjee P, Eakins KE. Biosynthesis of lipoxygenase products by ocular tissue. Exp
Eye Res 1983; 36:397-402.
32. Kass MA, Holmberg NJ. Prostaglandin and thromboxane synthesis by microsomes of rabbit ocular tissues.
Invest Ophthalmol Vis Sci 1979; 18:166-171.
33. Bazan HE. The synthesis and effects of eicosanoids in avascular ocular tissues. In: Bito LZ, Stjernschantz
J, eds. The Ocular Effects of Prostaglandins and Other Eicosanoids. New York: Alan R. Liss, 1989:73--E4.
34. Haynes WL, Proia AD, Klintworth GK. Effect of inhibitor of arachidonic acid metabolism on corneal ne-
ovascularization in the rat. Invest Ophthalmol Vis Sci 1989; 30: 1588-1593.
35. Oliw EH. Biosynthesis of 12(S)-hydroxyeicosatrienoic acid by bovine corneal epithelium. Acta Physiol
Scand 1993; 147:117-121.
36. Bhattacherjee P. The role of arachidonate metabolites in ocular inflammation. In: Bito LZ, Stjernschantz J,
eds. The Ocular Effects of Prostaglandins and Other Eicosanoids. New York: Alan R. Liss, 1989:211-228.
37. Gluud BS, Jensen OL, Krogh E, Birgens HS. Prostaglandin E2 level in tears during postoperative inflam-
mation of the eye. Acta Ophthalmol 1985; 63:375-379.
38. Napoli SA, Helm C, Insler MS, Ensley HE, Pretus HA, Feigen LP. External ocular inflammatory effects of
lipoxygenase enzyme products. Ann Ophthalmol1990; 22:30-34.
39. Eakins KE. Stier e, Bhattacherjee P, Greenbaum LM. Actions and interactions of bradykinin, prostagland-
ins, and nonsteroidal anti-inflammatory agents of the eye. Inflammation 1975; 1:117-125.
40. Woodward OF, Ledgard SE. Effect of LTD4 on conjunctival vasopermeability and blood-aqueous barrier
integrity. Invest Ophthalmol Vis Sci 1985; 26:481-485.
41. Srinivasan BD, Kulkarni PS. Inhibitors of arachidonic acid cascade in the management of ocular inflam-
mation. In: Bito LZ, Stjernschantz J, eds. The Ocular Effects of Prostaglandins and Other Eicosanoids.
New York: Alan R. Liss, 1989:229-249.
42. Flach AJ. eyclo-oxygenase inhibitors in ophthalmology. Surv Ophthalmol 1992; 36:259-284.
43. Kulkarni PS, Srinivasan BD. Steroidal and nonsteroidal anti-inflammatory drugs in ocular inflammation.
OcullnflamTher 1983; 1:11-18.
Cytochrome P450 and Arachidonic Acid Metabolism in the Corneal Epithelium 17
44. Spampinato S, Marino A, Bucolo C, Canossa M, Bachetti T, Mangiafico S. Effects of sodium Naproxen
eye drops on rabbit ocular inflammation induced by sodium arachidonate. J Ocul Pharm 1991; 7: 125--133.
45. Gupta AG, Hirakata A, Proia AD. Effect of inhibitors of arachidonic acid metabolism on corneal reepi-
thelialization in the rat. Exp Eye Res 1993; 56:701-708.
46. Beyer CF, Tepper OJ, Hill JM. Cryogenic induced ocular HSV-I reactivation is enhanced by an inhibitor of
the lipoxygenase pathway. Curr Eye Res 1989; 8: 1287-1292.
47. Kass MA, Neufeld AH, Sears ML. Systemic aspirin and indomethacin do not prevent the response of the
monkey eye to trauma. Invest Ophthalmol Vis Sci 1973; 14:604--606.
48. Gonzalez FJ. The molecular biology of cytochrome P450. Pharmacol Rev 1988; 40:275--288.
49. Nelson DR, Kamataki T, Waxman OJ, Guengerich FP, Estabrook RW, Feyereisen R, Gonzalez FJ, Coon
MJ, Gunsalus IC, Gotoh 0, Okuda K, Nebert OW. The P450 superfamily: update on the new sequences,
gene mapping, accession numbers, early trivial names of enzymes, and nomenclature. DNA Cell BioI
1993; 12:1-51.
50. Capdevila J, Parkhi11 L, Chacos N, Okita R, Masters BS, Estabrook RW. The oxidative metabolism of
arachidonic acid by purified cytochrome P-450. Biochem Biophys Res Commun 1981; 101: 1357-1363.
51. Oliw EH, Guengerich FP, Oates JA. Oxygenation of arachidonic acid by hepatic monooxygenase: isolation
and metabolism offour expoxide intermediates. J Bioi Chern 1982; 257:3771-3781.
52. Laethem RlI.1, Halpert JR, Koop DR. Epoxidation of arachidonic acid as an active-site probe of cytochrome
p-450 2B isoforms. Biochim Biophys Acta 1994; 1206:42-48.
53. Laethem RM, Koop DR. Identification of rabbit cytochrome P450 2CI and 2C2 as arachidonic acid epoxy-
genase. Mol Pharmacol 1992; 42:958-963.
54. Falck JR, Lumin S, Blair I, Dishman E, Martin MV, Waxman OJ, Guengerich FP, Capdevila JH. Cyto-
chrome P450-dependent oxidation of arachidonic acid to 16-, 17-, and 18-hydroxyeicosatrienoic acids. J
BioI Chern 1990; 265:10244-10249.
55. Shichi H. Microsomal electron transfer system of bovine retinal pigment epithelium. Exp Eye Res 1969;
8:6~8.
56. Shichi H, Atlas SA, Nebert W. Genetically regulated aryl hydrocarbon hydroxylase induction in the eye:
Possible significance of the drug-metabolizing enzyme system of the retinal pigment epithelium-choroid.
Exp Eye Res 1975; 21:557-567.
57. Das ND, Shichi H. Enzymes of mercapturate synthesis and other drug-metabolizing reactions-specific lo-
calization in the eye. Exp Eye Res 1981; 33:525--533.
58. Kishida K, Matsumoto K, Manabe R, Sugiyama T. Cytochrome P450 and related components of the mi-
crosomal electron transport system in the bovine ciliary body. Curr Eye Res 1986; 5:529--533.
59. Schwartzman ML, Masferrer J, Dunn MW, McGiff JC, Abraham NG. Cytochrome P450, drug metabo-
lizing enzymes and arachidonic acid metabolism in bovine ocular tissue. Curr Eye Res 1987: 6:623-629.
60. Abraham NG, Lin J, Dunn MW, Schwartzman ML. Presence of heme oxygenase and NADPH cytochrome
P450 (c) reductase in human corneal epithelium. Invest Ophthalmol Vis Sci 1987: 28:1464-1472.
61. Schwartzman ML, Pagano P, McGiff JC, Abraham NG. Immunochemical studies on the contribution of
NADPH cytochrome P450 reductase to the cytochrome P450-dependent metabolism of arachidonic acid.
Arch Biochem Biophys 1987; 252:635--{i45.
62. Matsumoto K, Kishida K, Manabe R, Sugiyama T. Induction of cytochrome P450 in the rabbit eye by phe-
nobarbital, as detected immunohistochemically. Curr Eye Res 1987; 6(7):847--E54.
63. Shichi H, Zhao C, Laniado-Schwartzman M. Immunocytochemical study of CYP4A I induction in mouse
eye. Invest Ophthalmol Vis Sci 1996; 37:S800.
64. Schwartzman ML, Abraham NG. Ocular cytochrome P450 metabolism of arachidonate: synthesis and bio-
assay. In: Murphy RC, Fitzpatrick FA, eds. Methods in Enzymology. New York: Academic Press Inc.,
1990:372-384.
65. Schwartzman ML, Abraham NG, Masferrer J, Dunn MW, McGiff JC. Cytochrome P450 dependent meta-
bolism of arachidonic acid in bovine corneal epithelium. Biochem Biophys Res Commun 1985;
132:343-351.
66. Lin MT, Wang Y-H, Chen Y-L, Chang EC. The effect of copper ion on arachidonic acid metabolism in the
porcine corneal epithelium. Biochem Biophys Res Commun 1993; 190:1122-1129.
67. Asakura T, Shichi H. 12(R)-hydroxyeicosatrienoic acid synthesis by 3-methylcholanthrene- and c1ofibrate-
inducible cytochrome P450 in porcine ciliary epithelium. Biochem Biophys Res Commun 1992;
187:455-459.
68. Asakura T, Matsuda M, Matsuda S, Shichi H. Synthesis of 12(R) and 12(S)-hydroxyeicosatrienoic acid by
porcine ocular tissue. J Ocul Pharm 1994; 10:525--536.
69. Masferrer J, Abraham NG, Dunn MW, McGiff JC, Schwartzman ML. Cytochrome P450 arachidonic acid
metabolism in human corneal epithelium. Invest Ophthalmol Vis Sci 1987; 28 (suppl):328.(abstract)
18 M. L. Schwartzman
70. Stoltz RA, Conners MS, Dunn MW, Schwartzman ML. Effect of metabolic inhibitors on arachidonic acid
metabolism in the corneal epithelium: Evidence for cytochrome P450 mediated reactions. J Ocul Pharm
1994; 10:307-317.
71. Yamamoto S, Nishimura M, Conners MS, Stoltz RA, Falck JR, Chauhan K, Laniado-Schwartzman M.
Oxidation and keto reduction of 12-hydroxy-5,8,10,14-eicosatrienoic acids in bovine corneal epithelial mi-
crosomes. Biochim Biophys Acta 1994; 1210:217-225.
72. Conners MS, Stoltz RA, Laniado-Schwartzman M. Chiral analysis of 12-hydroxyeicosatrienoic acid
formed by calf corneal epithelial microsomes. JOcular Pharmacol Ther 1996; 12: 19-26.
73. Conners MS, Stoltz RA, Webb SC, Rosenberg J, Dunn MW, Abraham NG, Schwartzman ML. A closed
eye-contact lens model of corneal inflammation. I. Induction of cytochrome P450 arachidonic acid metabo-
lism. Invest Ophthalmol Vis Sci 1995; 36:828--840.
74. Conners MS, Stoltz RA, Dunn MW, Levere RD, Abraham NG, Schwartzman ML. A closed eye-contact
lens model of corneal inflammation. II. Inhibition of cytochrome P450 arachidonic acid metabolism allevi-
ates inflammatory sequelae. Invest Ophthalmol Vis Sci 1995; 36:841-850.
75. Laniado Schwartzman M, Stoltz RA, Conners MS, Dunn MW, Mastyugin V, Abraham NG. Cytochrome
P450 (CYPIAI) expression and monooxygenase-derived arachidonic acid (AA) metabolism in the rabbit
corneal epithelium following injury. Invest Ophthalmol Vis Sci 1996; 37:S356.
76. Nishimura M, Schwartzman ML, Falck JR, Lumin S, Zirrolli JA, Murphy RC. Metabolism of 12(R)-hy-
droxy-5,8, I 0, 14-eicosatrienoic acid (I2(R)-HETE) in corneal tissues: formation of novel metabolites. Arch
Biochem Biophys 1991; 290:326--335.
77. Davis KL, Conners MS, Dunn MW, Schwartzman ML. Induction of corneal epithelial cytochrome P450
arachidonate metabolism by contact lens wear. Invest Ophthalmol Vis Sci 1992; 33:291-297.
78. Davis KL, Dunn MW, Schwartzman ML. Hormonal stimulation of 12(R)-HETE, a cytochrome P450
arachidonic acid metabolite in the rabbit cornea. Curr Eye Res 1990; 9:661--666.
79. Murphy RC, Falck JR, Lumin S, Yadagiri P, Zirrolli JA, Balazy M, Masferrer JL, Abraham NG,
Schwartzman ML. 12(R)-hydroxyeicosatrienoic acid: a vasodilator cytochrome P450 dependent arachi-
donate metabolite from bovine corneal epithelium. J Bioi Chern 1988; 263: 17197-17202.
80. Daikh BE, Laethem RM, Koop DR. Stereoselective epoxidation of arachidonic acid by cytochrome P450s
2CAA and 2C2. J Pharmacol Exp Ther 1994; 269:1130-1135.
81. Laethem RM, Laethem CL, Ding X, Koop DR. P-450-dependent arachidonic acid metabolism in rabbit ol-
factory microsomes. J Pharmacol Exp Ther 1992; 262:433-438.
82. Masferrer JL, Rios AP, Schwartzman ML. Inhibition of renal, cardiac and corneal Na+-K+-ATPase by
12(R) hydroxyeicosatrienoic acid. Biochem Pharmacol1990; 39:1971-1974.
83. Masferrer JL, Mullane KM. Modulation of vascular tone by 12(R)-, but not by 12(S) hydroxyeicosatrienoic
acid. Eur J Pharmacol 1988; 151 :487-490.
84. Webb RC, Bohr DF. Potassium-induced relaxation as an indicator of Na-K +-ATPase activity in vascular
smooth muscle. Blood Vessels 1978; 15:198-207.
85. Takahashi K, Capdevila J, Falck JR, Jacobson HB, Badr JF. Effects of cytochrome P450-derived metabo-
lites of arachidonic acid (AA) on renal hemodynamic and excretory functions in the rat. FASEB 1988;
2:5620.(abstract)
86. Quilley CP, McGiff JC. Renal effect of cytochrome P450-dependent metabolites of arachidonic acid: 12(R)
hydroxyeicosatrienoic acid. Br J Pharmacol 1989; 96:59P.
87. Edelhauser HF, Geroski DH, Woods WD, Holley GP, Schwartzman ML. Swelling in the isolated perfused
cornea induced by l2(R)-hydroxyeicosatrienoic acid. Invest Ophthalmol Vis Sci 1993; 34:2953-2961.
88. Kinoshita JH. Mechanism initiating cataract formation. Invest Ophthalmol Vis Sci 1974; 13:713-724.
89. Schwartzman ML, Dunn MW, Masferrer JL. 12(R) hydroxyeicosatrienoic acid a novel corneal metabolite,
lowers intraocular pressure. Invest Ophthalmol Vis Sci 1988; 29:85.
90. Socci RR, Bhattacherjee P, Delamere N. The influence of 12(R)-HETE upon ATPase activity in the rabbit
ciliary epithelium. Invest Ophthalmol Vis Sci 1989; (Suppl):447.
91. Bonting SL, Becker B. Studies on sodium-potassium activated adenosine-triphosphate: Inhibition of en-
zyme activity and aqueous humor flow in the rabbit eye after intravitreal injection of ouabain. Invest
Ophthalmol Vis Sci 1964; 3:523-533.
92. Simon KA, Bonting SL, Hawkins NM. Studies on a Na-K activated ATPase. II. Formation of aqueous hu-
mor. Exp Eye Res 1962; 1:253-261.
93. Krupin T, Becker B, Podos SM. Topical vanadate lowers intraocular pressure in rabbits. Invest Ophthalmol
Vis Sci 1980; 19:360-1366.
94. DeSousa RC, Grosso A. Vanadate blocks cyclic AMP-induced stimulation of sodium and water transport in
amphibian epithelia. Nature 1979; 279:803.
Cytochrome P4S0 and Arachidonic Acid Metabolism in the Corneal Epithelium 19
95. Crawford K, Kaufinan PL. Pilocarpine antagonizes prostaglandin F2-induced ocular hypotension in mon-
keys. Arch Ophthalmol1987; 105:1112.
96. Ping-yu L, Podos SM, Severin C. Effect of prostaglandin F2 on aqueous humor dynamics of rabbits, cat
and monkey. Invest Ophthalmol Vis Sci 1984; 25:1087-1093.
97. Masferrer J, Murphy RC, Pagano PJ, Dunn MW, Schwartzman ML. The ocular effects of a novel arachi-
donate metabolite fonned by bovine corneal epithelium. Invest Ophthalmol Vis Sci 1989; 30:454-460.
98. Masferrer JL, Rimarachin JA, Gerritsen ME, Falck JR, Yadagiri P, Dunn MW Schwartzman ML. 12(R)-hy-
droxyeicosatrienoic acid, a potent chemotactic and angiogenic factor produced by the cornea. Exp. Eye
Res. 1991 ;52:417-424.
99. Ford-Hutchinson AW, Bray MA, Doig MV, Shipley ME, Smith MJH. Leukotriene B, a potent chemokinetic
and aggregating substance released from polymorphonuclear leukocytes. Nature 1980; 286:264-265.
100. Folkman J, Klagsbum M. Angiogenic factors. Science 1987; 235:442-447.
101. Fromer CH, Klintworth GK. An evaluation of the role of leukocytes in the pathogenesis of experimentally
induced corneal vascularization. Am J Pathol 1976; 82: 157-168.
102. Laniado-Schwartzman M, Stoltz RA, Rapacon MM, Lavrovsky Y, Conners MS, Abraham NG. Rapid and
stereospecific activation of NFkB by a corneal angiogenic eicosanoic in microvesse! endothelial cells. In-
vest Ophthalmol Vis Sci 1995; 3S: 1466.
103. Stoltz RA, Conners MS, Gerritsen ME, Abraham NG, Laniado-Schwartzman M. Direct stimulation of Iim-
bal microvessel endothelial cell proliferation and capillary fonnation in vitro by a corneal-derived ei-
cosanoic. Am J Patho11996; 148: 129-139.
104. Stoltz RA, Abraham NG, Laniado Schwartzman M. NFkB activation is a key detenninant in the angio-
genic response of microvessel endothelial cells to 12(R)-HETrE. Proc Nat! Acad Sci USA 1996;
93:2832-2837.
lOS. Eliason JA, Elliot JP. Proliferation of vascular endothelial cells stimulated in vitro by corneal epithelium.
Invest Ophthalmol Vis Sci 1987; 28: 1963-1969.
106. Eliason JA. Angiogenic activity of the corneal epithelium. Exp Eye Res 1985; 41 :721-732.
107. Eliason JA. Leukocytes and experimental corneal vascularization. Invest Ophthalmol Vis Sci 1978;
17: 1087-1 09S.
108. Sholley MM, Gimbrone MA, Cotran RS. The effects of leukocyte depletion on corneal vascularization.
Lab Invest 1978; 38:32-40.
109. Camp RDR, Cunningham FM, Fincham NJ, Greaves MW, Black AK, Mallet AI, Woollard PM. Psoriatic
skin lesions contain a novel lipid neutrophil chemokinetic compound which is distinct from known
chemoattractant compounds. Br J Phannacol 1988; 94: I 043-1 OSO.
110. Camp RDR, Mallet AI, Woollard PM, Brain SD, Black AK, Greaves MW. The identification of hydroxy
fatty acids in psoriatic skin. Prostaglandins 1993; 26:431-448.
III. Van Wauwe J, Coene M-C, Van Nyen G, Cools W, Goossen J, Lejeune L, Lauwers W, Janssen PAJ.
NADPH-dependent fonnation of 15- and 12-hydroxyeicosatrienoic acid from arachidonic acid by rat epi-
dennal microsomes. Eicosanoids 1991; 4:155-163.
112. Wainwright S, Falck J.R. , Yadagiri P, Powell WS. Metabolism of 12(S)-hydroxy-5,8, 10, 14-eicosatrienoic
acid and other hydroxylated fatty acids by the reductase pathway in porcine polymorphonuclear leuko-
cytes. Biochem 1990; 29:10126--10135.
113. Cunningham FM, Woollard PM. 12(R)-hydroxy-5,8,1O,14-eicosanoic acid is a chemoattractant for human
polymorphonuclear leukocytes in vitro. Prostaglandins 1987; 34:71-78.
114. Bacon KB, Camp RDR, Cunningham FM, Woollard PM. Contrasting in vitro lymphocytes chemotactic ac-
tivity of hydroxyl enantiomers of 12-hydroxy-5,8,10,14-eicosatrienoic acid. Br J Phannacol 1988;
95:966--974.
liS. Evans JF, Leblanc Y, Fitzsimmons BJ, Charleson S, Nathaniel D, Leveille C. Activation of leukocyte
movement and displacement of [3H]leukotriene B4 from leukocyte membrane preparations by (12R)- and
(12S)-hydroxyeicosatrienoic acid. Biochim Biophys Acta 1987; 917:406--410.
116. Woollard PM, Murphy GM, Cunningham FM, Camp RDR, Greaves MW. Pro-inflammatory effects of
12(R)-hydroxy-S,8,1O,14-eicosatrienoic acid in human skin. Brit J Dennatol 1988; 118:277.(abstract)
117. Conners MS, Godfrey H, Chauhan K, Falck JR, Laniado-Schwartzman M. Enhancement of delayed-type
hypersensitivity inflammatory reactions by 12(R)-hydroxy-S,8,14-eicosatrienoic acid. J Invest Dennatol
1994; 104:1-5.
118. Capdevila J, Gil L, Orellana M, Marnett LJ, Mason JI, Yadagiri P, Falck JR. Inhibitors of cytochrome P450
dependent arachidonic acid metabolism. Arch Biochem Biophys 1988; 261 :257-263.
119. Testa B, Jenner P. Inhibitors of cytochrome P4S0s and their mechanism of action. Drug Metab Rev 1981;
12:1-117.
120. Maines MD, Kappas A. Metals as regulators of heme metabolism. Science 1977; 198: 1215-1221.
20 M. L. Schwartzman
121. Maines MD, Kappas A. Studies on the mechanism of induction of heme oxygenase by cobalt and other
metal ions. Biochem J 1976; 154: 125-131.
122. Abraham NG, Lin H-C, Schwartzman ML, Levere RD, Shibahara S. The physiological significance of
heme oxygenase. Int J Biochem 1988; 20:543-588.
123. Schwartzman ML, Abraham NG, Carroll MA, Levere RD, McGiff Jc. Regulation of arachidonic acid me-
tabolism by cytochrome P450 in rabbit kidney. Biochem J 1986; 238:283-290.
124. Sacerdoti D, Escalante B, Abraham NG, McGiff JC, Levere RD, Schwartzman ML. Treatment with tin pre-
vents the development of hypertension in spontaneously hypertensive rats. Science 1989; 243:388--390.
125. Kappas A, Maines MD. Tin: a potent inducer of heme oxygenase in kidney. Science 1976; 192:60-62.
126. Grant WM. Toxicology of the Eye. 2nd ed., CC Thomas, Springfield Illinois: 1974.
127. DaSilva J-L, Tiefenthaler M, Park E, Escalante B, Schwartzman ML, Abraham NG. Tin mediated heme
oxygenase gene activation and cytochrome P450 arachidonate hydroxylase inhibition in spontaneously hy-
pertensive rats. Am J Med Sci 1994; 70: 173-181.
2
Department of Ophthalmology
Emory University
Atlanta, Georgia
1. INTRODUCTION
The purpose of this presentation is twofold: (1) to summarize our studies as to the
direct effect of 12(R)HETE, 8(R)HHDTrE on the corneal endothelium, and (2) to illus-
trate the diffusion and metabolism of 12(R)HETE by the cornea. Our data has shown that
12(R)HETE causes corneal swelling in a dose-dependent manner when perfused directly
to the corneal endothelium. l As reported by Schwartzman et ai, 12(R)HETE is one of the
major cytochrome P450-dependent arachidonate metabolites of the corneal epithelium and
is an endogenous inhibitor of Na+/K+ ATPase. 2 Contact lens-induced hypoxic stress has
also been shown to stimulate the endogenous formation of l2(R)HETE by the corneal epi-
thelium. 3- s This increased formation of 12(R)HETE was found to correlate to increased
corneal thickness and to changes in the corneal endothelial structure. 3 These results sug-
gest that 12(R)HETE produced by the corneal epithelium might be capable of diffusing
across the cornea and affecting endothelial function.
2. REVIEW
The current review describes the direct effects of 12(R)HETE on the corneal endo-
thelium.
As shown in Table I, 12(R)HETE obtained from two sources (bovine and starfish)
has comparable effects on corneal swelling, Maximal effects are seen at concentrations
between 10-6 and 10-5 molli. The ultrastructure of the corneal endothelium after perfusion
of the 12(R)HETE also showed intact intercellular junctions and minimal endothelial cell
swelling. l The corneal endothelial ultrastructural appearance was comparable to that pre-
viously described after ouabain perfusion. After ouabain perfusion, the intercellular junc-
tions remained intact and endothelial paraceIIuIIar permeability was unaltered. 6 Endothe-
lial paracellular permeability is also unaffected by 12(R)HETE (Table 2), It can be con-
cluded that the corneal swelling that occurs when 12(R)HETE is perfused to the corneal
endothelium results from the inhibition of the endothelia pump function.
Our results agree with past studies that show that l2(R)HETE can lower intraocular
pressure by inhibiting ciliary body Na+/K+ ATPase 7 and that corneal epithelial Na+/K+ can
be inhibited by 12(R)HETE. 2 Reversal of corneal swelling can be obtained after the re-
moval of 12(R)HETE from the perfusate; whereas, swelling produced by an equal molar
concentration of ouabain does not reverse after removal of ouabain from the perfusate fol-
lowing 2 hours of washout. I If 12(R)HETE inhibits the Na+/K+ ATPase directly by binding
to the enzyme, our data suggests that ouabain binding to Na+/K+ ATPase is of higher affin-
ity than binding by l2(R)HETE.
HPLC analysis of the l2(R)HETE perfused to the corneal endothelium confirmed
that the compound remained in solution and was not significantly oxidized during the 2-3
hours of corneal endothelial perfusion; however, after 30 minutes of exposure to the cor-
neal endothelium, the perfusate contains a prominent peak with an HPLC elution time of
13 minutes. This peak was more polar than the 12(R)HETE and represents a metabolite of
the 12(R)HETE, called 8(R)HHDTrE.I,8 Such a product may also represent a metabolite
produced by the corneal endothelium to either enhance or blunt the effect of the
12(R)HETE to inhibit Na+/K+ ATPase.
When corneal swelling rates observed with 12(R)HETE perfusion are compared on a
molar basis to those of ouabain perfusion, greater corneal swelling occurs with ouabain
perfusionY However, purified Na+/K+ ATPase (from dog kidney) is inhibited to a greater
extent by 12(R)HETE compared to the inhibition produced by an equivalent concentration
of ouabain. J These findings suggest that when the 12(R)HETE is perfused directly to the
corneal endothelium, there is only partial inhibition of the Na+/K+ ATPase. The
12(R)HETE inhibition is also readily reversible. The rapid reversibility of 12(R)HETE,
compared to ouabain, might reflect the lower binding affinity of 12(R)HETE for the
Na+/K+ ATPase.) The appearance of the 12(R)HETE metabolite after endothelial perfusion
also suggests that metabolic breakdown might be a factor in the rapid reversibility of the
corneal swelling. Lower binding affinity of 12(R)HETE to native endothelial Na+/K+ AT-
Pase and endothelial metabolism of the compound could explain the lower swelling rates
(and greater inhibition of purified Na+/K+ ATPase) seen with 12(R)HETE perfusion, com-
pared to ouabain. It is also possible that the expression of Na+/K+ ATPase isoforms by the
corneal endothelium differs from the expression of isoforms by the kidney, and
12(R)HETE may affect the enzyme isoforms differently. '
The cytochrome P450 system of the corneal epithelium may have a role in the devel-
opment of contact lens-induced endothelial polymegathism. For example, contact lenses
may contain inflammatory mediators or limit the amount of oxygen delivered to the cor-
neal epithelium. This could increase the nicotinamide adenine dinucleotide phosphate gen-
erated by the hexose-monophosphate shunt and activate the cascade of arachidonic acid
metabolism through the epithelial P450 system (Figure I ).1.8 Our studies have shown that
l2(R)HETE may directly effect corneal endothelial Na+/K+ ATPase. The release of
l2(R)HETE by the corneal epithelium could impose a continual stress on the endothelial
cell Na+/K+ ATPase that ultimately affects the cell's ability to regulate volume. Such a
constant stress of Na+/K+ ATPase inhibition may result in a permanent cytoskeleton
change,IO·)) as has been reported in cultured cells when the Na+/K+ ATPase has been inhib-
ited with sultbydryl inhibitor (p-chloromercuriphenyl-sulfonic acid) or by increasing the
cell volume.
The second part of this review summarizes our studies on the diffusion and metabo-
lism of 12(R)HETE. We have been able to demonstrate that 12(R)HETE is capable of dif-
fusion or being transported from the epithelial to the endothelial side of the cornea
through the hydrophilic stroma. B In all cases: (A) intact cornea, (B) cornea having the epi-
thelium removed and (C) cornea having both epithelium and endothelium removed, the
cornea was permeable to l2(R)HETE and "Peak C" (8(R)HHDTrE). In addition, these
studies show that 12(R)HETE can be metabolized by all layers of the cornea (Figure I).
inhibits endothelial
NalKATPase
causes corneal swelling
Figure t. Diagram of corneal epithelial arachidonic acid metabolism by the corneal epithelium via the cyto-
chrome P450 pathway.
24 H. F. Edelhauser et aL
Our results suggest that 12(R)HETE produced in the corneal epithelium is capable
of diffusing posteriorly across the cornea, where it ultimately could affect the endothelial
metabolic pump as shown with direct endothelial perfusion. I Therefore, the ability of
12(R)HETE to inhibit the endothelial Na+/K+ ATPase as it diffuses across the cornea from
the epithelium may be a possible mechanism leading to the formation of polymegathism
in the corneal endothelium following contact lens wear.3 Our data 8 demonstrates that
12(R)HETE and its associated metabolites can diffuse across the cornea, even when the
epithelium is exposed to a very low concentration (8.2 nM).
Analysis of the radioactivity in the tear side reservoir of the perfusion chamber of the
isolated cornea shows a time-dependent decrease. When the epithelium is intact, 3H_
12(R)HETE rapidly disappears from the perfusion media, approximately 30% within the first
0.5 hour exposure. After 30 minutes, the rate levels off to a linear rate of approximately 5%
per hour. The data show that 12(R)HETE can be rapidly taken up into the corneal epithelium,
stroma and endothelium. 12(R)HETE is a very hydrophobic material that is easily absorbed
into the lipophilic epithelium of the cornea, a more compatible environment than the aqueous
side GBR perfusion media. Analysis of the tissue samples perfused for two hours (Table 3)
showed that when normalized by weight, the epithelial uptake of 12(R)HETE was greater
than in the endothelium or stroma. The concentration of total radioactivity (l2(R)HETE +
metabolites) was 16,500 d.p.m./mg in the epithelium as compared with approximately 600
d.p.m./mg ofGBR, which was initially added to the tear side reservoir. The data derived from
the time-dependent radioactivity (d.p.m.) in the aqueous side perfusion media showed that as
much as 15% of the radioactivity added to the tear side of the cornea can reach the aqueous
side reservoir in a 2 hour period when the epithelium remains intact. If the percentage of
12(R)HETE in the aqueous side reservoir is converted to a concentration of radioactivity per
mg to compare with the tissue uptake values, it translates to 90 d.p.m./mg GBR. When the
value is compared with the tissue uptake value of 3H-12(R)HETE into the endothelium, which
is 695±69 d.p.m./mg (Table 3), it can be inferred that about 7 times more 3H-12(R)HETE ac-
tually enters the endothelium than appears in the aqueous side reservoir. The greater retention
of 3H-12(R)HETE in the endothelium is most likely due to its hydrophilicity. Even when the
epithelium was removed, there was more radioactivity in the endothelium per unit weight
than in the GBR perfusate of the endothelial reservoir. Chromatographic analysis of the
12(R)HETE, however, determined that only a portion of the radioactive material that diffused
across into the endothelial reservoir was 12(R)HETE.8 The rest of the 12(R)HETE had been
converted to more polar compounds including the major metabolite previously described by
our group I a "Peak C." "Peak C" has not been definitely identified; however, through spectral
analysis and retention time values, "Peak C" is similar to 8(R)HHDTrE, which has also been
reported by Nishimura et al 12 to be a B-oxidation product of 12(R)HETE in the cornea. Analy-
sis by GC/Mass Spec will be necessary to confirm this identifY.
The metabolism of 12(R)HETE was greatest when the epithelium was intact but for-
mation of "Peak C" still occurred in the absence of the epithelium and to a lesser degree
when both the epithelium and endothelium were removed. 8 This observation suggests that
the majority of 12(R)HETE metabolism is occurring in the epithelial cells, although it can
also be metabolized by the endothelium and by stromal keratocytes. When the epithelium
is intact, approximately 3% of the l2(R)HETE, initially added to the tear side reservoir,
ends up as 12(R)HETE in the aqueous side reservoir. This represents a concentration of
approximately 2.5 x 10-10 M.
Following two hours of perfusion with an intact epithelium, "Peak C" is two-fold
greater than 12(R)HETE on the aqueous side. 8 Therefore, based upon the tissue uptake
(d. p.m.), one might expect at "Peak C" may also be concentrated in the endothelium. The
effects of "Peak C" upon the corneal tissues are as yet unknown. However, preliminary
studies by Woods et al l3 have shown that when "Peak C" is perfused directly to the endo-
thelium of the isolated rabbit corneas, swelling of the cornea will occur, suggesting that it
is also metabolically active.
In conclusion, our data has shown that l2(R)HETE placed on the tear side of the iso-
lated rabbit cornea is rapidly taken up by the epithelium where it is metabolized to the
compound "Peak C" and several other more polar metabolites. Some of this material dif-
fuses back into the tear side reservoir. The remainder of the metabolites, along with resid-
ual 12(R)HETE, can diffuse posteriorly across the cornea, through the stroma to the
endothelium where further metabolism of 12(R)HETE occurs.
Since it has been shown by Davis et al 5 and Conners et a1. 4 •5 that contact lenses acti-
vate the corneal epithelial cytochrome P450 pathway to produce both 12(R)HETE and
12(R)HETrE, understanding the diffusional and/or transport characteristics and metabo-
lism of 12(R)HETE in the cornea is of clinical importance for individuals wearing contact
lenses. This study has shown that 12(R)HETE and its metabolites preferentially remain in
singular
event
- low osmolality daily stress
- contacllens wear
- ouab n
I
lit
cellular - diabetes
.. swelling - inllammatlon
slight
cytoskeletal years 01
change volume regulation
_ _ _ volume permanent
cytoskeleton change
regulation
cell cytos eieton Polymegath sm
recovers
Figure 2. Summary diagram comparing the short-tenn corneal endothelial cellular changes to the long-term
changes that occur over many years that can cause corneal endothelial cell polymegathism.
26 H. F. Edelhauser et al.
the cornea as opposed to diffusing into the aqueous environment of tears and aqueous hu-
mor. Although 12(R»HETE is significantly metabolized in the corneal layers, there is
likely to be a significant amount of unmetabolized 12(R)HETE reaching the corneal endo-
thelium where it and/or one of its metabolites could ultimately affect the endothelial cells.
Thus, the endothelial metabolic pump can be inhibited to cause endothelial cellular and
corneal stroma swelling. I With repeated endothelial cell exposure to 12(R)HETE and its
metabolites that cause endothelial cell swelling, there is the possibility that l2(R)HETE
may induce a permanent change in the cellular cytoskeleton,14 leading to endothelial po-
lymegathism, which is a common occurrence in contact lens wear (Figure 2). It should be
mentioned that corneal endothelial polymegathism only occurs following long-term vol-
ume regulation stress such as contact lens wear and in diabetic patients.
3. ACKNOWLEDGMENTS
Supported in part by P30 EY06360 (Departmental Core Grant), EY00933, and Re-
search to Prevent Blindness, Inc. Henry F. Edelhauser, Ph.D., is a Research to Prevent
Blindness Senior Scientific Investigator.
4. REFERENCES
I. Edelhauser HF, Geroski DH, Woods WD, Holley GP, Schwartzman ML. Swelling in the isolated perfused
cornea induced by 12(R)hydroxyeicosatetraenoic acid. Invest Ophthalmol Vis Sci 1993;34:2953-2961.
2. Schwartzman ML, Balazy M, Masferrer J. Abraham NG, McGiff JC, Murphy RC. 12(R)-hydroxyei-
cosatetraenoic acid: A cytochrome-P450-dependent arachidonate metabolite that inhibits Na+/K+-ATPase in
the cornea. Proc Natl Acad Sci USA 1987;84:8125--8129.
3. Davis KL, Conners MS, Dunn MW, Schwartzman ML. Induction of corneal epithelial cytochrome P450
arachidonate metabolism by contact lens wear. Invest Ophthalmol Vis Sci 1992;33:291-297.
4. Conners MS, Stoltz RA, Webb SC, Rosenberg J, Dunn MW, Abraham NG, Schwartzman ML. A closed eye
contact lens model of corneal inflammation. Part I: Increase synthesis of cytochrome P450 arachidonic
acid metabolism. Invest Ophthalmol Vis Sci 1995;36:828--840.
5. Conners MS, Stoltz RA, Davis KL, Dunn MW, Abraham NG, Levere RD, Schwartzman ML. A closed eye
contact lens model of corneal inflammation. Part II: Inhibition of cytochrome P450 arachidonic acid meta-
bolism alleviates inflammatory sequelae. Invest Ophthalmol Vis Sci 1995;36:841-850.
6. Watsky MA, McCartney MD, McLaughlin BJ, Edelhauser HF. Corneal endothelial junctions and the effect
of ouabain. Invest Ophthalmol Vis Sci 1990;3 I :933-941.
7. Delamere NA, Socci RR, King KL, Bhattacherjee P. The influence of 12(R)-hydroxyeicosatetraenoic acid
on ciliary epithelial sodium, potassium-adenosine triphosphatase activity and intraocular pressure in the
rabbit. Invest Ophthalmol Vis Sci 1991;32:2511-2514.
8. Williams KK, Woods WD, Edelhauser HF. Corneal diffusion and metabolism of 12(R)-hydroxyei-
cosatetraenoic acid (12(R)HETE). Curr Eye Res 1996;15:852-859.
9. Geroski DH, Kies JC, Edelhauser HF. The effects of ouabain on endothelial function in human and rabbit
corneas. Curr Eye Res 1984;3:331-338.
10. Mills JW.. The cell cytoskeleton: Possible role in volume control. Curr Top Membrane Transport
1987;30:75--101.
II. Kleinzeller A, Zlyadeh FN. Cell volume regulation in epithelia-with emphasis on the role of osmolyles and
the cytoskeleton. Cell Volume Regulation, Beyenback KW (ed), Karger, Basel, 1990:59-86.
12. Nishimura M, Schwartzman ML, Falck JR, Lumin S, Zirrolli JA, Murphy RC. Metabolism of 12(R)hy-
droxy-5,8,IO,14-eicosatetraenoic acid 12(R)HETE) in corneal tissues: Formation of novel metabolites.
Arch Biochem Biophys 1991;290:326-335.
13. Woods WD, Holley GP, Edelhauser HF, Schwartzman ML. Comparable swelling is induced by corneal en-
dothelial perfusion with either 8(R)-hydroxyhexadecatrienoic acid or 12(R)HETE. Invest Ophthalmol Vis
Sci I 995;36(Suppl): 138.
14. Kim EK, Geroski DH, Holley GP, Urken SI, Edelhauser HF. Corneal endothelial cytoskeletal changes in F-
actin with aging, diabetes, and after cytochalasin exposure. Am J Ophthalmol 1992; 114:329-335.
3
Marjorie F. Lou
ABSTRACT
may play an important role in cataract formation while thioltransferase may be a repair
system for the lens.
aggregates containing proteins and membrane particles with sizes over 1 x 106 Dalton. 6
Although HMW components are also found in the normal lens, they are formed by hydro-
phobic interactions. 7 .8 The HMW aggregates in the cataractous lens differ not only be-
cause they occur in larger quantities but they are formed by covalent linkages. The light
scattering theory of Benedek9 predicts that when the lens protein aggregates reach a size
greater than 5 x 107 Dalton light scattering will occur, if a sufficient number of such ag-
gregates are formed, transparency is lost and opacification occurs.
What causes these macromolecular aggregates to increase in size and quantity?
Some believe the posttranslational glycation of protein by glucose or ribose is the major
factor. lo In recent years, ascorbate has been considered as the mediator for glycation ll .12
and also for the mixed function oxidation of proteins. 13 The role of ascorbate in glycating
lens protein was strengthened when Nagaraj et al. I2 recently demonstrated a correlation
between pentosidine, an ascorbate-derived protein crosslink, and the degree of pigmenta-
tion in human cataractous lenses. Another focus of protein modification research has been
the possible protein-thiol mixed disulfide formation, (PSSG, PSSC formation) or protein
thiolation. In the 1970's, Harding,4 Truscott and Augusteyn,5 and Anderson and Spector l4
all observed an elevated PSSG formation in human cataractous lenses. However Reddy
and Han ls did not find any increase of PSSG in the cataractous lens of a galactosemic rat.
Two reasons may have prevented the advancement of the idea of the importance of protein
thiolation in cataracts. First, there was no reliable method to differentiate and quantify the
protein-thiol mixed disulfides, thus the reported PSSG value varied considerably. Second,
interest in the protein-thiol mixed disulfide in cataracts was diminished after negative re-
sults in an osmotic induced cataract model were reported suggesting that the protein thio-
lation process did not appear to play an essential role in cataractogenesis. ls However, the
findings from our laboratory have rekindled interest in this field. A new method was de-
veloped in our laboratory in the mid 80's which not only identifies, but quantifies repro-
ducibly the non-protein thiol molecular species that thiolate the proteins. 16 This method
development allows further studies on the role of protein thiolation in cataractogenesis.
This review will cover only research progress in the lens, including the status of protein-
thiol mixed disulfides in normal and cataractous lenses, their proposed role in cataract for-
mation, and the possible regulation by thioltransferase, an enzyme recently demonstrated
to be present in the lens. The general subject of protein-thiol mixed disulfides in other tis-
sues can be found in several review articles. 17. 18
To prove the presence of protein-thiol mixed disulfides in the lens, the first step was
to develop a method which could quantify the degree of conjugation between the non-pro-
tein thiols and proteins. The method had to be sensitive and reproducible enough to detect
the basal level of the bound non-protein thiols in the normal tissue. Harding 4 and the oth-
ers 5 . 14 used NaBH4 to reduce the disulfide linkages in the lens homogenate and quantified
the released non-protein thiols with a nonspecific chromogen and reported all the colored
products as GSH. Reddy and Han l5 used a more sensitive amino acid analyzer to quantify
the released GSH. All these methods suffer from the risk of auto-oxidation of the released
thiols as well as possible incomplete removal of added NaBH4 or thiol protecting agents.
To circumvent the above shortcomings we have developed a new method by using a
strong oxidant (performic acid) to treat the lens homogenate, thereby opening the disulfide
bridges and releasing the non-protein thiols as stable sulfonic acid products. The unused
30 M. F. Lou
performic acid is inactivated with excess amount of water and then removed by lyophili-
zation. The dried powder is then reconstituted in 10% trichloroacetic acid (TCA) followed
by centrifugation to remove the unwanted oxidized proteins. The remaining acid soluble
non-protein sulfonic acids are saved for further analysis.
In biological samples, there are only a few non-protein thiol candidates which can
form protein-thiol mixed disulfides with proteins. These include GSH, cysteine and cys-
teamine. By complete oxidation, they will be derivatized to glutathione sulfonic acid, cys-
teic acid and taurine, respectively. Since they are strong acidic compounds with a free
a-amino group which can react with the ninhydrin reagent to form a colored product, any
amino acid chromatographic system with an anion exchange column can be used to sepa-
rate these sulfonic acids and to quantify the color intensity at 570 nm.
The initial analysis was developed using a single-column ion-exchange chroma-
tographic system constructed in Hamilton's laboratory.'9 This instrument was modified to
accommodate an anion-exchange resin of 9 ± 1 Ilm beads as described in Lou, Mckellar
and Chyan. '6 Later, a commercial automated Dionex BioLC system equipped with an
HPIC-AS8 anion-exchange column (Dionex Corp., Sunnyvale, CA, USA) was used. 20
Samples can be applied on the column manually either through an off-column injector
(LDC/Milton Roy Co., Riviera Beach, FL) or through an autoinjector. This system of sub-
nanomole sensitivity is computer-controlled and computes each peak by comparison with
the peak areas of standards using the PeakNet software (Dionex Corp.). This automated
system has shortened the analysis from 70 min to 20 mins.
Standards, GSH sulfonic acid and cysteic acid or taurine are available commercially
(Sigma Chemical Co., St Louis, MO). Typically, 5 nmoles each of the standards are in-
jected onto the column pre-equilibrated with IOmM Na2B40/120 mM NaOHI 2% metha-
nol mobile phase and eluted with a sodium acetate gradient (30% to 100% 0.5 M Na
acetate (pH 13.2) over 4 min, followed by 100% acetate buffer for 4 min and a regenera-
tion cycle with the first buffer. The flow rate is 1.0 ml/min. The column effluent is mixed
with the ninhydrin reagent (0.67 ml/min) and passes through a reaction coil heated to
130°C before the absorbance is monitored at 570 nm. The total time for each analysis is 20
min. A typical chromatogram for the standards is depicted in Figure 1.
E
c:
...
o
.
It)
..
o
c:
.0
.,o
2
.0
c(
The advantage of this method is its ability not only to differentiate the nature of
these non-protein thiols by the analytical system but also to provide a subnanomole sensi-
tivity and quantitation. Since the formation of protein-thiol mixed disulfides is nonenzy-
matic, a quick removal of the endogenous nonprotein thiols, reduced or oxidized, during
tissue processing is essential to prevent any artifact.
a 05
CS03H
0.04
I
i 0.03
i
0.02
Mixed disulfidest
Species (n) Age GSH GS03 H CS0 3 H
Monkey (3) 1 yr 3.33 ± 0.51 0.013 ± 0.001 0.010±0.001
Bird (emu) (3) 3 yrs 1.48 ± 0.08 0.023 ± 0.003 0.020 ± 0.001
Rat (6) 1.5 m 4.48 ± 0.43 0.014 ±0.01 0.320 ± 0.01
Mouse (3) 8m 5.33 ± 0.20 0.034 ± 0.005 0.430 ± 0.10
Squirrel (4) adult 6.75 ± 1.05 0.140 ± 0.02 0.060 ± 0.03
Dog (3) I yr 8.00 ± 0.55 0.260 ± 0.05 0.060 ± 0.01
Pig (4) 6m 6.94 ± 0.67 0.041 ± 0.01 0.035 ± 0.02
Guinea pig (3) 2 yrs 9.59 ± 0.50 0.406 ± 0.04 0.024 ± 0.002
Human (6) 19-20 yrs 2.62 ± 1.30 0.220 ± 0.15 0.040 ± 0.013
'Data is expressed as 11 mole/g wet weight, mean ± S.D.; n, number of samples.
tGS0 3 H represents protein·GSH mixed disulfide, CS0 3 H represents protein-cysteine mixed disul-
fide.
Protein-Thiol Mixed Disulfides and Thioltransferase in the Lens 33
The sensitivity of this method has been greatly improved over the previous tech-
niques used by Harding 4 and others,S.14 in which each analysis requires 2 human lenses
while our method only needs 2% of one human lens or 1/5 of one young rat lens. This new
method not only quantifies PSSG accurately but also reveals a new protein-thiol disulfide
species PSSc. In addition the technique opened the door to the investigation of the impor-
tance of PSSG and PSSC in cataractogenesis and to subsequent studies on control mecha-
nisms used by lenses to combat oxidative stress.
Table 3. Effect of age on the protein-thiol mixed disulfide distributions in rat lens protein fractions
0
:;- 0.20
;0
0
"
;0
0
eE 0.15 0
'-'"
I
I",0.10 0
0 8
00
o 0
fJl
o 0 0 0 0
co
0.05 0 0
0
10 20 30 40 50 60 70 80 90
AGE (years)
Figure 3. Protein-cysteine mixed disulfide (PSSC) in normal human lenses as a function of age. Individual lens
PSSC level (measured as CS0 3 H) is expressed in Ilmol g" wet weight. Linear regression line shown, r = 0.75, P <
0.0005.
0.8
o
o 0
o o
o
o
cP
o
o o
000 0 @o 0
0.2 Figure 4. Protein-GSH mixed disulfide (PSSG)
0 0
<0
(f0000 c9 0
0
in normal human lenses as a function of age. In-
00
0 0 0 dividual lens PSSG level (measured as GS03 H)
0 is expressed as Ilmol g" wet weight. Linear re-
0.0 L...----L_-'-_.l..----1_---L_-'-_'-----L_-'
o ID 20 30 40 50 60 70 80 90 gression line for data between zero and 19 years.
AGE (years) r = 0.72, P < 0.05.
Protein-Thiol Mixed Disulfides and Thioltransferase in the Lens 35
a
~
4
a 00
" ~
E
a a
00
a a
a
~
'"
"-
0 €l
aS a
E 3 a
-.:, a a
oar§! a
:r:
Vl
'-'
cB a 00
<lD
a rS>
2
a
a aBC{)
a 0
00
1
0 10 20 30 40 50 60 70 80 90
AGE (years)
Figure 5. Free GSH levels in normal human lenses as a function of age. Individual GSH level is expressed in
flmol g-' wet weight Linear regression line shown, r = -0.52, p < 0.0005.
yrs may be associated with a physiological function of PSSG in the lens while the late-
stage elevation is mostly pathological. It is also speculated that an age-dependent decrease
in free GSH, in particular in older lenses, may contribute to the accumulation of PSSG in
the same lenses (Figure 5). The effect of age on PSSG and PSSC in cortical and nuclear
regions of the lens and the WS and WI protein fractions are summarized in Tables 4 and 5
respectively. The source of PSSG will be further discussed in the next section.
~~~----------------------------------~
o 04 0·8 ,·0
Figure 6. Rat lens protein-thiol mixed disulfides as a function of H20, concentration. Lenses of I month-old rats
were exposed to various concentration ofHP, (0.05 mM. 0.1 mM, 0.5 mM and 1.0 mM) for 18 hr in culture. Each
data point represents four lenses which were pooled and processed for GS0 3H and cysteic acid quantitation. The
GSH data represent mean (n = 4) ± S.E.M.
Protein-Thiol Mixed Disulfides and Thioltransferase in the Lens 37
shows a reciprocal relationship to the gradually increased PSSG. GSSG however has not been
detected either in the lens or in the culture media. The formation of PSSG is also shown as a
function of time (Figure 7). PSSC level remains unchanged under both the time-dependent
and the HP2 concentration-dependent experimental conditions.
Recently I3C-NMR studies on cultured rabbit lens 26 have supported the hypothesis
that PSSG forms under oxidative stress as GSH is lost. Transport of 13C-cysteine into the
lens from a 13C-cysteine enriched culture medium and its incorporation into the cysteine
residue of endogenous GSH were observed. Subsequent exposure of the lens to increasing
concentration of an oxidant resulted in a proportional loss in 13C_NMR resonance intensity
in GSH. The I3C-NMR signal from GSSG only appeared when a high concentration of the
oxidant was used. The lost I3C-NMR signal in GSH and the delayed appearance of I3C_
GSSG implies that PSSG formation from GSH (as GSSG) may have taken place (NMR
can not detect bound GSH).
The source for PSSC is not as easily explained. If a cysteine pool is present in the
lens, it would be a reasonable source for the bound cysteine in PSSC. However, there is
limited work on this subject in the past. In 1962, Reddy and Kinsey 27 reported that only a
trace amount of cysteine could be detected in the rabbit lens. To clarify the possible pres-
ence of a cysteine pool in the lens, we have developed a sensitive technique for this pur-
pose, following the method of Fahey and Newton. 28 This technique employs an HPLC
system equipped with a fluorescent detector to monitor the thiols after they have been
derivatized by monobromobimane. Because of its high sensitivity (at the picomole range),
we have successfully quantified the cysteine pools in rae 9 and human lenses. 3o Free cyste-
ine is present in the lens at nanomole level, only 3-5% that of free GSH. It too has a con-
centration gradient but is opposite to that of GSH. The concentration is highest in the
epithelium but it is preferentially accumulated in the nucleus over the cortex. This distri-
bution pattern may help to explain the higher abundance of PSSC in the nucleus as men-
tioned earlier. Further work is needed to clarify the source for PSSc.
~5oor-----------------------------------~
_ ~4oo r o GSO,H
~
0'
I
o_______··~~Oz
i C>3OQ
/... ~o
a;
~
Table 6. Effect ofHP2 and xylose on the free GSH and the GSH- and
cysteine-protein mixed disulfide levels in cultured rat lens
Protein-thiol mixed disulfides*t
Free aSH* aS03 H Cysteic acid
Normal 3.766±0.193 0.0142 0.099
Xylose (30 mM) 2.582 ± 0.419 0.0200 0.1197
H,o, (0.5 mM) 1.891 ± 0.444 0.2700 0.1213
H,o, + xylose 1.316 ± 0.448 0.2714 0.1219
'Wet wt of lens ~lmol g.l; values represent mean ± S.D. of the four individual lenses (free GSH).
GS03 H represents the released protein bound GSH, cysteic acid represents the released protein bound
cysteine.
t Analysis data obtained from four pooled lenses
Table 7. Effect of the diabetic condition on the free GSH and the protein-GSH
and protein-cysteine mixed disulfide levels in rat lens
the PSSG or PSSC did not change at all at the early stage, in agreement with galactosemic
cataracts as observed by Reddy and Han,15 however a drastic loss in both PSSG and PSSC
was observed at a more advanced stage (6 weeks). This loss may be attributed to general pro-
tein loss due to the leaky membrane since it can be prevented when the diabetic rats are
treated with an ARl. 31
the cells may use the transient formation of protein-thiol mixed disulfide initially as a
means to protect protein SH groups. This modification may be removed quickly when the
natural defense systems against oxidative stress are still fully functioning. However, if the
stress becomes prolonged or in excess, the weakened defense system could result in per-
manant modification of the lens proteins. This thiolation process would introduce extra
charges and cause conformational changes to the proteins potentially rendering otherwise
buried functional groups to be exposed and capable of being modified. This process would
lead to further interaction between cytoplasmic proteins or between intracellular proteins
and membrane proteins to form additional disulfide bridges. Glycation and other covalent
crosslinking may also occur further potentiating the formation of HMW aggregates and
the cascading events of protein insolubilization, light scattering and cataract formation.
This scenario is feasible since Liang and Pelletier38 have demonstrated that the thiolation
process indeed led to conformation destablization and increased proteolysis when they
compared the native y-crystallin with y-crystallin-GSH mixed disulfide. These findings
were subsequently confirmed by Kono and Chakrabarti. 39 A similar hypothesis on the role
of protein-thiol mixed disufides in the oxidative damage to proteins from other tissues has
been suggested independently. 17 Further research is needed before a true understanding of
the significance, both physiological and pathiological, of protein-thiol mixed disulfide for-
mation in the lens can be attained.
2.5
~
~
~ 2.0
;0:
E 1.5
~
'"
"-
(; 1.0
J, 0.5
J:
U1
to 0.0
~ 0.15
OJ
;0:
~ 0.10
~'"
J, 0.05
0
J:
0'"
U1 0.00
()
cortex nucleus total
~ 0.8
~
OJ Figure 8. Effect of HP2 (0.5 mM) exposure on the hu-
E '" 0.6 man lens GSH, PSSG and PSSC in organ culture. A,
~ GSH; 8, CS03H; C, GS03H. The lens was dissected
'" 0.4
"-
(; into cortex and nucleus for analysis. Nucleus represents
E 25-30% lens wet weight with the balance as cortex.
2- 0.2
J: Data are the results ofthree separate experiments: error
0
U1
'" 0.0
bars are I S.D. (D) control lens in TC 199 media; (_)
to contralateral lens, in 0.5 mM Hp2-containingmedia.
Protein-Thiol Mixed Disulfides and Thioltransferase in the Lens 41
40
_ Control
!888l 24 hr. H202
B 48 hr. H202
3D
til
III
;:l 20
til
>-
10
Figure 9. Comparison of lens PSSG and PSSP levels after 48 hrs-exposure to Hp,stress. One-month old rat
lenses were exposed to 0.5 mM HP, for 24 and 48 hrs. PSSP crosslinks in the water soluble and urea soluble pro-
teins on the SDS-PAGE gels were quantified by using densitometry. PSSG was measured as the released GS0 3 H.
Control represents an average value from 24 and 48 hrs incubations.
42 M. F. Lou
5
......
.... 0 Control A
I!=
:>-
4 e H202
Recovery
'"
'tI
0
-El
till :I
"-
Q)
0
2
::I
"-'
::r:
t')
0
rn
t.!l 0
0 24 48 72 96
2.0
...... 0 Control B
~ e H202
1.5
:>- Cl Recovery
'"
'tI
-El
"-Q)
::I
1.0
"-'
::r:
t') 0.5
0
rn
u
0.0
0 24 4B 72 96
Figure 10. Changes of lens protein-thiol mixed disulfides and the recovery during long term incubation with
HP2' A. protein-GSH (measured as glutathione sulfonic acid, GS0 3H); B. protein-cysteine (measured as cysteic
acid, CS0 3 H). Five lenses of the same group were pooled and used for the analysis. Data is expressed as J..lmol g
dry wt', mean ± S.D., n = 5 (recovery n = 3). Please note that the S.D. values of the control group and the recovery
group in Figure 4 (A) are too small to show in the plot. (0) Control, (e) HP2' (0) recovery.
Protein-Thiol Mixed Disulfides and Thioltransferase in the Lens 43
A B
87 . 0
44 . 1
32 . 7
1 2 3 4
- 17 . 7
1 2 3 4
"igure II. Western Blot analysis of lens thioltransferase. (A). Ten micrograms of pig liver thioltransferase
(TTase) (lane 4) and 25 fig of lens TTase (lane 3) was each reacted with anti pig lever TTase (I :500 dilution). Lane
one represents the prestained standard protein mixture. Lane 2 represents the crude extract of lens homogenate
(non purified lens preparation). (B). Ten micrograms of pig liver TTase (lane 4) and 25 fig of lens TTase (lane 3)
was each reacted with antibody made against partially purified lens TTase (I: 100 dilution). Lane I represents ka-
leidoscope markers and lane 2 represents the crude extract of lens homogenate (non-purified lens preparation).
The arrow in each picture denotes the TTase protein.
late that it may be attributed to a certain enzyme which can dethiolate the protein-thiol
mixed disulfides to maintain the redox status of the lens.
c
o
E
1 2
Figure 12. Slot Blot hybridization of whole lens and rabbit lens epithelial cell RNA with the pig liver eDNA for
TTase. Samples of 5 fig RNA from (A) (B) (e) (D) bovine lens RNA; (E) blank without RNA; (F) rabbit epithelial
cell RNA (cell line NIN I 003) were applied to the wells in duplication I and 2 and hybridized with 32 p labeled eDNA.
44 M. F. Lou
shown that the lens from different animal species including humans all contains the thioltrans-
ferase activity (Table 8). We suggest that the lens thioltransferase has a significant physiologi-
cal role in sulfbydryl homeostasis in the lens by protecting the SH groups of the proteins from
S-thiolation. This would explain the low level ofprotein-thiol mixed disulfides in the normal
lens as well as the dethiolation process of these oxidant pre-exposed lenses under recovery. It
is therefore speculated that, lens thioltransferase may be a primary antioxidant in the lens
along with GSH and glutathione reductase in protecting the vulnerable lens proteins against
oxidative damage. Furthermore, PSSG is a preferred substrate over PSSC for this enzyme in
other tissues 50 which would explain the preferential and effective dethiolation of PSSG in the
recovery experiments described above. Studies are underway to purify this enzyme and to
study its physiological role in the lens. Work is also in progress to clone the cDNA of the lens
thioltransferase for use in future research.
7. ACKNOWLEDGMENTS
The author wishes to express her sincere appreciation to Keith Green and J. Sam
Zigler, Jr. for reading this manuscript. The superb technical assistances and stimulating
discussions provided at various stages of this research by Oliver Chyan, Robert McKellar,
Jaime E. Dickerson, Jr., Rekha Garadi, Sally Sheib, Xiao-Lan Cui, Guo-Tong Xu, Nalini
Raghavachari and Guo-Ming Wang are greately appreciated. This research was supported
by Alcon Laboratories and National Institute of Health.
8. REFERENCES
I. Delaye M. and Tardieu A. Short-range order of crystallin proteins accounts for eye lens transparency. Na-
ture (London) 1983; 302: 415-17.
2. Spector A. The search for a solution to senile cataract. Invest Ophthalmol Vis Sci 1984; 25: 130-46.
3. Spector A. Oxidative stress-induced cataract: mechanism of action. FASEB J 1995; 9: 1173-82.
4. Harding JJ. Free and protein-bound glutathione in normal and cataractous human lenses. Biochem J 1970;
117: 957--60.
5. Truscott RJW and Augusteyn RC. The state of sulphhydryl groups in normal and cataractous human lenses.
Exp Eye Res 1997; 25: 139-48.
6. Spector A and Roy D. Disulfide-linked high molecular weight protein associated with human cataract. Proc
Natl Acad Sci USA 1978; 75:3244-8.
7. Jcdziniak JA, Kinoshita JH. Yates EM and Benedek GB. On the presence and formation of heavy molecu-
lar weight aggregates in human normal and cataractous lenses. Exp Eye Res 1973; 15: 245-52.
Protein-Thiol Mixed Disulfides and Thioltransferase in the Lens 45
8. Ortwerth BJ and Olesen PRo Studies on the nature of the water-insoluble fraction from aged bovine lenses.
Exp Eye Res 1989; 48: 605.
9. Benedek GB. Theory of transparency of the eye. Appl Optics 1971; 10: 459--73.
10. Stevens VJ, Rouzer CA, Monnier VM and Cerami A. Diabetic cataract formation: Potential role of glyco-
sylation fo lens crystallins. Proc Natl Acad Aci USA 1978;75: 2918--22.
II. Ortwerth BJ and Olesen PRo Ascorbic acid-induced crosslinking of lens proteins: evidence supporting a
maillard reaction. Biochim Biophys Acta 1988; 956: I 0-22.
12. Nagaraj RH, Sell DR, Prabhakaram M, Ortwerth BJ and Monnier VM. High correlation between pentosid-
ine protein crosslinks and pigmentation implicates ascorbate oxidation in human lens senescence and catar-
actogenesis. Proc Nat! Acad Sci USA 1991; 88:10257-61.
13. Garland D, Zigler JS Jr. and Kinoshita JH. Structural changes in bovine lens crystallins induced by ascor-
bate, metal and oxygen. Arch Biochem Biophys 1986; 251: 771-6.
14. Anderson E and Spector A. The state of sulfhydryl groups in normal and cataractous human lens proteins
inuclear region. Exp Eye Res 1978; 26: 407-17.
15. Reddy VN and Han RF. Protein-bound glutathione in mammalian lenses and in galactose cataract. Doc
Ophthalmol Proc Ser 1976; 8: 153.
16. Lou MF. Mckellar Rand Chyan O. Quantitation of lens protein mixed disulfides by ion-exchange chroma-
tography. Exp Eye Res 1986; 42: 607-16.
17. Brigelius R. Mixed disulfides:Biological functions and increase in oxidative stress. In: Sies H, ed. Oxida-
tive Stress, New York, 243-72,1985, Academic Press.
18. Inoue M. Dynamic aspect of protein mixed diusulfide formation. In: Dolphin D, Poulson Rand Avramovic
0, eds. Glutathione: Chemical, Biochemical and Medical Aspects, New York, 613-44,1989, Wiley.
19. Hamilton PB. Ion exchange chromatography of amino acidsa single column. high resolving, fully automat-
ic procedure. Anal Chern 1963; 35: 2055-64.
20. Lou MF, Dickerson Jr., JE and Garadi R. The role of protein-thiol mixed disulfides in cataractogenesis.
Exp Eye Res 1990; 50: 819--26.
21. Moore S. On the determination of cystine as cysteic acid. J Bioi Chern 1963; 238: 235--7.
22. Hirs CHW. The oxidation of ribonuclease with performic acid. J BioI Chern 1956; 219: 611-21.
23. Dickerson Jr., JE and Lou MF. A new mixed disulfide species in human cataractous and aged lenses. Bio-
chim Biophys Acta 1993; 1157: 141-6.
24. Giblin F, Padgaonkar VA, Leverenz VR, Lin L-R, Lou MF, Unakar NJ, Dang L. Dickerson Jr.• JE and
Reddy VN. Nuclear light scattering, disulfide formation and membrane damage in lenses of older guinea
pigs treated with hyperbaric oxygen. Exp Eye Res 1995; 60: 219--35.
25. Lou MF and Dickerson Jr., JE. Protein-thiol mixed disulfides in human lens. Exp Eye Res 1992; 55:
889--96.
26. Willis A and Schleich T. l3C NMR spectroscopic measurement of glutathione synthesis and antioxidant
metabolism in the intact ocular lens. Biochem Biophys Res Commun 1992; 186: 931-5.
27. Reddy DVN and Kinsey VE. Studies on the crystalline lens IX Quantitative analysis of free amino aicds
and related compounds. Invest Ophthalmol 1962; 1: 635--41.
28. Fahey RC and Newton GL. Determination of low-molecular-weight thiols using monobromobimane fluo-
rescent labeling and high-performance liquid chromatography. Meth Enzymol 1987; 143: 85--96.
29. Veltman J and Lou MF. Quantitation of free cysteine in the lens. Invest. Ophthalmol Vis Sci 1993; 34
(Supp!.): 758.
30. Lou MF and Dickerson Jr., JE. Quantitation of free cysteine in normal and cataractous human lenses. In-
vest Ophthalmol Vis Sci 1995; 36: S885.
31. Lou MF. Quantitation of protein mixed disulfides in the lens. US-CCRG abstract 1985; Honolulu, HI.
32. Xu GT, Zigler Jr., JS and Lou MF. The possible mechanism of naphthalene cataract in rat and its preven-
tion by an aldose reductase inhibitor (ALOI576). Exp Eye Res 1992; 54: 63-72.
33. Zigman S, Paxhia T, McDaniel T, Lou MF and Yu N-T. Effect of chronic near- ultraviolet radiation on the
gray squirrel lens in vivo. Invest Ophthalmol Vis Sci 1991; 32: 1723-32.
34. Lou MF and Zigler Jr., JS. The effect of photooxidation of protein mixed disulfide in lens. Proc. ISER
1986; IV. 87.
35. Yu LM. Bhuyan DK, Lou MF, Dickerson Jr., JE and Bhuyan KC. Redox-active thiols of rabbit lens as al-
tered by diquat in vivo . Invest Ophthalmol Vis Sci 1991; 32 (Supp!.):749.
36. Bhuyan DK, Lou MF, Wang G-M, Kuriakose G, Game.r WH and Bhuyan KC. Mixed disulfide crosslinking
of proteins and Hp21evel in emory mouse cataract Invest Ophthalmol Vis Sci 1996; 37 (Supp!.): S559.
37. Lou MF, Dickerson Jr., JE, Wolfe JK,Tung Band Chylack Jr., L. Correlation ofprotein-thiol mixed disul-
fide level and the pigmentation in human lens nucleus. Invest Ophthalmol Vis Sci 1993; 34 (Suppl.): 987.
46 M. F. Lou
38. Liang IN and Pelletier MR. Destabilization of lens protein conformation by glutathione mixed disulfide.
Exp Eye Res 1988; 47: 17-25.
39. Kono M and Chakrabarti B. Changes in the conformation stability of gamma-crystallins upon glutathione
reaction. Biochemistry 1990; 29: 464--70.
40. Spector A. and Gamer WHo Hydrogen peroxide and human cataract. Exp Eye Res 1985; 33: 673--81.
41. Huang QL and Hu TS. Aqueous hydrogen peroxide and senile cataract. Invest Ophthalmol Vis Sci 1990;
31: 350.
42. Bhuyan KC and Bhuyan OK. Regulation of hydrogen peroxide in eye humor, effect of3-amino-I-H-I,2,4-
triazole on catalase and glutathione peroxidase of rabbit eye. Biochim Biophys Acta 1997; 497: 641-51.
43. Cui X-L and Lou MF. The effect and recovery of long term H,G, exposure on rat lens morphology and bio-
chemistry. Exp Eye Res 1993; 57: 157-67.
44. Lou MF, Xu G-T and Cui X-L. Further studies on the dynamic changes of glutathione and protein-thiol
mixed disulfides in H,O, induced cataract in rat lenses: Distributions and effect of aging. CUIT Eye Res
1995; 14: 951--8.
45. Lou MF and Xu G-T. Recovery of oxidative damage in human lenses. Invest Ophthalmol Vis Sci 1994; 35
(Supp1.): 1569.
46. Gan ZR and Wells WW. The primary structure of pig liver thioltransferase. J BioI Chern 1987; 262:
6699-705.
47. Padilla A, Martinez-Galisteo E, Antonio-Barcena J, Spyrou G and Holmgren A. Purification from placenta,
amino acid sequence, structure comparisons and cDNA cloning of human glutaredoxin. Eur J Biochem
1995; 227: 27-34.
48. Mieyal JJ, Starke OW, Gravina SA and Hocevar BA. Thioltransferase in human red blood cells: Purifica-
tion and properties. Biochemistry 1991 ;30: 6088-97.
49. Raghavachari N and Lou MF. Evidences for the presence ofthioltransferase in the lens.Exp Eye Res 1996;
63:433--41.
50. Gravina SA and Mieyal JJ. Thioltransferase is a specific glutathionyl mixed disulfide oxidoreductase. Bio-
chemistry 1993; 32: 3368-76.
4
1. INTRODUCTION
Subjects were young adult male and female beagle dogs received form Marshall
Farms USA Inc, North Rose, New York; dogs were housed individually, fed Agway Pro-
lab R Canine 1600 (Agway Inc. St. Mary's, Ohio) and were allowed tap water ad libitum.
All dogs were acclimated to the laboratory environment (71° +/- 2° F; 50% +/- 10% rela-
tive humidity; 10 to 15 air changes per hour in a fixed 12 hour photo period) for a mini-
mum of 55 days prior to dosing.
Two studies were conducted with different dosing regimens. The first experiment was a
cyclic dosing study, in which five dogs/sex/dose were dosed orally once daily with the follow-
ing schedule: 5-EU, 3 mglkg, alone for one day; then 5-EU at the same dosage followed about
30 minutes later by 5-FU for 5 days; then 5-EU alone for one day; then no treatment for 21 to
23 days, to allow recovery from treatment related effects. This dose schedule was repeated
three more times. 5-FU doses were 0, 0.3, 0.6, and 0.9 mg/ kg. Two to three days after the last
cycle of treatment, two dogs/sex/dose were euthanized for necropsy; remaining dogs were al-
lowed to recover for another 25 days, then euthanized and necropsied.
The second experiment was a 3-month study, in which 5 dogs/sex/dose were dosed
orally once daily for three months. 5-EU and 5-FU doses for each group were 0.0/0.0;
3.0/0.0; 0.6/0.1; 0.6/0.2; and 0.6/0.4 mg/ kg respectively. After three months, three
dogs/sex/dose were euthanized and necropsied; remaining dogs were allowed to recover
for two months, then euthanized and necropsied.
Observations and measurements included regular evaluation of clinical signs; body
weight; food consumption; clinical pathology; urinalysis; drug plasma concentration de-
termination; electrocardiographic examination; and ophthalmic examination.
In the cyclical study, ophthalmic examinations were performed five days prior to the
initiation of dosing and 18 to 19 days after termination of dosing for each cycle. In the 3-
month study, ophthalmic examinations were performed prior to initiation of dosing, on
days 27, 41, 63, and 85 during dosing, and 26 and 61 days after dosing stopped.
Ophthalmic examinations consisted of dilation of the pupils with topical 1.0% tropicamide,
followed by biomicroscopic examination of the anterior segment and indirect ophthalmoscopy of
the fundus. The lesions were documented by biomicroscopic photography.
At necropsy, eyes were removed, immersion-fixed in Bouin's solution, and proc-
essed routinely with paraffin embedding and hematoxylin and eosin staining for light mi-
croscopic examination.
3. RESULTS
In the 3-month study, all high-dose dogs were euthanized for humane reasons be-
cause of treatment-related effects. In general, clinical signs were similar to those observed
with the cyclic dosing schedule with a clear dose response that persisted until the end of
the recovery period.
Figure 1. A. Group 3 female, left eye, Study Day 27 of the continuous dosing experiment. Within the temporal
palpebral fissure. brown pigmentation of the perilimbal conjunctiva (black arrow) and black-brown pigmentation
of the adjacent cornea (white arrow head) was noted. B. Continuous dosing experiment, Study Day 85 , group 3 fe-
male: A swirling brown pigmentation of the corneal epithelium was noted, with adjacent epithelial edema. C.
Continuous dosing experiment. Study Day 61. Group 4 female: Superficial corneal neovascularization was pre-
sent. D. Continuous dosing experiment, Study Day 61. Group 4 female: The corneal changes in this animal were
accompanied by a lipid degeneration, with a disposition of subepithelial cholesterol in a curvilinear pattern.
50 R. L. Peiffer, Jr., and J. E. Dillberger
Figure 2. Pigmentary changes include hyperplasia of subepithelial and epithelial melanocytes, with melanin gran-
ules present within all layers of epithelial cells of the conjunctiva (a) and cornea (b). Hematoxylin and eosin, origi-
nal magnification 400X.
Corneal Lesions in Beagle Dogs Given Oral 5-Ethynyluracil 51
3.4. Pathology
Histopathologic changes were identical in both studies, dose-related, and correlated
with the clinical findings. The hyperpigmentation of the conjunctiva was due to a mild hy-
perplasia of subepithelial and epithelial melanocytes, with an increase in intracytoplasmic
melanin in all layers of the epithelial cells. As the pigmentation progressed onto the cor-
nea, pigmentation was limited to melanin granules within all epithelial cells (Figure 2).
The corneal epithelium was characterized by irregularity in thickness and loss of
normal polarity. The axial epithelium was normal or 1 to 2 cell layers thicker than normal;
the peripheral epithelium was markedly thinned to 1 to 3 layers, with loss of basal cell dif-
ferentiation (Figure 3). The underlying stroma was characterized by mild scarring and
Figure 3. Epithelial changes included thinning of the corneal epithelium adjacent to the limbus. with axial thin-
ning. (a) hematoxylin and eosin, original magnification 40X; (b) the axial epithelium was reduced to as thin as I to
2 cell layers, while axially, the epithelium was somewhat thickened with loss of polarity (c). Original magnifica-
tion (a) 40X. (b and c) 400X.
52 R. L. Peiffer, Jr., and J. E. Dillberger
Figure 4. An occasional dog demonstrated epithelial edema with loss of polarity and a vascular keratitis with both
acute and inflammatory cell infiltrate. Hematoxylin and eosin. original magnification 200X.
edema, vascularization from the limbus to the depths of the mid stroma, and in an occa-
sional eye, acute and chronic inflammatory cell infiltration (Figure 4). The globes were
otherwise unremarkable.
5-Ethynyluracil reduced the toxicity and steepened the dose-response curve, and 5-
EU usage prevented dose-related neurotoxicity, allowing greater 5-FU exposure.
4. DISCUSSION
The conjunctival and corneal changes seen in dogs administered 5-EU followed by 5-
FU in both cyclical and continuous dosing are species specific; no ocular lesions were ob-
served in mice administered equivalent mg/kg doses of the combination, and ocular lesions
were not seen in rodents or dogs treated with 5-EU or 5-FU alone at similar mg/kg dosage.
One might speculate possible mechanisms for this species, specifically including differences
in epithelial cell turnover rates; greater concentration of the drugs in canine tears; or species
differences in tissue concentration of metabolic enzymes. The clinical and histopathologic
findings suggest a direct effect on the surface epithelium of the conjunctiva and cornea and
the resident melanocyte population with stimulation of the latter and disruption of matura-
tional processes in the former, as well as the elaboration of va so genic stimuli by corneal epi-
thelium and/or stroma. Regrettably, our data limit mechanistic discussion to speculation.
The drug combination has the potential for enhancing the usefulness of 5-FU as an
anticancer agent and clinical studies are underway. Our observations in dogs justify mak-
ing ophthalmic examinations a component of the clinical study and we anticipate learning
if the human eye is different in it sensitivity to this drug combination compared to the dog.
5. REFERENCES
I. Bonadonna A, Brusamolino E, Peruccia V et al. Combination chemotherapy as an adjutant treatment in op-
erable breast cancer. N Eng J Med 1976;294:405-410.
Corneal Lesions in Beagle Dogs Given Oral 5-Ethynyluracil 53
2. Fraunfelder FT, Meyer MS. Ocular toxicity of anti-neoplastic agents. Ophthalmology 1983;90: 1-3.
3. Insler MS, Helms CJ. Ankyloblepharon associated with systemic 5-Fluorouracil treatment. Ann Ophthal-
mol 1987;19:374-375.
4. Caravella LP Jr, Burns A, Zangmeister M. Punctal-canalicular stenosis related to systemic fluorouracil
therapy. Arch Ophthalmol 1981 ;99:284-286.
5. Heuer DK, Gressel MG, Parrish RK II et al. Topical fluorouracil. Postoperative administration in an animal
model of glaucoma filtering surgery. Arch Ophthalmol 1986; I 04: 132-136.
6. Spector T, Porter DJT, Nelson DJ et al. 5-Ethynyluracil (776C85), a modulator of the therapeutic activity
of 5-tluorouracil. Drugs of the Future 1994; 19:565--571.
7. Shapiro MS, Thoft RA, Friend J et al. 5-Fluorouracil toxicity to the ocular surface epithelium. Invest
Ophthalmol Vis Sci 1985;26:580-583.
5
1. INTRODUCTION
Preclinical safety studies by the intravenous route, via an indwelling catheter, are
performed on a wide range of pharmaceutical articles in order to mimic the proposed clini-
cal route. We have observed an increased incidence and severity of corneal lesions in in-
travenously infused rats in such studies. In particular, linear medial (nasal) opacities are
commonly seen. Since the origin of this increased ocular damage remains unknown, we
performed a study to determine if it was related to anaesthesia, infusion material or infu-
sion method. In addition, we tested the efficacy of a potential preventative measure during
the study.
Ninety Sprague-Dawley rats (Crl CD® (SD) BR, Caesarian Obtained, Barrier Sus-
tained-Virus Antibody Free (COBS- V AF®» were allocated to 5 groups as shown in the
table below.
At the beginning of the study, the animals were approximately 8 weeks old.
Group I animals were absolute controls and did not receive anaesthesia or sub-
sequent manipulations. Group 2 animals were anaesthetized for approximately 60 minutes
using an association of acepromazine (Vetranquil®, Sanofi Sante animale) and ketamine
hydrochloride (ImalgEme®, RhOne-Merieux). Group 3 animals were anaesthetized in the
same way and were fitted with the infusion jacket but did not receive any surgical inter-
vention. Group 4 animals were anaesthetized and surgically prepared for continuous intra-
venous infusion by location of an indwelling catheter in the vena cava caudalis through
one of the femoral veins. After location of the catheter, group 4 animals were continu-
ously infused (24 h/24 h) with sterile isotonic saline, at a rate of 1 ml/kg/hour. Group 5
animals were prepared in the same way as group 4 animals; in addition they received 5 in-
stillations daily of eye drops (Lacrypos®, Alcon Laboratories) on the day of surgery (be-
fore surgery, and immediately, 30, 60, and 120 minutes afterwards) and the following day
(regularly during the day).
The study extended over a period of 13 weeks. Gross examinations, indirect ophthal-
moscopy and slit-lamp examinations were performed before surgery, weekly for the first
four weeks and every 2 weeks thereafter (i.e., in weeks -1, 1,2,3,4,6,8, IO and 12).
Group 1M IF 2M 2F 3M 3F 4M 4F 5M 5F
OPO 30% 33.3% 20% 40% 0% 0% 45% 10% 30% 10%
lAC 30% 16.7% 60% 40% 50% 30% 25% 40% 30% 20%
OCO 20% 0% 10% 0% 30% 20% 15% 5% 0% 0%
LMO 0% 0% 0% 0% 0% 10% 10% 25% 10% 10%
CNY 0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
ICT 0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
Group 1M IF 2M 2F 3M 3F 4M 4F 5M 5F
OPO 30% 20% 20% 40% 10% 10% 24% 10% 20% 10%
lAC 0% 0% 60% 20% 50% 30% 12% 28% 20% 10%
OCO 0% 0% 0% 0% 0% 10% 8% 2% 0% 0%
LMO 0% 0% 10% 10% 10% 30% 10% 16% 10% 10%
CNY 0% 0% 0% 0% 0% 0% 2% 0% 0% 0%
ICT 0% 0% 0% 0% 0% 0% 0% 0% 0% 0%
Group 1M IF 2M 2F 3M 3F 4M 4F 5M 5F
OPO 30% 20% 10% 20% 20% 10% 50% 45% 20% 20%
lAC 0% 10% 40% 30% 50% 10% 30% 20% 20% 0%
OCO 0% 0% 0% 0% 30% 10% 10% 5% 0% 0%
LMO 0% 0% 30% 20% 10% 30% 35% 45% 10% 0%
CNY 0% 0% 0% 0% 10% 0% 10% 0% 0% 0%
ICT 0% 0% 0% 0% 10% 0% 10% 0% 0% 0%
Group 1M IF 2M 2F 3M 3F 4M 4F 5M 5F
OPO 30% 20% 0% 20% 10% 10% 35% 25% 70% 20%
lAC 0% 0% 70% 20% 50% 10% 22% 20% 0% 0%
OCO 0% 0% 0% 0% 0% 10% 5% 0% 0% 0%
LMO 0% 0% 30% 20% 50% 30% 15% 28% 0% 0%
CNY 0% 0% 0% 0% 10% 0% 0% 0% 0% 0%
ICT 0% 0% 0% 0% 10% 0% 0% 0% 0% 0%
stromal and epithelial layers, with corneal thickening in a number of animals, were seen
from week I in group 4 and from week 2 in groups 2 and 3. These corneal changes oc-
curred with a similar incidence in groups 2 to 4 and were sometimes associated with cor-
neal neovascularization (CNV) in group 3 males from week 3 and in group 4 males from
week 2.
Severe opacities may progress to CNV accompanied by inflammatory changes; in
this study CNV was particularly associated with LMO and was seen in some males of
group 3 (from week 3) and group 4 (from week 2). Local increases in corneal thickness
(lCT) were observed either accompanying or following CNV.
Severe corneal damage (LMO, CNV and ICT) was usually medially located and the
temporal upper and lower peripheries generally remained unaffected.
Other lesions (OCO) related to severe corneal damage, such as dry cornea (ORC),
lacrimation (LCR), blepharitis (8PT) and chromodacryorrhea (CDR) were only observed
in groups 3 and 4.
Males were more frequently and more severely affected than females. Reversibility was
rarely observed (except in group 5 and for OPO and lAC). Corneal damage generally does not
recover after reaching the level ofCNV, although we noted that corneal vascularization which
was observed in one group 3 male from week 4 to week 10, disappeared in week 12.
2IJI'
'8'
,n
,....
I~
,OtI
8'
....
n
~
0tI
~ ~ ; , .. N [
w.u
III I ~
2IJI'
'5"10
lOtI
5"10
a..
- ...-
~
I &i i i i ; ..
~ II:!
~ ~ N [w.u
~
i I lis
i
~
tonic saline (two instillations during anaesthesia) had been unsuccessful, we decided to
perform five daily instillations of eye drops on the day of surgery and the following day.
For this we used chondroitin sulphate eye drops which are useful for hypolacrimation and
corneal dryness. As seen above, this approach was successful in reducing the incidence
and severity ofLMO and eliminating other lesions (CNY, DRC, LCR, BPT and CDR).
-- -
-,.........
a.... _ _
..........
Photo 3. Cornea: Note sinuslike dilatation probably of vascular origin lined with flattened epith,
62 O. Loget et al.
4. CONCLUSIONS
The typical corneal opacity in anaesthetized animals consists of a line most often lo-
calized in the medial region of the palbebral fissure. This opacity was observed only after
anaesthesia and may therefore be either a direct (palpebral fissure staying opened resulting
in the absence of corneal wetting by tears)7 or indirect (lack of tear secretion) effect of
ketamine/acepromazine.
The slight increase in frequency of lAC and the severity of the other corneal lesions
noted among group 2 to 4 animals may well be related to an exacerbation of normal back-
ground lesions '--{) as a result of the anaesthetic used, possibly in conjunction with the im-
plantation and infusion procedure in group 4. Interestingly, corneal lesions were mainly
observed in groups 2 to 4 suggesting an anaesthesia-related effece which could be attenu-
ated by regular instillation of Lacrypos® eye drops.
Since the presence of anaesthesia-related lesions is now well established, compari-
sons with historical data should take account of this fact, and historical data should be
based on infused animals. Even using such data, the presence of anaesthesia-related le-
sions may mask test-compound induced lesions, and consideration should be given to the
routine use of preventative measures such as those described above.
5. REFERENCES
I. Bellorn R, Korte GE, Abrutyn D. Spontaneous corneal degeneration in the rat. Lab Anim Sci
1988;38:46-50.
2. Fabian RJ, Bond JM, Drobeck HP. Induced corneal opacities in the rat. Br J Ophthalmol 1967;51: 124-129.
3. Losco PE, Troup CM. Corneal dystrophy in Fisher 344 rats. Lab Anim Sci 1988;38:702-710.
4. Taradach C, Regnier B, and Perraud J. Eye lesions in Sprague-Dawley rats: type and incidence in relation
to age. Lab Anim 1981; I 5:285-287.
5. Wegener A, Jochims K. Clinical, histological and ultrastructural characteristics of a spontaneous corneal
opacity in Sprague Dawley rats. Ophthalmic Res 1994;26:296-303.
6. Weisse I, Kreuzer H, Stender E, Frolke W, Meyer D. Band keratopathy in rats due to increased dietary con-
tent of vitamin D3. Concepts Toxicol 1987;4: 164-178.
7. Anderson DA, Braun TW. Eye injury during general anaesthesia for oral and maxillofacial surgery: etiol-
ogy and prevention. J Oral MaxiIlofac Surg 1995:53:321-324.
6
!Ciba Vision
Barcelona, Spain
2Department of Pharmacology and Therapeutics
University of Alicante, Spain
3Department of Optic
School of Optics
University of Alicante, Spain
ABSTRACT
The variations of the spectral irradial of sunlight and the increasing use of artificial
sources of light that emits UV -radiation, have increased the incidence of nocive effects of
UV -radiation. Preliminary studies demonstrated the photochemical damage to ocular tis-
sues caused by solar radiation and other UV sources. Since ocular tissues do not develop
tolerance to UV exposure, excessive exposure of the ocular surface might be nocive to the
eye. Corneal epithelium is the external layer of the eye directly exposed to the UV radia-
tion of the sun. To determine the effect of UV radiation on corneal epithelium, rabbit cor-
neal epithelial cells in culture were irradiated during 6 days at 370 or 525 nm with
different intensities and the role of an absorption filter was also investigated.
1. INTRODUCTION
The principal source of electromagnetic radiation is the sun, and the near ultraviolet
(UV from 300 nm to 400 nm) is the more important component of solar ultraviolet. The
atmosphere acts like a filter, since ozone absorbs wavelengths below 300 nm, effectively
preventing the biologically injurious UVC-radiation, while the CO 2 and water absorb a
greater portion of IR-radiation. The variations of the spectral irradial of sunlight (reduc-
tion of the thickness of atmosphere or the ozone layer, pollution, etc.) and the increasing
use of artificial sources of light that emits UV-radiation, have increased the incidence of
nocive effects of UV -radiation. I Photochemical damage to ocular tissues caused by solar
radiation and other UV sources has been reviewed by several authors. 2 ,3
Ocular damage is produced by the UV-radiation absorbed in the tissue; and the se-
verity of the effects depends on the wavelength and photonic energy. UV-radiation con-
sists of short-wavelengths and high energy photons, and IR-radiation consists of
large-wavelengths and low-energy photons. Ultraviolet photons may lead to higher vibra-
tional states or electronic excitation and can break biochemical bonds; however, infrared
photons may alter lower-energy excitational levels of a molecule by affecting rotational
and vibrational states producing a thermic effect. I ,4
UV-C and UV-B radiation (wavelengths from 200 to 310 nm) are absorbed mainly
by the cornea and could modify the morphology of the cells. While UV-A radiation is ab-
sorbed by the cornea and the lens, and its biological effects are seen predominantly in both
tissues. 5- 7 Only I % of the ambient UV -radiation can penetrate to retina. 8
Excessive exposure of the ocular surface to UV -radiation may be dangerous to the
eye. A single exposure to significant levels of UV -radiation in animal and man elicits a re-
sponse, after a latent period of 6 to 12 hours, affecting all layers of cornea with concomi-
tant conjunctival hyperemia, lacrimation and anterior uveitis which is known as actinic
keratoconjuctivitis. 9,lo The acute symptoms usually last from 12 to 24 hours, and almost
all discomfort disappears within 48 hours. The corneal pain can be very severe, but rarely
does the exposure result in permanent damage. The latent period is due to decrease of cor-
neal sensitivity. I I Unlike the skin, the ocular system does not develop tolerance to re-
peated ultraviolet exposure. Chronic exposure or UV dose greater than the threshold for
photokeratitis produces swelling or shrinking of groups of corneal epithelial cells. The
surface of these cells show nuclear fragmentation and mid-epithelial cells show vacuole
formation and inhibition of mitosis in basal cells. 12- 15 It has been shown that UV -A radia-
tion can be lethal and mutagenic to mammalian cells. 16,17 The targets of near UV radiation
include DNA, RNA, proteins, cell membranes and respiratory electron transport sys-
tems. 18- 21 The contribution of a neurogenic mechanism to anterior segment inflammation
induced by UV exposure is modest. 22
In these experiments, we investigated the cytolitic action on cultured rabbit corneal
epithelial cells of a range of wavelength, radiation-time and intensity of UV -radiation and
the protective effect of UV -absorption filters.
2.1. Chemicals
Culture media and antiobiotics and all additional chemicals used were from Sigma
(St. Louis, Mo., USA). Culture materials were from Costar (Cambridge).
2.2. Animals
Corneas from male New Zealand albino rabbits were used. Animals were treated in
accordance with The Guiding Principles in the Care and Use of Animals. 23 Animals were
sacrificed by intravenous injection of sodium pentobarbital overdose, and the corneas
were circumferentially excised just within the limbus.
Ultraviolet Light-Induced Damage in Rabbit Corneal Epithelial Cells 65
Three hours post-subculture, the cells were exposed to irradiation. The light source
was a spectral xenon-mercury lamp HG 100-Phywe (Phywe Systeme GMBH,
Deutschland), with a wide spectrum. The color special filters used were ultraviolet and
green filters (Phywe) placed at 4 cm of the lamp. The ultraviolet filter has a spectrum of
absorption between 320 nm to 400 nm with a maximum of at Am= 370 nm and 28% of
transmission. The green filter has a spetrum of absorption between 480 nm to 570 nm with
a maximum at Am= 525 nm and 45% of transmission.
Monolayers of corneal epithelial cells monolayer in fresh medium-B were exposed
to the radiation directed perpendicular to culture plate, with control cultures covered by an
aluminum foil. The cell cultures were radiated for I or 2 hours/day during 6 days. The ra-
diation intensity ranged from 1.22 to 25 11 W /cm 2 • All exposures were performed with and
without protection absorption-filter: ESSILOR UVX (this filter has a total protection at
UV -B and UV -A radiation until 380 nm and 95% from 380 to 400 nm). Filters were
placed at 1 cm from the light source, between the lamp and color special filters. The le-
thality of the radiation was determined on the basis of the protein concentration (Bradford
method) 25 from I to 6 days treatment.
The experimental groups were assigned as follows:
Statistical analysis: Data were expressed as mean±S.D. Each group was compared
versus control using unpaired t-test. A value lower than 0.05 was used as a significance p
value.
66 M. Palmero et al.
3. RESULTS
A nocive effect of radiation on corneal epithelial cells was observed. When cells
were radiated with Am 370 nm for 2 hours daily at different intensities (1.2 and 2.0
f.L W/cm 2 ) a notorious decrease in cellular protein concentration was detected with respect
to control (non radiated cells). An augmentation of intensity did not represent a magnifica-
tion of cell death (Fig. lA). However, the increase in exposure time from 1 to 2 hours at
2.0 fJ.W/cm 2 showed an increase in the lethal effect (Fig IB).
To study the protective role of absorption filter on cellular death, the cells were radiated
in the presence and absence of this filter. The toxic effect observed when the radiation (Am
370 nm) was performed at 1.2 fJ.W/cm 2 during 1 hour daily for 6 days, was increased when
the radiation was made during 2 hours daily in absence of absorption filter. In both cases, the
presence of absorption filter (ESSILOR UVX) had an obvious protective effect (Fig. 2).
o Control
0.4
A B l.21lW/cm2
I!ll 2.01lW/cm2
o 24 48 72 96 120 144
o Control
0.4
B ri:iI I=lh
o 1=2h
'"
=
~
.
o
=-
o 24 48 72 96 120 144
Time (hours)
Figure 1. Effect of 370 nm radiation on the protein concentration of corneal epithelial cell culture. (A) Cells were
radiated for 2 hours/day at 1.2 and 2.0 f.lwl cm'.(B) Cells were radiated 2.0 f.lW/ cm' for I and 2 hours daily dur-
ing 6 days. «+) p<0.05 vs control; (*) p<0.05 I h vs 2h).
Ultraviolet Light-Induced Damage in Rabbit Corneal Epithelial Cells 67
When the corneal epithelial cells were radiated at 2.0 IlW/cm2 during 1 or 2 hours/day
the decrease of cellular protein concentration was more accentuated at 2 hours exposure. In
this case, the same protective effect of absorption filter was observed at I and 2 hours (Fig. 3).
In our experimental conditions, the effect of Am 525 nm radiation (visible light) us-
ing a very intense exposure schedule (25 11 W/cm2 during 2 hours/day) has also a nocive
effect on corneal epithelial cells. This effect, which was lower than that produced by the
Am 370 nm radiation, was clearly reduced by the coloured absorption filter (Fig. 4).
Pathologic signs of cytolysis were evidenced by scanning electron microscopy in ra-
diated cells (370 nm at 2.0 flW/cm 2/2 hours for 3 days) (Fig. 5). It was possible to notice
alterations of plasma membrane such as swelling or shrinking, and an increased number of
-
vacuoles in radiated cells.
Control
---0- NonUVX
A ---0-- UVX
0.4
0.3
e
Oil
!
'" 0.2
'O=J
=
...
"" 0.1
••
0.0
0 24 48 72 96 120 144
B
0.4
0.3
e
Oil
!
0.2
'"
=
]
...0
"" 0.1
** .* •• ** **
0.0
0 24 48 72 96 120 144
Time (hours)
Figure 2. Effect of 370 nm radiation on the protein concentration of corneal epithelial cell culture. Cells were ra-
diated at 1.2 J.L W/cm' for I hour (A) and 2 hours (B) per day during 6 days in presence/absence of absorption fil-
ters. Immediately after the exposure, the protein concentration was determined at specific times. •~. control;
0- -0 cclls radiated with absorption filter; O~O cells radiated without absorption filter. «*) p<0.05 vs con-
trol; (**) p<O.OI vs control).
68 M. Palmero et al.
4. DISCUSSION
Ultraviolet radiation is divided into UV-A(400 to 320 run), UV-B (320 to 290 run) and UV-
e (below 290 nm). This division is not phenomenologically exact, and although the spectrum for
cytotoxicity in cornea and lens of UV radiation is maximal between 320-290 run 2, 21, 26, 27 it has
been shown that near UV radiation can be lethal or mutagenic to mammalian cells. 16, 18, 19,28
In our experimental conditions radiation with 370 nm produced a nocive effect on cor-
neal epithelial cells in culture. A little nocive effect was also evidenced when the cells were ir-
radiated with Am = 525 nm (visible light spectrum), probably due to the very high radiant
exposure dose (25 IlW/cm2). The cell damage (determined by the reduction in proteins) in-
-- Control
--<>-- Non UVX
- - 0 - - UVX
A
0.4
0.3
0.2
0.1
••
O.O+--~--r-~-"T""'"~--"--~-"-~---r-~-'--
o 24 48 72 96 120 144
B
0.4
0.3
0.2
0.1
**
O.O+--~-.-~--r-~-r--~-.-~-.--~-'--
o 24 48 72 96 120 144
Time (hours)
Figure 3. Effect of370 nm radiation on the protein concentration of corneal epithelial cell culture. Cells werc ra-
diated at 2.0 fl W/cm 2 for I hour (Al and 2 hours (B) per day during 6 days in presence/absence of absorption fil-
ters. Immediately after the exposure, the protein concentration was determined at specific times. e--e control;
0 - 0 cells radiated with absorption filter; 0 - 0 cells radiated without absorption filter. ({*) p<0.05 vs con-
trol; (**) p<O.OI vs control).
Ultraviolet Light-Induced Damage in Rabbit Corneal Epithelial Cells 69
.e
!
0.3
0.2
'"
=
]
.
~
"-
0.1
0.0
0 24 48 72 96 120 144
Time (hours)
Figure 4. Effect of 525 nm radiation on the protein concentration of corneal epithelial cell culture. Cells were ra-
diated at 25 ~ W/cm' for 2 hours per day during 6 days in presence/absence of absorption filters. Immediately after
the exposure, the protein concentration was determined at specific times. e--e control; 0--0 cells radiated
with absorption filter; 0--0 cells radiated without absorption filter. «*) p<0.05 vs control; (**) p<O.OI vs con-
trol).
Figure 5. Scanning electron microscopy (JEOL JSM Scanning Microscopy). Scanning electron microscopy of
cultured epithelial cells non radiated (A and B) and radiated (C and OJ with A= 370 nm at 2.0 ~ W/cm'l2h for 3
days. Note the alteration of plasmatic membranes and an increased number of vacuoles in radiated cells.
70 M. Palmero et al.
duced by 370 nm radiation during 2 hours/day was only greater at 24 hours when comparing
2.0 11 W/cm2 of intensity to 1.22 IlW/cm2• However, the cell damage clearly increased with
the time of radiation when a similar intensity (2.011 W/cm2) was used. So, we have shown that
the lethal effect of near UV radiation on corneal epithelial cell culture depends on the spec-
trum used, and the relative intensity/time of exposure. In our experience the biggest nocive ef-
fect was observed at 370 nm with an intensity of2.0 11 W/cm2for 2 hours every day.
It is not possible to compare the quantitative damage obtained in this study with any
other cell type. In other studies, the UV sources employed provided different spectral dis-
tributions of energy, and the studies performed with UV-A on cultured rabbit lens epi-
thelial cells showed cell death in a dose dependent manner. 19, 28
The biochemical effects of UV radiation on corneal epithelial cells are now well
known, although a photodynamic action could also account for the lethal action on these
cells. We measured cellular protein concentration as a parameter to define the lethal effect
of radiation regardless of its mechanism. Targets of UV photons are proteins and DNA
primarly.'9.26,28.29 UV-A and UV-B induce the formation of pyrimidine- and thymine-di-
mers in corneal epithelial cells. '9 •3o However, a wavelength- and dose-dependent decrease
of metabolic activity has been shown in corneal epithelial cells after UV radiation (de-
crease of corneal phosphocreatine concentration and oxygen uptake).31.32 By the way, it
has been demostrated that UV -radiation promotes an increase of enzymatic intermediates
(Heme oxygenase-I) of oxidative-stress in rabbit corneal epithelial cells. 33
The reduction in the number of cells could be due to a lack of growth produced by a
decrease of cell division, but also caused by the lysis of cells. Scanning electron micros-
copy revealed signs of cell necrosis such as swelling, shrinking of cell membrane and ap-
pearance of vacuoles and granules in the cytosol, as well shape modifications. Similar
results have been previously reported by Doughty.34
In this preliminary study, we did not look at the possible biochemical mechanisms
by which UV-A radiation induces nocive effects on corneal epithelial cells, which will be
the aim of further studies. Our purpose was to study the possible protective effect of col-
ored absorbing filters. In our study, we also demonstrate that appropriate UV absorption
filter protected the lethal action of UV radiation. Similar results have been observed in
rabbits exposed to UV radiation and protected with UV -absorbing contact lenses. 35 In this
way, UV absorbing contact lenses provided complete protection to all corneal layer
against UV-radiation-induced ultrastructural changes in r;abbits cornea. However, regular
soft contact lenses offered no protection. 36
Corneal epithelium is the external layer in the eye, and it is directly exposed to the UV
radiation of the sun. UV radiation can be transmitted through this layer, and could also dam-
age intraocular tissues such as the lens or the retina. Since a protective effect of the absorption
filter has been evidenced in the current study, the use of this filter is recommended.
5. ACKNOWLEDGMENTS
The authors wish to thank to Mr. M. A. Compaiiy and Mr. M. Sanz for technical assistance.
6. REFERENCES
I. Parrish JA, Anderson PR, Urbach F, Pitts D, Optical properties of the skin and eyes. pp. 5~4. In: UV-A.
Biological effects of ultraviolet radiation with emphasis on human responses to longwave ultraviolet. Ple-
num Press. New York, J978.
Ultraviolet Light-Induced Damage in Rabbit Corneal Epithelial Cells 71
2. Zuchlich JA. Ultraviolet-induced photochemical damage in ocular tissues. Health Phys. 1989; 56:671-682.
3. Wittenberg S. Solar radiation and the eye: a review of knowledge relevant to eye care. Am. J. Optom.
Physiol. Opt. 1986; 63:676-689.
4. Parrish JA, Anderson PR, Urbach F, Pitts D. Effects of ultraviolet radiation on the eye. In: UV-A. Biologi-
cal effects of ultraviolet radiation with emphasis on human responses to longwave ultraviolet. Plenum
Press. New York, 1978, pp. 177-219.
5. Pitts DG, Cullen AP, Hacker PD. Ocular effects of ultraviolet radiation from 295 to 365 nm. Invest.
Ophthalmol. Vis. Sci. 1977; 16: 932-939.
6. Pitts DG. The ocular effects of ultraviolet radiation. Am. J. Optom. 1978; 55: 19-35.
7. Hightower KR. The role of the lens epithelium in development of UV cataract. Curro Eye Res. 1995;
14:71-78.
8. Liu X, Yanoff M, Li W. Characterization of lethal action of near ultraviolet on retinal pigment epithelial
cells in vitro. Curro Eye Res. 1995; 14:1087-1093.
9. Cullen AP, Chou BR, Hall MG, Jang SE. Ultraviolet B damages corneal endothelium Am. J. Optom.
Physiol. Optics. 1984; 61 :473-478.
10. Thorrington L. Spectral, irradiance, and temporal aspects of natural and artificial light. Ann. NY. Acad. Sci.
1985; 453: 28-54.
II. Millodot M, Earlam RA. Sensitivity of the cornea after exposure to ultra-violet light. Ophthalmic Res.
1984; 16:325-328.
12. Cullen AP, Chou BR, Hall MG, Jang SE. Ultraviolet B damages corneal endothelium. Am. J. Optom.
Physiol. Optics. 1984; 61: 473-478.
13. Cullen AP, Chou BR. Blue free fundus examination. Can. J. Optom. 1984; 46: 808-814.
14. Pitts DG, Tredeci JT. The effects of ultraviolet on the eye. Am. Ind. Hygiene Assoc. J. 1971; 32:235-246.
15. Pitts DG, Bergmanson JPG, Chu LWF. Ultrastructural analysis of corneal exposure to UV radiation. Acta
Ophthalmol. 1987; 65:263-273.
16. Tyrrell RM, Moares EC, Werfelli P. Lethal action of ultraviolet and visible (blue-violet) radiation at de-
fined wave-lengths on human Iympho-blastoid cells: action spectra and interactions sites. Photochem. Pho-
tobiol. 1984; 39: 183-189.
17. Tyrrell RM. Damage and repair from non-ionizing radiation. In: Repairable lesions in microorganisms.
Eds. Hurst A and Nasim A. pp. 85-124. Academic Press. London, 1985.
18. Rosenstein BS, Mitchell DL. Action spectra for the induction of pyridine photoproducts and cyclotane
pyrimidine dimers in normal human skin fibroblasts. Photochem. Photobiol. 1987; 45:775-780.
19. Sidjanin D, Zigman S, Reddan J. DNA damage and repair in rabbit lens epithelial cells following UVA ra-
diation. Curro Eye Res. 1993; 12:773-781.
20. Eggset G, Volden G, Krokan H. UV-induced DNA damage and its repair in human skin in vivo studied by
sensitive immuno-histochemical methods. Carcinogenesis. 1983; 4:745-750.
21. Bjerknes R, Kranck K. Early cell kinetic effects of a single dose of narow-banded ultraviolet B irradiation
on the rat corneal epithelium. Photochem. Photobiol. 1993; 57:663-666.
22. Gallar J, Garcia P, Gonzalez GC, Belmonte C. Irritation of the anterior segment of the eye by ultraviolet ra-
diation: influence of nerve blockade and calcium antagonists. Curro Eye Res. 1995; 14:827-835.
23. Guide for the Care and Use of Laboratory Animals. US Department of Health and Human Services. Public
Health Service, National Institutes of Health, publication No. 86-23, revised 1985.
24. Xie L, Gebhardt BP. A simplified technique for the short-term tissue culture of rabbit corneal cells. In vitro
cellular and developmental biology. 1989; 25:20-22.
25. Bradford MM. A rapid sensitive method for the quantification of microgram quantities of proteins utilizing
the principle of protein-dye binding. Annal. Biochem 1976; 72: 465-470.
26. Andley UP, Walsh A, Kochevar IE, Reddan JR. Effect ofultraviolet-B radiation on protein synthesis in cul-
tured lens epithelial cells. Curro Eye Res. 1990; 9: 1099-1106.
27. Andley UP, Lewis RM, Reddan JR, Kochevar IE. Action spectrum of cytotoxicity in the UVA and UVB-
wavelength region in cultured lens epithelial cells. Invest. Ophthalmol. Vis. Sci. 1994; 35:367-373.
28. Jones CA, Cunningam ML, Huberman E, Peak MJ. Mutagenesis and cytotoxicity in human epithelial cells
by far- and near-ultraviolet radiations: action spectra. Radiat. Res. 1987; 110:244-254.
29. Tyrrell RM. UVA (320-380 nm) radiation as an oxidative stress. In Oxidative Stress: Oxidants and Antioxi-
dants. Ed. Sies H. pp. 2311-251. Academic Press. London, 1991.
30. Ley RD, Applegate LA, Freeman SE. Photorepair of ultraviolet radiation-induced pyrimidine dimers in
corneal DNA. Mutat. Res. 1988; 194:49-55.
31. Lattimore MR. Effect of ultraviolet radiation on the energy metabolism of the corneal epithelium on the
rabbit. Photochem. Photobiol. 1989; 194:49-55.
72 M. Palmero et al.
32. Lattimore MR. Effects of ultraviolet radiation on the oxygen uptake rate of the rabbit cornea. Opto. Vis.
Sci. 1989; 66:117-122.
33. Abraham NG, da Silva JL, Lavrosky Y, Stolz RA, Dun MW, Schwartman ML. Adenovirus-mediated Heme
oxygenase-I gene transfer into rabbit ocular tissues. Invest. Ophthalmol. Vis. Sci. 1995; 36:2202-2210.
34. Doughty MJ. Morphometric analysis of the surface cells of rabbit corneal epithelium by scanning electron
microscopy. Am. J. Anal. 1990; 189:316-324.
35. Cullen AP, Dumbleton KA, Chou BR. Contact lenses and acute exposure to ultraviolet radiation. Optom.
Vis. Sci. 1989; 66:407-411.
36. Bergmanson JP, Pitts DG, Chu LW. The efficacy of UV-blocking soft lens in protecting cornea against UV
radiation. Acta Ophthalmol. Copenh. 1987; 65:279-286.
7
1. INTRODUCTION
initiation, rats were 35 days old and weighed 104 to 195 grams. They were housed in indi-
vidual suspended stainless wire cages, based on a random allocation scheme. Animal
rooms were air-conditioned and maintained at a temperature of 22 ± 1°C and a relative hu-
midity of 40 to 70%. There were approximately 15 air-changes per hour and a 12-hour
light cycle with a light intensity at I m above the floor of approximately 100 to 250 Lux.
All rats had ad libitum access to tap water.
3. RESULTS
C::=::J GROUP 1 (Ad libitum) GROUP 2 (- 25% DR) 1:11._ GROUP 3 (- 55% OR)
4. DISCUSSION
-==
===:I GAOUP' (Ad
CIAOUP 2 (- 2& DA)
Figure 2. (Continued) c. Percentage of slight opacities in females. D. Percentage of moderate opacities in fe-
males. E. Percentage of marked opacities in females.
78 M.-F. Hubert et al.
'--_-' GROUP 1 (Ad Ibtum ) GROUP 2 (- 25% DR) a._ GROUP 3 (- 55% DR)
creases in serum alkaline phosphatase activity in group 3 females were associated with a
decrease in the incidence of corneal dystrophy. Therefore, this corneal disorder in SD rats
did not appear to be related to increases in serum alkaline phosphatase activity .
A number of mechanisms have been postulated to account for the reduction of age-
related senescence and degeneration by DR. The most widely held hypotheses are that DR
prevents long-term damage to DNA and other macromolecules by either free radicals or
......._~ GROUP 1 (Ad libitum) GROUP 2 (- 25% DR) • • • GROUP 3 (- 55% DR]
c:::=::r GROUP 1 (Ad /btum ) GROUP 2 (- 25% DR) _ _ GROUP 3 (- 55% DR)
glycation.7 The routine laboratory parameters analyzed in this study could not verify the
free radical theory. The glycation theory, that proposes glucose as a mediator of aging,
was not supported because serum glucose levels were similar or slightly higher in the 55%
DR females, compared to AL-fed females, throughout the study.
5. CONCLUSION
The present study confirms the favorable effect of DR on degenerative processes in
rats. It also pointed out the necessity to reach a certain percentage of DR according to the
target organ or tissue and to the gender. A DR of approximately 55% of AL food con-
sumption decreased the incidence of corneal dystrophy in non-pregnant and non-lactating
female SD rats, but did not affect the lesion in males.
Mechanisms by which DR lessened corneal degenerative process remained un-
known; additional parameters should be analyzed to verify the possible role of calcium
and free radicals.
6. REFERENCES
I. Bellhorn RW. Korte GE. Abrutyn D. Spontaneous corneal degeneration in the rat. Lab Anim Sci
1988;38:46-50.
2. Wegener A. Jochims K. Clinical, histological and ultrastructural characteristics of a spontaneous corneal
opacity in Sprague-Dawley rats. Ophthal Res 1994;26:296-303.
3. Losco PE. Troup CM. Corneal dystrophy in Fischer 344 rats. Lab Anim Sci 1988;38:702-710.
4. Bruner RH, Keller WF, Stitzel KA. Sauers LJ, Reer PJ , Long PH, Bruce RD, Alden CL. Spontaneous cor-
neal dystrophy and generalized basement membrane changes in Fischer-344 rats. Toxicol Pathol
1992;20:357- 366.
80 M.-F. Hubert et al.
5. Fabian RJ, Bond JM, Drobeck HP. Induced corneal opacities in the rat. Brit J Ophthal 1967;51: 124-129.
6. Weisse I, Kreuzer H, Stender E, Frolke W, Meyer D. Band keratopathy in rats due to increased dietary con-
tent of vitamin D3. Concepts Toxicol 1987;4: 164-178.
7. Keenan KP, Soper KA, Hertzog PR, Gumprecht LA, Smith PF, Mattson BA, Ballam GC, Clark RL. Diet,
overfeeding, and moderate dietary restriction in control Sprague-Dawley rats: II. Effects on age-related
proliferative and degenerative lesions. Toxicol Pathol 1995; 23:287-301.
8. Mittl R, Galin MA, Opperman W, Camerini-Davalos RA, Spiro D. Corneal calcification in spontaneously
diabetic mice. Invest Ophthalmol 1970;9:137-145.
9. Gillet JP, De Burlet G, Molon-Noblot S, Wise LD, Durand-Cavagna G, Duprat P. Exacerbation of sponta-
neous rat corneal opacities in a polymorphonuclear elastase inhibitor toxicity study. Ocular Toxicology, ed-
ited by I. Weisse, O. Hockwin, K. Green and R.C. Tripathi, Plenum Press, New York, 1995, pp. 335-342.
8
1. INTRODUCTION
Studies of cells from cultured intraocular tissues have been generally carried out at
atmospheric pressure. However, normally in vivo, these tissues are conditioned to hydro-
static pressure which ranges from 10 to 21 mmHg above atmospheric pressure, and are
subjected to a several-fold higher pressure in glaucomatous states of the eye. The purpose
of this study was to investigate the response of trabecular cells exposed to varying levels
of hydrostatic pressure to simulate normal and abnormal conditions of the eye.
* Corresponding author: Professor Brenda J. Tripathi, Ph.D., Department of Pathology and Vision Research
Laboratory, The University of South Carolina School of Medicine, 6439 Garners Ferry Road, Columbia, SC
29209.
Figure 2. Phase-contrast micrographs of trabecular cells, x 135. A: Cells conditioned for 2 hours at zero hydro-
static pressure. Note a spindle-shaped morphology and uniform intercellular spaces. B: The same field of view af-
ter 24 hours' exposure to a hydrostatic pressure of 15 mmHg. Cells maintained their morphology and intercellular
spaces.
3. RESULTS
4. DISCUSSION
The exclusive use of human eyes is necessarily limited because of the inadequate
supply of suitable fresh normal tissue from donors of comparable age and because post-
84 B. J. Tripathi et aL
Figure 3. Phase-contrast micrographs of trabecular cells, x 135. A: Cells conditioned for 2 hours at zero hydro-
static pressure. Note a spindle-shaped morphology and uniform intercellular spaces. B: The same field of view af-
ter 24 hours' exposure to a hydrostatic pressure of 30 mmHg. Note the contraction of cells, widening of
intercellular spaces, and elongated cellular profiles.
mortem changes occur not only in the aqueous humor, but also in the tissues bordering the
anterior chamber that liberate proteases, ribonucleases, and other endogenous enzymes
which degrade both polypeptide growth factors and mRNAs within the cells. We routinely
obtain freshly enucleated porcine eyes within minutes of decapitation from a local abat-
toir. Significant similarities exist in the morphologic organization of the trabecular system
as well as their cellular and extracellular matrix components in porcine and primate
eyes.l.3-!i In addition, comparison of confluent cultures of human and porcine trabecular
cells reveals significant similarities in cell morphology, immunoreactivity, biochemistry,
and pharmacology,u Furthermore, the normal intraocular pressure as well as the composi-
tion of aqueous humor of normal porcine eyes are similar to those of humans. Therefore,
the results obtained in this experiment may be extrapolated to reflect conditions in human
eyes.
Our present experiment demonstrated that, unlike vascular endothelium which has
the capacity to withstand variations in diastolic and systolic blood pressures, trabecular
Figure 4. Phase-contrast micrographs of trabecular cells. x 135. A: Cells conditioned for 2 hours at zero hydro-
static pressure. Note a spindle-shaped morphology and uniform intercellular spaces. B: The same field of view af-
ter 24 hours of exposure to a hydrostatic pressure of 50 mmHg. The cells became elongated with widened
intercellular spaces. Compare the changes with those observed at the pressure of 30 mmHg as seen in Figure 3.
Barotoxic Effects on Morphology and Viability of Trabecular Cells 85
cells which are conditioned to a relatively steady-state intraocular pressure, are sensitive
to the changes in the ambient pressure within their environment. When cultured at a sus-
tained pressure of 15 mmHg, which is the pressure that trabecular cells are exposed to un-
der normal conditions in vivo, the cells maintain a spindle shaped morphology similar to
that in vivo. As the pressure was increased, the trabecular cells became elongated, the in-
tercellular spaces widened, and pseudopodal movements increased. The exact mechanisms
which are responsible for these effects are not clear. Previous studies have demonstrated
that mechanical stress modulates the composition of cytoskeleton as well as the amount of
various forms of microfilaments such as actin. 7.8 Increased pressure may also regulate the
production of extracellular matrix components and adhesion molecules. These cffects
could account for the observed alterations in the morphology and movement of the cells.
The effects of mechanical stretching on the morphology of cardiac fibroblasts have
been investigated by using a three-dimensional culture system (sail-sheet cultures).7.8 Low
frequency mechanical stress resulted in a decrease in the number and size of cellular mi-
croprocesses; the locomotory activity of the fibroblasts was also reduced, and the cells be-
came rounded but mitotic activity increased. 7.s These observations are distinct from our
findings because of the difference in the types of cells studied, the culture systems and
stress perimeters that we used.
The alteration of trabecular cell morphology in response to hydrostatic pressure ap-
pears to be a function of pressure magnitude. At higher pressures, the alterations in cell
morphology and viability were more marked. Increased hydrostatic pressure also resulted
in the degeneration and death of trabecular cells in monolayer culture. Our results demon-
strate that an acute rise in intraocular pressure, such as that occurs in acute onset glau-
coma, can cause deleterious changes in the trabecular cells.
Our system provides a model for studying barotoxicity in other tissues of the eye as
well as those tissues of the body which are normally subjected to hydrostatic pressure,
such as vascular endothelium, choroid plexus, cerebral cortex and spinal cord, and possi-
bly cells of the inner ear. A constant flow of medium through the chamber provides a
means of effluent collection for biochemical analyses. Future studies will focus on the
precise morphometric analyses, modulation of cytoskeleton elements, mRNA expression
and protein synthesis for various growth factors and ECM molecules by trabecular cells
exposed to normal and abnormal levels of hydrostatic pressure.
5. ACKNOWLEDGMENTS
This study was supported in part by a grant from the NEI, EY-03747, Research to
Prevent Blindness, Inc., the Vision Research Foundation, and The Glaucoma Foundation,
New York, NY.
6. REFERENCES
1. Tripathi RC, Park JK, Tripathi BJ, Tsao C, Tissue plasminogen activator synthesis by trabecular cells and
its implications for fibrinolytic therapy of the eye. Drug Dev Res 1989; 18:245-54.
2. Ziegler T, Nerem RM. Effect of flow on the process of endothelial cell division. Arteriosclerosis and
Thrombosis 1993; 14:636-{)43.
3. Tripathi Re, Tripathi BJ. Functional anatomy of the anterior chamber angle. In: Duane T, Jaeger EA, eds.
Biomedical Foundations of Ophthalmology, Harper and Row, Philadelphia, 1-88, 1982.
4. McMenamin PG, Steptoe RJ. Nonnal anatomy of the aqueous outflow system in the domestic pig eye. J
Ana! 1991;178:65-77.
86 B. J. Tripathi et al.
5. Tripathi B1, Tripathi RC, Yang C. Millard C8, Dixit VM. Synthesis of a thrombospondin-like cytoadhesion
molecule by cells of the trabecular meshwork. Invest Ophthlmol Vis Sci 1991 ;32: 177-84.
6. Tripathi RC, Li 1, Tripathi B1. Immunolocalization of bFGF in the trabecular meshwork and detection of
its mRNA in trabecular cells. Exp Eye Res 1994;58:503-7.
7. Tripathi SC. A possible role of actin in the mechanical control of the cell cycle. Biology of the Cell
1989;67:351-3.
8. Tripathi SC, Kerr 1. Effect of mechanical stress on cellular morphology. Tissue & Cell 1989;21 :747-52.
9
Department of Ophthalmology
University of South Carolina School of Medicine
Four Richland Medical Park, Suite 300
Columbia, South Carolina 29203
1. INTRODUCTION
2. CASE REPORT
At a local eye clinic, a sixty-two year old African American female with a history of
hypertension and diabetes had a routine eye examination which included applanation
tonometry after topical instillation of fluorescein. Three days later the patient complained
of ocular pain and haloes. She was referred promptly to our institution for raised intraocu-
lar pressure. Our examination revealed a visual acuity of 20/30 in both eyes. Extraocular
motility was full and external eye examination was normal. On slit lamp examination, we
noted a one plus conjunctival injection, a slight corneal haze with an intact epithelium and
no stromal disease. A few keratic precipitates were noted centrally (Fig. I). The anterior
chamber was deep but showed a three plus flare and one plus cell. The intraocular pres-
* Address correspondence to: Professor Ramesh C. Tripathi, M.D., Ph.D., F.A.C.S., Department of
Ophthalmology, The University of South Carolina School of Medicine, South Carolina Eye Institute, Four
Richland Medical Park, Suite 300, Columbia, South Carolina 29203.
sure, measured by applanation tonometry, was 39 mmHg in both eyes. Gonioscopic ex-
amination revealed grade 3 to 4 open angles. Fundus examination with a dilated pupil was
within normal limits. The cup/disc ratio was 0.5, with a healthy rim 360 degrees in both
eyes. Our presumptive diagnosis was bilateral iritis and secondary glaucoma.
The patient was treated with topical drops of f3-blocker (Betoptic-s), one drop twice
daily, prednisolone acetate (Pred-forte), one drop every six hours, and a non-steroidal
anti-inflammatory drug (Voltaren) one drop every six hours, and systemically with Aceta-
zolamide sequel (Diamox) 500 mg, twice daily. On return visit next day, the patient's an-
terior chamber reaction had decreased and showed one plus flare, and intraocular
pressures were 27 mmHg in the right eye, and 31 mmHg in the left eye. After one week of
treatment, the intraocular pressure normalized (21 mmHg in the right eye, and 20 mmHg
in the left eye); also the anterior chamber reaction had subsided. On consecutive follow-up
examinations at intervals of7 to 14 days, the intraocular pressure was measured with topi-
cal instillation of a drop of Fluorocaine, which contained 0.25% fluorescein sodium and
0.5% proparacaine hydrochloride. On each occasion, the patient complained of pain, pho-
tophobia, redness and discomfort in both eyes one to three days later. On each examina-
tion, ciliary injection and 2-3 plus flare and cells were noted in the anterior chamber. The
intraocular pressure by applanation tonometry varied from 29 to 43 mmHg. About three
weeks after the initial referral, when we became suspicious of fluorescein as an incrimi-
nating agent, we applied a fluorescein strip topically to test the ocular reactions of the pa-
tient. Three days later when the patient was examined, she had flare-up of iritis and
elevated intraocular pressure (37 mmHg). Because of our correlation of fluorescein with
inflammatory glaucoma, we began to measure the intraocular pressure with the use of
proparacaine as a topical anesthetic and a Tonopen. With this regimen, the patient mani-
fested no recurrence of iritis nor increase in intraocular pressure.
3. DISCUSSION
NaO
sified as mild, moderate, severe, and death. 2 A mild reaction is characterized as a transient
effect which does not require treatment because it has a rapid and complete resolution
without sequelae. Examples of mild reactions include nausea, vomiting, sneezing, and
pruritus. Moderate reactions are also transient in nature although some form of medical
treatment may be required. This reaction has a complete but gradual resolution without se-
quelae and presents no threat to the safety of the patients. Urticaria, syncope, thrombo-
phlebitis are categorized as moderate adverse reactions. A severe reaction is one which
exhibits prolonged effects and requires intense treatment. It also poses as a threat to the
patient's safety and results in variable recovery. This type of reaction involves the respira-
tory, cardiac or nervous systems, and includes bronchospasm, anaphylaxis, circulatory
shock, myocardial infarction and cardiac arrest. Death is attributed to the use of fluo-
rescein if symptoms are noted within 24 hours and the incident occurs within 48 hours of
the use. In a survey of 221,781 fluorescein angiograms performed, the frequency rates
were 1:63 for a moderate reaction, and 1: 1900 for a severe reaction. 2 The most common
side effects are nausea and vomiting, followed by itching. 5.8- lo Rare cases of hemolytic
anemia, I I carpopedal spasm,12 photoallergic skin sensitivity,13 generalized seizure,14.15 and
acute pulmonary edema l6 after intravenous fluorescein administration have been reported.
The life threatening reactions to intravenous injection results from anaphylactic shock
with hypotension, bronchospasm, and death. 2
The precise pathogenetic mechanisms for reactions to intravenous fluorescein are
not clearly understood. Some proposed mechanisms 2 include: (l) a vaso-vagal phenome-
non which results in bradycardia, hypotension, and reduced cardiovascular perfusion; (2) a
drug allergy of the immediate hypersensitivity type (immunological response to a hapten-
protein configuration); (3) a histamine release of a non-allergic nature in the absence of an
antigen-antibody reaction; (4) an effect of a contaminant in the formulation of the drug;
(5) an anxiety-related medullary sympathetic discharge which elicits tachycardia and
myocardial stress; and (6) any combination of these potential factors.
Hypersensitivity reactions to topical application of fluorescein are not well docu-
mented. In one report, two cases of iritis were related to fluorescein drops used in the im-
mediate postoperative examination of patients who underwent radial keratotomy and had
microperforations. 3 We inadvertently challenged our patient's reaction by allowing the
eyes to quieten, then instilled fluorescein topically. The incidence of repeated iritis and in-
creased intraocular pressure led to our suspicion that topical fluorescein was a causative
factor.
According to manufacturer's package insert, Fluorocaine contains fluorescein so-
dium 0.25%, proparacaine hydrochloride 0.5%, thimerosal 0.01%, and other inactives in-
cluding povidone, glycerine, boric acid, polysorbate 80, sodium hydroxide and/or
hydrochloric acid, and purified water. In our patient, the association of topical fluorescein
90 C. Bryan et al.
application with the occurrence of iritis and increased intraocular pressure was established
on the following findings: (I) the same sequence of events occurred on three consecutive
occasions, pointing to the possibility of a causative effect of fluorescein; (2) no iritis, con-
junctival injection or elevated intraocular pressure were observed when fluorescein was
not used; (3) a similar effect induced by the use of fluorescein strip and no adverse reac-
tions with the use of proparacaine and Tonopen tonometry demonstrated that other com-
ponents in fluorocaine, such as proparacaine hydrochloride, were not the incriminating
factor. The preservative used in Fluorocaine (thimerosal) could not have been the inciting
factor because a similar effect occurred even with the use of fluorescein strip which con-
tained no preservative.
Our patient may represent a case of fluorescein-induced glaucomatocyclitic crisis
similar to Posner-Schlossman syndrome. 17 This syndrome, initially, has the characteristics
of elevated intraocular pressure, open iridocorneal angle, small keratic precipitates, and
normal visual fields and optic disc. The increased intraocular pressure may not be solely
the result of iris inflammation, as the possibility exists of a direct inflammatory effect on
the trabecular meshwork, which leads to a trabeculitis. 18
To our knowledge, our case is the first report of iritis caused by topical instillation
of fluorescein onto an intact cornea. Two cases of iritis related to fluorescein use after ra-
dial keratotomy were reported earlier. 3 However, it may be argued that after radial kerato-
tomy and microcorneal perforation, penetration of fluorescein is enhanced because of a
breach in the integrity of the cornea. The extent of iritis that may be related to the dose of
fluorescein entering the anterior chamber is unknown. However, our case illustrates that
even a single drop of fluorocaine or application of fluorescein strip can cause iris reaction
and raised intraocular pressure, thereby signifying the possible hypersensitivity side effect
of fluorescein. In such patients, applanation tonometry with the use of fluorescein should
be avoided. Instead, Tonopen or electronic tonometry (which do not require instillation of
fluorescein) could be used as alternative devices for the measurement of intraocular pres-
sure.
4. SUMMARY
Various side effects have been reported after intravenous use of fluorescein. How-
ever, the adverse reactions to topical administration of this chemical have not been well
documented. We report a case of unusual reaction to fluorescein when used topically on
an intact cornea for the measurement of intraocular pressure. The patient manifested pho-
tophobia, conjunctival injection, iritis and elevated intraocular pressure, which necessi-
tated medical management. These symptoms and signs occurred on each of four separate
occasions when fluorescein was used topically, but disappeared upon discontinuing the
use of this chemical agent. We believe that our patient exhibited an unusual hypersensi-
tivity response to topical application of fluorescein akin to Posner-Schlossman syn-
drome.
5. ACKNOWLEDGMENTS
This study was supported in part by a grant from Research to Prevent Blindness,
Inc., and the Vision Research Foundation.
Topical Fluorescein Application Can Induce Iritis and Glaucoma 91
6. REFERENCES
I. Gifford H. Use of fluorescein intravenous as an aid to ophthalmic diagnosis and treatment. Arch Ophthal-
mol 1940;24: 122-31.
2. Yannuzzi LA, Rohrer KT, Tindel LJ, Sobel RS, Con stanza MA, Shields W, Zang E. Fluorescein angiogra-
phy complication survey. Ophthalmology 1986;93:611-7.
3. Brodsky ME, Bauerberg 1M, Sterzovsky A. Case report: probable fluorescein-induced uveitis following ra-
dial keratotomy. 1 Refract Surg 1987;3:28-9.
4. Fraunfelder FT, Grove lA. Drugs used primarily in ophthalmology. In: Fraunfelder FT, Grove lA, Eds.
Drug-Induced Ocular Side Effects, 4th edition, p. 460-3, 1996. Williams & Wilkins, Philadelphia, PA.
5. Grant WM, Schuman IS. Chapter II. Encyclopedia of chemicals, drugs, plants, toxins, and venoms, and
their effects on the eye or vision; also, drugs used in treating eye diseases, and their general side effects. In:
Grant WM, Ed. Toxicology of the Eye, 4th edition. p. 691--4,1993, Charles C. Thomas, Springfield, IL.
6. Levacy RA, lustice 1 lr. Adverse reactions to intravenous fluorescein. Int Ophthalmol Clin 1976; 16:53-61.
7. Stein MR, Parker Cw. Reactions following intravenous fluorescein. Am 1 Ophthalmol 1971 ;72:861-8.
8. Greene GS, Bell LW, Hitching RA, Spaeth GL. Adverse reaction to intravenous fluorescein: evidence for
sex difference. Ann Ophthalmo1. 1976;8:533-6.
9. lennings Bl, Mathews DE. Adverse reactions during retinal fluorescein angiography. 1 Am Optometric As-
sociation 1994;65:465-71.
10. Kwiterovich KA, Maguire MG, Murphy RP, Schachat AP, Bressler NM, Bressler SB, Fine SL. Frequency
of adverse systemic reactions after fluorescein angiography. Results of a prospective study. Ophthalmology
1991 ;98: 1139--42.
II. Munizza M, Kavitsky D, Schainker BA, Poyserr A, Peek C, Nance S. Hemolytic anemia associated with
injection of fluorescein. Transfusion 1993;33:689-92.
12. Mole T. Carpopedal spasm: an unusual reaction to fluorescein? 1 Audiovisual Media in Medicine
1991;14:54.
13. Hochsattel R, Gall H, Weber L, Kaufman R. Photoallergic reaction to fluorescein. Contact Dermatitis
1990;22:42--4.
14. Gombos GM, Lieberman RM. Seizures associated with fluorescein angiography. Ann Ophthalmol
1989;21 :89-90.
15. Kelly SP, MacDermott Nl, Saunders DC, Leach FN. Convulsion following intravenous fluorescein angiog-
raphy. Br 1 Ophthalmol. 1989;73:655-6.
16. Hess JB, Pacurariu Rl. Acute pulmonary edema following intravenous fluorescein angiography. Am 1
Ophthalmol 1976;82:567-70.
17. Posner A, Schlossman A. Syndrome of unilateral recurrent attacks of glaucoma with cyclitic syndromes.
Arch Ophthalmol 1948;39:517-20.
18. Krupin T, Feitl ME, Karalekas D. Glaucoma associated with uveitis. In: Ritch R, Shields MB, Krupin T,
eds. The Glaucomas. Pp 1235-7, 1996. Mosby-Year Book: St. Louis.
10
1. INTRODUCTION
The tapetum is a specialized cellular layer in the posterior ocular segment located
between the choriocapillaris and the pigment epithelium of the retina. The tapetal layer is
located only in the superior aspect of the fundus. The tapetum may be present in all do-
mestic animals except the pig. The tapetal layer in the the dog and cat is called the
tapetum cellulosum and in ungulates and the horse the tapetum fibrosum. The tapetum
may be absent from some animals especially color dilute dogs or animals with incomplete
melanin pigment formation such as the Siamese cat. The tapetum lucidum is composed of
5- \0 layers of non pigmented reflective cells that contain a high content of zinc. The zinc
content of the tapetum in the dog has been determined to be 64.9 +/- 18.4 mg/g tissue. 4 •5. 12
In this study with DDTC, Beagle dogs were administered sodium DDTC (sodium
diethylthiocarbamate) intravenously at doses of 50,100 and 200 mg/kg/day. Ocular toxic-
ity developed and was dose-dependent. In the high dose group (200mg/kg) ocular find-
ings developed bilaterally withing 24 hours of drug administration. Clinical findings were
characterized by severe pan-uveitis with the posterior choroid in the area of the tapetum
cellulosum being most severely affected. There was chemosis (conjunctival edema), peri-
orbital swelling, inflammatory cells and flare (protein) in the vitreous cavity, marked reti-
nal edema, exudative retinal detachment in the area of the tapetum cellulosum, tapetal cell
necrosis, photophobia, blepharospasm and pain. Acute visual loss developed withing 24
hours of administration of the high dose of the DDTC. Because of the acute visual loss
and severe pain, animals receiving the high dose of DDTC were euthanized and tissues
submitted for histologic evaluation.
At the medium dose (100 mg/kg) DDTC produced focal tapetal cell necrosis, focal
choroiditis with exudative retinal detachments and transient visual disturbances. These
changes were evident 24 hours after initial dosing. Severe pain and more generalized
uveal inflammation was not present. Evaluation of these animals at a 7 day interval fol-
lowing initial dosing indicated that the focal tapetal cell inflammation and exudative reti-
nal detachment had resolved producing focal retinal atrophy and changes in pigment
epithelium and tapetal cell pigmentation. Because the choroiditis and retinal inflammation
was focal only in the tapetal area and resolved, permanent generalized visual alteration did
not appear to develop.
Low dose administration of DDTC at 50 mg/kg/day resulted in changes in tapetal
cell pigmentation but no choroiditis or focal retinal detachments. The tapetum lost its nor-
mal pastel reflective appearance and became blue-purple in color.
Histologic evaluation of the globes from dog's receiving 200 mg/kg DDTC revealed
a panuveitis characterized by exudative retinal detachments. The tapetal layer was edema-
tous and infiltrated with inflammatory cells mainly small lymphocytes. Severe choroiditis
was present and the retina was detached with necrosis of the photoreceptor cell layer.
Examination of histologic sections of eyes from Beagle dogs receiving 100mg/kg
DDTC revealed alteration of the tapetal cellular layer characteriazed by tapetal cell atro-
phy and pigment cell hyperplasia of the pigment epithelium of the retina. The retinal pho-
toreceptor cell layer was normal in the area of the tapetal cell atrophy.
3. DISCUSSION
The tapetum has been reported to be the target organ in a number of toxicologic
studies in which the compound under investigation is a zinc chelator. Included in these
studies are the componds: ethylenediamine derivatives including ethambutol, dithizone
(diphenylthiocarbazone), diethylthiocarbamate, hydroxypyridinethione, dilevalol (SCH-
19927) a non-selective beta adrenergic blocker. '-3.5-'O Dithizone and pyrithione, when ad-
ministered systemically to dogs at the appropriate doses, produce ocular changes similar
to that seen with DDTC. Both agents, known to chelate metals such as zinc, cause decol-
oration of the tapetum and edema, hemorrhage, and exudative detachment of the retina de-
pending upon the dose of drug administered to the animal. 5 In the case of pyrithione,
histology shows the zinc-rich, tapetum to be necrotic, and both tapetum and choroid to be
edematous and infiltrated with inflammatory cells.In contrast, pyrithione administered
systemically to rats, rabbits, and monkeys produced no ocular changes such as those seen
in dogs. Delahunt et al. 3 concluded that, like dithizone, pyrithione-induced retinal detach-
ment was specific to dogs.
Although it is a metal chelator, ethambutol, an ethylenediamine derivative, affects of
the eyes of animals and humans somewhat differently than dithizone and pyrithione de-
pending upon the species. Ethambutol causes decoloration of the tapetum in dogs; in cats
it causes tapetal decoloration and retinal detachment. In monkeys (which lack a tapetum),
ethambutol at doses approximately 50 to 100 times the recommended human dose (i.e., 15
mg/kg) produces demyelination of optic nerve fibers and proliferation of microglia in the
center of the optic chiasm. In humans (which lack a tapetum), ethambutol produces signs
and symptoms of retrobulbar neuritis, (at doses of 25 mg/kg or higher per day) the inci-
dence of which is proportional to the dose given. This neuritis in humans is reversible
upon withdrawal of the drug. Grant5 states that "the effects of ethambutol on the tapetum
peculiar to dogs and cats very likely has no relevance to the retrobulbar neuritis observed
Ocular Toxicity of Sodium Diethylthiocarbamate, DDTC, in the Beagle Dog 95
in human beings, but the lesions observed in monkeys in nerve fibers in the chiasm seem
more likely to have a relationship to the optic neuropathy in the retrobulbar portion of the
nerve observed in human beings". Another ethylenediamine derivative, sodium edetate,
also causes decoloration of the tapetum in dogs presumably due to removal of zinc. 5.11
4. CONCLUSIONS
The ocular changes in the Beagle dog produced by the systemic administration of
DOTe is specific to animals with a tapetum. Dogs are a species with a zinc-rich tapetum.
DOTe is a chelator of zinc. Acute ocular lesions are characterized by a panuveitis with
acute choroiditis resulting in exudative retinal detachment. Ocular inflammatory lesions
asssociated with DOTe are dose-related and the most severe inflammatory changes are in
the area of the tapetum cellulosum. No ocular changes were found in rats, rabbits, or mon-
keys who were given DOTe.
5. REFERENCES
I. Budinger 1M: Diphenylthiocarbazone blindness dogs. Arch Path 1961; 71: 304-310.
2. Capeillo VP, Layton WM lr: A one year study of the toxicity of ethambutol in dogs; results of gross and
histopathologic examinations. Toxicol Appl Pharmacol 1965; 7: 844-849.
3. Delahunt CS, Stebbins RB, et al: The cause of blindness in dogs given hydroxypyridinethione. Toxicol
Appl Pharmacol 1962; 4: 286--291.
4. Figueroa R, Wess H, Smith lC: Effects of Ethambutol on the ocular zinc concentration in dogs. Am Rev
Resp Disease 1971; 104: 592-594.
5. Grant WM, Schuman lS: Toxicology o.fThe Eye 4th Edition; Charles C. Thomas, Springfield. 1993.
6. Kaiser lA, Paino AF. A one year study of the toxicity of ethambutol in dogs; results during life. Toxicol
Appl Pharmacol 1964; 6: 557-567.
7. Massa T., Davis G 1., Schiavo D. et.al. Tapetal changes in beagle dogs, [I. Ocular changes after intravenous
administration ofa macrolide antibiotic - Rosaramicin. Toxicol Appl Pharmacol.72: pp. 195-200. 1984.
8. Place VA, Peets EA, et al. Metabolic and special studies of ethambutol. Ann NY Acad Science 1966; 135:
775-795.
9. Sato S: Toxic effects on the visual system of diaminodipenoxybutane, quinine, and ethambutol in con-
scious dogs. Fundam Appl Toxicol1985; 5: 777-784.
10. Schiavo DM., Sinha DP., Black L. et al: Tapetal changes in beagle dogs. I. Ocular changes after oral ad-
ministration of a beta-adrenergic blocking agent- SCH 19927. Toxicol Appl Pharmacol 1984; 72: 187-194.
II. Schiavo DM. Retinopathy from administration of an imidazo quinazolin to beagles. Toxicol Appl Pharma-
col 1972;23: 782-783.
12. Wietzel G, Strecker Fl, et al. Zinc in the tapetum lucidum. Z Physiol Chern Hoppe-Seyler's 1954; 296:
19-30.
11
1. INTRODUCTION
• Address for Correspondence to: Dr. Dinesh Talwar, Assistant Professor, Dr. R.P. Centre for Ophthalmic
Sciences, All India Institute of Medical Sciences, New Delhi-I 10029, India.
evaluated color vision by Farnsworth-Munsel 100 hue test and visual evoked responses
(VERs) in patients on ethambutol in order to detect evidence of subclinical optic nerve
toxic effects, if present in patients on a daily dose of ethambutol of IS mg/kg body weight.
3. RESULTS
Of the total of 15 patients who were enrolled for the study, 9 were males and 6 were
females. The mean age of the patients was 26.06 years (range, 19 years to 42 years). None
of the patients were on any other potential neuro or ocular toxins. All patients had an in-
itial visual acuity of 6/6 in both eyes. All patients continued to retain the same visual acu-
ity during the subsequent follow-ups.
The mean latency of the pattern VER of the 15 patients two weeks after the initia-
tion of therapy was 112.15 msec. (range 80 msec. to 142 msec.). Three of the 15 patients
Do Therapeutic Doses of Ethambutol Cause Optic Nerve Toxicity? 99
(6 eyes) had an increased latency (i.e., greater than 120 msec.) at the initial follow up visit
(i.e., 2 weeks after the initiation of ethambutol therapy). Seven of the IS patients did not
turn up for a follow up 4 weeks later. Of the eight patients, who had a repeat evaluation of
VER after an interval of at least 4 weeks, the VER latency increased markedly in both
eyes of two of these patients (latency in these two patients had increased from a mean of
94 msec on I st follow up to a mean of IS4.17 msec.). The 3 patients, whose initial VER
latencies were high and who also came for follow up, were found to maintain similar
slightly high values of VER latency. As a result, the mean VER latency increased from
107A±IS.99 msec. at the initial visit to 129±27.7 msec. during follow up visits more than
four weeks later (in the seven patients for whom follow up data was available)
[p=0.0494]. Four to eight weeks following cessation of therapy in 4 patients (in whom
ethambutol therapy was stopped during the study), the VER was repeated. It was found
that the VER latency had decreased from 119.6±11A msec. to 113.7±10.69 msec. (which
was statistically significant [p<O.OS]).
F-M 100 hues total error scores of patients were also similarly tabulated and ana-
lyzed. The mean error scores of patients 2 weeks after the initiation of ethambutol was
2S2.84. Repeat evaluation of colour vision on a subsequent follow-up could be carried out
in 7 of the IS patients. The mean error scores of these seven patients had marginally in-
creased from 163.6±44.778 to 168.6±9S.83 (which was not statistically significant). In the
four patients in whom, 4 to 6 weeks following cessation of ethambutol therapy, the F-M
100 hue test was repeated, the error scores had decreased significantly from 273± 112.6 to
202.9±lS8.8 which was statistically significant (P=0.0237 on Wilcoxon single ranked
test).
On comparing the VER latency and the total error scores on Farnsworth-Munsell
100 hues test, there was no direct correlation of the increase in VER latency with the total
error scores found during the course of ethambutol therapy. However, there was definite
correlation of changes in VER latency with changes in total error scores on F-M 100 hues
test with a decrease in both the parameters following cessation of ethambutol therapy.
4. DISCUSSION
rence of onset of symptoms may be of great value in preventing extensive optic nerve
damage and consequently in preventing or minimizing the extent of vision loss. There is,
therefore, a need for screening procedures which are capable of detecting optic nerve tox-
icity in its subclinical stage before the onset of the non-reversible visual deficit. Among
the tests recognized to be of value in diagnosing subclinical optic nerve dysfunction, pat-
tern VER and color vision evaluation by F-M 100 hue test are most useful. We, therefore,
evaluated these parameters in our patients.
Most of the studies available on ethambutol toxicity have evaluated the visual acu-
ity, color vision and visual fields in their patients to detect drug toxicity earlier to enable
physicians to discontinue the drug if needed before the onset of irreversible visual loss.
The reversibility of the ocular toxic effects of ethambutol has been described pre-
viously.I3·14 However VER was not evaluated in either of these studies.
Few reports are available regarding the role of VER in the diagnosis of optic nerve
toxicity following the ethambutol therapy. Van Lith l5 and Bourquin and Korol '6 found that
flash VER may be considerably disturbed at a stage when there is relatively little evidence
of optic neuritis. Yiannikas et al observed increase latency of P 100 wave of pattern VER
in 6 out of 14 patients who were put on 15 to 25 mg/kg body weight. '2 In three out of
these six patients, VER latency returned back to normal after cessation of ethambutol.
They also found that 5 of these 6 patients had decreased visual acuity (in 4 patients in both
eyes, and in the remaining one only in one eye).
In our study also, 33.3% of patients (5 out of 15 patients) had an increased latency
during therapy with 15 mg/kg body weight of ethambutol but none of these patients had
either any subjective complaints or any effect on Snellen visual acuity. When the VER
was repeated 4-8 weeks following the cessation of therapy, the latency had recovered
from 119 .6± 11.4 msec.l to 113. 7± 10.69 msec. which was also statistically significant. The
color vision as assessed by F-M 100 hue test also revealed similar observations. While the
total error scores had not significantly increased on ethambutol, they were definitely
higher than the expected in the normal population. On stopping the drug, the total error
scores had decreased significantly from 273±112.6 to 202.9±158.8 confirming the revers-
ibility of these subclinical changes.
We thus conclude that in patients who are on ethambutol at the dose of 15 mg/kg
body weight, optic nerve toxicity can still occur in a significant number of patients though
the changes are subtle and remain in the subclinical stage. These changes are reversible af-
ter stopping the drug. Since only a few individuals on therapy with 15 mg/kg of ethambu-
tol develop overt signs of ethambutol toxicity, it is likely that some other factor, e.g.,
vitamin deficiencies and especially vitamin B 12, may have a role in the pathogenesis of
this condition. It has been previously postulated that ethambutol induced optic nerve tox-
icity may be caused by a mechanism similar to one responsible for tobacco-alcohol am-
blyopia which is attributed to Vitamin BI2 deficiency.s There is need for further research
to determine the relationship of the nutritional status to the likelihood of a deleterious re-
sponse to ethambutol.
It is, however, clear that the pattern VER and colour vision assessment by
Farnsworth-Munsell 100 hue test are more sensitive parameters for detecting this form of
optic neuritis. While there is still a need for a bigger study with a larger number of pa-
tients in whom the VER and colour vision values before, during and after stopping etham-
butol treatment has been evaluated to determine the likelihood of developing optic nerve
toxicity in an individual patient, we recommend these tests as routine screening proce-
dures, though they are more time consuming, in all patients who are on therapeutic doses
of ethambutol.
Do Therapeutic Doses of Ethambutol Cause Optic Nerve Toxicity? 101
5. REFERENCES
1. Doster B, Murray Fl, Newman R, Woolpert SF: Ethambutol in the initial treatment of pulmonary tubercu-
losis. U.S. Public Health Service tuberculosis therapy trials. Am Rev Respir Dis 1973; I 07: 177-190.
2. Yiannikas C, Walsh JC, McLeod JG: Visual evoked potential in the detection of subclinical optic toxic ef-
fects secondary to ethambutol. Arch Neurol 1983;40:645-648.
3. Citron KM, Thomas GO: Ocular toxicity from ethambutol (editorial). Thorax 1986;41 :737-739.
4. Corr RE, Henkind P: Ocular manifestations of ethambutol. Arch Ophthalmol 1962; 67(5):566.
5. Citron KM: Ethambutol: A review with special reference to ocular toxicity. Tubercle 1969;50:32.
6. Chatterjee VKK, Buchanan DR, Friedman AI, Green M: Ocular toxicity following ethambutol in standard
dosage. Br J Dis Chest 1986;80: 288-291.
7. Barron GJ, Tepper L, Iovine G: Ocular toxicity from ethambutol. Am J Ophthalmol 1974;77:256-260.
8. Roberts SM: Review of papers on the ocular toxicity of ethambutol hydrochloride, myambutol, an antitu-
berculous drug. Am J Optom Physiol Optics 1974;51 :987.
9. Fraunfelder FT, Grove JA: Antitubercular agents. In: Fraunfelder FT, Grove JA (Eds.), Drug Induced Ocu-
lar Side Effects, 4th edition, Williams & Wilkins, Baltimore, 1996, pp. 69-80.
10. Bobrowitz ID: Ethambutol-isoniazid vs streptomycin-ethambutol-isoniazid in original treatment of cavi-
tary tuberculosis. Am Rev Respir Dis 1974; I09:548-553.
II. Liebold JE: The ocular toxicity of ethambutol and its relation to dose. Ann NY Acad Sci
1966; 13 5:904-914.
12. Hopewell CP, Bloom BR: Tuberculosis and other mycobacterial diseases. In: Murray JF, Nadel JA (Eds.)
Textbook of Respiratory Medicine, 2nd edition, w.B. Saunders Company, Philadelphia 1994; pp.
1123-1124.
13. Aquinas M, Citron KM: Rifampicin, ethambutol and capreomycin in pulmonary tuberculosis previously
treated with both first and second line drugs The results of 2 years chemotherapy. Tubercle
1972;53: 15~5.
14. Hong Kong Tuberculosis Treatment Services, Brompton Hospital, British Medical Research Council: A
controlled clinical trial of daily and intermittent regimens ofrifampicin plus ethambutol in the retreatment
of patients with pulmonary tuberculosis in Hong Kong. Tubercle 1974;55: 1-27.
15. Van Lith GHM: Electrophthalmology and side-effects of drugs. Doc Ophthalmol 1977;44: 19-21.
16. Bourquin C, Korol S: Toxicite oculaire du Myambutol. Ophthalmologica 1976; 172:264-265.
12
NAmSA
Northwood, Ohio
For some years, the owl monkey was the animal of choice for evaluating the inflamma-
tory potential of visco elastics used in ocular surgery. The low viscosity of the vitreous body in
this species was felt to be a critical part of the test because it permits partial removal of vitreous
fluid by aspiration and replacement with the viscoelastic test material by injection. The growth
of the ocular viscoelastic products industry coupled with a critical shortage of owl monkeys
and the lack of an alternative in vitro assay forced investigators to examine other animal spe-
cies. The rabbit is probably the most frequently used alternative. Regulatory agencies now ac-
cept rabbit intravitreal tests for lot to lot evaluation of finished products, but little has been
published about typical white blood cell counts that should be expected. White blood cell
counts obtained from vitreous before treatment and after treatment with a clinically acceptable
viscoelastic will be presented. Various methods have been tried that include the injection of vis-
coelastic fluid without aspiration of vitreous, and the exchange of fluids at different volumes.
Regardless of the technique employed, and as long as the volume of vitreous removed or ex-
changed was 0.5 ml or less, consistent cell counts less than SO/mm3 were obtained.
16
(t) 14
E
~ 12 _.
=ai
~
C
::J 10 .
0
0
0
CD
:s:c 8·
<tI
QJ
:::!: 6·
4
2 3 4 5 6 7 8
Study
Figure 1. Baseline WBC counts consistenly low prior to treatment. No. of Eyes=1 0 I.
50.-------------------------------------~----,
45-·····__···_··_··········
(t)
E 40
__
",E
35
=ai
~
~
~ 15
~ 10
:::!:
5
O~r--r--.--.--.--.--.--.--.--.--.-~~.--,--.-~~
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Study
Figure 2. Direct vitreal injection of up to 0.2ml can be tolerated without prior aspiration. WBC counts remain
less than 50 cells. No. of Eyes=63 control, 192 test. Test=HA viscoelastic. Control=Balanced salt solution.
Testing of Ocular Viscoelastics by Injection into the Rabbit Vitreous 105
60
55
c;)
E 50
E
"'
'ijj
,£.
45
40
C
::J
0
35
0
0 30
[J:J
3: 25
c
III
Q) 20
~
15
10
2 3 4 5
Study
Figure 3. Traditional O.5ml aspiration followed by O.5ml dosing with flexible cannula did not show an increase
in the control eyes. Compilation of various treatments to test eyes produced acceptable counts as well. This is an
official release test. No. of Eyes=31 control, 27 test. Test=HA viscoelastic. Control=Balanced salt solution.
60
c;)
50
E
.§.
en
'ijj 40
,£.
C
::J
0 30
0
0
[J:J
20
3:
c
III
Q)
~
10
0
2 3 4 5 6 7 8 9 10 11 12 13
Study
Figure 4. Traditional O.Sml aspiration followed by O.Sml dosing with standard stationary syring/needle produced
consistently low WBC counts. No. ofEyes=77 control, 76 test. Test=HA viscoelastic. Control=Balanced salt solution.
106 P. J. Upman et aL
60.-------------------------------------------.
W50
~
!c
40 ..·..............·..............·
:3u
30 ......................·
In 20 ..............................
3:
~
~ 10
OL-.----.----.----.----,----r----.----r----~
2 3 4 5 6 7 8 9
Study
Figure 5. Screening assay with 0.2ml aspirated and 0.2ml dosed. WBC counts consistently low. At this volume.
the aspiration step may be omitted. No. of Eyes=51 control. 50 test; Test=Cellulose based viscoelastic; Con-
trol=Balanced salt soultion.
2 3 4
Method
Figure 6. I = 0.2ml dosed/no aspiration. 2 = 0.5ml aspirated/0.5ml dosed, flex connector. 3 = 0.5ml aspi-
rated/0.5ml dosed. 4 = 0.2ml aspirated/0.2ml dosed.
13
!Department of Ophthalmology
2Institute of Neurophysiology
University of Cologne
D-50931 Koln, Germany
1. INTRODUCTION
indicator of the integrity of the photoreceptors, the first retinal synapse and higher retinal
neurons.
We studied the effects of acyclovir, ganciclovir and foscamet on the parameters of
the electroretinogram in the isolated perfused bovine retina. Ganciclovir is a nucleoside
analog of acyclovir with an additional 3'carbon and hydroxyl group. These antiviral
agents are potent inhibitors of the herpes group viruses.
To distinguish drug effects on the photoreceptors and effects on the postsynaptic bi-
polar layer, we studied the influence of the antiviral drugs on the a-wave amplitude. The
retina was perfused with the standard perfusate and stimulated repeatedly until the ERG
amplitudes showed a steady-state. Then aspartate [1 mMolJl] was added to the solution
and the responses to the standardized flashes were observed. As aspartate is an inhibitor of
the synaptic transmission on the level of the first retinal synapse the b-wave was sup-
pressed making it possible to record unmasked a-wave amplitudes. Thereupon the antivi-
ral drug was added to the aspartate-standard medium for 45 minutes and light responses
were recorded at five minutes intervals. Then the retina was reperfused with the aspartate-
standard solution for 45 minutes. Finally the preparation was perfused with standard me-
dium to acertain that the effects on a- and b-waves were reversible. The percentage of
a-wave amplitude reduction was calculated.
The antiviral drugs were added to the perfusate in the following concentrations:
Acyclovir: 0.3,0.9,3,9 mMolJI
Ganciclovir: 0.38,0.8, 1.2,3.8 mMolJl
Foscarnet: 0.001,0.003,0.01,0.03 mMolJl.
3. RESULTS
3.1. Acyclovir
A concentration of 0.3 mMolJl acyclovir did not affect the b-wave of the ERG dur-
ing 45 minutes of drug application. Concentrations of 0.9 mMolll or higher were found to
reduce the b-wave amplitude. Time course and percentage of b-wave reduction correlated
with the concentration of acyclovir. Acyclovir in a concentration of 9 mMolJl reduced the
b-wave completely. These effects on the b-wave amplitude proved to be reversible within
the recovery time (Figure 2). The a-wave amplitude was not changed by the applied con-
centrations between 0.3 to 9 mMoll1.
Acyclovir
~2 ~ -",A. "JV'~"'-~
Gl
~ '0,'>.. ~"'~l>.-("'/T~
,
"0
o
:§ 15
~O.3mMo.
Q. ~ 0
E
01 '0-0....<:>,,<:1
~ 1
~o-
~
.c
5 \ - 0
-<>- 0.9 mMoli1
\ / -0- 3mMoin
0 0-0-0-0-0-0-0 -0- 9 mMoln
0 20 40 60 80 100
time [min]
Figure 2. B-wave amplitude of the isolated bovine retina during and after perfusion with acyclovir in different
concentrations.
110 C. Luke et al.
r_____ f\
standard medium
~~----------
sta~ard medium +
2c10Vir [3 mMoln)
V~ ~ standard medium
after 25 min
~I L~~~--,-----,----~
0,0
stimulus
0,5 1,0
standard medium
1,5
after 45 mi
2,0
time[s)
Figure 3. ERGs of a single isolated human retina before, during and after application of acyclovir.
We had the opportunity to study the effects of acyclovir on a single isolated human
retina that could be prepared from a globe enucleated during surgery of a retroorbital tu-
mor. In an earlier investigation of an anamnestically identical case 2 both the in situ elec-
troretinogram before the operation and the electroretinogram from the enucleated
superfused preparation exhibited identical near normal responses. The acyclovir concen-
trations reducing the b-wave amplitude of the isolated human retina were found to coin-
cide with the concentration range that showed toxic effects in the isolated bovine retina
(Figure 3).
3.2. Ganciclovir
During application of ganciclovir in a concentration of 0.38 mMoll1 the electroreti-
nogram remained unchanged. Concentrations of 0.8 mMoll1 and higher led to a dose-de-
pendent b-wave reduction. During reperfusion with standard solution the b-wave showed a
complete revovery. There was no effect on the a-wave amplitude within the examined
concentration range. The toxic concentrations of acyclovir and ganciclovir are similar as it
is shown in Figure 4.
3.3. Foscarnet
Effects on the b-wave amplitude were found at lower molar concentrations for fos-
camet compared with acyclovir and ganciclovir. Concentrations of 0.003 mMolll or less
did not show any effect on the b-wave. Higher concentrations of 0.01 mMol/1 or more re-
duced the b-wave amplitude (Figure 4). The reduction of the b-wave was dose-dependent.
Application of foscarnet in a concentration of 0.03 mMoll1 led to a complete b-wave re-
duction. The a-wave remained unaffected by the applied concentrations. All effects on the
electroretinogram were reversible within the reperfusion time of 45 minutes.
Effects of Antiviral Agents on Retinal Function in Vertebrate Retina 111
10
80
:gc 60
:::l
"0
~
CIl 40
>
m
~
.c
20
" Ganciclovir
• Acyclovir
o • Foscarnet
0.01 0.1 10
concentration [mMol!11
Figure 4. Dose-response curves of the antiviral agents on the b-wave amplitude in the isolated perfused bovine
retina. Average ± standard deviation of representative drug series.
4. CONCLUSIONS
The aim of our study was to evaluate safe concentrations of common antiviral agents
for intravitreal injection and vitrectomy infusion fluid using a suitable test system. The
electroretinogram reflects photoreceptor and higher order neuron activity allowing the ef-
fects of potentially toxic drugs on retinal function to be monitored. One advantage of the
model is the continuous application of a drug in precisely known concentrations.
Our experiments revealed toxic effects of acyclovir in concentrations of 0.9 mMoll1
and higher. This concentration is about five times higher than the concentration recom-
mended for vitrectomy infusion solution (0.17 mMol1l = 40 Jlg/ml)]. The ERG showed
toxic effects only at concentrations distinctly exceeding the ID50 (inhibitory doses) of acy-
clovir against HSV (3.6 IlMolll) and CMV (0.15 mMol/l)4 . The single human retina that
could be used in our study showed toxic effects of acyclovir in a concentration range that
is comparable with the concentrations found toxic in the isolated bovine retina. The coin-
cidence of effective concentrations in bovine and human retina would seem to qualify the
experimental approach for toxicological statements in human.
Ganciclovir is mainly used for treatment of cytomegalovirus retinitis. Toxic effects
of ganciclovir were found in concentrations of 0.8 mMol/1 and higher. Kao et a1. 5 reported
toxic effects on the ERG in rabbits for ganciclovir in vitrectomy infusion fluid with con-
centrations exceeding 0.13 mMolll (30 Jlg/ml). This difference may be due to species dif-
ferences. The ID50 of ganciclovir against CMV is 7.2 IlMolll6 which is a hundred times less
than the toxic concentrations reported here.
Foscarnet showed toxic effects in concentrations of 0.01 mMolll and higher whereas
the IDso of foscarnet against sensitive CMV strains vary between 0.015 and 0.3 mMollI 7•8 •
Therefore we do not exclude toxic side effects of foscarnet on retinal function within the
therapeutic concentration range. Intraocular application of foscarnet should be limited to
cases with ganciclovir resistant CMV -retinitis. Patients receiving intraocular foscarnet
should be followed up at short intervals.
The experiments on aspartate preincubated retinas showed that the photoreceptor
layer is not affected. Therefore we attribute the toxic effects to postsynaptic interaction of
the antiviral drugs.
112 C. Liike et al.
5. REFERENCES
I. Sickel W. Retinal metabolism in dark and light. In: Fuortes MGF, ed. Handbook of Sensory Physiology
VoI.VIII2, Berlin, Heidelberg, New York, Springer, 1972,667-727.
2. Haschke W, Sickel W. Das Elektroretinogramm des Menschen bei Ausfall der Ganglienzellen und partieller
Schiidigung der Bipolaren. Acta Ophthalmol 1962; Supplement 70: 164-167.
3. Small GH, Peyman GA, Srinivasan A, Smith RT, Fiscella R. Retinal toxicity of combination antiviral drugs
in an animal model. Can J Ophthalmol 1987; 22:300-303.
4. Smee DF, Martin JC, Verheyden JPH et al. Anti-herpes virus activity of the acyclic nucleoside 9-( 1,3-dihy-
droxy-2-propoxy-methyl) guanine. Antimicrobial Agents Chemother 1983; 23: 676--682.
5. Kao GW, Peyman GA, Fiscella R, House B. Retinal toxicity of ganciclovir in vitrectomy infusion solution.
Retina 1987; 7: 80-83.
6. Tocci MJ, Livell TJ, Perry HC et al. Effects of the nucleoside analog 2'-Nov-2' Deoxyguanosine on human
cytomegalovirus replication. Antimicrobial Agents Chemother 1984; 25: 247-252.
7. Jabs DA, Dunn JP, Enger C, Forman M, Bressler P, Charache P. Cytomegalovirus retinitis and viral resis-
tance. Arch Ophthalmol1996; 114:809--814.
8, Dunn JP, MacCumber MW, Forman MS, Carache P, Apuzzo L, Jabs DA. Viral sensitivity testing in patients
with cytomegalovirus retinitis clinically resistant to foscamet or ganciclovir. Am J Ophthalmol 1995;
119:587-596,
14
Department of Ophthalmology
University of Cologne
Joseph Stelzmann Str.9
D-5093I Cologne, Germany
1. INTRODUCTION
SIGNAl·
VlDEO~HIP CONVERTER
I PLANT: RECEIVER
AND SnMULAnNG
ELECTRODES
RenNA·
ENCODER
2.1. Implant
The implant consists of a soft carrier of silicone rubber and a foil of silicium embed-
ded in the carrier (Fig.2). The weight of the device is 9.4 mg. The implants were designed
and produced in the Fraunhofer Institute for Microelectronic Circuits and Systems in
Duisburg, Germany.
'--..L._ _ _~_~__ "'_ _ _ _...... _"""_.L;,I silicium foil embedded in the rubber. The overall
size is 6.5 mm x 5 mm.
Experimental Implantation of Devices for Electrical Retinal Stimulation in Rabbits 115
2.3. Electrophysiology
The electroretinograms (ERG) were recorded after a dark adaptation period of 30
minutes with the pupil maximally dilated. The rabbits were sedated with an i.m. injection
of ketamine 12.5 mg/kg body weight and xylazine 3.2 mg/kg body weight. An at 1 Hz
flashing white Xenon tube Ganzfeld stimulus was used with a luminance of 0.6 cds/m 2
and a duration of 15 IlS. The recordings were obtained with a corneal electrode (Type
'Henke', Medical Workshop, Groningen, The Netherlands) and a reference electrode at
the forehead. Both connected earlobes served as ground. Visual evoked potentials were re-
corded simultaneously with the ERG using the same stimulus. EEG surface electrodes
were placed with adhesive paste on the skull's surface with the active electrode at the oc-
cipital pole and the reference electrode at the forehead. The amplifier's bandpass was set
between 2 and 100 Hz. 32 responses for both ERG and VEP were averaged and the data
was processed with a digital signal acquisition and analysis system.
3. RESULTS
3.1. Anatomy
The primary procedure (lensectomy, vitrectomy, endolaser) was successful in 11114
cases. As intra- and postoperative complications a retinal detachment due to a dialysis, a
severe endophthalmitis, and a PVR-detachment occurred. The second procedure was char-
acterized by a higher rate of intra- and postoperative complications: in two cases a dialysis
with a complete retinal detachment occurred. In one case we had a massive choroidal
bleeding which was untreatable. In seven rabbits we implanted the stimulator onto the
retinal surface. In one case the implant dislocated into the anterior chamber at the first
postoperative day. In the remaining six rabbits the implant's position was stable for one
week. The three implants not covered with fibrin glue dislocated after 1~2 weeks into the
116 P. Walter et al.
Figure 3. The implant dislocated in the inferior vitreous cavity three weeks after implantation.
inferior half of the vitreous cavity (Fig.3). The others were still in their original position
embedded in a fibrous network (FigA). However, the follow-up period of these cases was
between 2 and 4 weeks. The eye with the primary dislocated implant found in the anterior
chamber developed a traction retinal detachment. In the other six rabbits the retina is still
attached (follow-up: 2-20 weeks after implantation). The papilla in the two rabbits with a
follow-up of20 weeks after implantation showed a narrowing of the optic nerve vessels.
3.2. Electrophysiology
The ERG revealed no signs of toxicity in the successful implanted rabbits with the
retina being attached (Fig. 5).
The ratio of the b-waves between the operated and the not operated eye were around
1.0 (Fig. 6). The eyes with intra- or postoperative complications immediately showed a
considerable decrease in the b-wave amplitude.
However, the YEP although normal after a short interval after the implantation
(Fig. 7) disclosed a considerable reduction of the P2 component after 10 weeks of implan-
tation (Fig.S and 9).
Figure 4. The implant located on the visual streak three weeks after implantation with fibrin glue (Tissue-Col®)
and fluid-air exchange. Note the fibrous membranes in which the implant is embedded. The retina is still attached.
Experimental Implantation of Devices for Electrical Retinal Stimulation in Rabbits 117
IJaYJ IJaYJ
27060 OS 2706000
1110 1110
100 100
110 110
0 0
.aJ .aJ
0 110 100 1110 2IlII [mil 0 110 100 1110 2IlII [mil
Figure 5. Scotopic electroretinograms (0.6 cds/m2) recorded from the left operated eye (left) and from the right
eye (right) in one rabbit 12 weeks after implantation of the device.
3,5~-------------------------------------------------'
3,0
2,5
2,0
1,5
0,5
Figure 6. Time course of the ratio between the b-wave amplitudes of the scotopic electroretinogram of the oper-
ated eye and the unoperated fellow eye. Arrows indicating the surgical procedures: OPI, lensectomy. vitrectomy,
endolaser; OP2, re-vitrectomy, implantation of the device. Horizontal line at y = 1.0 indicating same b-wave am-
plitudes in operated and unoperated eyes of one individual animal.
10 10
o o
-10 -10
Figure 7. Visual evoked potentials (VEP) after flash stimulation in one rabbit six weeks after implantation. Ar-
rows indicating the P2 component of the VEP. Left, operated eye; right, unoperated fellow eye.
1t8 P. Walteretal.
~v.-------------------------------------------~
30
20
10
OP1 OP2
i 5 10 15 weeks
Figure 8. Time course of P2 amplitudes of flash evoked cortical potentials (YEP) after stimulation of the oper-
ated eye. OP I and OP2 indicating the surgical procedures.
4. DISCUSSION
The development of a neuroprosthesis for the blind providing these patients with some
kind of visual perception has been a challenge for researchers, ophthalmologists and techni-
3,0~-------------------------------------------'
2,5 /
2,0 ~
1,5
1,0 -Hr*-~:-""':~-::::==--oo::::::-~'-----=~----------I
0,5
weeks
OP1 OP2
Figure 9. Time course of the ratio between the P2 amplitude of the flash evoked cortical potential (YEP) after
stimulation of the operated eye and the un operated fellow eye. Arrows at OPI and OP2 indicating the surgical
procedures. Horizontal line at y = 1.0 indicating same P2 amplitudes in operated and unoperated eyes of one indi-
vidual animal.
Experimental Implantation of Devices for Electrical Retinal Stimulation in Rabbits 119
cians over the years. Mechanical devices working with tactile stimulationS had been discussed
as well as cortical6•7 and optic nerve stimulators8 . In recent years with the progress in vitre-
oretinal surgery electrical stimulation at more distal parts of the visual pathway i.e. at the
level of retinal neurons gained more interest. The general feasability of a technical retinal
prosthesis using today available high tech resources has been shown earlier9 . Mark Humayun
and colleagues demonstrated that electric stimulation of the inner retinal surface may lead to
visual perception in blind RP patients I.
A multielectrode stimulation array implanted onto the retinal surface maybe a possible con-
cept to supply optic nerve fibers with electrical information related to the visual world yielding to
some kind of spatial resolution in blind patients. Of course, at the moment there are still a lot more
questions than answers: is it possible to implant such a device into an animal's eye, and if so, does
the implant remains at its original position which is necessary to ensure a stable electrode-axon
relation. Does the animal's eye tolerate the implant and its materials? The very preliminary results
shown here indicate that we can give answers to a few ofthese questions. The technical advances
in vitreoretinal surgery ensure that the implantation of devices for electrical epiretinal stimulation
onto the retinal surface of rabbits is possible although the anatomical differences compared to the
human eye cause a number of surgical problems, e.g. the necessity oflensectomy due to it's thick-
ness and diameter. Surgical fixation procedures such as fluid-air exchange and use of fibrin glue
or others are necessary to achieve a better position stability of the stimulator than just putting it
onto the retina. The ERG data disclosed that the device used here seemed to be non-toxic at the
level of photo receptors and the postsynaptic retinal network even 20 weeks after implantation.
However, the YEP data showed, that after 10 weeks of implantation the optic nerve function de-
creased, indicating a toxic effect of the device on ganglion cells or its axons. This finding con-
firmed the clinical impression of narrowed vessels at the papilla 20 weeks after implantation of
the stimulator.
These data indicate, that for further experiments a modified silicone for the carrier and a
modified material for the microcontact foil should be used or to encapsulate them in a more
biocompatible material exhibiting similar mechanical features.
Taking all these positive and negative experiences together, we still believe that the
development of a retinal prosthesis for the treatment of progressive retinal degenerations,
for which we today have no treatment available, is still feasable and worthwhile to con-
tinue with the development.
5. ACKNOWLEDGMENTS
This work was done in collaboration with the German "EPI-RET" research team and
funded by the German Minister for Education, Science, Research, and Technology (01 IN
50 I K). The animal experiments were done in accordance with institutional guidelines for
animal welfare and according to the principles for laboratory animal care.
6. REFERENCES
I. Humayun MS, de Juan E, Dagnelie G, Greenberg RJ, Propst RH, Phillips H. Visual perception elicited by
electrical stimulation of the retina in blind humans. Arch Ophthalmol1996; 114:40-46.
2. Stone JL, Barlow WE, Humayun MS, de Juan E, Milam AH. Morphometric analysis of macular photore-
ceptors and ganglion cells in retinas with retinitis pigmentosa. Arch Ophthalmol 1992; 110: 1634-1639.
3. Eckmiller R. Electronic simulation of the vertebrate retina. IEEE Trans Biomed Eng 1997; 22(4): 305-311.
4. Eckmiller R. Towards Retina Implants for Improvement of Vision in Human with RP - Challenges and first
results. Proc WCNN'95, INNS Press, Lawrence Earlbaum Assoc., Hillsdale 1995, Vol. I, pp.228-233.
120 P. Walter et at.
5. Bach-y-Rita P, Collins CC, Saunders FA, White B, Scadden L. Vision substitute by tactile image projec-
tion. Nature 1969;221 :963-964.
6. Brindley GS, Lewin WS. The sensations produced by electrical stimulation of the visual cortex. J Physiol
(Lond) 1968; 196:47~93.
7. Dobelle WH, Mladejovsky MG,Girvin JP. Artificial vision for the blind: electrical stimulation of the cortex
offers hope for a functional prosthesis. Science 1978;183: 1-39.
8. Shandurina AN, Lyskov EB. Evoked potentials to contact electrical stimulation of the optic nerves. Hum
Physiol 1986; 12:9-16.
9. Eckmiller R et at. Neurotechnologie Report 1975, Bundesministerium fur Forschung und Technik.
15
1. INTRODUCTION
Nitric oxide (NO) is a highly reactive radical which is produced by many cell types.
The amino acid arginine is converted to citrulline by a specific enzyme, NO synthase
(NOS), with the release of NO. Even though NO is important to physiological functions
such as neurotransmission and regulation of blood pressure, its overproduction has been
implicated in pathophysiological conditions. NO is freely diffusible in the cell and is not
stored. When released into the bloodstream, NO is quickly bound to hemoglobin and con-
verted to nitrate which is eliminated in the urine. Intense research over the past few years
has uncovered a wide variety of NO-mediated biologic activities, both beneficial and dele-
terious. Differential competitive inhibition of isoforms of NOS by N G -substituted L-ar-
ginine analogues allows selective intervention to inhibit production of NO as well as to
further determine its functions.
Two of the enzymes (neuronal and endothelial) are constitutively expressed (cNOS),
and mediate numerous physiological processes, such as vasodilation and neurotransmis-
sion 6•7 (Table I). Small amounts of NO are generated by these two Ca2+/calmodulin-de-
pendent cNOS enzymes. 8.9 NO activates soluble guanylate cyclase leading to synthesis of
cGMP and subsequently relaxation of vascular smooth muscle cells and neuronal commu-
nication in the central nervous system. 10
Inducible NOS (iNOS) produces sustained, high levels of NO over longer periods
which can be cytotoxiC. I !.12 Inducible NOS is not dependent on Ca 2+ but contains cal-
modulin tightly bound to each subunit of the enzyme, resulting in permanent activation. In
contrast to cNOS, iNOS is synthesized by many immunocompetent cells, such as macro-
phages J3 · 14 and neutrophils,15.16 and diverse cell types including hepatocytes,17 cardiac
myocytes,18 fibroblasts,19 and vascular smooth muscle cells. 20 Resting cells do not express
iNOS, but nearly all cells have the capacity to express it after challenge with immunologic
or inflammatory stimuli. I I After stimulation, NO production is delayed, usually two to
eight hours, as the iNOS enzyme and possibly essential cofactors are synthesized. 21 Over-
production of iNOS-generated NO has been implicated in the pathogenesis of inflamma-
tory diseases, such as septic shock. 22- 25 Increased levels of NO are found in inflammatory
conditions, such as arthritis and ulcerative colitis, and many inflammatory symptoms are
reversed by NOS inhibitors. Inflammatory stimuli can generate production of iNOS and
NO in a variety of tissues. For example, administration of endotoxin induces hepatotoxic-
ity in rats which is associated with increased NO. 26 Also, there may be a strong synergy
between the stimulating agents which is both species- and cell-dependent manner. For ex-
ample, lipopolysaccharide (LPS) and interferon-gamma «yIFN) induce maximal levels of
NO in murine macrophages whereas human hepatocytes require a cocktail of LPS, yIFN,
TNF-a, and IL-I P for a similar induction to occur.
tors have been instrumental in investigating the role of NO in physiological and patho-
physiological processes. Development of more selective inhibitors will be required in or-
der to reduce excessive NO production indicative of many pathophysiological conditions
while sparing NO production during normal physiological processes (eNOS regulated).
Ideally, such inhibition should reduce excessive NO production but not interfere with the
important physiological actions of constitutive NO release.
4. ACTIONS OF NO
Both beneficial and deleterious consequences have been attributed to NO.34 For ex-
ample, NO exhibits antimicrobial and antitumor activities in the immune system and aids
blood flow and peristalsis in the gut. On the other hand, NO contributes to inflammation
(erythema, vascular leakiness) and tissue damage (respiratory burst) in several organ sys-
tems (Table II).
The mechanism(s} by which reactive nitrogen metabolites cause tissue injury during
inflammation are still unclear. NO acts as a signaling molecule by activating soluble
guanylate cyclase resulting in an increase in cellular cGMP levels. Increased levels of
cGMP mediates relaxation of smooth muscle, neutrophil chemotaxis, and inhibition of
platelet aggregation. NO exerts its tumoricidal and microbicidal activities by inhibition of
protein synthesis and cellular respiration by inactivation of certain enzymes such as aconi-
tase of the Krebs cycle and NADPH/Succinate:ubiquinone oxidoreductase of the electron
transport chain. NO can also react rapidly with superoxide anion to form peroxynitrite, a
strong oxidative molecule that can react with several biological molecules. 35 Peroxynitrite
is a stronger oxidant than NO, and may account for much of the toxicity previously associ-
ated with NO. After protonation at biologic pH, peroxynitrite can decompose to an even
more potent tissue-damaging hydroxyl-like radical and nitrogen dioxide. 35
5. REGULATION OF NO LEVELS
There are potentially many sites for regulation of NOS function, such as limiting the
activity or availability of cofactors, inducing inhibitors, altering transcription, increasing
degradation of mRNA, and impairing translation. 4 For example, since arginine is the pre-
cursor of NO, limiting this substrate in vitro (with special media preparations) or in vivo
(such as occurs in wounds where macrophages release arginase) would decrease synthesis
of NO. Another possible regulatory pathway involving an endogenous hormone-like mole-
124 J. B. Allen et al.
cule is suppression of iNOS expression by TGF - p. Studies using murine elicited perito-
neal macrophages show that TGF-p reduced iNOS expression by decreasing stability and
translation of iNOS mRNA and by increasing degradation of iNOS protein. 36 Other studies
suggest that TGF-p-dependent suppression of induced murine macrophage NO production
by LPS correlates with a direct effect on iNOS gene expression. 37 In addition, Ding et al.
have shown that TGF-p 1, -P2, and -P3 block yIFN-induced production of NO by mouse
peritoneal macrophages. 38 Since TGF-p is produced by virtually every cell, these three
studies underscore the important wide-spread regulatory actions of TGF-p on iNOS ex-
pression and function. In addition, transcription of the iNOS gene is under control of the
transcription factor NF-kB; therefore, inhibitors of NF-kB activation would provide an-
other means of altering iNOS levels.
Several assays are currently being used to detect NOS activity. Depending on the
particular NOS isoform under investigation, certain criteria for the assay must be consid-
ered, such as the presence! absence of Mg2+ required for iNOS activity, the inclusion of
Ca 2 +!calmodulin, addition of F AD, and the concentration of arginine. The synthesis of NO
activity can be determined spectrophotometrically by taking advantage of the reaction be-
tween NO and oxyhemoglobin to form methemoglobin and measuring the absorbance at
401 nm.J9 In an alternate assay, the amount of radio labelled-citrulline converted from radi-
olabelled L-arginine can be determined by utilizing a cation-exchange column.40 One of
the most frequently used methods is the spectrophotometric Greiss assay. A magenta-col-
ored azo dye is formed from the reaction of nitrite, a stable end product of NO metabo-
lism, and the Greiss reagent (equal volumes of 0.1 % naphthylethylenediamine
dihydrochloride and 1% sulfanilamide in 5% phosphoric acid);41 this reading is compared
to a sodium nitrite standard curve. Since nitrate is preferentially formed under physiologi-
cal conditions where superoxide is present and reacts with NO, it may be necessary to
convert the nitrate to nitrite with nitrite reductase before measurement in the Greiss reac-
tion. In addition, NOS activity has been colocalized with NADPH diaphorase staining. 42
Furthermore, immunohistochemical analysis with antibodies specific for either the cNOS
or iNOS isoforms not only allows for identification of the protein, but also localization of
expression. 42
disrupted resulting in protein extravasation and cellular infiltration into the aqueous hu-
mor. Studies of the pathogenesis of uveitis from human uveitis and animal models of ocu-
lar inflammation suggest that cytokines play a major role. Increased cytokine
(interleukin-l p, tumor necrosis factor-a, and interleukin-6) levels are found in inflamed
ocular fluids and tissues when compared to normal. 48 - 50 In addition, intraocularly injected
cytokines induce uveitis. 47.51 Although cytokines are considered to be primary mediators
of LPS-induced inflammation, treatment with anti-TNFa antibodies 52 or soluble IL-l re-
ceptors 53 has not inhibited LPS-induced uveitis. Cytokines can induce and/or modulate the
expression of other cytokines through a complex network of interactions; therefore inhib-
iting only one portion of the network may not be an effective approach. As mentioned
above, iNOS is induced by LPS and certain inflammatory cytokines, making LPS-induced
uveitis a good model system for the study of NO in inflammation.
LPS-induced ocular inflammation in rabbits is induced by an intravitreal injection of
LPS (lOng) into one eye of New Zealand white male rabbits (2-3 kg) and the appropriate
vehicle into the contralateral eye, which serves as a matched control. 46 Evidence of uveitis
is monitored by direct examination of the anterior segment with a slit-lamp biomicro-
scope. This method of clinical observation at various time intervals allows temporal as-
sessment of the inflammation such that the appropriate time for termination of the
experiments can be determined. 43 Entry of protein through disrupted barriers (flare) is
graded from 0-4, with 0 being normal and 4 being maximal light diffraction. 54 In addition
to aqueous flare (AF), iridal hyperemia (IH; a measure of iridal blood vessel dilation in
the iris-ciliary body) is also assessed. The more subjective slit-lamp observations can be
quantitatively confirmed at the experiments' termination by analyzing aliquots of aqueous
humor for protein content and presence of inflammatory cells. 46 In addition, at the termi-
nation of the experiments, the eyes can be removed and processed for histologic/molecular
analyses. Hematoxylin and eosin-stained sections of eye, removed 18-24 hours after LPS
injection, show a predominantly polymorphonuclear infiltrate with some mononuclear
cells and a proteinaceous exudate in the anterior chamber.
Several publications in the past three years evaluated the importance of NO in ocu-
lar inflammation/ 5.55-57 as shown in Table III. Parks et al. found that intraperitoneal ad-
ministration of L-NAME, an inhibitor of both cNOS and iNOS, before and after
subcutaneous injection of LPS in rats significantly decreased aqueous protein and total
cell numbers in the anterior chamber. 58 Tilton et al. induced uveitis in rats by footpad in-
jection of LPS and found that multiple injections of amino guanidine, a relatively selective
iNOS inhibitor, reversed LPS-increased plasma nitrate levels and cellular infiltration into
the anterior uvea. 57 Using the same rat model, Mandai et al. administered N G -nitro-L-ar-
ginine (L-NNA), a relatively selective cNOS inhibitor, at the same time as LPS and also
six hours later, and saw a decrease in aqueous protein, total cell numbers, and iNOS activ-
ity.55 In another study, they compared the effects ofL-NNA (used in earlier study) and N-
iminoethyl I-ornithine (L-NIO), a relatively selective iNOS inhibitor, in the same rat
model of LPS-induced uveitis and correlated the inhibition of uveitis with the inhibition of
cNOS. 56 Wang and Hakanson induced ocular inflammation in rabbits by infrared irradia-
tion of the iris, and showed that intravenous administration of L-NAME prior to irradia-
tion reduced ocular inflammation, probably by inhibiting the NO induced release of
transmitters/mediators, such as calcitonin gene-related peptide, from the C-fibres. 59 Bellot
126 J. B. Allen et al.
et al. found that L-NAME inhibited LPS-induced uveitis in rats as well as levels of TNF
and NO in the intraocular fluids. 60 In an additional study, they inhibited LPS-induced
uveitis in rabbits by inhibiting both NOS and cyclooxygenase (COX) inflammatory path-
ways; the effects of these inhibitors were additive. 61 Goureau et al. showed that increased
iNOS mRNA expression in the iris/ciliary body and retina and aqueous NO production are
important in LPS-induced uveitis in rats, and that L-NAME decreased uveitis and aqueous
NO levels. 62 In all of these studies, uveitis was induced in rats or rabbits and treated with a
variety of NOS inhibitors with differential specificities for iNOS and cNOS. In addition,
different routes of administration were utilized. All of the studies concluded that inhibi-
tion of NOS decreased ocular inflammation. Therefore it is clear that NO is an important
mediator of anterior uveitis in these models. It is likely then that NO is also involved in
human uveitis and NOS inhibitors are potential therapeutic compounds for treatment of
this disease.
In additional studies, we used relatively selective cNOS and iNOS inhibitors to dis-
sect the contribution of each NOS isoform in intraocular inflammation. 63 Rabbits were in-
jected in one eye with LPS (lOng) and endotoxin-free saline in the contralateral eye.
Nitric Oxide in Ocular Inflammation 127
Either at the same time or various times after LPS injections, rabbits also received in-
travitreal injections of L-NAME (10 and 100 mM), D-NAME (100 mM), or AG (114
mM). Twenty-four hours after LPS, aqueous flare and iridal hyperemia were assessed by
slit lamp. As seen in Table IV, eyes injected with LPS exhibited a marked disruption of
the blood-aqueous barrier (AF), and dilation of iridal blood vessels (IH). Histopathologi-
cal analysis confirmed thes!;: observations (data not shown). Uveitis in rabbits coinjected
with LPS and L-NAME (100 mM) was almost completely abolished. As a control, D-
NAME, the inactive enantiomer, was coinjected with LPS and did not suppress uveitis. No
inflammation was observed in control eyes receiving coinjections of saline and L-NAME
or D-NAME. Using a more objective measurement of ocular inflammation, aqueous hu-
mors were analyzed for protein content and total cell count. Coinjection of LPS and L-
NAME reduced the number of inflammatory cells entering the aqueous humor, in a
dose-related manner. In eyes co-injected with D-NAME and LPS, there was no significant
decrease in cell number when compared to eyes injected with LPS alone. In addition, the
protein content in aqueous humors from eyes coinjected with LPS and L-NAME, but not
D-NAME, was significantly reduced.
We then co-injected LPS and AG, a more selective inhibitor of iNOS, to determine
the importance of this isoform in the uveitic response. Rabbits co-injected with AG (114
mM) and LPS (lOng) were observed for changes in anterior inflammation. AG signifi-
cantly suppressed all parameters of LPS-induced ocular inflammation (Table IV), suggest-
ing the importance of iNOS in ocular inflammation.
Do NOS inhibitors have an effect on LPS-induced inflammation that is already in
progress? To answer this question, L-NAME (100 mM) was injected 0.5 hr and 6 hr after
LPS administration. L-NAME still inhibited uveitis when injected 0.5 hr after LPS. 63
However, when injected 6 hr after LPS, L-NAME did not reduce LPS-induced uveitis (Ta-
ble V). In fact, injection of L-NAME at this time point after LPS significantly increased
cellular infiltration. In contrast, injection of AG 6 hr after LPS did inhibit cellular infiltra-
tion and protein extravasation.
To determine whether the reduced cellular infiltration into the anterior uvea is due to
a direct effect of L-NAME on leukocyte migration, rabbit peripheral blood polymor-
phonuclear leukocytes were incubated, in vitro, with and without L-NAME (1 mM) for
four hours at 3rc and assessed for their ability to respond to a positive chemotactic
stimulus. The chemotaxis assay was carried out in a 48-well microchamber plate (Neuro-
probe, Inc) with endotoxin-activated rabbit serum (C5a).64 L-NAME-treated cells were
found to migrate much less in response to C5a than untreated cells (data not shown).
128 J. B. Allen et al.
9. CONCLUSIONS
The studies of uveitis described above underscore the importance of NO and its me~
tabolites in ocular physiology and pathophysiology. As our knowledge of the role of NO
in diseases increases, development of even more selective inhibitors will be necessary in
Nitric Oxide in Ocular Inflammation 129
order to distinguish the involvement of the different NOS isoforms, thus allowing for con-
trol of harmful, induced NO production, while not affecting the physiological actions of
constitutive NO release.
to. REFERENCES
I. Marietta MA. Nitric oxide synthase structure and mechanism. J Bioi Chem 1993;268: 12231-12234.
2. Steuhr OJ, Griffiths Ow. Mammalian nitric oxide synthases. Adv Enzymol Mol Relat 1992;65:287-346.
3. Moncada S, Palmer RM. Higgs EA. Nitric oxide: physiology, pathophysiology, and pharmacology. Phar-
macal Rev 1991;43:109-142.
4. Nathan C, Xie Q- W. Regulation of biosynthesis of nitric oxide. J Bioi Chem 1994;269: 13725--13728.
5. Nathan C, Xie Q-W. Nitric oxide synthases: roles, tolls. and controls. Cell 1994;78:915-918.
6. Gibaldi M. What is nitric oxide and why are so many people studying it? J Clin Pharmacol
1993;33:488--496.
7. Snyder SH. Nitric oxide: First in a new class of neurotransmitters. Science 1992;257:494-496.
8. Lorsbach RB, Murphy WJ, Lowenstein CJ, Snyder SH, Russell Sw. Expression of the nitric oxide synthase
gene in mouse macrophages activated for tumor cell killing. J Bioi Chem 1993;268: 1908--1913.
9. Pollock JS, Forstermann JA, Mitchell TO. Warner HHHW, Schmidt HH, Nakane M, Murad F. Puritication
of a soluble isofonn of guanylyl cyclase-activating-factor synthase. Proc Natl Acad Sci USA
1991 ;88: 10480--10484.
10. Bredt OS, Hwang PM, Glatt CEo Lowenstein C. Reed RR. Snyder SH. Cloned and expressed nitric oxide
synthase structure resembles cytochrome P-450 reductase. Nature 1991 ;35 I :7 I4-7 18.
I I. Nathan C. Nitric oxide as a secretory product of mammalian cells. FASEB J 1992;6:3051-3064.
12. Nussler AK, Billiar TR. Inflammation. immunoregulation, and inducible nitric oxide synthase. J Leuk Bioi
1993;54: I 71-1 78.
13. Hibbs Jr JB,Taintor RR, Vavrin Z. Macrophage cytotoxicity: role of L-arginine deiminase and imino nitro-
gen oxidation to nitrite. Science 1987;235:473--476.
14. Steuhr OJ, Cho HJ, Kwon NS, Weise MF, Nathan CF. Purification and characterization of the cytokine-in-
duced macrophage nitric oxide synthase: an FAO- and FMN-containing flavoprotein. Proc Natl Acad Sci
USA 1991 ;88:7773-7777.
15. Grisham MB, Yamada T. Neutrophils, nitrogen oxides and inflammatory bowel disease. NY Acad Sci
1992;664: 103--115.
16. McCall TB, Boughton-Smith NK, Palmer RMJ, Whittle BJR, Moncada, S. Synthesis of nitric oxide from
L-arginine by neutrophils. Biochem J 1989;261 :293-296.
17. Curran RD, Billiar TR. Steuhr OJ, Hoffman K, Simmons RL. Hepatocytes produce nitrogen oxides from L-
arginine in response to inflammatory products from Kupffer cells. J Exp Med 1989; 170: 1769-1 774.
18. Schulz R, Nava E, Moncada S. Induction and possible biological relevance ofa Ca'+ -independent nitric ox-
ide synthase in the myocardium. Br.l Phannacol 1992; I 05:577-580.
19. Werner-Felmayer G, Werner ER, Fuchs 0, Hausen A, Reibnegger G, Trentz O. Tetrahydrobiopterin-de-
pendent formation of nitrite and nitrate in murine fibroblasts. J Exp Med 1990; 172: I 599-1607.
20. Beasley 0, Schwartz JH, Brenner BM. Interleukin- I induces prolonged L-arginine-dependent cyclic
guanosine monophosphate and nitrite production in rat vascular smooth muscle cells. J Clin Invest
1991 ;87:602--608.
21. Morris Jr SM, Billiar TR. New insights into the regulation of inducible nitric oxide synthase, Am Physiol
Soc 1994;94:E829-E839.
22. Anon. Nitric oxide in the clinical arena. Lancet 1991 ;338: 1560--1 562.
23. Billiar TM. Hoffman RA, Curran RO, Langrehe JM, Simmons RL. A role for inducible nitric oxide biosyn-
thesis in the liver in inflammation and in the allogeneic immune response. J Lab Clin Med
1992; 120: 192-197.
24. Ki Ibourn RG, Griffith OW. Overproduction of nitric oxide in cytokine-mediated septic shock. J Natl Can-
cer Institute 1992;84:827--831.
25. Nava E, Palmer RMJ, Moncada S. Inhibition of nitric oxide synthesis in septic shock: how much is benefi-
cial? Lancet 1991;338:1555--1557.
26. Laskin JO, Heck DE, Laskin DL. Multifunctional role of nitric oxide in inflammation. TEM
1994;5:377-382.
27. Moncada S, Higgs A. The I-arginine-nitric oxide pathway. N Eng J Med 1993;329:2002-2012.
130 J. B. Allen etal.
28. Hasan K, Heeson BJ, Corbett JA, McDaniel ML, Chang K, Allison W, Wolffenbuttel BHR, Williamson JR,
Tilton RG. Inhibition of nitric oxide formation by guanidines. Eur J Pharmacol 1993;249: I 0 I-I 06.
29. Rees DD, Palmer RMJ, Schultz R, Hodson HF, Moncada S. Characterization of three inhibitors of endothe-
lial nitric oxide synthase in vitro and ill vivo. Br J PharmacoI1990;101:746--752.
30. Misko TP, Moore WM, Kasten TP, Nickols GA, Corbett JA, Tilton RG, McDaniel ML, Williamson JR,
Currie MG. Selective inhibition of the inducible nitric oxide synthase by aminoguanidine. Eur J Pharmacol
1993;233: 119-125.
31. Cross AH, Misko TP, Lin RF, Hickey WF, Trotter JL, Tilton RG. Aminoguanidine. an inhibitor of inducible
nitric oxide synthase, ameliorates experimental autoimmune encephalomyelitis in SJL mice. J Clin Invest
1994;93:2684--2690.
32. O'Connor KJ, Moncada S. Glucocorticoids inhibit the induction of nitric oxide synthase and the related
cell damage in adenocarcinoma cells. Biochein Biophys Acta 1991; I 097:227-231.
33. Howes EL, Wong KL, Hartiala, KT, et al. Complement and polymorphonuclear leukocytes do not deter-
mine the vascular permeability induced by intraocular LPS. Am J Pathol 1985; 118:35--42.
34. Anggard E. Nitric oxide: mediator, murderer, and medicine. Lancet 1994;343: 1199-1206.
35. Beckman, JS, Beckman TW, Chen J, Marshall PM, Freeman BA. Apparent hydroxyl radical production by
peroxynitrite: implications for endothelial injury from nitric oxide and superoxide. Proc Natl Acad Sci
USA 1990;87:1620-1624.
36. Vodovotz Y, Bogclan C, Paik J, Xie QW, Nathan e. Mechanisms of suppression of macrophage nitric oxide
release by transforming growth factor p. J Exp Med 1993;178:605-613.
37. Hausmann EHS, Hao S-Y, Pace JL, Parmely MJ. Transforming growth factor $1 and gamma interferon provide
opposing signals to lipopolysaccharide-activated mouse macrophages. Infect Immun 1994;62:3625-3632.
38. Ding A, Nathan CF, Graycar J, Derynck R, Steuhr DJ, Srimal S. Macrophage deactivating factor and trans-
forming growth factors-p" -p" and -P l inhibit induction of macrophage nitrogen oxide synthesis by inter-
feron-y. J Immunol 1990; 145:940-944.
39. Murphy ME, Noack E. Nitric oxide assay using hemoglobin method. Meth Enzymol 1994;233:240-250.
40. Hevel JM, MarIetta MM. Nitric oxide synthase assays. Meth Enzymol 1994;233:250-258.
41. Green LC, Wagner DA. Glogowski J, Skipper PL, Wishnok JS, Tannenbaum SR. Analysis of nitrate. ni-
trite, and ["Njnitrate in biologic fluids. Anal Biochem 1982;126:131-138.
42. Kobzik L, Bredt DS, Lowenstein CJ, Drazen J, Gaston B, Sugarbaker D, Stamler JS. Nitric oxide synthase
in human and rat lung: immunocytochemical and histochemical localization. Am J Respir Cell Mol BioI
1993;9:371-377.
43. Fleisher LN, Ferrell JB, McGahan Me. Ocular inflammatory effects of intravitreally-injected tumor ne-
crosis factor-alpha and endotoxin. Inflammation 1990;14:325-335.
44. Fleisher LN, Ferrell JB. Smith MG, McGahan Me. Lipid mediators of tumor necrosis factora-induced
uveitis. Invest Ophthalmol Vis Sci 1991 ;32:2393-2399.
45. Fleisher LN, Ferrell JB, McGahan MC. Synergistic uveitic effects of tumor necrosis factor-a and inter-
leukin I-p. Invest Ophthalmol Vis Sci 1992;33:2120-2127.
46. Fleisher LN, McGahan Me. Endotoxin-induced ocular inflammation increases prostaglandin E, synthesis
by the lens. Exp Eye Res 1985;40:711-719. -
47. De Vos AF, Hoekzema R, Kijlstra A. Cytokines and uveitis, a review. Exp Eye Res 1992; 11 :581-597.
48. Davis JL, Jalkh AE, Roberge F, Caspi R, Flynn Jr HW, Schepens CL, Nussenblatt RB. Subretinal fluid
from human retinal detachment contains interleukin I. Invest Ophthalmol Vis Sci 1988;29:65.
49. Limb GA, Little BC. Meager A, Ogilvie JA, Wolstencroft RA, Franks WA, Chignell AH, Dumonde De.
Cytokines in proliferative vitreoretinopathy. Eye 1991 ;5:686--693.
50. Murray PI, Hoekzema R, Van Haren MAC, De Han FD, Kijlstra A. Aqueous humor interleukin-6 levels in
uveitis. Invest Ophthalmol Vis Sci 1990;31 :917-920.
51. Wakefield D, Lloyd A. The role of cytokines in the pathogenesis of inflammatory eye disease. Cytokine
1992;4: 1-5.
52. Kasner L, Chan CC, Whitcup SM, Gery I. The paradoxical effect of tumor necrosis factor (TNFa) in endo-
toxin-induced uveitis. Invest Ophthalmol Vis Sci 1993;34:2911-2917.
53. Rosenbaum JT, Boney RS. Use of a soluble interleukin-l receptor to inhibit ocular inflammation. Curr Eye
Res 1991;10:1137-1139.
54. Hogan MJ, Kimura SJ, Thygeson P. Signs and symptoms of uveitis I. Anterior uveitis. Am J Ophthalmol
1959;47: 155-170.
55. Mandai M, Mittag T, Yoshimura N, Honda Y. Role of nitric oxide synthase isozymes in endotoxin induced
uveitis. Invest Ophthalmol Vis Sci 1995;36:S546.
56. Mandai M, Yoshimura N, Yoshida M, Iwaki M, Honda Y. The role of nitric oxide synthase in endotoxin-in-
duced uveitis: effects ofNG-nitro L-arginine. Invest Ophthalmol Vis Sci 1994;35:3673-3680.
Nitric Oxide in Ocular Inflammation 131
57. Tilton RG, Chong K, Corbett JA, Misko TP, Currie MG, Bora NS, Kaplan HJ, Williamson JR. Endotoxin-
induced uveitis in the rat is attenuated by inhibition of nitric oxide production. Invest Ophthamol Vis Sci
1994;35:3278--3288.
58. Parks OJ, Cheung MK, Chan CC, Roberge FG. The role of nitric oxide in uveitis. Arch Ophthalmol
1994; 112:544-546.
59. Wang ZY, Hakanson R. Role of nitric oxide (NO) in ocular inflammation. Br J Pharmacol
1995;116:2447-250.
60. Bellot J, Goureau 0, Thillaye B, Chatenoud L, De Kozak Y. Inhibition of endotoxin-induced uveitis by Ni-
tric Oxide (NO) synthase inhibitor; effect on intraocular TNF and NO synthesis: In: Nussenblatt RB,
Whitcup SM, Caspi RR, Gery I, editors. 1994; Advances in Ocular Immunology. Elsevier Sci., pp 225-8.
6 I. Bellot JL, Palmero M, Garcia-Cabanes C, Espi R, Hariton C, Orts A. Additive effect of nitric oxide and
prostaglandin-E, synthesis inhibitors in endotoxin-induced uveitis in the rabbit. Inflamm Res
1996;45:203-208.
62. Goureau 0, Bellot J, Thillaye B, Courtois Y, de Kozak Y. Increased nitric oxide production in endotoxin-
induced uveitis. Reduction of uveitis by an inhibitor of nitric oxide synthase. J Immunol
1995; 154:65 I 8--6523.
63. Allen 18, McGahan MC, Ferrell 18, Adler KB, Fleisher LN. Nitric oxide synthase inhibitors exert differen-
tial time-dependent effects on LPS-induced uveitis. Exp Eye Res 1996;62:21-28.
64. Wahl SM, Allen 18, Ohura K, Chenoweth DE, Hand AR. lFNy inhibits inflammatory cell recruitment and
the evolution of bacterial cell wall-induced arthritis. J lmmunol 1990;146:95-100.
65. Sery TW, Petrillo R. Superoxide anion radical as an indirect mediator in ocular inflammatory disease. Curr
Eye Res 1984;3:243-252.
66. Moreno JJ, Pryor WA. Inactivation of a I-protease inhibitor by peroxynitrite. Chern Res Toxicol
1992;5:425-431.
67. Rachmilewitz 0, Stamler JS, Karmel F, Mullins ME, Singell OJ, Loscalzo J, Xavier RJ, Podolsky OK. Per-
oxynitrite-induced rat colitis- a new model of colonic inflammation. Gastroenterology
1993; I 05: I 68 1-1688.
68. Van Oykke K, Antonini JM, Wu L, Ye Z, Reasor MJ. The inhibition of silica-induced lung inflammation by
dexamethazone as measured by broncheoalveolar lavage fluid parameters and peroxynitrite-dependent
chemiluminescence. Agents Actions 1994;4 1:44-49.
69. Ischiropoulos H, Zhu L, Beckman JS. Peroxynitrite formation from macrophage derived nitric oxide. Arch
Biochem Biophys 1992;298:446--45 I.
16
Ronald G. Tilton
1. INTRODUCTION
Nitric oxide (NO) has emerged as a small, membrane-permeable free radical modulat-
ing numerous physiological functions in the cardiovascular, immune, and nervous systems l- 5•
Within the cardiovascular system NO, originally described as endothelium derived relaxing
factor, is released by a variety of agonists and cytokines as well as by pulsatile flow and shear
stress, and is an important physiological modulator of systemic vascular tone and intravascu-
lar fluid volume 6.7. Vascular NO functions constitutively to prevent adhesion of platelets and
leukocytes and platelet aggregation thereby modulating coagulation and fibrinolyis. In addi-
tion, NO affects vascular function by modulating endothelial cell barrier function and vascu-
lar smooth muscle proliferation. Derangements in NO production have been linked to a
variety of vascular pathologies including septic shock, hypertension, atherosclerosis, acute in-
flammation, and diabetes mellitus, and the overproduction of NO has been implicated in the
pathogenesis of a number of autoimmune diseases.
As detailed below, it has become established in the last two years that NO plays a
major proinflammatory role in the initiation and development of ocular inflammatory dis-
eases, including endotoxin-induced uveitis. This is a widely used experimental animal
model for a number of ocular inflammations, collectively termed uveitis, which are associ-
ated with several human inflammatory and autoimmune diseases 8 • It is noteworthy that
uveitis has a significant vascular component characterized by hemodynamic and hyperper-
meability changes which now appear to be mediated by NO. Increased vascular perme-
ability and edema formation also occur in experimental allergic conjunctivitis, an ocular
hypersensitivity response recently linked to NO since these changes can be attenuated by
blocking NO formation 9 . lo .
In the retinal and uveal vasculatures of the eye, short duration diabetes induces in-
creased blood flow, microvascular hypertension, and increased capillary extravasation of
2.1.1. Neuronal NOS. Nitric oxide produced by NOS I (a 150 to 160 kD soluble en-
zyme) functions as a neurotransmitter, and participates in retrograde neurotransmission, or
long term potentiation associated with learning. Immunochemical staining using specific
antibodies to NOS I suggest that it is also present in certain areas of the spinal cord, in
sympathetic ganglia, adrenal glands, epithelial cells of lung, uterus, and stomach, in kid-
ney macula densa cells, pancreatic islet cells, and smooth muscle cells I6 •17 • Although con-
stitutively expressed, there is growing evidence for expressional regulation of NOS I,
including upregulation of NOS I mRNA and increased activity following estrogen therapy,
axotomy, brain ischemia induced by occlusion of the middle cerebral artery, and chronic
salt loading l &-21.
2.1.2. Endothelial NOS. Nitric oxide produced by NOS III (135 kD, >90% particu-
late enzyme) appears to be relatively specific for endothelial cells although immunohisto-
Role of Nitric Oxide in Vascular Dysfunction Associated with Ocular Diseases 135
chemical studies have localized this enzyme to kidney tubular epithelial cells, interstitial
cells in the colon, and in neurons of the rat hippocampus 22- 24 . Nitric oxide generated by
NOS III interacts with guanylate cyclase within the vascular smooth muscle cell, increas-
ing intracellular cGMP which produces relaxation via decreases in intracellular calcium.
Like NOS I, NOS III can be expressional regulated. Shear stress not only increases the ac-
tivity but also increases the expression of this enzyme 25 • TNFa can downregulate l5 while
estrogens and hypoxia upregulate NOS III gene expression I5 .20 •
~
II NHZ
C I
~/ 'NH-(CHZ)n- CHCOO -~
Inhibitor R, RZ R3 n Selectivity
L-arginine analogs
NG-methyl-L-arginine eNOS;,iNOS
NH Z NCH 3 H 3
(L-NMA)
eNOS» iNOS
NG-Nitro-L-arginine NH Z NNO Z H 3
(L-NNA)
NG-Nitro-L -arginine- eNOS» NOS
NH Z NNO Z CH 3 3
methyl ester (L-NAME)
NG-amino-L-arginine eNOS ;,iNOS
NH Z NNH Z H 3
(L-NAA)
a few among this group (aminoguanidine, isothioureas, and some imidazoles and inda-
zoles) demonstrate significant isoform selectivity. Of these, aminoguanidine has received
the most attention to date due to the early recognition of its iNOS selectivity, its apparent
low toxicity, and its potential clinical usefulness. We have reported that aminoguanidine,
which contains the two chemically equivalent guanidino nitrogen groups in addition to a
hydrazine moiety, is a potent and selective inhibitor of the inducible isoform of NO syn-
thase with much less impact on blood pressure when administered in vivo than L-arginine
analogue inhibitors of NO synthase 34 • Aminoguanidine has been studied extensively in
diabetes (see below), where it prevents IL-l~-induced decreases in glucose secretion in is-
lets of Langerhans and glucose- and streptozotocin-induced vascular dysfunction in rats
via NOS-mediated mechanisms. Other effects of aminoguanidine linked to its inhibitory
effect on iNOS include inhibiting progression of experimental autoimmune encephalo-
myelitis 35 , transplant rejection 36 , uveitis 37, stroke 3 S, inflammatory bowel disease 39 , and en-
dotoxemia40 • S-substituted isothioureas are potent inhibitors of NOS with variable
138 R.G. Tilton
selectivity for the different isoforms 33 . While S-methyl isothiourea shows preference for
iNOS, some bis-isothioureas (sulfur atoms of two isothiourea units are linked by a carbon
chain) show marked selectivity for human iNOS41. Unfortunately, these compounds have
poor cellular uptake and demonstrate potent cytotoxicity. 7-Nitroindazole has received
considerable attention as a selective NOS I inhibitor and may function in this manner via
multiple mechanisms. Homopiperidine amidines have been reported to be selective iNOS
inhibitors 42 .
In addition to enzyme inhibition described above, other approaches for limiting NO
production include limiting enzyme co-factors (interfering with tetrahydrobiopterin syn-
thesis or its binding site, flavoprotein inhibitors, calmodulin binders such as gangliosides
or calmodulin antagonists such as chlorpromazine and W-7, iron-heme binders, calcium
chelators) and NO scavenging (hemoglobin, superoxide generators)43.44. Additional ap-
proaches for limiting NO via iNOS include inhibiting membrane signal transduction (pyr-
rolidine dithiocarbamate, an inhibitor of NFkB activation), transcription (giucocorticoids,
colchicine), and RNA/protein synthesis (actinomycin D, cycloheximide). TGF13, bFGF,
IL-4, and IL-IO have been reported to inhibit expression of iN OS by macrophages44 .
3.1. Uveitis
Systemic or intraocular injection of small doses of lipopolysaccharides induces an
acute ocular inflammation which has been associated with vascular hemodynamic and per-
meability changes in the affected tissues. A number of specific cytokines, prostaglandins,
leukotrienes, and other factors have been implicated in the pathogenesis of EIU, and a
number of substances have been shown to inhibit or ameliorate the condition. These stud-
ies have provided evidence that the secreted products of neutrophils, macrophages, and
mast cells played a central role in the pathogenesis of EIU, but their relationship to the
vascular dysfunction associated with uveitis remained enigmatic until recently. It is now
becoming evident that the upregulation of different cytokines, such as TNFa and IL-I13 ,
during endotoxin-induced uveitis can lead to the induction of NOS activity.
Recently, there have been numerous reports of induction of iNOS after LPS and cy-
tokine stimulation37.45-49 as well as in pathologic conditions such as CMV retinitis in AIDS
patients50 . Mandai et al. 49 reported that endotoxin-induced uveitis increased NOS activity.
Goureau et al. 46 reported increased nitrite (an oxidation product of NO measured spectro-
photometrically) in aqueous and vitreous fluid and this correlated with reverse transcrip-
tase-PCR evidence of increased iNOS mRNA in the iris/ciliary body and retina. In their
experiments, a nonspecific NOS inhibitor (L-NAME) blocked the clinical symptomology
and increased nitrite levels which is consistent with the findings of other laboratories us-
ing the same NOS inhibitor45 ,48. Subsequently, this research group reported that the epi-
thelial cells of the iris/ciliary body as well as infiltrating cells in the anterior segment and
in the retina were the source of N047. We have reported37 that footpad injection of endo-
toxin can induce a significant increase in plasma nitrite levels as well as albumin permea-
tion in all ocular vasculatures including the retina, anterior uvea, choroid, and in the
aqueous fluid. Blood flow is increased only in the anterior uvea. A selective inhibitor of
iNOS, aminoguanidine, normalized plasma nitrite levels and prevents or significantly at-
tenuates the blood flow and albumin permeability changes (Figure 2). In addition, ami-
Role of Nitric Oxide in Vascular Dysfunction Associated with Ocular Diseases 139
...
..c:
200
Retina
...• *
• ...
.
..c:
700
Anterior uvea
,..... *
-::-
.~ 150
l- .S!' 600
....
.....iI: ....
.....**
iI:
.~ 500
• • It
•• •
~ 100 • L
iI:
400 •
-+
•• •
Cl
•
"'Ec:: ~ .........
• ••• "''Ec:: 300 ••
50
•• I • •••• •
""'-
Cl "'"'-
Cl 200
0
control LPS
or control LPS
AG + + AG + +
Figure 2. Effects of the NOS inhibitor, aminoguanidine on LPS-induced increases in '25[_albumin permeation in
retina and anterior uvea. LPS dose = 100 Ilg (50 Ilg/foot pad); aminoguanidine dose = 100 mg/kg body weight in-
jected subcutaneously at 0, 3, 6, 12, 18, and 20 to 24 hours after LPS injection. Bars represent mean values for
each group. Significantly different from untreated controls by least square means: *p<O.OOO 1. Significantly differ-
ent from LPS-treated rats: 'p<O.OOOI; t p<0.02.
3.2. Conjunctivitis
At least one research group9,IO, using a guinea pig model of allergic conjunctivitis,
has reported that NO mediates the increased conjunctival blood flow and hyperpermeabil-
ity following administration of antigen, compound 48/80, or histamine to the eyes of sen-
sitized guinea pigs. Increased levels of nitrite, an oxidation product of NO, was measured
in lavage fluid 30 min after challenge. Pretreatment of eyes with L-NAME or ami-
noguanidine significantly inhibited the clinical score of conjunctivitis, edema formation,
and the hyperpermeability following antigen stimulation. Consistent with data reported
below for diabetes, the topical application of sodium nitroprusside (which releases NO as
an active moiety) dilated conjunctival blood vessels and produced a dose-dependent in-
crease in vascular permeability.
chemical imbalances that mediate sorbitol pathway-linked vascular functional and structural
changes in the eye remains unclear. Postulated mechanisms have included intracellular os-
motic stress due to accumulation of sorbitol, myo-inositol depletion with impaired phospha-
tidylinositol metabolism, alterations in PKC isoform activities, and a cascade of metabolic
imbalances initiated by decreased availability ofNADPH (which is oxidized to NADP+ cou-
pled to reduction of glucose to sorbitol by aldose reductase) as well as increased cytosolic free
NADH (resulting from increased oxidation of sorbitol to fructose by sorbitol dehydro-
genase)51-53.58.59.
Hemodynamic changes and hyperpermeability of vessels are the two most charac-
teristic changes observed in poorly controlled human diabetics. These changes particularly
affect ocular tissues, peripheral nerves, and kidney. Early in the course of diabetes, retinal
(and renal) blood flow is increased, but later in the disease process, blood flow is reduced
and falls below normal with progression to end-stage proliferative retinopathy. There is
general agreement that in animal models of diabetes several different aldose reductase in-
hibitors prevent early vascular functional abnormalities in ocular tissues. We have re-
ported that blood flow and vascular permeation of radiolabeled albumin are increased in
all ocular tissues (retina, anterior uvea, and choroid) of rats with diabetes of 4 to 6 weeks
duration, and that these changes can be markedly reduced or prevented by a variety of
structurally different inhibitors of aldose reductase 54-56.
On the other hand, reports by different investigators on effects of aldose reductase
inhibitors on retinal structural changes remain controversiaI6o,61. More importantly, the use
of aldose reductase inhibitors has not resolved the issue of which of the above listed puta-
tive mechanisms mediate diabetes-induced, sorbitol pathway-linked retinal vascular func-
tional and structural changes, since they prevent all redox and metabolic imbalances
associated with the first and second steps of the pathway. In contrast, the use of sorbitol
dehydrogenase inhibitors, which block the second half of the sorbitol pathway, has pro-
vided some important insights, Our observation62 that a putative inhibitor of sorbitol dehy-
drogenase, S-0773, increases retinal sorbitol levels several fold above the already elevated
levels in diabetic tissues, but significantly attenuates the vascular blood flow and hyper-
permeability changes, together with the finding that sorbitol dehydrogenase inhibitors in-
crease retinal sorbitol levels in nondiabetic tissues to levels observed in diabetic tissues
without producing vascular injury, provides documentation against involvement of sorbi-
tol accumulation per se, in the pathogenesis of diabetes-induced vascular dysfunction,
Since the increased blood flow in retinas of diabetic rats is not associated with any
change in mean arterial pressure, this finding implies that resistance arterioles have di-
lated. Nitric oxide continues to be a potential candidate for mediating early diabetes-in-
duced increases in blood flow and permeability since its effect on the vasculature mimics
those of diabetes. First, NO is a potent vasodilator and increases blood flow. Second,
agonist stimulation of NOS using histamine or bradykinin increases blood flow and vascu-
lar permeability. Third, topical application of NO donors (i.e., nitroglycerine, sodium ni-
troprusside) increases blood flow and vascular permeability in the skin chamber
granulation tissue model (described below). We have shown that glucose- and diabetes-in-
duced increases in albumin hyperpermeability andlor blood flow can be attenuated or pre-
vented in vivo by NOS inhibitors such as L-NMMA and aminoguanidine 34 ,66,67, suggesting
that these vascular changes are mediated by a relative or absolute increase in NO produc-
tion. Aminoguanidine also attenuates albuminuria68 , and prevents vascular structural
changes occurring early after the onset of diabetes in the retina69 . In addition to inhibiting
NOS, aminoguanidine also inhibits diamine oxidase, prevents advanced glycation endpro-
duct formation, and oxidative modification of low density lipoproteins via NOS-inde-
pendent mechanisms 70 ,71. Therefore, it is noteworthy that methylguanidine, another
guanidine NOS inhibitor which does not share these additional effects of aminoguanidine,
is equally effective as aminoguanidine at blocking diabetcs-induced vascular dysfunction
(Figure 3)66.
We have previously reported that increases in blood pressure induced by intravenous
bolus injections of aminoguanidine are markedly blunted in diabetic rats and that this
blunted pressure response is normalized by treatment with an aldose reductase inhibitorn .
These observations suggest an increased nitric oxide synthesis in resistance arterioles of
diabetics and a need for higher aminoguanidine levels to block nitric oxide synthesis suffi-
ciently to increase blood pressure.
In contrast to the implications of an increase in NO production in selected tissues of
diabetic rats described above, numerous observations in streptozotocin- and genetically-
induced diabetic rats and in alloxan-diabetic rabbits indicate that endothelium-dependent
relaxation in response to acetylcholine (as well as other agonists that mediate their vasodi-
lating effect via nitric oxide) is impaired in large vessels such as the aorta 73- 80 • Exposure
of normal arteries to elevated glucose levels in vitro impairs acetylcholine-induced relaxa-
150
:;::
01
'0;
3:
~ 100
~
~E 50
"'
<II
Q.
~ 0
Control Control Control Diabetic Diabetic Diabetic
+AG +MG +AG +MG
Figure 3. Effects of aminoguanidine and methyl guanidine on diabetes-induced increases in 12sl_albumin permea-
tion in the retina. Bars represent mean values + I SD and are measured in ~g plasma/min/g wet weight. Untreated
controls. n = 25; control + AG, n = 8; control + MG, n = 6; untreated diabetic, n = 25; diabetic + AG, n = 9; dia-
betic + MG, n = 9. Significantly different from untreated controls by least square means: *p<O.OOO I. Significantly
different from untreated diabetic rats: 'p<O.OOOI.
142 R.G. Tilton
tion 81 •82 • While diabetes-induced changes in large conduit vessels such as the aorta may
not reflect the impact of diabetes on terminal arterioles regulating tissue perfusion, it is
noteworthy that both acute (1-2 weeks) and chronic (-2 months) hyperglycemia com-
pletely suppressed acetylcholine-induced vasodilation of intestinal arterioles 83 • Mayhan et
al. 84 have reported that NO-mediated relaxation of cerebral arterioles in vivo also is sup-
pressed in diabetic animals. If these results in the intestine and brain can be extrapolated
to the vasculature of tissues affected by diabetes such as retina, peripheral nerve, and kid-
ney, these findings suggest that a specific deficit of acetylcholine-induced NO action can
develop in the small arterioles responsible for modulating flow. On the other hand, numer-
ous observations (found in the above cited reports dealing with impaired vascular re-
sponses to acetycholine) indicate that relaxation of diabetic vessels in response to
nitroglycerin, which activates the cytosolic form of cGMP within vascular smooth muscle
cells via a NOS-independent mechanism, does not differ from controls. Taken together,
these observations suggest that the ability of vascular smooth muscle cells to vasodilate is
not compromised and that the loss of vasodilator response to acetylcholine is primarily of
endothelial origin. Various factors have been implicated in this defective relaxation of dia-
betic arteries and arterioles, including: 1) decreased production and/or release of nitric ox-
ide, 2) increased destruction of nitric oxide by oxygen-derived free radicals such as
superoxide, 3) increased production of endothelium-derived constricting factors, and 4)
impaired agonist (i.e., acetylcholine) interaction/activation of its receptor resulting in im-
paired post-receptor signal transduction. At this time, the mechanism(s) mediating this de-
crease in arterial resistance remains unclear.
It has been reported that the impaired vasodilating response to acetylcholine can be
restored by dietary supplementation with L-arginine 84 • While the mechanism by which
dietary L-arginine restores NO-mediated vasodilating response in diabetes remains un-
known, these observations suggest impaired NOS activity. However, a deficient supply of
L-arginine for NO synthesis in diabetic vessels cannot explain the observation that the
vascular response to the calcium ionophore A23187, which stimulates cNOS activity by
increasing intracellular calcium, remains normal in diabetic rats 84 • In addition, most stud-
ies reporting impaired NO production in arteries are based on experiments in which ves-
sels are incubated for several hours in media lacking L-arginine as substrate for NO
synthesis, suggesting that the results obtained may be an artifact of the in vitro milieu. Im-
paired acetylcholine-induced relaxation of arteries from hypercholesterolemic rabbits ap-
pears to be an in vitro artifact of increased free radical production catalyzed by trace
metals in the buffer85 ; such an effect would be even greater in hyperglycemic media be-
cause of autoxidation of glucose. Finally, it should be pointed out that some investigators
have reported that the magnitude of the acetylcholine-induced relaxation to precontracted
arteries does not differ between controls and diabetics or is actually increased in diabet-
ics86.87 •
sue glucose levels, interferes with NO prior to its interaction with vascular smooth muscle.
This interpretation is consistent with the conclusions of Langenstroer and Pieper 78 , who
have reported an enhanced NO release in diabetic rat aorta which is unmasked by the ad-
dition of superoxide dismutase (SOD); SOD produced a signficantiy greater relaxation in
diabetic aorta, both in vitro and in vivo, suggesting an increased production of NO which
was destroyed at an increased rate by the generation of more oxygen-derived free radicals
in the diabetic milieu.
Acute hyperglycemia of only a few hours duration has been reported by many inves-
tigators to increase retinal blood flow in humans and animals. We have demonstrated that
blood flow changes (identical to those observed in diabetic animals) can be produced in
nondiabetic ocular tissues (retina, anterior uvea, and choroid) made hyperglycemic by
acute intravenous glucose infusion (25 % glucose solution infused at the rate of 0.0 1
mUmin) for -5 hours 67 . These glucose-induced increases in ocular blood flow can be
blocked by aldose reductase inhibitors and by pyruvate (Figure 4). We have interpreted
these observations to suggest that sorbitol pathway-linked retinal blood flow increases in-
duced by elevated tissue glucose levels are linked to an increased rate of reduction of
NAD+ to NADH coupled to oxidation of sorbitol to fructose by sorbitol dehydro-
genase 59 .90 , and that the increased activity of this enzyme results in a "hypoxia-like" cy-
tosolic reductive stress. The observation that pyruvate normalized glucose-induced blood
flow changes in the retina and anterior uvea would be consistent with this interpretation
since elevated tissue pyruvate would drive oxidation of NADH to NAD+ (coupled to re-
duction of pyruvate to lactate by lactate dehydrogenase) as rapidly as NADH is formed by
the action of sorbitol dehydrogenase. Using this animal model, we have demonstrated that
increases in ocular blood flow caused by glucose can be prevented by co-infusion of NOS
inhibitors such as aminoguanidine and L-NMMA (Figure 4). These findings suggest that
retinal vascular dysfunction resulting from glucose-induced reductive stress is mediated in
some way via NO.
Figure 4. Effects of NOS inhibitors, aldose reductase inhibitors, and pyruvate on blood flow responses to acute
glucose elevation in ocular tissues. Bars represent mean values + I SO and are measured in mllmin/g wet weight.
NOS inhibitors include aminoguanidine (AG) and NG-methyl-L-arginine (L-NMA; see Figure I). ARi = aldose re-
ductase inhibitor; Pyr = pyruvate. Significantly different from control saline infusion by least square means:
*p<O.OOOI.
144 R.G. Tilton
hypoxia _ tNADHINAD+
/' +-- pyruvate
(cytosolic)
•
tVEGF
• VEGF Ab
Figure 5. Scenario for linkage of metabolic imbalances that
mediate effects of hyperglycemia on vascular dysfunction.
tintracellular Ca··
Compounds which are shown in boldface produce vascular in-
jury and include glucose, sorbitol, hypoxia, VEGF, and so-
dium nitroprusside (SNP). Interventions which block
•
t Nitric oxide synthase activity
• L·NMMA/AG
glucose-induced vascular injury are shown in italics and in-
clude aldose reductase inhibitors (ARi), sorbitol dehydro-
SNP _ t nitric oxide
genase inhibitors (SDi), pyruvate, VEGF antibodies, and NOS
inhibitors.
•
VASCULAR DYSFUNCTION
cellular caIcium96 • These observations are consistent with reports that: I) VEGF stimulates
proliferation of coronary venular endothelium through production of nitric oxide and
cGMP, and that the VEGF proliferative effect could be blocked with NOS inhibitors 97 ,
and 2) systemic administration of VEGF in rats decreases blood pressure, cardiac output,
and stroke volume without affecting myocardial contractility and these changes are attenu-
ated with NOS inhibitors98 . Our finding that VEGF mediates vascular dysfunction via in-
creased NO production is consistent with our previous evidence that a relative or absolute
increase in NO mediates vascular dysfunction in the retina of diabetic rats and in nondia-
betic rats with acute hyperglycemia.
Taken together, these observations suggest that retinal vascular dysfunction induced
by diabetes is mediated by increased NO production as a consequence of "hypoxia-like"
cytosolic reductive stress linked to increased flux of glucose via the sorbitol pathway as
shown in Figure 5. This scenario is supported by evidence that true hypoxia increases vas-
cular permeability and blood flow and endothelial NOS mRNA, NOS enzyme activity,
and increased NO production 99 •
5. ACKNOWLEDGMENTS
The research collaborations of Drs. J.R. Williamson (Department of Pathology;
Washington University School of Medicine; St. Louis, MO), Khalid S. Hasan (Pediatric
Endocrinology and Diabetes Section; Phoenix Childrens's Hospital; Phoenix, AZ), and
T.A. Brock and C.C. Stephan (Department of Pharmacology; Texas Biotechnology Corpo-
ration; Houston, TX) for the granulation tissue chamber and VEGF data reported here are
gratefully acknowledged.
6. REFERENCES
I. Moncada S, Higgs E, Hodson H, Knowles R, Lopez-Jaramillo P, McCall T, Palmer M, Redomski M, Rees
D, Schulz R. The L-arginine: nitric oxide pathway. J Cardiovasc Pharmacol 1991: 17:S I-S9.
146 R.G. Tilton
2. Moncada S, Higgs EA. Molecular mechanisms and therapeutic strategies related to nitric oxide. FASEB J
1995;9: 1319-1330.
3. Sessa Wc. The nitric oxide synthase family of proteins. J Vase Res 1994;31: 131-143.
4. Farrell AJ, Blake DR. Nitric oxide. Annals Rheumatic Diseases 1996;55:7-20.
5. Schroeder RA, Kuo Pc. Nitric oxide: physiology and pharmacology. Anesth Analg 1995;81: 1052-1059.
6. Schini-Kerth VB, Vanhoutte PM. Nitric oxide synthases in vascular cells. Exp Physiol 1995;80:885-905.
7. Marsden PA, Brenner MB. Nitric oxide and endothelins: novel autocrine/paracrine regulators of the circu-
lation. Semin NephroI1991;11:169-185.
8. Rosenbaum n, Mac Devitt HO, Guss RB, Egbert PRo Endotoxin-induced uveitis in rats as a model for hu-
man disease. Nature (Lond) 1980;286:611-613.
9. Meijer F, Ruijter JM, Van Delft JL, Van Haeringen NJ. Nitric oxide induces vascular permeability changes
in the guinea pig conjunctiva. Eur J Pharmacol 1995;284:61-67.
10. Meijer F, Van Delft JL, Garrelds 1M. Van Haeringen NJ, Kijlstra A. Nitric oxide plays a role as a mediator
of conjunctival edema in experimental allergic conjunctivitis. Exp Eye Res 1996;62:359-365.
II. Stevens MJ. Dananberg J, Feldman EL, Lattimer SA, Kamijo M. Thomas TP, Shindo H. Sima AAF,
Greene DA. The linked roles of nitric oxide, aldose reductase and (Na+,K+)-ATPase in the slowing of
nerve conduction in the streptozotocin diabetic rat. J Clin Invest 1994;94:853-859.
12. Calver A, Collier J, Vallance P. Inhibition and stimulatioin of nitric oxide synthesis in the human forearm
arterial bed of patients with insulin-dependent diabetes. J Clin Invest 1992;90:2548-2554.
13. Smits P, Kapma J-A. Jacobs M-C, Lutterman J, Thien T. Endothelium-dependent vascular relaxation in pa-
tients with type I diabetes. Diabetes 1993;42:148-153.
14. Stuehr DJ, Kwon NS, Nathan C, Griffith O. Feldman P, Wiseman J. Nw-Hydroxy-L-arginine is an interme-
diate in the biosynthesis of nitric oxide from L-arginine. J BioI Chern 1991 ;266:6259-6263.
15. Forstermann U, Kleinert H. Nitric oxide synthase: expression and expressional control of the three iso-
forms. Naunyn-Schmiedeberg's Arch Pharmacol 1995;352:351-364.
16. Dun NJ, Dun SL, Wu SY, Forstermann U. Nitric oxide synthase immunoreactivity in rat superior cervical
ganglia and adrenal glands. Neurosci Lett 1993;158:51-54.
17. Schmidt HHHW, Gagne GD. Nakane M, Pollock JS, Miller MF, Murad F. Mapping of neural nitric oxide
synthase in the rat suggests frequent co-localization with NADPH diaphorase but not with soluble guanylyl
cyclase, and novel paraneural functions for nitrinergic signal transduction. J Histochem Cytochem
1992;40: 1439-1456.
18. Weiner CP, Lizasoain I, Baylis SA, Knowles RG, Charles IG, Moncada S. Induction of calcium-dependent
nitric oxide synthases by sex hormones. Proc Natl Acad Sci USA 1994;91:5212-5216.
19. Herdegen T, Brecht S, Mayer B, Leah J, Kummer W, Bravo R, Zimmermann M. Long-lasting expression of
JUN and KROX transcription factors and nitric oxide synthase in intrinsic neurons of the rat brain follow-
ing axotomy. J Neurosci 1993; 13(4130-4145)
20. Zhang ZG, Chopp M, Zaloga C, Pollock JS, Forstermann U. Cerebral endothelial nitric oxide synthase ex-
pression after focal cerebral ischemia in rats. Stroke 1993;24:2016-2021.
21. Kadowaki K, Kishimoto J, Leng G, Emson Pc. Up-regulation of nitric oxide synthase (NOS) gene expres-
sion together with NOS activity in the rat hypothalamo-hypophysial system after chronic salt loading - ar-
ginine vasopressin and oxytocin secretion. Endocrinology 1994; 134: I 011-1 017.
22. Pollock JS, Nakane M, Buttery LK, Martinez A, Springall D, Polak JM, Forstermann U, Murad F. Charac-
terization and localization of endothelial nitric oxide synthase using specific monoclonal antibodies. Am J
PhysioI1993;265:CI379-CI387.
23. Tracey WR, Pollock JS, Murad F, Nakane M, Forstermann U. Identification of type III (endothelial-like)
particulate nitric oxide synthase in LLC-PKI kidney tubular epithelial cells. Am J Physiol 1994;266:C22-
C26.
24. Dinerman JL, Dawson TM, Schell MJ, Snowman A, Snyder SH. Endothelial nitric oxide synthase local-
ized to hippocampal pyramidal cells: implications for synaptic plasticity. Proc Nail Acad Sci
1994;91 :4214-4218.
25. Nishida K, Harrison DG, Navas JP, Fisher AA, Dockery SP, Uematsu M, Nerem RM, Alexander RW, Mur-
phy TJ. Molecular cloning and characterization of the constitutive bovine aortic endothelial cell nitric ox-
ide synthase. J Clin Invest 1992;90:2092-2096.
26. Stuehr DJ, Cho HJ, Kwon NS, Weise MF, Nathan CF. Purification and characterization of the cytokine-in-
duced macrophage nitric oxide synthase: an FAD- and FMN-containing flavoprotein. Proc Natl Acad Sci
1991 ;88:7773-7777.
27. Bandaletova T, Brouet I, Bartsch H, T. S, Esumi H, Ohshima H. Immunohistochemical localization of an
inducible form of nitric oxide synthase in various organs of rats treated with propionibacterium-acnes and
lipopolysaccharide. APMIS 1993; 101 :330-336.
Role of Nitric Oxide in Vascular Dysfunction Associated with Ocular Diseases 147
28. Cho HJ, Xie QW, Calaycay J, Mumford RA, Swiderek KM, Lee TD, Nathan C. Calmodulin is a subunit of
nitric oxide synthase from macrophages. J Exp Med 1992; I 76:599-{;04.
29. Dimmeler S, Lottspeich F, Brune B. Nitric oxide causes ADP-ribosylation and inhibition of glyceralde-
hyde-3-phosphate dehydrogenase. J BioI Chern 1992;267:16771-16774.
30. Clancy RM, Levartovsky D, Leszczynska-Piziak J, Yegudin J, Abramson SB. Nitric oxide reacts with intra-
cellular glutathione and activates the hexose monophosphate shunt in human neutrophils: evidence for S-
nitrosoglutathione as a bioactive intermediary. Proc Natl Acad Sci USA 1994;91:3680-3684.
31. Stamler JS, Singel DJ, Loscalzo J. Biochemistry of nitric oxide and its redox-activated forms. Science
1992;258: 1898-1902.
32. Salvemini D, Misko TP, Masferrer JL, Seibert K, Currie MG, Needleman P. Nitric oxide activates cy-
c100xygenase enzymes. Proc Natl Acad Sci USA 1993;90:7240-7244.
33. Southan GJ, SzabO C. Selective pharmacological inhibition of distinct nitric oxide synthase isoforms. Bio-
chern Pharmacol 1996;51 :383-394.
34. Corbett JA. Tilton RG, Chang K. Hasan KS, Ido Y, Wang JL, Sweetland MA, Lancaster JR, Williamson JR,
McDaniel ML. Aminoguanidine inhibits nitric oxide formation and prevents diabetic vascular dysfunction.
Diabetes 1992;41 :552-556.
35. Cross AH, Misko TP, Lin RF, Hickey WF, Trotter JL. Tilton RG. Aminoguanidine, an inhibitor of inducible
nitric oxide synthase, ameliorates experimental autoimmune encephalomyelitis in SlL mice. 1 Clin Invest
1994;93:2684-2690.
36. Worrall NK, Chang K, Suau GM, Allison WS, Misko TP, Sullivan PM, Tilton RG, Williamson lR, Fer-
guson TBI. Inhibition of inducible nitric oxide synthase prevents myocardial and systemic vascular barrier
dysfunction during early cardiac allograft rejection. Cire Res 1996;78:769-779.
37. Tilton RG, Chang K, Corbett lA, Misko TP, Currie MG. Bora NS, Kaplan HI, Williamson JR. Endotoxin-
induced uveitis in the rat is attenuated by inhibition of nitric oxide production. Invest Ophthalmol Vis Sci
1994;35:3278-3288.
38. ladecola C, Zhang F, Xu X. Inhibition of inducible nitric oxide synthase ameliorates cerebral ischemic
damage. Am 1 Physiol (Regulatory Integrative Comp Physio\) 1995;268:R286-R292.
39. Miller MIS, Thompson JH, Zhang X-I. Sadowska-Krowicka H, Kakkis JL. Munshi UK, Sandoval M,
Rossi JL. Eloby-Childress S, Beckman JS. Ye YZ. Rodi CP, Manning PT, Currie MG. Clark DA. Role of
inducible nitric oxide synthase expression and peroxynitrite formation in guinea pig ileitis. Gastroenterol-
ogy 1995;109:1475-1483.
40. Wu C-C, Ruetten H. Thiemermann C. Comparison of the effects of aminoguanidine and Nw-nitro-L-ar-
ginine methyl ester on the multiple organ dysfunction caused by endotoxaemia in the rat. Eur 1 Pharmacol
1996;300:99-104.
41. Garvey EP, Oplinger lA. Tanoury GJ, Sherman PA, Fowler M. Marshall S, Harmon MF. Patih IE. Furtine
ES. Potent and selective inhibiton of human nitric oxide synthases. 1 BioI Chern 1994;269:26669-26676.
42. Hansen OW, Peterson KB, Trivedi M. Moore WM. Currie MG, Manning PT, Connor lR. Pitzele BS. Ho-
mopiperidine amidines as selective inhibitors of inducible nitric oxide synthase. Endothelium 1995;3:S80.
43. Marietta MA. Approaches toward selective inhibition of nitric oxide synthase. 1 Med Chern
1994;37: 1899-1907.
44. Pfeilschifter 1, Ederhardt W, Hummel R. Kunz D, Muhl H. Nitsch D, Pluss C, Walker G. Therapeutic
strategies for the inhibition of inducible nitric oxide synthase - potential for a novel class of anti-inflamma-
tory agents. Cell Bio Internat 1996;20:51-58.
45. Parks Dl, Cheung MK. Chan C-C, Roberge FG. The role of nitric oxide in uveitis. Arch Ophthalmol
1994;112:544-546.
46. Goureau 0, Bellot 1, Thillaye B, Courtois Y. de Kozak Y. Increased nitric oxide production in endotoxin-
induced uveitis. 1 Immunol 1995; 154:6518-6523.
47. lacquemin E. de Kozak Y. Thillaye B. Courtois Y, Goureau O. Expression of inducible nitric oxide
synthase in the eye from endotoxin-induced uveitis rats. Invest Ophthalmol Vis Sci 1996;
37:1187-1196.
48. Allen 18, McGahan MC, Ferrell 18. Adler KB, Fleisher LN. Nitric oxide synthase inhibitors exert differen-
tial time-dependent effects on LPS-induced uveitis. Exp Eye Res 1996;62:21-28.
49. Mandai M, Yoshimura N, Yoshida M, Iwaki M. Honda Y. The role of nitric oxide synthase in endotoxin-in-
duced uveitis: effect ofNG-nitro L-arginine. Invest Ophthalmol Vis Sci 1994;35:3673-3680.
50. Dighiero p. Reux I. Hauw JJ, Fillet AM, Courtois Y, Goureau O. Expression of inducible nitric oxide syn-
thase in cytomegalovirus-infected glial cells of retinas from AIDS patients. Neurosci Letter
1994; 166:31-38.
51. Ruderman NB. Williamson JR, Brownlee M. Glucose and diabetic vascular disease. FASEB 1
1992;6:2905-2914.
148 R. G. Tilton
52. Williamson JR, Kilo C, Tilton RG, Mechanisms of glucose· and diabetes-induced vascular dysfunction. in
Hyperglycemia. Diabetes. and Vascular Disease. N.B. Rudennan, J.R. Williamson, and M. Brownlee, Edi-
tor. 1992, Oxford University Press: New York. p. 107-132.
53. Winegrad AI. Does a common mechanism induce the diverse complications of diabetes? Diabetes
1987;36:396-406.
54. Williamson JR, Tilton RG, Kilo C, The Polyol Pathway and Diabetic Vascular Complications. in Diabetic
Complications: Epidemiology and Pathogenetic Mechanisms. D. Andreani, et al., Editor. 1991, Raven
Press: New York. p. 45-58.
55. Tilton R, Chang K, Pugliese G, Eades D, Province M, Shennan W, Kilo C, Williamson J. Prevention on
hemodynamic and vascular albumin filtration changes in diabetic rats by aldose reductase inhibitors. Dia-
betes 1989;38: 1258-1270.
56. Tilton RG, Chang K, Weigel C, Kilo C, Williamson JR. Increased ocular blood flow and "'I·albumin per-
meation in galactose-fed rats: Inhibition by sorbinil. Investigative Ophthalmology & Visual Science
1988;29:861-868.
57. Williamson JR, Ostrow E, Eades DM, Chang K, Allison W, Kilo C, Shennan WR. Glucose-induced mi-
crovascular functional changes in nondiabetic rats are stereospecific and are prevented by an aldose reduc-
tase inhibitor. J Clin Invest 1990;85: 1167-1172.
58. Williams B. Glucose-induced vascular smooth muscle dysfunction: the role of protein kinase C. J Hyper-
tension 1995;13:477-486.
59. Williamson JR, Chang K, Frangos M, Hasan KS, Ido Y, Kawamura T, Nyengaard JR, Van den Enden M,
Kilo C, Tilton RO. Hyperglycemic pseudohypoxia and diabetic complications. Diabetes 1993;42:801-813.
60. Kador PF, Akagi Y, Takahashi Y, Ikebe H, Wyman M, Kinoshita JH. Prevention of retinal vessel changes
associated with diabetic retinopathy in galactose-fed dogs by aldose reductase inhibitors. Arch Ophthalmol
1990; I 08: 1301-1309.
61. Engennan RL, Kern TS. Aldose reductase inhibition fails to prevent retinopathy in diabetic and galacto-
semic dogs. Diabetes 1993;42:820-825.
62. R.G. Tilton, K. Chang, J.R. Nyengaard, M. Van den Enden, Y. Ido, J.R. Williamson. Inhibition of sorbitol
dehydrogenase. Effects on vascular and neural dysfunction in streptozocin-induced diabetic rats. Diabetes.
1995; 44:234-242 .
63. M.J. Stevens, J. Dananberg, E.L. Feldman, S.A. Lattimer, M. Kamijo, T.P. Thomas, H. Shindo. A.A.F.
Sima, D.A. Greene. The linked roles of nitric oxide, aldose reductase and (Na+,K+)-ATPase in the slowing
of nerve conduction in the streptozotocin diabetic rat. J. Clin. Invest., 1994; 94: 853-859.
64. M.F. Lou, J.R.J. Dickerson, R. Garadi, B.M. York. Glutathione depletion in the lens of galactosemic and
diabetic rats. Exp. Eye Res., 1988;46: \7-530.
65. I. Obrosova, J. Marvel, A.M. Faller, and J.R. Williamson. Reductive stress is a very early metabolic imbal-
ance in sciatic nerve in diabetic and galactose-fed rats. Diabetologia. 1995;38: A8.
66. Tilton RG, Chang K, Hasan KS, Smith SR, J.M. P, Misko TP, Moore WM, Currie MG, Corbett JA.
McDaniel ML, Williamson JR. Prevention of diabetic vascular dysfunction by guanidines. Inhibition ofni-
tric oxide synthase versus advanced glycation end-product fonnation. Diabetes 1993;42:221-232.
67. Hasan KS, Chang W, Allison A, Faller A, Santiago JV, Tilton RG, Williamson JR. Glucose-induced in-
creases in regional blood flow are prevented by aminoguanidine and L-NMMA. inhibitors of nitric oxide
synthase. FASEB J 1993;7:105.
68. Soulis·Liparota T, Cooper M, Papazoglou D, Clarke B, Jerums G. Retardation by aminoguanidine of devel-
opment of albuminuria, mesangial expansion, and tissue florescence in streptozocin-induced diabetic rat.
Diabetes 1991;40:1328-1334.
69. Hammes H-P, Martin S, Federlin K, Geisen K, Brownlee M. Aminoguanidine treatment inhibits the devel-
opment of experimental diabetic retinopathy. Proc Nat! Acad Sci. USA 1991 ;88: 11555-11558.
70. Brownlee M, Vlassara H, Kooney A, Ulrich P, Cerami A. Aminoguanidine prevents diabetes-induced arte-
rial wall protein cross-linking. Science 1986;232: 1629-1632.
71. Picard S, Parthasarathy S, Fruebis J, Witztum JL. Aminoguanidine inhibits oxidative modification of low
density lipoprotein protein and the subsuquent increase in uptake by macrophage scavenger receptors. Prox
Natl Acad Sci USA I 992;89:6876-{)880.
72. Hasan KS, Chang K, Allison W, Santiago JV, Tilton RG. Aminoguanidine-induced increases in mean arte-
rial blood pressure are suppressed in diabetic rats and nonnalized by sorbinil. Diabetes 1992;41: 195A.
73. Kamata K, Miyata N, Kasuya Y. Impainnent of endothelium-dependent relaxation and changes in levels of
cyclic GMP in aorta from streptozotocin-induced diabetic rats. Brit. J. of Phannacol. 1989;97:614--618.
74. Pelligrino DA, Albrecht RF. Chronic hyperglycemic diabetes in the rat is associated with a selective im-
painnent of cerebral vasodilatory responses. J Cerebral Blood Flow and Metabolism 1991; II :667-677.
Role of Nitric Oxide in Vascular Dysfunction Associated with Ocular Diseases 149
75. Cameron NE, Cotter MA. Impaired contraction and relaxation in aorta from streptozotocin-diabetic rats:
role of polyol pathway. Diabetologia 1992;35: 1011-1019.
76. Heygate KM, Lawrence IG, Bennett MA, Thurston H. Impaired endothelium-dependent relaxation in iso-
lated resistance arteries of spontaneously diabetic rats. Brit J Pharmacol 1995; 116:3251-3259.
77. Pieper GM, Moore-Hilton G, Roza AM. Evaluation ofthe mechanism of endothelial dysfunction in the ge-
netically-diabetic BB rat. Life Sciences I 996;58:PL 147-PLl52.
78. Langenstroer P, Pieper GM. Regulation of spontaneous EDRF release in diabetic rat aorta by oxygen free
radicals. Am J Physiol I 992;263«(Heart Circ. Physiol. 32)):H257-H265.
79. Hattori Y, Kawasaki H, Abe K, Kanno M. Superoxide dismutase recovers altered endothelium-dependent
relaxation in diabetic rat aorta. Am J Physiol 1991 ;261 «Heart Circ. Physiol. 30»:H I 086-H I 094.
80. Mayhan WG. Impairment of endothelium-dependent dilatation of the basilar artery during diabetes melli-
tus. Brain Res 1992;580:297-302.
81. Tesfamariam B, Brown ML, Deykin D, Cohen RA. Elevated glucose promotes generation of endothelium-
derived vasoconstrictor prostanoids in rabbit aorta. 1 Clin. Invest. 1990;85:929-932.
82. Tesfamariam B, Cohen RA. Free radicals mediate endothelial cell dysfunction caused by elevated glucose.
Am. J. ofPhysiol. 1992;263 (Heart Circ. Physiol. 32):H321-H326.
83. Lash JM, Bohlen HG. Structural and functional origins of suppressed acetylcholine vasodilation in diabetic
rat intestinal arterioles. Circ Res 1991 ;69: 1259-1268.
84. Mayhan WG, Simmons LK, Sharpe GM. Mechanism of impaired responses of cerebral arterioles during
diabetes mellitus. Am J Physiol 1991 ;260:H319-H326.
84. Pieper GM, Peltier BA. Amelioration by L-arginine of a dysfunctional arginine/nitric oxide pathway in dia-
betic endothelium. J Cardiovasc Pharmacol 1995;25:397-403.
85. Simon BC, Cohen RA. EDTA influences reactivity of isolated aorta from hypercholesterolemic rabbits. Am
J Physiol 1992;262:HI606-HI610.
86. Fortes ZB, Leme G, Scivoletto R. Vascular reactivity in diabetes mellitus: role of the endothelial cell. Br J
Pharmacol 1983;79:771-781.
87. Head R, Longhurst P, Panek R, Stitzel R. A contrasting effect of the diabetic state upon the contractile re-
sponses of aortic preparations from the rat and rabbit. Br J Pharmacol 1987;91:275-286.
88. Tesfamariam B, Cohen RA. Role of superoxide anion and endothelium in vasoconstrictor action of pro-
staglandin endoperoxide. Am. J. of Physiol. 1992;262:H 1915-H 1919.
89. Bohlen HG, Lash JM. Topical hyperglycemia rapidly suppresses EDRF-mediated vasodilation of normal
rat arterioles. Am J PhysioI1993;265(Heart Circ. Physiol. 34):H219-H225.
90. Tilton RG, Baier LD, Harlow JE, Smith SR, Ostrow E, Williamson JR. Diabetes-induced glomerular dys-
function: Links to a more reduced cytosolic ratio ofNADH/NAD+. Kidney International 1992;41 :778-788.
91. Cavallini L, Valente M, Rigobello MP. The protective action of pyruvate on recovery of ischemic rat heart:
comparison with other oxidizable substrates. J Mol Cell Cardiol 1990;22: 143-154.
92. Minchenko A, Bauer T, Salceda S, Caro 1 Hypoxic stimulation of vascular endothelial growth factor ex-
pression in vitro and in vivo. Lab Invest 1994;71 :374--379.
93. Aiello LP, Northrup JM, Keyt BA, Takagi H, Iwamoto MA. Hypoxic regulation of vascular endothelial
growth factor in retinal cells. Arch Ophthalmol 1995; 113: 1538-1544.
94. Williamson JR, Chang KC, Stephan CC, Brock TA, Tilton RG. Links between retinal vascular dysfunction
induced by elevated glucose levels and VEGF. Invest Ophthalmol Vis Sci 1996;37:S969.
95. Sone S, Kawakami Y, Okuda Y, Kondo S, Hanatani M, Suzuki H, Yamashita K. Vascular endothelial
growth factor is induced by long-term high glucose concentration and up-regulated by acute glucose depri-
vation in cultured bovine retinal pigmented epithelial cells. Biochem Biophys Res Comm
1996;221: 193-198.
96. Ku DD, Zaleski JK, Liu S,.Brock TA. Vascular endothelial growth factor induces EDRF-dependent relaxa-
tion in coronary arteries. Am J Physiol 1993;265:H586-H592.
97. Morbidelli L, Chang C-H, Douglas JG, Granger HJ, Ledda F, Ziche M. Nitric oxide mediates mitogenic ef-
fect ofVEGF on coronary venular endothelium. Am J Physiol 1996;270:H411-H415.
98. Yang R, Thomas GR, Bunting S, Ko A, Ferrara N, Keyt B, Ross J, Jin H. Effects of vascular endothelial
growth factor on hemodynamics and cardiac performance. J Cardiovasc Pharmacol 1996;27:838-844.
99. Arnet UA, McMillan A, Dinerman JL, Ballermann B, Lowenstein Cl Regulation of endothelial nitric ox-
ide synthase during hypoxia. J Bioi Chem 1996;271: 15069-15073.
17
ABSTRACT
LO-metabolites plays a role in the development ofEIU, and we speculate that interactions
between NO and oxygen free radicals could modulate the inflammatory response.
1. INTRODUCTION
Nitric oxide (NO) is a free radical synthesized by constitutive (cNOS) and inducible
(iNOS) nitric oxide synthase.1.2 In the eye, vascular endothelial cNOS release low levels
of NO that regulates the vascular tone in normal conditions. 3 Greater amounts of NO from
iNOS have been implicated in inflammatory disorders,4 and are released not only by
macrophages and neutrophils, but also by several ocular resident cells after in vitro stimu-
lation with lipopolysaccharide (LPS) and/or cytokines.4-7
Footpad or intraocular injection of LPS in animals elicits an intraocular inflamma-
tory response termed endotoxin induced-uveitis (EIU).8 There is a growing body of evi-
dence that NO is implicated in the pathogenesis of EIU: 1) NOS has been localized in
ocular tissues during EIU;9-11 2) the level of nitrites (NO derivatives) increases in aqueous
humor and vitreous with the time-course of EIU; 10.11 3) the severity of EIU signs, protein
leakage into the aqueous humor, and cellular infiltration has been shown to be reduced by
treatment with NOS-inhibitors such as NG-nitro-L-arginine methyl ester (L_NAME).10-14
However, conflicting results have been reported and the role of NO in EIU is still
unclear. Interactions of NO with other pro-inflammatory mediators have been described
and could modulate the inflammatory response. It has been shown that NO activates cy-
clooxygenase (COX),15 and that a dual inhibition of NOS and COX pathways attenuates
inflammation in dermatitis 16 and EIU 17 models. On the other hand, the inhibition of 5-
Jipoxygenase (5-LO) by nordihydroguayaretic acid (NDGA) inhibited NO synthesis in
cultured macrophages. 18 LTB4 is a arachidonate metabolite synthesized by 5-LO with po-
tent chemoattractant properties that has also been implicated as a mediator ofEIU. The in-
hibition of 5-LO by NDGA has been shown to reduce cellular infiltration during EIU in
rabbits 19 or S-antigen-induced uveitis in rats20 .
It has been reported that treatment with L-NAME could induce the adhesion of leu-
kocytes to vascular endothelium/I and could enhance the LPS-induced inflammation in
other models. 22 With these facts in mind, we hypothesize that the concomitant inhibition
of 5-LO could improve the anti-inflammatory activity ofL-NAME on EIU.
2.1. Animals
Male New Zealand albino rabbits (2-2.5kg of weight) were provided by an animal
breeding center, and were housed under standard conditions. All the procedures adhered
to the European Community rules on the use and care of animals in research.
2.2. Drugs
Nordihydroguayaretic acid (NDGA), purchased from Sigma (St. Louis, MO, USA),
was diluted to 20 mglml in 50%dimethylsulfoxide (DMSO) in saline (v:v). Nw-Nitro-L-
arginine methyl ester (L-NAME, Sigma) was diluted to 100 mg/ml in phosphate buffered
saline (pH 7,3). Salmonella typhimurium lipopolysaccharide (5 IJ.g/ml LPS, Sigma) was
diluted in sterile saline.
Effects of the Inhibition of Nitric Oxide Synthase and Lipoxygenase 153
2.3. Treatments
Animals were randomized in four separated groups of treatments, as follows.
2.3.1. Positive Control. Six rabbits received i.p. 0.5 mllkg 50% DMSO 12 h before
and simultaneously to the uveitis induction, and 0.5 mllkg i.p. of PBS (pH 7.3) I h before,
1 hand 3 h after the induction.
2.3.2. NDGA. Five rabbits were treated with i.p.IO mg/kg NDGA 12 h before, and at
the time of the uveitis induction. The dose was chosen from a previous report l9 •
2.3.3. L-NAME. Five rabbits were treated with i.p 50 mg/kg L-NAME Ih before,
and I and 3h after the uveitis induction.
2.3.4. NDGA+L-NAME. Five rabbits were treated with a combination ofi.p. NDGA
and i.p. L-NAME following the dosages and schedules described for the separated treat-
ments.
2.5. Sampling
Six hours after uveitis induction, animals were killed by an intravenous sodium pen-
tobarbital overdose. Aqueous humor (AH) was sampled by paracentesis, then centrifuged,
and supernatant ali quoted in separated tubes. The anterior segment of the eye was excised,
and iris/ciliary bodies (ICB) were kept in tubes containing Iml DMEM (Sigma culture
media) and 5 mM phenylmethylsulfonyl fluoride (PMSF, Sigma). All samples were kept
frozen at -80°e.
2.7.1. Nitrite (NO) Levels. Were determined using the GriessE reagent as previously
described. 12 Na 2NO was used as a standard.
154 J. L. Bellot et al.
2.7.2. PGE2 and LTB4 Levels. Were determined by using commercial specific ra-
dioimmunoassay (RIA) kits (NEN Research Products, Dupont Inc., Boston, MA, VSA).
3. RESULTS
In the positive control group, the inflammatory response evidenced 3 h after LPS in-
jection significantly increased (p < 0.01) three hours later (Fig. I). No significant increase
in the inflammatory score from 3 to 6 h was found in NDGA or L-NAME groups; how-
ever, NDGA or L-NAME alone did not change significantly the inflammatory score when
compared to the positive control group. The combination of both drugs significantly re-
duced the intensity of uveitis at 3h (p < 0.0 I) or 6h (p < 0.000 I) when compared to the
positive control group, and prevented the increase in the inflammatory score from 3 to 6 h.
As it is shown in Fig. 2A, an intense reduction ofMPO activity in ICB was produced by
NDGA treatment (4.2 ± 0.5 V/IOO mg, p < 0.0025) as compared to the positive control group
(7.2 ± 0.6 Units/lOO mg). A moderate reduction in MPO was achieved by the combination
NDGA+L-NAME (4.95 ± 0.4 V/I 00 mg, p < 0.025). L-NAME induced only slight changes in
MPO activity (5.6 ± 0.4 U/lOO mg, p < 0.05). In negative control eyes, 2.5 ± 0.5 U/lOO mg of
MPO activity were detected 6 h after the intravitreal injection of saline.
Traces of proteins in AH (0.37 ± 0.1 mg/ml) were found in normal intact eyes. Six
hours after intravitreal injection of saline (negative control group), protein levels signifi-
cantly increased to 1.65 ± 0.42 mg/ml (p < 0.05). In the positive control group, 21.2 ± 3.3
mg/ml AH proteins were determined 6 h after LPS injection (Fig. 2B). L-NAME alone
significantly changed the protein levels in AH (12.7 ± 2 mg/ml, p < 0.05) when compared
to the control group. A more pronounced effect was achieved by the combination
7 #
o 3 hours after LPS
_ 6 hours after LPS
t'
o
1 1
1; 3 Figure 1. Inflammatory score of endotoxin-
.
E
E induced uveitis 3 and 6 hours after LPS in-
jection into the vitreous of rabbits. The
~ 2
.1. severity of intraocular inflammation was
graded under slit lamp examination. The
mean score ± SEM from each group (n=IO,
except for positive control where n=12), was
'--'-- -'-- '--- -- '--- compared by statistical analysis: (#) p < 0.01
Negative Control NDGA L-NAME NDGA (3h vs 6h); (*) p < 0.0025; (**) P < 0.00 I (3h
control + L-NAME or 6h vs control).
Effects of the Inhibition of Nitric Oxide Synthase and Lipoxygenase 155
NDGA+L-NAME, which AH protein level (6.5 ± 1.3 mg/ml) was significantly lower than
those of positive control (p < 0.001) and L-NAME (p < 0.025) groups. In contrast, the pro-
tein level in NDGA group was not reduced compared to the positive control.
As is shown in figure 3, levels of nitrites (9.5 ± 1.5 11M), LTB4 (2.6 ± 1.7 ng/ml),
and PGE2 (4.5 ± 0.7 ng/ml) in the positive control group were higher than in the negative
control group. The mean level of nitrite was 3 ± I 11M in the negative control group but
levels of LTB4 and PGE 2 were not detectable.
NDGA significantly reduced L TB4 levels in AH (0.5 ± 0.4 ng/ml, p < 0.05), as ex-
pected, but did not change significantly the level of nitrites or PGE 2 when compared to
their respective controls (Fig. 3). L-NAME reduced nitrites (4.7 ± 0.6 11M, P < 0.01) and
PGE2 (0.9 ± 0.2 ng/ml, p < 0.0 I), and changed LTB4 levels (0.6 ± 0.5 ng/ml, p < 0.05) as
compared to positive control values. The combination NDGA + L-NAME significantly re-
duced nitrites (5.6 ± 111M, P < 0.05), LTB4 (0.5 ± 0.3 ng/ml, p < 0.05), and PGE2 (1.6 ±
0.5 ng/ml, p < 0.0 I) compared to the positive control group, as reported in Fig.3.
8
A
7
Oil
S
<:>
<:>
6
:::....
'" 5
'a~
'-'
.0 4
'>'-C
...
01
3
0
~ 2
~
25
B
20
:=
<
.5
--.
Figure 2. Myeloperoxidase activity in e 15
Oil
the iris/ciliary body (A), and proteins
levels in the aqueous humor (8) of
!';j'
=
..
rabbit eyes 6 hous after LPS injection. 10
They were used to assess the effects of
!0
treatments on the cellular infiltration ~
and on the blood-aqueous barrier in-
tegrity, respectively. MPO activity and 5
protein levels from each treatment
group (mean ± SEM, n=IO) were com-
pared to that found in the positive con- 0
trol group (n=12). (*) p < 0.05; (**) p Negative Control NDGA L-NAME NDGA
< 0.025; (***) p < 0.0025. control +L-NAME
156 J. L. Bellot et al.
4. DISCUSSION
In the present study we have demonstrated that the concomitant treatment with NDGA
(a 5-lipoxygenase inhibitor) clearly improves the anti-inflammatory activity of L-NAME dur-
ing the early phase ofEIU. The reciprocal is also true, since L-NAME also improved the anti-
inflammatory activity ofNDGA. Thus, the specific modifications in the pathophysiology of
EIU produced by the combination of the drugs is more efficacious than either drug alone.
The beneficial effect ofNDGA on L-NAME treatment could be due to the reduction
in MPO activity (i.e., leukocyte infiltration) in iris/ciliary body, correlating with the re-
duction of LTB4 levels. A scavenging action of hydroxyl radicals by DMS0 25 should also
be considered. This suggests that L-NAME could induce leukocyte chemotaxis or adhe-
sion, and the release of oxygen free radicals, as has been previously evidenced. 16.2&--28 Oxy-
12
10
=
< 8
.5
~
~
ell
6
:s
~
4
~
2
B o LTB4 levels
- PGE2 levels
=
<
5
.5
:::::- 4
5
'BiI
E-
M
~ 3
(,;I
e:...
0
~ 2
...!Xl
::.
Figure 3. Nitrite (A), LTB4 and PGE,
(B) levels in aqueous humor 6 hours after
intravitreal injection of LPS in rabbits.
Bars represent mean ± SEM. Each pa-
rameter from treatment groups (n= I 0)
o -'----------'-- were compared to their corresponding
Negative Control NDGA L-NAME NDGA positive controls (n=12). (*)p < 0.05;
control +L-NAME (**)p < 0.01.
Effects ofthe Inhibition of Nitric Oxide Synthase and Lipoxygenase 157
gen free radicals could induce leukocyte adhesion, vascular endothelial injury/6--28 and are
one of the principal mediators ofEIU. 29
In addition, it has been shown that the inhibition of NO could promote leukocyte ad-
hesion through the oxygen free radical-induced expression ofCDlllCD18 adhesion mole-
cules,27 and that treatment with CDll/CD18 MoAb attenuated EIU in rabbits. 30 The
reduction of infiltrating leukocytes achieved by NDGA was not followed by a parallel at-
tenuation in the inflammatory score or in the protein leakage, suggesting that a leukocyte-
independent mechanism contributes to the microvascular dysfunction during the early
phase of EIU. The mechanism could be mediated by NO and/or PGE 2 because they were
not reduced by NDGA at the doses used.
This conclusion is supported by the fact that L-NAME alone reduced the protein
leakage into the aqueous humor, correlating with the reduction of nitrite and PGE2 levels.
Thus, NO could activate -or L-NAME could inhibit- cyclooxygenase during the early
phase of EIU; however, during the peak of EIU (18 to 24 h after LPS), L-NAME did not
inhibit PGE2 synthesis. 17 This discrepancy has been previously found in vitro by Swierk-
osz et al./ I and could explain the diversity of effects obtained by differential inhibition of
cNOS or iNOS, or the time-dependent variability of these treatments on EIU .11.14 In addi-
tion, pathophysiological interactions between NO and PGE217.32 could also explain the
changes in vascular permeability.
Although MPO activity and LTB4 levels were slightly decreased, the inflammatory
score was not significantly reduced by pretreatment with L-NAME. Because the combina-
tion of NDGA in DMSO with L-NAME achieved a greater anti-inflammatory effect than
other treatments, we speculate that the imbalance between NO and oxygen free radicals is
a key factor in the pharmacological treatment of EIU.
However, this consideration should be confirmed, and further studies are still needed
to elucidate the possible interactions of NO with the complex network of mediators in-
volved in EIU.
5. ACKNOWLEDGMENTS
The authors would like to thank the technical assistance of Miguel A. Company and
Miguel A. Diez. This study has been supported in part by Ciba-Vision, Barcelona, Spain,
and Ciba-Vision, Basel, Switzerland.
6. REFERENCES
I. Nathan C: Nitric oxide as a secretory product of mammalian cells. FASEB J. 1992;6:3051-64.
2. Moncada S, Palmer RMJ, Higgs EA: Nitric oxide: physiology, pathophysiology, and pharmacology. Phar-
maco!. Rev. 1991;43:109-142.
3. Donati G, Poumaras CJ, Munoz JL, Poitry-Yamate CL, Tsacopoulos M: Nitric oxide controls arteriolar
tone in the retina of the miniature pig. Invest. Ophthalmol. Vis. Sci. 1995;36:2228-2237.
4. Nussler AK, Billiar TR: Inflammation, immunoregulation, and inducible nitric oxide synthase. J. Leuko-
cyte Bio!. 1993; 54:171-178.
5. Chakravarthy U, Stitt AW, McNally J, Bailie JR, Hoey EM, Duprex P: Nitric oxide synthase activity and
expression in retinal capillary endothelial cells and pericytes. Curro Eye Res. 1995; 14:285-294.
6. Goureau 0, Hicks D, Courtois Y: Human retinal pigmented epithelial cells produce nitric oxide in reponse
to cytokines. Biochem. Biophys. Res. Comm. 1994; 198: 120-126.
7. Goureau 0, Hicks D, Courtois Y, de Kozak Y: Induction of nitric oxide synthase in retinal Miiller glial
cells. J. Neurochem. 1994; 633:3 I 0-3 I 7.
IS8 J. L. Bellot el aL
8. Rosenbaum JT, McDevitt HO, Guss RB, ;:;gbert PR: Endotoxin-induced uveitis in rats as a model for hu-
man disease. Nature (London). 1980; 286:611--613.
9. Jacquemin E, de Kozak Y, Thillaye B, Courtois Y, Goureau 0: Expression of inducible nitric oxide syn-
thase in the eye from endotoxin-induced uveitis rats. Invest. Ophthalmol. Vis. Sci. 1996; 37: 1187-1196.
10. Mandai M, Mittag TW, Kogishi J, Iwaki M, Hangai M, Yoshimura: Role of nitric oxide synthase isozymes
in endotoxin induced uveitis. Invest. Ophthalmol. Vis. Sci. 1996;37:826-832.
II. Goureau 0, Bellot J, Thillaye B, Chatenoud L. de Kozak Y: Increased nitric oxide production in endotoxin-
induced uveitis. Reduction of uveitis by an inhibitor of nitric oxide synthase. J. Immunol. 1995;
154:6518--6523.
12. Parks DJ, Cheung MK, Chan CC, Roberge FG: The role of nitric oxide in uveitis. Arch. Ophthalmol. 1994;
112:544--546.
\3. Tilton RG, Chang K. Corbett JA, Misko TP, Currie MG, Bora NS, Kaplan HJ, Williamson JR: Endotoxin-
induced uveitis in the rat is attenuated by inhibition of nitric oxide production. Invest. Ophthalmol. Vis.
Sci. 1994; 35:3278--3288.
14. Allen JB. McGahan MC. Ferrell JB. Adler KB, Fleisher LN: Nitric oxide synthase inhibitors exert differen-
tial time-dependent effects on LPS-induced uveitis. Exp. Eye Res. 1996; 62:21-28.
15. Salvemini D. Misko TP. Masferrer JL, Seibert K, Currie MG, Needleman P: Nitric oxide activates cy-
clooxygenase enzymes. Proc. Natl. Acad. Sci. USA. 1993; 90:7240-7244.
16. Salvemini D. Manning PT. Zweifel BS. Seibert K, Connor J. Currie MG. Needleman P, Masferrer JL: Dual
inhibition of nitric oxide and protaglandin production contributes to the antiinflammatory properties of ni-
tric oxide synthase inhibitors. J. C1in. Invest. I 995;96:301-308.
17. Bellot JL. Palmero M. Garcia-Cabanes C. Esp! R. Hariton C, Orts A: Additive effect of nitric oxide and
prostaglandin-E, synthesis inhibitors in endotoxin-induced uveitis in the rabbit. Inflamm. Res. 1996;
45:203-208.
18. Imai Y. Kolb H. Burckart V: Nitric oxide production from macrophages is regulated by arachidonic acid
metabolites. Biochem. Biophys. Res. Comm. 1993; 197:105--109.
19. Ruiz JM. Sifre J. Alia JL. Ruiz O. Bellot JL: Nordihydroguayaretic acid and arachidonic acid in the endo-
toxin-induced uveitis. Arch. Soc. Esp. Oftalmol. 1995; 68:31-38.
20. Chen F. Pararajasegaram G, Sevanian A, Rao N: Treatment of S antigen uveoretinitis with lipoxygenase
and cyclo-oxygenase inhibitors. Ophthalmic Res. 1991;23:84--91.
21. Kubes P, Granger DN: Nitric oxide: an endogenous modulator of leukocyte adhesion. Proc. Natl. Acad. Sci
USA. 1991 ;88:4651--4655.
22. Boughton-Smith NK, Evans SM. Laszlo F. Whittle BJR, Moncada S: The induction of nitric oxide syn-
thase and intestinal vascular permeability by endotoxin in the rat. Br. J. Pharmacol. 1993; II 0: 1189-1195.
23. Bradford MM: A rapid sensitive method for the quantification of microgram quantities of protein utilizing
the principle of protein-dye binding. Anal. Biochem. 1976; 72:465-470.
24. Williams RN. Paterson CA. Eakins KE. Bhattacherjee P: Quantification of ocular inflammation: Evalu-
ation of polymorphonuclear infiltration by measuring myeloperoxidase activity. Curro Eye Res.
1983;2:465-470.
25. Bonne C, Latour E, Muller A, de Kozak Y. Faure JP. Malet F, Colin J, Tissot M, Giroud JP, Maghni K. Si-
rois P, Griswold DE. Coquelet C: 2-(2-hydroxy-4-methylphenyl) aminothiazole hydrochloride as a dual in-
hibitor of cyclooxygenase/lipoxygenase and a free radical scavenger: antiinflammatory activity.
Arzneim.Forsch.lDrug Res. 1989;39:1246-1250.
26. Beckman JS, Beckman TW, Chen J, Marshall AM, Freeman BA: Apparent hydroxyl radical production by
peroxynitrite: implications for endothelial injury from nitric oxide and superoxide. Proc. Natl. Acad. Sci.
USA. 1990; 87:1620-1624.
27. Kurose I, Wolf R, Grisham MB, Aw TY, Specian RD, Granger DN: Microvascular responses to inhibition
of nitric oxide production. Role of active oxidants. Circ. Res. 1995;76:30-39.
28. Rotzinger S, Aragon CM. Rogan F. Amir F. Amit Z: The nitric oxide synthase inhibitor NW-L-nitro ar-
ginine methyl ester attenuates brain catalase activity in vitro. Life Sci. 1995;56: \321-\324.
29. Rao NA, Romero JL. Fernandez MAS, Sevanian A, Marak GE Jr: Role of free radicals in uveitis. Surv.
Ophthalmol. 1987;32:209-2 \3.
30. Rosenbaum JT, Boney RS. Efficacy of antibodies to adhesion molecules CDlla or CDI8 in rabbit models
of uveitis: Curro Eye Res. 1993; 12:827--831.
31. Swierkosz TA, Mitchell JA, Tomlinson A, Botting RM, Warner TD, Vane JR: Relationship between differ-
ent isoforms of nitric oxide synthase and cyclooxygenase in various cell types. Pol. J. Pharmacol.
1994;46:587-592.
32. Hiraki S, Zhang XV, Hayasaka S: Effects of a nitric oxide synthase inhibitor on prostaglandin-induced
aqueous flare elevation in pigmented rabbits. Ophthalmic Res. 1996;28:260-264.
18
1. ABSTRACT
Polymethylmethacrylate (PMMA) is the standard intraocular lens (IOL) optic mate-
rial based upon its long history of clinical use, stability and biocompatibility. Regulatory
guidelines require IOLs that utilize novel lens material to be evaluated for biocompatibil-
ity in a long term (I year) ocular implant study using an appropriate species and a refer-
ence lOt, e.g., PMMA. The purpose of this study was to evaluate the rabbit as a potential
model for conducting long-term ocular implant studies that satisfy regulatory biocompati-
bility guidelines. In this study, lensectomy by phacoemulsification with implantation of
PMMA IOLs in the capsular bag was performed on one group of rabbits, while a second
group received only implantation of PMMA wafers (3 x 5 x O.4mm) in the anterior cham-
ber without lensectomy. Animals received ocular examinations at selected intervals during
the study, e.g., slit-lamp biomicroscopy. Except for the expected postoperative inflamma-
tion, no signs of inflammation were observed in either rabbit model for up to approxi-
mately 3 months following implantation of the PMMA IOLs. However, after about 3
postoperative months rabbits with the PMMA IOLs in the capsular bag developed poste-
rior synechiae, posterior capsular opacification (pearls and fibrosis), uveitis, and secon-
dary cataracts. In some rabbits extrusion/displacement of the IOL haptics caused
mechanical trauma to the iris with subsequent hyphema. In contrast, the rabbits that were
implanted with the PMMA wafers in the anterior chamber did not display any signs of in-
flammation during a one year study period. Histopathology confirmed the presence of the
ocular inflammation which was observed in the rabbits with the IOLs implanted in the
capsular bag. These results demonstrate that studies using the rabbit to evaluate IOLs fol-
lowing lensectomy and capsular bag implantation should be limited to a shorter study du-
ration of 6 months or less. In contrast, a longer study duration of up to one-year may be
conducted in rabbits following implantation of IOL wafers into the anterior chamber. Fur-
thermore, these studies demonstrate that the site of implantation and duration of study can
influence results obtained in an ocular biocompatibility study.
2. INTRODUCTION
Cataracts are a pathologic cloudiness of the crystalline lens that may be congenital
or caused by trauma, disease, or age. Surgical removal of the opaque lens followed by re-
placement with an intraocular lens (IOL) has become a common medical procedure.
A key clinical and regulatory concern exists with the biocompatibility of the IOL,
which may be derived from both the chemical composition and physical design of the
IOL. To address this concern, a long term (one year duration) ocular implantation study in
an appropriate animal model is required for regulatory submissions for IOLs with a novel
chemical composition or physical design. This ocular implantation study should be con-
ducted with the final physical and chemical form of the test IOL and utilize a reference
IOL. Polymethylmethacrylate (PMMA) is the standard intraocular lens (IOL) optic mate-
rial based upon its long history of clinical use, stability and biocompatibility and, there-
fore, is typically used as the reference IOL.
The rabbit is a species that is commonly utilized in the area of ocular toxicology. A
five month IOL implant study with limited data as well as several reports of shorter dura-
tion (less than or equal to three months) have been reported in the rabbit. '- 5 However,
these studies do not provide adequate background data to fully determine the usefulness of
the rabbit for conducting longer term ocular implant studies with IOLs. Therefore, the pur-
pose of this study was to evaluate two different ocular implantation techniques in the rab-
bit and determine their suitability for conducting a long term ocular implant study that
satisfies regulatory biocompatibility guidelines.
prednisolone acetate were instilled topically onto the eyes four times a day for one week.
Two drops of topical ocular I % atropine were instilled twice a day from the second
through the fourth weeks following the capsular implantation of the PMMA IOL.
Postoperative examinations included biomicroscopy of the IOL and surrounding tis-
sues. Examinations occurred at approximately I week and at 1,2,3, and 6 months follow-
ing IOL implantation. Three rabbits implanted with the PMMA IOLs were sacrificed prior
to the six month examination. Following the six month examination, all remaining animals
were sacrificed. Eyes were removed and immersion-fixed in Davidson's fixative and pre-
served in 10% buffered formalin. The eyes were processed for light microscopy using step
sectioning and staining with hematoxylin and eosin.
4. RESULTS
(a) (b)
(c)
Figure I. Photographs from biomicroscopic (slit-lamp) examinations of a rabbit eye implanted with a PMMA
IOL in the capsule at approximately the following times post-implantation: (a) I week; (b) I month; (c) 3 months:
and (d) 6 months. Diffuse corneal edema was present at I week. which later localized to the superior limbal inci-
sion. Fibrin in the anterior chamber and iridal edema were typically present by 6 months. For a color repre-
sentation of this figure. see the color insert facing this page.
(a)
Figure 2. Photographs from biomicroscopic (slit-lamp) examinations of rabbit eyes implanted with PMMA IOLs
in the capsule at approximately 6 months post-implantation. Posterior capsular opacification, conjunctival conges-
tion. posterior synechiae and deposits on the anterior surface of the IOL (a) as well as hyphema (b) are present at
6 months. The hyphema was attributed to extrusion of an IOL haptic from the capsular bag and subsequent me-
chanical erosion into the iridal tissue. For a color representation of this figure. see the color insert facing this page.
a b
c d
Chapter IS, Figure I. Photographs from biomicroscopic (slit-lamp) examinations of a rabbit eye implanted with
a PMMA IOL in the capsule at approximately the following times post-implantation: (a) I week; (b) I month; (c)
3 months: and (d) 6 months. Diffuse corneal edema was present at I week. which later localized to the superior
limbal incision. Fibrin in the anterior chamber and iridal edema were typically present by 6 months.
a b
Chapter IS, Figure 2. Photographs from biomicroscopie (slit-lamp) examinations of rabbit eyes implanted with
PMMA IOLs in the capsule at approximately 6 months post-implantation. Posterior capsular opacification, con-
junctival congestion, posterior synechiae and deposits on the anterior surface of the IOL (a) as well as hyphema
(b) are present at 6 months. The hyphema was attributed to extrusion of an IOL haptic from the capsular bag and
subsequent mechanical erosion into the iridal tissue.
a b
c d
Chapter 18, Figure 3. Photographs from biomicroscopic (slit-lamp) examinations of a rabbit eye that received a
sham surgery at approximately the following times post-surgery: (a) I week; (b) I month; (c) 3 months; and (d) 6
months. Corneal edema localized to the superior limbal incision at I week and later resolved. A conical-shaped
secondary cataract (arrow) is present at 6 months.
a b
Chapter 18, Figure 4. Photographs from biomicroscopic (slit-lamp) examinations of two rabbit eyes, (a) and (b),
implanted with PMMA wafers in the anterior chamber at approximately I year post-implantation. The PMMA wa-
fers localized to the inferior anterior chamber throughout the study. Eyes were non-reactive throughout the study
duration.
Evaluation of Two Rabbit Ocular Implantation Models 163
(a) (b)
Figure 3. Photographs from biomicroscopic (slit-lamp) examinations of a rabbit eye that received a sham surgery
at approximately the following times post-surgery: (a) I week; (b) 1 month; (c) 3 months; and (d) 6 months. Cor-
neal edema localized to the superior limbal incision at I week and later resolved. A conical-shaped secondary cata-
ract (arrow) is present at 6 months. For a color representation of this figure, see the color insert facing this page.
ate conjunctival congestion and swelling as well as corneal edema were observed during
the first week post-surgery (Figure 3a). No overt signs of ocular inflammation were ob-
served in rabbit eyes that received sham surgeries for up to about 3 months post-surgery
(Figure 3b and 3c). During 3 t06 months following the sham surgeries, rabbit eyes devel-
oped "doughnut" and conical-shaped secondary cataracts (Figure 3d). In general, ocular
inflammation in rabbit eyes that received sham surgeries was similar in nature to eyes im-
planted with the PMMA IOLs but of a lesser degree.
Histopathologic analysis of the ocular tissues following the animal sacrifice revealed
changes consistent with observations from the biomicroscopic examinations (Table 1).
The histopathologic changes were evident in the iris, vitreous, anterior chamber and cor-
nea of all PMMA IOL-implanted eyes and sham-operated eyes. Incidental areas of focal
retinal degeneration also were observed in eyes implanted with PMMA IOLs and sham
surgery eyes.
week postimplantation. Thereafter no ocular inflammation was present in any eye for the
duration of the study (Figure 4). Only a small focal area of corneal edema was present in
rabbit eyes due to contact between the implanted wafer and the corneal endothelium. No
other ocular abnormalities were noted in the implanted eyes during the study.
Histopathologic analysis of the ocular tissues following the animal sacrifice revealed
significant changes consistent with observations from the biomicroscopic examinations
(Table I). A thickened descements membrane in two of the rabbit eyes was observed,
while one eye had a minimally reactive endothelium. No other lesions were observed in
the eyes implanted with the PMMA wafers.
5. DISCUSSION
The rabbit is a species that is commonly utilized in the area of ocular toxicology, in-
cluding IOL biocompatibility. A five month IOL implant study with limited data as well
as several reports of shorter duration (less than or equal to three months) have been re-
ported in the rabbit.'-s However, these studies do not provide adequate background data to
fully determine the usefulness of the rabbit for conducting longer duration ocular implant
Evaluation of Two Rabbit Ocular Implantation Models 165
(a) (b)
Figure 4. Photographs from biomicroscopic (slit-lamp) examinations of two rabbit eyes implanted with PMMA
wafers in the anterior chamber at approximately 1 year post-implantation. The PMMA wafers localjzed to the in-
ferior anterior chamber throughout the study. Eyes were non-reactive throughout the study duration. For a color
representation of this figure, see the color insert facing p. 163.
studies with IOLs. Therefore, the purpose of this study was to evaluate two different ocu-
lar implantation techniques in the rabbit and determine their suitability for conducting a
long term ocular implant study that satisfies regulatory biocompatibility guidelines.
Ocular inflammation, as an adverse effect following IOL implantation, has been as-
sociated with implant surface irregularities,6-8 the sterilization procedure,9-" and the IOL
material composition.1 2-1 5 Potential inflammatory mediators were minimized during this
study by using a common sterilization procedure and the typical reference IOL material
(i.e., PMMA). Therefore, the study design allowed a better evaluation of the rabbit model
dependent on implant location in the eye.
Biomicroscopic (slit-lamp) examinations during the two studies revealed differences
in the reactivity of the rabbit eye to the IOL implants. Observations support that the site of
implantation and the duration of the study directly influence interpretation of results in an
ocular implantation study. Knowledge of the animal model and the issues being addressed
are vital to the design of such studies.
Similar to previous work conducted in the nonhuman primate,I6-17 this study demon-
strated that the implantation of an anterior chamber PMMA IOL wafer was well tolerated
by the phakic rabbit eye. The anterior chamber implantation in the phakic rabbit eye elic-
ited a minimal, if any, inflammatory response, while a persistent inflammation during a
two-year anterior chamber implantation in the phakic primate eye was reported by Peiffer
et al. I6 Furthermore, Peiffer et al. reported iris alterations, such as focal atrophy, haptic
erosion and pupillary irregularities,16 which were not observed in this rabbit study. The
dissimilarity in observed responses between the rabbit and primate may stem from shape
and size differences of the IOL implants. However, the observed differences in the rabbit
and primate are not considered due to the IOL composition since both implants were made
fromPMMA.
The capsular placement of the PMMA IOL was tolerated in the rabbit model for up
to approximately 6 months following implantation. Other investigators have utilized the
rabbit for capsular IOL implantation studies of shorter duration (typically three months or
less) with similar findings of ocular compatibility. 1-5 However, after about 3 postoperative
months, rabbits with the PMMA IOLs in the capsular bag began to develop posterior
166 J. N. Norton etal.
synechiae, posterior capsular opacification (pearls and fibrosis), fibrosis, uveitis and sec-
ondary cataracts. The development of an uveitis-glaucoma-hyphema (UGH) syndrome
was apparent in several of these rabbits/,I 8-2 I whereby displacement of the TOL haptics
from the capsule caused mechanical erosion to the iridal tissues with subsequent hyphema.
A similar displacement of IOL haptics from the capsular bag in rabbits has been reported,
whereby the lens luxation was attributed to the unique anatomy and rapid inflammatory
response of the rabbit eye. I If a capsular implantation study of greater than 6 months dura-
tion is necessary, the use of an alternative species, such as the cat and nonhuman primate,
should be considered. 22 •23
In conclusion, the results from this study demonstrate that using the rabbit to evalu-
ate the biocompatibility of IOLs following lensectomy and capsular bag implantation
should be limited to a shorter term study duration of 6 months or less. In contrast, longer
term studies of up to a one-year duration may be conducted in rabbits with implantation of
IOL wafers into the anterior chamber.
6. ACKNOWLEDGMENTS
We thank Wendy Martin, who provided excellent technical expertise and coordi-
nated the data for this study.
7. REFERENCES
I. Cook CS, Peiffer RL. Mazzocco TR. Clinical and pathologic evaluation of a flexible silicone posterior
chamber lens design in a rabbit model. J Cataract Refract Surg 1986;12:130-4.
2. Hansen SO, Solomon KD, McKnight GT, Wilbrandt TH, Gwin TD, O'Morchoe DJC. Tetz MR. Apple DJ.
Postcrior capsular opacification and intraocular lens decentration Part I: Comparison of various posterior
chamber lens designs implanted in the rabbit model. J Cataract Refract Surg 1988; 14:605-13.
3. Tetz MR. O'Morchoe DJC. Gwin TD, Wilbrandt TH, Solomon KD, Hansen SO. Apple DJ. Posterior cap-
sular opacification and intraocular lens decentration Part II: Experimental findings on a prototype circular
intraocular lens design. J Cataract Rerract Surg 1988; 14:614--23.
4. Solomon KD, Gwin TD. O'Morchoe DJC. Tetz MR. Hansen SO, Sugita A, Imkamp EM, Apple DJ. Protec-
tive effect of the anterior lens capsule during extracapsular cataract extraction. Ophthalmology
1989;96:591-7.
5. Tamura M, Kanagawa R, Saika S, Ohmi S. Nakao T. Kinoshita C. Uenoyama K. Comparison of the cellu-
lar response on intraocular lenses implanted in rabbit eyes with and without extracapsular lens extraction. J
Cataract Refract Surg 1990; 16:746-50.
6. Keates RH, Ehrlich DR. "Lenses of chance" Complications of anterior chamber implants. Ophthalmology
1978;85:408-14.
7. Olson RJ. Intraocular lens manufacturing quality. In: Rosen ES, Haining WM, Arnott EJ, eds. Intraocular
Lens Implantation. St Louis. 9(}-8. 1984, CY Mosby.
8. Mamalis N, Apple DJ, Brady SE, Notz RG, Olson RJ. Pathological and scanning electron microscopic
evaluation of the 91Z intraocular lens. Am Intra-Ocular Implant Soc J 1984; I 0: 191-9.
9. Boyaner D, Solomon LD. Ocular reaction to the use of wet-pack versus dry-pack intraocular lenses. Am
Intra-Ocular Implant Soc J 1980;6:252-4.
10. Stark WJ. Wet-pack vs dry-pack lenses. Am Intra-Ocular Implant Soc J 1978;4:213.
II. Stark W J, Rosenblum P. Maumenee AE, Cowan CL. Postoperative inflammatory reactions to intraocular
lenses sterilized with ethylene-oxide. Ophthalmology 1980;87:385--9.
12. Kincaid MC, Green WR, Iliff W J. Granulomatous reaction to Choyce style intraocular lens. Ophthalmic
Surg 1982; 13:292-9.
13. Sievers H, von Domarus D. Foreign-body reaction against intraocular lenses. Am J Ophthalmol
1984;97:743-51.
Evaluation of Two Rabbit Ocular Implantation Models 167
14. Stinson NE. Tissue reaction induced in guinea pigs by particulate polymethylmethacrylate, polythene and
nylon of the same size range. Br J Exp Pathol 1965;46: 135-46.
15. Wolter JR. Foreign both giant cells on intraocular lens implants. Graefes Arch Clin Exp Ophthalmol
1982;219: 103-11.
16. Peiffer RL, Porter DP, Eifrig DE, Boyd J. Experimental evaluation ofa phakic anterior chamber implant in
a primate model Part I. Clinical observations. J Cataract Surg 1991; 17:335-41.
17. Porter DP, Peiffer RL, Eifrig DE, Boyd J. Experimental evaluation ofa phakic anterior chamber implant in
a primate model Part II. Pathology. J Cataract Surg 1991;17:342-52.
18. Ellingson FT. Complications with the Choyce Mark VII anterior chamber lens implant (uveitis-glaucoma-
hyphema). Am Intra-Ocular Implant Soc J 1977;3: 199-205.
19. Choyce DP. Complications of the AC implants of the early 1950's and the UGH or Ellingson syndrome of
the late 1970's. Am Intra-Ocular Implant Soc J 1978;4:22-9.
20. Ellingson FT. The uveitis-glaucoma-hyphema syndrome associated with the Mark V\II anterior chamber
lens implant. Am Intra-Ocular Implant Soc J 1978;4:50-3.
21. Percival SP, Das SK. UGH syndrome after posterior chamber lens implantation. Am Intra-Ocular Implant
Soc J 1983;9:200-1.
22. Buchen SY, Richards SC, Solomon KD, Apple OJ, Knight PM, Christ R, Ph am LT, Nelson DL, Clayman
HM, Karpinski LG. Evaluation of the biocompatibility and fixation of a new silicone intraocular lens in the
feline model. J Cataract Refract Surg 1989; 15:545-53.
23. Fogle JA, Blaydes JE. Fritz KJ. Peiffer RL, Cook C, Wright E. Clinicopathologic observations of a silicone
posterior chamber lens in a primate model. J Cataract Refract Surg 1986; 12:281-4.
19
CHARACTERIZATION OF IMMORTALIZED
LENS EPITHELIAL CELLS AS A POTENTIAL IN
VITRO ALTERNATIVE MODEL FOR THE
CELLULAR TOXICITY AND THE EFFICACY
EVALUATIONS OF OCULAR DRUG
CANDIDATES
1. ABSTRACT
Immortalized rabbit lens epithelial cell lines were established by transfection of pri-
mary lens epithelial cells with origin defective Simian virus DNA (OrrSV 40) containing
large T antigen coding sequence. Lens epithelial cell line, RLEpi-T3 (DF-12), has a con-
tinual expression of the SV40 large T antigen in the nuclei indicating these cells were
truly immortalized. Positive immunofluoresence staining confirmed the presence of both
cytokeratin AE5 and vimentin intermediate filaments in the cytoskeleton of these cells. In-
direct immunocytochemical staining with anti-a crystallin monoclonal antibody demon-
strates that these lens epithelial cells do express a-crystallin, a specific marker for lens
epithelial cells, in the cytoplasm of these cells. In serum-free, HEPES buffered DF-12 (a
1:1 ratio of Dulbecco's Modified Eagle's Medium and Ham's F-12 nutrient mixture) cul-
ture medium, these cells have a cobblestone-like morphology before reaching confluency
and further develop into multi layers at the overconfluent stage. The average 24-hr plating
efficiency ranges from 35.0 ± 6.6% to 127.0 ± 25.6% and the average population doubling
time ranges from 15.0 ± 1.1 to 20.7 ± 2.1 hrs for these lens epithelial cell lines. Myco-
plasma test results indicate that all established immortalized lens epithelial cell lines are
free of mycoplasma contamination. Our results demonstrate that these lens epithelial cells
retain their key lens-specific differentiation characteristics in addition to their faster
growth rate which makes them an attractive in vitro alternative model for lens-specific
cellular toxicity and efficacy evaluations of ocular drug candidates.
2. INTRODUCTION
Cataract, an ocular disease characterized by progressive lens opacity and visual loss,
is a major cause of blindness in man l - 3 • The etiology of cataracts is not certain. Studies
have demonstrated that cataracts may be related to risk factors such as oxidative stress4-{"
diabetes 7-8, ultraviolet radiation 9 , medicines 10- 13 and malnutrition. I4-18 However, the real
causes and initiating factors have not yet been identified. Therefore, a suitable in vitro cell
culture model will be useful for screening the safety profiles of new ocular drug candi-
dates for lens-associated adverse effects and for studying the mechanism of cataractogene-
sis. One important criterion is that this in vitro cellular model must retain the key lens
epithelial cell differentiation characteristics that distinguish it from other ocular cells or
somatic cell lines in culture. Accordingly, the lens-specific responses to therapeutic or
toxic agents can best be reflected when a study is conducted in this lens-specific model.
Logically, primary cultures of lens epithelial cells may best serve as a potential in vitro al-
ternative model since they are the only live biological system in the ocular lens. However,
the paucity of obtainable lens epithelial cells and the low proliferation potential of these
cells limits this model. Primary lens epithelial cells are slow growing in culture. There-
fore, the number of available cells depends on the availability of the source animals. A
more attractive and useful in vitro model will need to be more uniform in its responses
and constantly available. To achieve this end, several difficulties must be overcome. First,
the batch to batch differences are observed in studies using cultured cells; This can be cir-
cumvented by introducing the immortalized gene (SV40 DNA containing T antigen cod-
ing sequence, clone 6-1) that will confer in these cells an indefinite lifespan and allow it
to select cells derived from the same clone. Conceivably, the responses from cells derived
from the same clone will be much more uniform when treated with toxins or drug candi-
dates. A second advantage of this approach is that large quantities of cells with consistent
biochemical characteristics can be raised in a short period of time for meaningful large
scale in vitro experimentations. In our laboratory, we have established several immortal-
ized lens epithelial cell lines from primary rabbit lens epithelial cells as the initial step in
developing in vitro cellular models which can be used to address lens-specific safety as-
sessment. In this communication, we present the data on cell immortalization and the pre-
liminary characterization of these immortalized lens epithelial cell lines.
every 48 hrs until the cultures were ready for transfection. Serum-free DF-12 medium l9
with the same supplements as in serum-free M-199 medium was also used in maintaining
the primary cultures of lens epithelial cells and the medium change schedule is the same
as serum-free M-199 medium.
3.3.2. Staining ofCytokeratin AE5. Vimentin. T Antigen and a-Crystallin. The method
of Tsai and Tseng 22 was followed for the staining of anti-a-crystallin (AaC), anti-keratin
(or AE5), anti-SV40 T antigen (AT), anti-vimentin (AV) and anti-a-crystallin antibodies.
Briefly, cells grown on glass coverslips were fixed with cold ethanol containing 3% acetic
acid for 10 minutes, rinsed with balanced salt solution (BSS) and blot dried. Primary anti-
bodies, AaC, AE5, AV, or AT at a concentration of 2 ug/ml in phosphate buffered saline
(PBS) containing 0.1 % BSA were added to the cells and incubated in a humidified cham-
ber at room temperature for 2 hrs. Cells were then rinsed thoroughly to remove unbound
primary antibodies. The secondary antibody, anti-mouse IgG(H+L)-FITC or IgG(H+L)-
Texas Red at a concentration of 2 ug/ml in PBS, was applied to the primary antibody-
treated cells and incubated in a humidified chamber for an additional 2 hrs. The resulting
coverslips were rinsed extensively to remove unbound fluorescein or Texas Red, mounted
to a microscope slide with Fluoromount to retard the photobleaching of dyes and the fluo-
rescent image of cells recorded microscopically with a Leica Fluovert-FU Fluorescence
172 C. Vao et al.
microscope equipped with proper fluorescent filters and a motorized camera attached to
the Leica Wild MPS46/52 photo automat photographic control system.
3.3.4. Cytogenetic Analysis. The procedures for cell preparation and chromosome
staining was mainly derived from HSU 24 and Benn and Perle 25 , respectively, with minor
modifications. Briefly, cells at 80% con fluency were selected. Fresh medium containing
0.06 ug/ml Colecemid was added to arrest the cells in culture and subsequently incubated
for additional 6 hours before harvest. Cells were then harvested by Trypsin/EDTA. The su-
pernatant was aspirated and a hypotonic solution (I :2, v/v, culture medium: water) was
added to gently suspend cells and let stand (7 mins). Cells were then centrifuged at 200g
for 3 minutes. Fresh cold fixative (3 parts of methanol and I part of glacial acetic acid was
added dropwise and the cells were maintained at room temperature for 20 minutes before
dropping or smearing onto a ethanol-cleaned glass slide. Smears were stained with 0.5%
quinacrine dihydrochloride according to Benn and Perle 25 Briefly, slides were immmersed
in the quinacrine dihydrochloride solution for 10 minutes, then quenched for 2 minutes in
MacIlvaine's buffer. Slides were then mounted with MacIlvaine's buffer or allowed to dry
and observed under oil immersion objective (lOOOx) using a Leica Flurovert FU fluores-
cence microscope with G/R fluorescent filter cube and photographs were taken with an at-
tached Wild MPS46 photoautomat for fluorescent image of cells.
3.3.5. Mycoplasma Testing. Rabbit lens epithelial cell lines were tested to ensure
that mycoplasma contaminations are not present in our cultured cells by American Type
Culture Collection (ATCC, Rockville, Maryland) Applied Sciences27 • Briefly, cells were
cultured in original serum-free medium deprived of any antibiotic and antimycotic for at
least 2 weeks before sending the cultures for mycoplasma contamination test. Both indi-
rect (Hoechst stain) and direct (agar) methods were used to confirm the results. The proce-
dures for both methods can be acquired from ATCC Quality Control Methods for Cell
Lines 26 .
4. RESULTS
•
•
Bubble the calcium phosphate solution containing DNAs
Figure 1. Procedures in transfection of primary lens epithelial cell in serum-free culture condition.
after this actively growing primary culture was transfected with the origin defective SV40
DNA. Four immortalized lens epithelial cell lines were established. It is apparent that our
serum-free culture condition has provided a favorable condition for DNA transfection. Af-
ter transfection, 70% of the lens epithelial cells in culture assume a quiescent appearance.
These cells did not seem to increase in number significantly which was probably due to
the available growing space being occupied by calcium phosphate crystals and the trans-
fection shock. However, cell colonies began to appear between 8-12 weeks after initial
transfection.
4.2.2. Immortalized Lens Epithelial Cell Lines. Figure 2 shows the growth curves of
RLEpi-Tl, -T2 and -T3. Growth analysis shows the plating efficiencies range from 35.0 ±
6.6 to 127.0 ± 25.6 and the population doubling times were between 15.0 ± 1.1 to 20.7 ±
2.1 hours, respectively. These cells had a similar morphology to that of the primary cul-
tures of rabbit lens epithelial cells. Multinucleated giant cells were observed in these cul-
tures (data not shown) as were other morphological difference. For example, RLEpi-Tl
cells were smaller and form a sheet consisting of several layers of cells which may detach
as a sheet when it is overconfluent. Table 1 shows the estimated cell size of these lens epi-
thelial cell lines. RLEpi-T2 and RLEpi-T3 cells were similar in size (cell diameter: 33.9
174 C. Yao et al.
0
LEpTl P20
(1 ) LEpT2p6
LEpT3p8-DF
(2) LEpT3p34-YDF
a;
;;:
N (3)
,
a;
....Cl~ (4)
§
e
"- (5)
c:
--'
(6)
(7)
0 24 48 72 96 120 144
Hours in Culture
vs. 32.9 11m) and both were larger than the RLEpi-Tl cells (cell diameter is 22.4 11m). A
ruffled border, typical of lens epithelial cells, was observed in these immortalized cells in
culture and was more apparent at subconfluent stage than at confluent stage.
Table 1. Estimated cell size and cell diameter of rabbit lens epithelial cell lines
in serum-free culture conditions
(C)
Figure 3. Immunofluorescence staining of RLEpi-13 cells (A) a-crystallin, (B) T antigen, (C) Vimentin, (D) Cy-
tokeratin AE5. For a color representation of this figure, see the color insert facing p. ! 80.
fection is established. To confirm the presence of SV40 large T antigen in the nucleus of
RLEpi-T3 cells, we stained these cells with a SV40 large T antigen-specific monoclonal
antibody (AT) and then stained with a secondary FITC- or Texas Red-conjugated antibody
specific for the primary antibody. Figure 3(B) shows Positive nuclear staining for the pres-
ence of SV 40 large T antigen in RLEpi-Tl cells. Vimentin has been selected because it
has been reported as a marker for lens epithelial cells. Primary lens epithelial cells were
not stained but their immortalized counterpart, RLEpi-T3 cells, were all positive when
stained with anti-vimentin monoclonal antibody. Figure 3(C) shows the positive staining
of vim entin intermediate filament in RLEpi-T3 cells. Since corneal epithelial and lens epi-
thelial cells are from the same embryonic origin, cornea-specific cytokeratin AES interme-
diate filaments expression in lens epithelial cells was examined. Figure 3(D) shows the
positive staining of AES (against keratin AES) intermediate filaments in the cytoskeleton
of RLEpi-Tl cells. Our results demonstrated that our immortalized lens epithelial cells
continued to express keratin AES and confirmed that lens epithelial cells have the same
embryonic origin as corneal epithelial cells.
4.4. EM Analysis
RLEpi-Tl and RLEpi-T2 cells had a patchy distribution over the surface of
coverslip. In most areas, space were present between cells. Figure 4(A) shows the surface
characteristics of lens epithelial cells. Microplicaes were present across the surface of
these cells. Cells extended multiple processes onto the bare areas of coverslip. TEM ob-
176 C. Yao et al.
(A)
Figure 4. EM photomicrograph of lens epithelial cell lines (Al T3-cells have overlapped cell border with ruffle.
(8) T3-cells have short microplicae and desmosome, (Cl TI-multilayers of cells and junctional complex.
Table 2. Distribution of chromosomes per cell in rabbit lens epithelial cell lines in serum-free
culture conditions
Total Cells
RLEpi-Tl Chromosomes per cell 44 48 3
(DF-12)" Cells 2 Mean=45.3
servations show that junctional complex exist in areas of cell overlapping (Figure 4(B)).
Desmosomes were also observed between cells. In less confluent area, only a monolayer
of cells was observed. However, the culture is becoming multilayered in overgrown areas
(Figure 4(C)).
5. CONCLUSIONS
Several immortalized lens epithelial cell lines were established as an in vitro model
for cytotoxicity evaluation of drug candidates for lens indications. First, these epithelial
cell lines were shown to be free of mycoplasma contamination. Second, cytogenetic analy-
sis showed that the chromosome numbers were not significantly different from that of the
178 C. Yaoetal.
laboratory rabbit (2N=44). Third, these cells have reasonably good plating efficiency
ranges from 35.0 ± 6.6 to 127.0 ± 25.6 and a relatively fast population doubling time
ranges from 15.0 ± 1.1 to 20.7 ± 2.1 hrs. These results indicate that these cells can be
raised in large quantity for meaningful large scale in vitro experimentations. Fourth, the
morphology of these immortalized lens epithelial cells resembles that of the primary lens
epithelial cells. Fifth, immunocytochemical staining results demonstrate that these lens
epithelial cells retain key differentiation markers such as a-crystallin, vimentin and kera-
tin AE5. Moreover, the identification of large T antigen in the nucleus of these lens epi-
thelial cells confirmed that they are truly immortalized and will be constantly available for
in vitro toxicity screening.
6. ACKNOWLEDGMENT
The authors wish to express their thanks to Ms. D. Grimm for her technical assistance.
7. REFERENCES
I. Goldstein H The reported demography and causes of blindness throughout the world. Adv Ophthahnol
1980;40: 1-99.
2. Clayton RM, Cuthbert J, Duffy J. Seth J, Phillips CI, Bartholomew, RS, Reid, JMcK. Some risk factors as-
sociated with cataract in S.E. Scotland: a pilot study. Trans Ophthalmol Soc UK 1982; I 02:331-6.
3. Leske MC, Sperduto RD. The epidemiology of cataract: a review. Am J Epidemol 1983; 118: 152-165.
4. Garner MH, Spector A. Selective oxidation of cysteine and methionine in normal and senile cataractous
lenses. Proc Nat! Acad Sci USA 1980;77:1274-7.
5. Spector A, Garner WHo Hydrogen peroxide and human cataract. Exp Eye Res 1981 ;33:673-81.
6. Zigler JS, Hess HH. Lipid peroxidation products as possible initiators of cataract. Invest Ophthalmol Vis
Sci 1983;24:75.
7. Varma SO, Schocket SS, Richards RD. Implications of aldose reductase in cataracts in human diabetes. In-
vest Ophthalmol Vis Sci 1979; 18:237-41.
8. Ederer F, Hiller E, Taylor H. Senile lens changes and diabetes in two population studies. Am J Ophthalmol
1981;91:381-95.
9. Hu TS, Zhen Q, Sperduto RD. Age-related cataract in the Tibet Eye StUdy. Arch Ophthalmol
1989;107:666-9.
10. Shichi H, Gaasterland E, Jensen NM, Nebert DW. Ah Locus: Genetic differences in susceptibility to cata-
racts induced by acetaminophen. Science 1978;200:539-441.
II. Lubek BM, Avaria M, Basu PK, Wells PG. Phannacological studies on the in vivo cataractogenicity of
acetaminophen in mice and rabbits. Fundam Appl Toxicol 1988; I 0:596-606.
12. Caballero LG Auditory, Visual and neurologic toxicity of deferoxamine. Revista Medica De Chile
1989; 117:557-61.
13. Jaanus SO. Drug-related cataract. Optometry Clinics 1991;1:143-57.
14. Halevi HS, Landau J. hospitalized senile cataract in different jewish communities in Israel. Br J Ophthal-
mol 1962;46:285---90.
15. Hollows FC, Waterford J, Jones D et a1. The National Trachoma and Eye Health Programme Report. Royal
Australian College of Ophthalmologists. 1980.
16. Chatterjee A, Milton RC, Thyle S. Prevalence and aetiology of cataract in Punjab. Br J Ophthalmol
1982;66:35-42.
17. Hiller R, Sperduto RD, Ederer F. EpidemiOlogic associations with cataract in the 1971-1972 National
Health and Nutrition Examination Survey. Am J Epidemiol 1983; 118:239-49.
18. Gibson JM, Shaw DE, Rosenthal AR. Senile cataract and senile macular degeneration:an investigation into
possible risk factors. Trans Ophthalmol Soc UK 1986; I 05:463-8.
19. Yao C, Acosta D. Surfactant cytotoxicity potential evaluated with primary cultures of ocular tissues: a
method for the cultures of rabbit conjunctival epithelial cells and initial cytotoxicity studies. Toxicology
Methods 1992;2: 199-218.
Immortalized Lens Epithelial Cells as a Potential in Vitro Alternative Model 179
,~'i-': ~.
'. .: .',
'~. '
..... .
Q( . . ~\~. ...
. - .
:.
c d
Chapter 19. Figure 3. Immunofluorescence staining of RLEpi-T3 cells (A) a-crystallin. (8) T antigen, (C) Vimentin,
(0) Cytokeratin AE5.
a b
c d
D
• ,. • • , V
$ ...! " 0' \ ......, I l
~ v. . • ~ •• • , . '* ,p
~ 0 0 , i ·..'
~ . ..
~
~. .S : .. .'
.~ . .... ..,
. ~ . .. .~
..<a> _:')
.. .... -• II
- • .
Chapter 20, Figure 3. Immunofluorescent staining of primary and immortalized corneal epithelial cells. (Al Primary
cells·AE5 Keratin, (B) ReEpi-T3 dual staining for AE5 Keratin and T antigen, (e) ReEpi-T3-T antigen staining, (D)
phase contrast image of (el.
20
IMMORTALIZATION AND
CHARACTERIZATIONS OF RABBIT CORNEAL
EPITHELIAL CELL LINES AS POTENTIAL IN
VITRO ALTERNATIVE MODELS FOR
EVALUATING THE CELLULAR TOXICITY OF
OCULAR DRUG CANDIDATES
1. ABSTRACT
Several immortalized rabbit corneal epithelial cell lines were established as potential
in vitro alternative models for cellular toxicity and efficacy evaluations by transfection of
serum-free primary corneal epithelial cells with origin defective Simian virus (OrrSV 40,
clone 6--1) DNA. Indirect immunofluorescence staining results show that these corneal
cells express corneal epithelial cell-specific cytokeratin intermediate filament, AE5, but
not vimentin in their cytoskeleton confirming the epithelial origin of these cells. SV40
large T antigen is expressed in the nuclei of these corneal epithelial cells indicating that
they were truly immortalized. These cells appear to have a cobblestone-like epithelial cell
morphology in serum-free culture media and SEM analyses show that these cells have
both short and long microplicae across their cell surface. TEM analyses show that these
cells grow as a monolayer at subconfluent or confluent stages but form multi layers at
overconfluent stages. Cell junctions and desmosomes are occasionally observed. The
mean estimated cell diameter ranges from 25 to 50 11m and the average 24-hr plating effi-
ciency is 30% while the average population doubling time is 18.3 ± 2.0 hrs for these cell
lines. Mycoplasma tests show that all established immortalized corneal epithelial cell lines
are free of contamination. Cytogenetic analyses of these corneal epithelial cell lines also
reveal that these cells have chromosome number not significantly different from the chro-
mosome number of the laboratory rabbits Our results demonstrate that our immortalized
corneal epithelial cell lines do retain key corneal epithelium-specific differentiation char-
acteristics which makes them useful as in vitro models for screening xenobiotics for cellu-
lar toxicity.
2. INTRODUCTION
The corneal epithelium, the outermost layer of ocular tissue, has been one of the
three major ocular tissues evaluated in in vivo rabbit eye irritancy tests. One of the great-
est challenges that an in vitro ocular toxicologist encounters is to develop a functional in
vitro ocular cell model that provides a true assessment of ocular drug toxicity potential. In
order for an in vitro ocular cell model to best represent the in vivo response of targeted
ocular tissue, the biological relevance of the cell model to the target tissue must be estab-
lished. Hence, it is logical to use the primary cultures of rabbit corneal epithelial cells in-
stead of other cell models (e.g. MOCK cells) as a better in vitro model to evaluate the
corneal epithelium-specific responses to toxic agents. Culturing of primary cells may
sound like a good solution but it is very labor-intensive. In addition, the batch to batch dif-
ference observed between cultures imposes problems in the response uniformity of these
cells to toxic stimuli.
The methodology for culturing corneal epithelial cells has been establishedl.2. The
coexistence of corneal endothelial, epithelial cells and keratocytes in the cornea makes
culturing pure corneal epithelial cells difficult. In particular, corneal fibroblast contamina-
tion is a constant problem. The rabbit corneal cell line, SIRC, which had been used to pre-
dict eye irritation potential 3 was later shown to be fibroblastic instead of epithelial in
origin4. As a consequence, the response from these corneal fibroblasts may not reflect the
true corneal epithelial responses. The availability of a cell model is also important because
high throughput toxicity screening can be an effective line in selecting non-toxic drug
candidates and the availability of cells should not be the limiting factor. One solution to
the problem of response uniformity and the constant availability of cell models is to con-
fer the in vitro biological system with an indefinite life span. To do so, we established
methods for culturing rabbit corneal epithelial cells in a chemically defined serum-free
culture medium and then immortalized corneal epithelial cells by transfecting the origin
defective SV40 DNA containing large T antigen. Our laboratory has further characterized
these immortalized corneal epithelial cell lines as to whether the corneal epithelial-spe-
cific differentiation characteristics were retained.
and were incubated at 37 °c in a humidified incubator with 5% CO 2-95% air. After 4 hrs,
SCM was replaced with a serum-free medium (SFM)5, which is a 1:1 ratio ofDMEM me-
dium and Ham's F-12 nutrient mixtures in which 5% FBS was omitted. SFM was changed
every 48 hr until the cultures were ready for transfection (at 60% confluency with active
growth).
3.3.1. Morphological Characterization. All corneal epithelial cell cultures for SEM
(scanning) and TEM (transmission) analyses were grown on the Thermanox if coverslips
(13 mm, round) and fixed for one hour and fifteen minutes in a O.lM Cacodylate buffered
solution (pH=7.3) containing 2% paraformaldehyde and 2.5% glutaraldehyde8 • TEM sam-
ples were subsequently rinsed in a 0.1 M sodium cacodylate buffer, osmicated, stained en
bloc with a 2% uranyl acetate, dehydrated in an ascending series of alcohols and embed-
ded in a polyBed/Araldite mixture 7• TEM samples were then thin-sectioned and examined
in a Zeiss CEM 902 transmission electron microscope. SEM samples were rinsed in a
0.1 M sodium cacodylate buffer, osmicated, treated with 1% tannic acid, stained with 2%
uranyl acetate, dehydrated through a descending series of acetone and placed in ace-
tone:Peldri II (1: 1) followed by 100% peldri II. The Peldri II was solidified (room temp.)
and then sublimed under vacuum (approx. 75 kPa at room temp. for 24 hours). Specimens
were subsequently sputtered with 20 nm of gold and examined in a Zeiss DSM-940 scan-
ning electron microscope.
3.3.3. Growth Analysis-Plating Efficiency and Population Doubling TIme. Growth cha-
racteristics include plating efficiency (PE) and population doubling time (PDT) were deter-
mined by the mean ofthree 6-day growth curves of each cell line based on the number of cells
attached in the first 24 hours and the increase in cell numbers or protein contents over time,
respectively. The day the cells were plated was designated as day O. Plating efficiency was
calculated by the number of cells attached 24 hours after plating divided by the initial number
of cells plated. The cell numbers in day 0 were not included in calculating the slope of growth
curves9 for population doubling time. The following formula is used to calculate the popula-
tion time (T d): Td=ln2/J.l (J.l is defined as the growth rate); The J.l has a unit of Ilhr and is de-
rived from the equation: J.lt=ln NINo where No is the initial cell number and N is the final cell
number after a period oft (hr).
3.3.4. Cytogenetic Analysis. The procedures for cell preparations and staining of
cells were from Hsu 'O and Benn and Perle ", respectively, with minor modifications.
Briefly, cells at 80% con fluency were used. Fresh medium containing 0.06 uglml Cole-
cemid was added to arrest the cells in culture and subsequently incubated for additional 6
hours before harvest. Cells were then harvested by Trypsin/EDTA. The supernatant was
aspirated and a hypotonic solution (l :2, vlv, culture medium:water) was added to gently
suspend cells and let stand (7 minutes). Cells were then centrifuged at 200g for 3 minutes.
Fresh cold fixative (3: 1, v:v, methanol:glacial acetic acid) was added dropwise and the
cells were maintained at room temperature for 20 minutes before dropping or smearing on
a ethanol-cleaned glass slide. Smears were stained with 0.5% quinacrine dihydrochloride
according to Benn and Perle". Briefly, slides were immmersed in the quinacrine dihydro-
chloride solution for 10 minutes, then quenched for 2 minutes in MacIlvaine's buffer.
Slides were then mounted with MacIlvaine's buffer or allowed to dry and observed under
oil immersion objective (I OOOx) using a Leica Flurovert FU fluorescence microscope with
G/R fluorescent cube and photographs were taken with an attached Wild MPS46 pho-
toautomat for fluorescent images of these cells.
3.3.5. Mycoplasma Testing. Immortalized rabbit corneal epithelial cell lines were
tested for mycoplasma contamination by ATCC (American Type Culture Collection,
Rockville, Maryland) Applied Sciences. Briefly, corneal epithelial cells were cultured in
their original serum-free culture medium deprived of any antibiotics and antimycotics for
at least 2 weeks before sending the culture for mycoplasma contamination analysis. Both
indirect (Hoechst) and direct (agar) methods were used to confirm the results of each
other. The procedures '2 for both direct and indirect methods can be acquired from ATCC
Quality Control Methods for Cell Lines
4. RESULTS
neal epithelial cells at 60% con fluency were transfected with origin defective SV40 DNA,
clone 6-1, in a totally serum-free medium at a 1to 10 ratio of DNA mixture to culture me-
dium (v/v). At the time of transfection, primary corneal epithelial cells appeared healthy
and assumed an active growth appearance. However, majority of the cells assumed a qui-
escent appearance 24 hours posttransfection. Due to a decrease in cell number, feeding
medium was reduced to 50% of its original volume to allow the cells to condition the cul-
ture medium faster and cells were fed every other day. After transfection, horizontal ex-
pansion of cell colonies were obstructed by the calcium phosphate crystals in culture
wells. Although changing the medium did remove some of the crystals, horizontal cell
colony expansion was still hindered. Actively growing foci were numbered at the button
of the culture well and selectively picked for subculture in 12-well muticluster plate. After
the selected cells had developed into a monolayer in a 12-well plate, these cells were
again subcultured into a 6-well plate. The ceIl expansion continues to culture substratum
with larger growth area, from 6-well plate to 60 mm round dish, etc., for increasing the
cell number for characterizations.
(A) (8)
(C) (D)
Figure 1. Photomicrographs showing the morphology of primary and immortalized corneal epithelial cells. (A)
Primary cells at day 2. (8) Primary cells at day 5. (C) RCEpi-T3 at day 1. (D) RCEpi-T3 at day 2.
186 C. Yao etal.
hibited a squamous cuboidal morphology with darken nucleoli (Figure I(B)). These im-
mortalized cells had a similar morphology as seen in primary culture of rabbit corneal epi-
thelial cells. Figures I(C) and Figure 1(D) show the cell appearance at day 1 l;lnd day 2,
respectively. Some multinucleated giant cells exist in cultures (data not shown). Morpho-
logical differences between these immortalized corneal epithelial cell lines were also ob-
served.
0
CEpTlp35
(1) CEpT2p33
CEpT3p35
(2)
Ii
:t (3)
N
,
~ (4)
~
~
... (5)
5
(6)
(7)
0 24 48 72 96 120 144
Hours in Culture
Table 1. Estimated cell size and eell diameter of rabbit corneal epithelial cell lines in
serum-free culture media
Mean L Axis' Mean S. Axisb Cell Size' Cell Diameterd
Cell Line (~m) (~m) (~m2) (~m)
(A) (8)
(C) (0)
Figure 3. Immunofluorescent staining of primary and immortalized corneal epithel ial cells. (A) Primary cells-
AE5 Keratin. (B) RCEpi-T3 dual staining for AE5 Keratin and T antigen. (C) RCEpi-T3-T antigen staining, (D)
phase contrast image of (C). For a color representation of this figure. see the color insert facing p. 18 1.
188 C. Yaoetal.
ing these corneal cells are truly epithelial in origin, DNA transfection experiments were
performed on these corneal epithelial cells. To verify the continual expression of corneal
specific keratins in immortalized corneal epithelial cell lines, we stained these cells with
AE5 monoclonal antibodies. Positive staining was observed in these cells. Figure 3(B)
shows the AE5 staining of immortalized corneal epithelial T3 (RCEpi-T3) cells. The
vimentin was used as a negative control since vimentin is a marker for cells from meschy-
mal but not epithelial origin. Primary corneal epithelial cells and their immortalized coun-
terpart, RCEpi-Tl cells, were both negative when stained with anti-vimentin monoclonal
antibody (data not shown). SV40 large T antigen is typically expressed in the nucleus of
SV40 transfected cells after stable transfection. To confirm the presence of SV40 large T
antigen in the nucleus of RCEpi-T3 cells, we stained these cells with a SV40 large T anti-
gen-specific monoclonal antibody (AT) and then stained with a secondary FITC-conju-
gated antibody specific for the primary antibody. Figure 3(C) shows the result of
RCEpi-T3 cells stained for SV40 T antigen. Positive nuclear staining suggests the pres-
ence ofSV40 T antigen. Figure 3(D) is the phase contrast view of corneal epithelial cells
in figure 3B. Primary cultures of rabbit corneal epithelial cells were negative for AT stain-
ing (data not shown).
4.5. EM Analysis
SEM analysis of individual immortalized rabbit corneal epithelial cell line reveals a
confluent sheet of cells with a cobblestone appearance for RCEpi-T I cell line (Fig-
ure 4(A». The surface of these cells were all covered with both long and short microvilli.
TEM results show that cultures were mixed in numbers of cell layers. In some areas, a
monolayer was present with overlapped cell edges. Other areas of the culture had 2 to 3 or
more layers of cells Cell junctions and desmosomes were occasionally observed in areas
where cell borders overlapped (Figure 4(B»).
Figure 4. (A) SEM photo of RCEpi-T1 cells showing cell border and microplicaes across cell surface. (8) TEM
photo of RCEpi-TI cells reveals desmosome and overlapped cell contacts.
Rabbit Corneal Epithelial Cell Lines as Potential in Vitro Alternative Models 189
5. CONCLUSIONS
We have established and characterized several corneal epithelial cell lines as poten-
tial in vitro models for cellular toxicity screening of ocular drug candidates. First, these
corneal epithelial cell lines were shown to be free of mycoplasma contamination. Second,
cytogenetic analysis showed that the chromosome numbers were not significantly differ-
ent from the chromosome number for laboratory rabbit (2N=44). Third, these corneal epi-
thelial cells have plating efficiency ranges from 24.3 ± 3.2 to 46.3 ± 12.7 and a relatively
fast population doubling time ranges from 15.0 ± 1.1 to 25.4 ± 4.7 hrs. These results indi-
Table 2. Distribution of chromosomes per cell in rabbit corneal epithelial cell lines in
serum-free culture conditions
Total Cells
RCEpi-TI Chromosomes Per Cell 43 44 45 47 50 5
Cells Mean=45.8
cate that these corneal epithelial cells can be raised in large quantity for meaningful large
scale in vitro experimentations in a short period of time. Fourth, the morphology of these
immortalized lens epithelial cells resembles that of the primary corneal epithelial cells.
Fifth, immunocytochemical staining results demonstrate that these corneal epithelial cells
do retain corneal epithelial-specific differentiation marker, the presence of cytokeratin
AE5 and the absence of vimentin intermediate filament in their cytoskeleton. Moreover,
the presence of large T antigen in the nucleus of these corneal epithelial cells confirmed
that these cells are truly immortalized and will be constantly available for in vitro cellular
toxicity and efficacy screening. It is evident that we established functional corneal epi-
thelial cell lines that have met our objectives and that these immortalized corneal epi-
thelial cell lines will offer us a better opportunity to further select and validate assays that
address ocular irritation or other corneal epithelium-related toxicity.
6. ACKNOWLEDGMENT
The authors wish to thank Ms. D. Grimm for her technical assistance.
7. REFERENCES
I. Grant RL, Yao C, Gabaldon D., Acosta D. Evaluation of surfactant cytotoxicity potential by primary cul-
tures of ocular tissues: I. Characterization of rabbit corneal epithelial cells and initial injury and delayed
toxicity studies. Toxicology 1992;76: 153-176.
2. Chan K Y, Haschke RH. Isolation and culture of corneal cells and their interactions with dissociated
trigeminal neurons. Exp Eye Res 1982;35: 137-156.
Rabbit Corneal Epithelial Cell Lines as Potential in Vitro Alternative Models 191
3. North-Root H, Yackovich F, Demetrulias J, Gacula M Jr., Heinze JE. Evaluation of an in vitro cell toxicity
test using rabbit corneal cells to predict the eye irritation potential of surfactants. Toxicol Lett 1982; 14:
207-212.
4. Niederkorn JY, Meyer DR, Ubelaker JE, Martin JH. Ultrascultural and immunohistochemical charac-
terization ofthe SIRC corneal cell line. In Vitro Cell Dev Bioi 1990;26: 923--930.
5. Yao C, Acosta D. Surfactant cytotoxicity potential evaluated with primary cultures of ocular tissues: a
method for the cultures of rabbit conjunctival epithelial cells and initial cytotoxicity studies. Toxicology
Methods 1992;2: 199-218.
6. Gibco technical procedure #8306SA-DNA transfection procedure.
7. Cantu-Crouch D, Howe WE, McCartney M. Comparison of SEM processing methods for cultured human
lens epithel ial cells grown on flat and microcarrier bead substrates. Microsc Res & Tech.
1995;30:419-426.
8. Tsai RJ-F,Tseng SCG. Substrate modulation of cultured rabbit conjunctival epithelial cell differentiation
and morphology. Invest Ophthalmol Vis Sci 1988;29: 1565-1576.
9. Butler M. The characteristics and growth of cultured cells. In: Butler M, ed. Mammalian Cell Biotechnol-
ogy-A Practical approach. London and Washington, DC, 14-16, 1991 IRL Press.
10. Hsu Te. Karyology of Cells in Culture. In: Krause P, Patterson MK, eds. Tissue Culture: Methods and Ap-
plications, New York, 764-768, 1973, Academic Press.
11. Benn PA, Perle MA. Chromosome staining and banding techniques. In: Rooney DE, Czepulkowski BH,
cds. Human Cytogenetics-A Practical Approach Volume I Constitutional Analysis, London and Washington
D.C,96-97, 1992, IRL Press.
12. American Type Culture Collection (ATCC) Quality Control Methods for Cell Lines, 23--27, 1992.
21
Allergan, Inc.
Irvine, California 92713-9534
1. INTRODUCTION
Surfactants may interact with cell membranes of the cornea during use of ophthal-
mic products or as a result of accidental exposure. Inflammatory mediators including the
eicosanoids, prostaglandins and hydroxyeicosatetraenoic acids (HETE), are released from
the cornea in vivo as a response to chemical and mechanical injury.1-4 The production of
eicosanoids by cultured corneal cells has been used as an in vitro model for the ocular re-
sponse to acute injury caused by surfactants. 5- 7 The present study compared the in vitro
production of eicosanoids to two other in vitro measurements of cell membrane damage,
vital dye staining S and intracellular calcium ion concentrations. 9
2. METHODS
surfactant; 100 /lg/mL sodium dodecyl sulfate (SDS), an anionic surfactant; 0.1 % (w/v)
polyoxyethylene (24) hydrogenated lanolin (PEG-24), a non-ionic surfactant; 100 J.1g/mL
digitonin, a non-ionic surfactant that is a steroidal saponin. The vehicle for the test solu-
tions was phosphate-buffered saline, pH 7.4, which also served as the negative control.
Control
10 ug/mLBAK
50 ug/mL BAK . . . . .
lOOug/mLsDs • • • • • • •
0.1% PEG-24!~!!~!~~~J
o 5000 10000 15000
Figure 1. Release of radioactive metabolites of['Hlarachidonic acid into the medium by rabbit corneal epithelial
cells following 5 min treatment with surfactant.
dual excitation filter wheel, a 40x Fluor oil immersion objective, a 37° C heated stage and
a digital CCD camera (Hamamatsu C4742). Under computer control using MetaMorph
software (Universal Imaging), images were taken every 10 sec at excitation wavelengths
of 340 and 380 nm and an emission wavelength of 512 nm, the 340/380 ratio image was
calculated and the ratio values for 12 to 20 selected cells were stored. Calcium ion con-
centrations were determined using the calibration method of Grynkiewicz et aJ. II .
3. RESULTS
Radioactivity
(cpm)
BAK
(ug/mL)
." ." :r >-
:r ~
51 ~ ~ :r
'":rS2.
~
<a ~ 0:
0
""
""i';0:
-E' Q.
~ ~.
Figure 2. Thin Layer Chromatography separation of radioactive metabolites of arachidonic acid released by rab-
bit corneal epithelial cells following 5 min treatment with BAK.
controljiii=:-------l
0.1% PEG-24
o 500 1000
PGE2 (pg/mL)
Figure 3. Release of PGE, into the medium by rabbit corneal epithelial cells following 5 min treatment with sur-
factant.
In Vitro Measurements of Membrane Damage to Corneal Epithelial Cells 197
Control
lOug/mLBAK
100ug/mLBAK . . . . .II----~
0.1% PEG-24
o 50 100
12(S}-HETE (ng/mL)
Figure 4. Release of 12(S)-HETE into the medium by rabbit corneal epithelial cells following 5 min treatment
with surfactant.
of 12(S)-HETE (Fig. 4). There was no significant increase in the production of 12(S)-
HETE after treatment with PEG-24. In no case was the production of 15-HETE, LTB4 or
PGD 2 by corneal epithelial cells observed following surfactant treatment.
Figure 5. Effect of Inhibitors on eicosanoid metabolism by rabbit corneal epithelial cells following 5 min treat-
ment with I 00 ~g/mL digitonin.
some cases repeated increases and decreases in the intracellular calcium ion concentra-
tions were observed. PEG-24 did not cause elevations in the intracellular calcium ion con-
centrations (Fig. 10). Treatment with 100 I!g/mL BAK gave a result similar to that for 50
I!g/mL BAK and treatment with 10 I!g/mL BAK did not cause an elevation in the intracel-
lular calcium ion concentration (data not shown).
4. DISCUSSION
Brief in vitro treatment of rabbit corneal epithelial cells with surfactants caused in-
creased release of metabolites of arachidonic acid including material co-chromatographing
with phospholipid, prostaglandins and HETEs. Immunoassay showed that PGE z' l2(S)-
HETE were produced. A dose-dependent response was observed for BAK and other sur-
factants, such as digitonin and SDS. For the case of the non-ionic surfactant, PEG-24, the
discrepancy between the relatively larger amount of total radioactivity released and
smaller amounts of PGE 2 and 12(S)-HETE released may be explained by a greater amount
of radioactive material released which co-chromatographed with arachidonic acid.
The production ofPGE" 12(S)-HETE and other eicosanoids by the cornea occurs in
vivo following injury'-4. The in vitro production of PGE 2 and 12(S)-HETE by corneal epi-
In Vitro Measurements of Membrane Damage to Corneal Epithelial Cells 199
comro,jiiiiiiiiiii.l
50 ug/mLBAK
~
100 ug/mL BAK
Figure 6. Neutral red dye retention by rabbit corneal epithelial cells following 5 min treatment with surfactant.
SOug/mL BAK
2.5
0.5 +---+----+----+----+------~--_+_--_+_--_...--_I
o 200 400 600 800 1()()() 1200 1400 1600 1800 2()()()
seconds
Figure 7. Intracellular calcium ion concentrations in rabbit corneal epithelial cells during treatment with 50
flglmL BAK. Microscopic digital imaging was carried out with rabbit corneal epithelial cells loaded with FURA2
during treatment with the indicated test solutions. Each line on a graph represents the intracellular calcium ion
concentration for a single cell. The 340/380 ratio values correspond to the following calcium ion concentrations: I
= 125 nM, 1.5 = 363 nM, 2 = 755 nM.
200 S. S. Matsumoto et al.
0.5.L---------+----t-----+-----+------<-----+----4
o 200 400 600 800 1000 1200 1400 1600
seconds
Figure 8. Intracellular calcium ion concentrations in rabbit corneal epithelial cells during treatment with 100
IlglmL SDS.
thelial cultures simulated the in vivo response. The amount of 12(S)-HETE produced in vi-
tro exceeded that of PGE 2 by approximately 20 to 100 fold. Additional metabolites of
arachidonic acid also may have been produced in vitro. However, production of 15-HETE,
LTB4 or PGD 2 was not observed in this in vitro model.
The five min pulse treatment with surfactants used in the present study induced a
larger and/or more consistent production of eicosanoids compared to the continuous treat-
ment for 1 hr used in a previous studl. In addition, the shorter 5 min treatment was a
more accurate simulation of in vivo corneal exposure times in the presence of tear flow.
An inhibitor of phospholipase decreased the amount of total eicosanoid release in-
duced by surfactant. An inhibitor of cyclooxygenase decreased the production of PGE 2 •
An inhibitor of lipoxygenase slightly decreased the production of 12(S)-HETE. These re-
3.5.,-------------------------------,
20 uglmL Digitonin
0.5 L -_ _+-_ _+-_ _-+--_ _-+--_ _-I-_ _-I-_ _-I-_ _-+-_ _-+-_ _--l
o 200 400 600 800 1000 1200 1400 1600 1800 2000
seconds
Figure 9. Intracellular calcium ion concentrations in rabbit corneal epithelial cells during treatment with 20
~(g/mL Digitonin.
In Vitro Measurements of Membrane Damage to Corneal Epithelial Cells 201
0.1% PEG-24
2.5
o
".; 2
~
-=00
~L5
O.5+-------------------+-------------------+-------------------+-~
o 500 1000 1500
seconds
Figure 10. Intracellular calcium ion concentrations in rabbit corneal epithelial cells during treatment with 0.1%
PEG-24.
suits suggest that surfactant induced the release of arachidonic acid from phospholipids
and the subsequent metabolism of the arachidonic acid.
The surfactant effect on Neutral Red dye retention did not correlate with eicosanoid
release. This suggests that a degree of membrane disruption milder than that required to
release dye from lysosomes was sufficient to increase eicosanoid release. Surfactants
caused an increase in the intracellular calcium ion concentration in a manner that corre-
lated with eicosanoid release. This is consistent with calcium activation of phospholipase.
The following mechanism for eicosanoid release by the in vitro model is consistent
with the data. Surfactants caused membrane disruption and increased the membrane per-
meability to extracellular calcium ion. The elevated intracellular calcium ion concentra-
tion activated phospholipase and released arachidonic acid from membrane phospholipids.
The arachidonic acid was then metabolized to PGE 2 and 12(S)-HETE. Thus, the in vitro
model appears to accurately simulate some aspects of the physiological response of the
cornea to injury.
5. REFERENCES
I. Bazan HEP. Corneal injury alters eicosanoid formation in the rabbit anterior segment in vivo. Invest
Ophthalmol Vis Sci 1987;28:314-319.
2. Bazan HEP, Birkle DL. Beuerman R, Bazan NG.lnflammation-induced stimulation of the synthesis of pro-
stagIandins and Iipoxygenase reaction products in rabbit cornea. Curr Eye Res 1985;4: 175-179.
3. Oliw EH. Biosynthesis of 12(S)-hydroxyeicosatetraenoic acid by bovine corneal epithelium. Acta Physiol
Scand 1993; 147: 117-121.
4. Murphy RC, Falck JR, Lumin S, Yadagiri P, Zirrolli JA, Balazy M, Masferrer JL, Abraham NG,
Schwartzman ML. 12(R)-hydroxyeicosatrienoic acid: a vasodilator cytochrome P450-dependent arachi-
donate metabolite from bovine corneal epithelium. J Bioi Chern 1988;263: 17197-17202.
5. Klocking HP, Schlegelmilch U, Klocking R. ['H)-Arachidonic acid release as an alternative to the eye irri-
tation test. "Ocular Toxicology" (Weisse I, Hockwin 0, Green K, Tripathi RC, eds.), p 255, Plenum Press,
New York (1995).
6. Matsumoto SS, Sabatine SC, Cheng JW, Anger CB. Prostaglandin production by cultured corneal epithelial
cells as an in vitro measure of ocular toxicity. Presented at the World Congress on Alternatives and Animal
Use in the Life Sciences, Baltimore, MD, Nov. 1993.
202 S. S. Matsumoto et al.
7. Tcheng MF, Faquet B, Cottin M, Leclaire 1. Metabolism of arachidonic acid in the endothelial and epi-
thelial cells of the cornea following treatment by surfactants and other xenobiotics. Invest Ophthalmol Vis
Sci 1996;37:S78.
8. Borenfreund E, Puerner J. Toxicity determined in vitro by morphological alterations and neutral red ab-
sorption. Toxieol Lett 1985;24:119-124.
9. Yang W, Acosta D. A digitized fluorescence imaging study of intracellular Ca2+, pH and mitochondrial
function in primary cultures of rabbit corneal epithelial cells exposed to sodium dodecyl sulfate. In Vitro
Cell Dev Biol- Animal 1995;31 :499-507.
10. Jumblatt MM. Corneal epithelial and endothelial cell culture. in "Methods in Toxicology" (Tyson CA, Fra-
zier 1M, eds.), Vol. 1, p. 94, Academic Press, San Diego (1993).
II. Grynkiewicz G, Poenie M, Tsien RY. A new generation of ci+ indicators with greatly improved fluores-
cence properties. J Bioi Chern 1985;260:3440-3450.
22
I. ABSTRACT
Iris colorimetry was required for a 52-week primate toxicity study with topical ad-
ministration of eye drops with ocular hypotensive activity in order to evaluate possible
color changes in the Cynomolgus monkey. The conventional method for color determina-
tion is based on the comparison of the object with a standard color scale. To ensure GLP-
compliant documentation, photographs have to be taken and archived. Development of the
photographs over a 6-months period creates the problem that due to changes in the devel-
oper, each photograph must be developed manually and calibrated to a color scale which
must be included on the photograph along with the object. This method is expensive, com-
plicated, time consuming and unreliable. Therefore, we decided to develop a computer-
ized method (patent pending). A 3-CCD (3 chip) video camera is attached to a slit lamp.
The signals are transferred to a computer and a single frame-depicting one eye from the
front-is "frozen" by means of a "grabber card". The computer and the light source of the
slit lamp are calibrated in advance. The picture of the iris is evaluated using an image
analysis software package. The program displays the average color as numerical ROB
(red, green, blue)-values. The pictures are stored as raw data on optical disc. This method
provides a fast and accurate means of measuring color changes in the iris of experimental
animals. It can also be applied as standardized method in humans.
2. INTRODUCTION
dark-brown irises. The first changes of iris pigmentation occurred after three months of
treatment. Occasionally, photographs were taken to document the color changes4 • The
long-term effect and mechanism of this adverse reaction is not known 2 • Unpublished stud-
ies indicate that stimulation of melanin production rather than proliferation of melano-
cytes may be the cause for this unusual finding. In monkeys with yellowish-greenish eyes,
a similar change of pigmentation was noted: a dose-dependent increase was observed after
3 to 12 months of topical administration. In a recent 52-week primate toxicity study with
topical administration of eye drops against glaucoma, iris colorimetry was required in or-
der to evaluate possible color changes in the Cynomolgus monkeys.
3. METHOD
The conventional method for color determination was based on the comparison of
the object with a standard color scale. To ensure GLP-compliant documentation, photo-
graphs had to be taken and archived. Development of the films and paper prints over a 6-
months or longer period creates the problem that due to possible changes in the
developing chemicals, variation of temperature, exposure times and other dark chamber
variants, each photograph must be developed manually and calibrated to a color scale
which must be included in the single photograph along with the object. This method is ex-
pensive, complicated, time consuming and unreliable. Further, there are individual differ-
ences in color perception and it is almost impossible to assure a standardized evaluation
over a long period; evaluation of approximate iris color by comparison against color scale
cannot be more than a rough visual integration of the evaluation area.
Therefore, we decided to develop a computer-assisted method (patent pending). A 3-
CCD ("3 chip") video camera was attached to a slit lamp. The signals are transferred to a
computer and a single frame---depicting one eye from the front-is "frozen" by means of
a "grabber card". The camera, computer and light source of the slit lamp are calibrated in
advance. The experimental set-up is shown in Fig. 1.
4. RESULTS
The picture of the iris is evaluated using an image analysis software package, The
area for evaluation is fixed: nasal, as high as the pupil, not touching the inner or outer bor-
Figure 2. Color analysis of the iris; values are given as Red-Green-Blue-components ranging from 0 to 255 in
each color.
der of the iris. When this area is marked as a rectangle with the cursor, the program dis-
plays the average color and the numerical RGB (red, green, blue)-values of the averaged
area (Fig. 2). Repeated recording and measurements in the same eye of one animal led to a
coefficient of variation of less than 5%. The human eye can barely note this color differ-
ence. Longitudinal measurements in 20 Cynomolgus monkeys over 6 months did not re-
veal changes in the color of untreated irides.
5. CONCLUSION
This computer-assisted method, developed for evaluation of iris color changes, pro-
vides a fast and accurate means of measuring color changes in the iris of experimental ani-
mals. The pictures of the eyes can be stored permanently as raw data on optical disc or
other media but can also be shipped as compressed data files on floppy discs. The method
can also be applied as standardized method in human clinical trials.
6. REFERENCES
I. AIm A, Stjemschantz J, The Scandinavian Latanoprost Study Group. Effects on intraocular pressure and
side effects of 0.005% latanoprost applied once daily, evening or morning: A comparison with timolol.
Ophthalmology 1995; I 02: 1743-1752.
2. Stewart We. Prostaglandin analogues: Their usage and benefits in glaucoma treatment. Ophthalmic Prac-
tice 1996;14:48-55.
3. Camras C B, The United States Latanoprost Study Group. Comparison of latanoprost and timolol in pa-
tients with ocular hypertension and glaucoma: A six-month, masked, multicenter trial in the United States.
Ophthalmology 1996;103:138-47.
4. Watson P, Stjemschantz J, The Scandinavian Latanoprost Study Group. A six-month, randomized, double-
masked study comparing latanoprost with timolol in open-angle glaucoma and ocular hypertension. Oph-
thalmology 1996; 103: 126-37.
23
IDepartment of Ophthalmology
Emory University School of Medicine
Atlanta, Georgia
2Departments of Ophthalmology, and Physiology and Endocrinology
Medical College of Georgia
Augusta, Georgia
1. INTRODUCTION
ing is performed. Although some exceptions have been reported, where skin but not ocular
tissue irritation occurs, this is not a reliable approach. The next level of testing would be
in vivo rabbit eye irritation.
Some in vitro test may be found that will replace the animal model. Ideally, we would
like to go directly from in vitro testing into clinical trials although this is unlikely. Presently,
the initial toxicity screening still uses the Draize rabbit model as an in vivo test. There is no
singular in vitro test that can substitute for the Draize test and in vitro tests are used as com-
plementary ones to enhance available drug toxicity information. There is very little guidance
for defining the mechanism of ocular irritation that may be caused by any product.
To assess the toxicity of a product to the eye, the application protocol must be de-
fined. There must be a very clearly defined protocol, ocular grading system and interpreta-
tion. The amount of the test solution added to the ocular surface was originally suggested
to be 0.1 ml (l00 Ill); this is in excess of tear volume in a rabbit or human which is about
7 to 10 Ill. If 100 III is placed on the eye and a blink occurs, then a portion will fall off
onto the cheek or is squeezed down the punctum. The amount of test substance being
tested becomes uncontrollable and remains an unknown quantity.
Another important factor affecting test substance distribution is the blink rate. The
blink rate is less than once per minute in rabbit. In humans the blink rate is about 15 times
a minute. Thus, the distribution caused by blinking is very different in the rabbit eye than
in the human eye. If the test substance causes irritation, then the volume of tears produced
dilutes the test substance and washes drug away from the ocular surface more quickly than
in the absence of irritation. The test substance availability, as altered by these factors, will
affect the potential ocular irritation scores of the cornea and conjunctiva. Similarly, the
human conjunctival area is two-fold larger than that of the rabbit, thereby allowing a much
greater area for drug absorption and dilution of the applied dose. Thus, the cornea:con-
junctiva ratio is quite different in rabbits and humans. The number of animals used per
evaluation of a substance or chemical is still evolving down to I to 3 from what was origi-
nally 9 per test substance.
The original Draize test required 100 III with or without rinse but with ophthalmic
products this will be modified because rinsing is not a prerequisite as this would defeat the
purpose of drug application. Rinsing with ophthalmics is redundant since one applies the
topical drug to treat a disease or disorder, thus the object is to have as much of the active
ingredient on the ocular surface for as long as possible. Originally the examination times
were 24 hours, 48 hours, 72 hours and 7 days; now I hr is used frequently to see what oc-
curs between the time of drug application and 24 hours.
Interpretation of the Draize scores can be in terms of: (l) a positive or non-positive
response, or (2) a derived score ranging between 0 and 110 (Table I). The cornea is con-
sidered the most important ocular tissue and accordingly 80 of the 110 points assigned to
the scoring system is derived from corneal effects. The cornea score is derived by adding
the degree of edema (score A) and the surface area of involvement (score B) then multi-
plying the sum by 5 to give a total possible score of 80. Iris has a total score of 10. The
conjunctival score is derived by adding the redness score plus chemosis score plus dis-
charge score and the sum multiplied by two for a total score of 20. In the rabbit model ex-
posed to ophthalmic products that have already passed initial toxicity evaluation, there is
usually no corneal response. Typically, a Draize score of greater than 0 is caused by con-
junctival redness and occasionally discharge. Yet, the conjunctival and iris scoring is only
27% of the total Draize score (30 of 110 total points).
Kay and Calandra re-evaluated the Draize scoring system to provide a descriptive
interpretation of the degree of irritation, persistency and overall consistency of the data.
210 H. F. Edelhauser and K. Green
They grouped numbers into eight small steps and if the score fell in a category where you
had an immediate (one hour) total score between 0 and 0.5 out of 110 (total pooled Draize
numbers: 80 + 20 + 10), and 24 hours after applying the drug it was 0, the test substance
was designated as non-irritating (Table 2). If you had a score of 15 to 25 out of the 11 0
and all the scores at 96 hours returned to 0, then this was classified as mildly irritating. If
it did not return to 0, then the next level, moderately irritating, was invoked. The interpre-
tations become more complex with higher Draize scores, yet most ophthalmic prepara-
tions rarely cause large Draize scores.
There is a need for a more sensitive evaluation or test of ophthalmic products be-
cause test compounds generally are categorized as non-irritating to minimally irritating.
To grade a Draize test room light and visual examination of the animal eye is all that is
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 211
needed, Hopefully, the investigators will be consistent in their evaluation and interpreta-
tion of the ocular irritation. The McDonald-Shadduck test includes evaluations of the cor-
nea and aqueous in addition to the standard Draize test. Aqueous flare, corneal pannus and
a fluorescein evaluation (the amount of stain in the cornea) are performed with the assis-
tance of a slit-lamp. Extra descriptions are provided for grading the cornea and conjunc-
tiva. The slit-lamp provides a higher magnification and better resolution of the tissue
pathology.
The rabbit model often will have some baseline corneal staining that is observed as
fine streaks or multiple streaks coalesced into patches. This will confound data collection
in small groups of test animals. Furthermore, a more elaborate grading system is necessary
to record the multiple stain patterns that are possible. Merely using the grades of intensity
of stain uptake and percent distribution is inadequate. The investigator must clearly define
the percent distribution. Is it the percentage of corneal surface stained, or the location of
the stain on the surface that is most important? Would a distribution be better defined as
location, i.e., central or peripheral cornea? Holden has extended the stain grading for hu-
mans into an elaborate scheme. He evaluates contact lens wear in patients by dividing the
cornea into five zones in order to get more insight into the origin of the epithelial staining.
Each of the five zones receives all gradings, which creates a great deal of data from each
eye.
There has been data published which states that the human response is over-esti-
mated using the rabbit model. That is, the rabbit response is more severe than in human.
Thus, the rabbit model is hypersensitive relative to the responses of human ocular tissues.
The Draize test has many deficiencies in quantifying the ocular toxicity of a test
substance. One of the difficulties is the subjective interpretation and scoring of the inflam-
mation and epithelial fluorescein staining. Utilizing a singular investigator will improve
the test consistency. Also, the test gives an incomplete picture because of the bias of 72%
of the score (80 of 110) being generated by the cornea. In my experience, most ophthalmic
products will not cause loss of corneal transparency as scored in the Draize test. There
may be conjunctival redness without chemosis which may be affected by the method of
application of the test substance.
There are several questions to be asked. If the test substance is applied to the inferior
cul-de-sac then does it have good distribution to the cornea and superior cul-de-sac? Will
the inflammatory response be altered if the test substance is applied directly to the cornea
and not in the cul-de-sac? Does the test substance cause pain without inflammation? A
pain response is not included in the Draize test, nor is there a repeatable quantitative
212 H. F. Edelhauser and K. Green
measure of this parameter. I record the rabbit's response after test substance application,
such as if the animal holds its lids partially or completely closed or is pawing at its face.
The Draize test records a mixture of toxicity, irritation and inflammation. Toxicity
refers to the agent that is being tested, irritation is a description of what happened at the
tissue, and inflammation is how it happened. Corneal irritation is a histopathological event
caused by a cellular mechanism. Clinically we may be able to look at the epithelial inju-
ries as noted by the fluorescein staining and even corneal opacity. Histologically these
epithelial injuries could be shown as cell damage or breakdown of membranes. Corneal
opacity, i.e., stromal edema, will be associated with epithelial or endothelial cell damage
and keratocyte necrosis. The mechanisms are referred to as cytotoxicity. The conjunctiva
can be considered in the same manner, with clinical events occurring that will be ex-
pressed as conjunctival redness, i.e., vascular engorgement, and chemosis, i.e., tissue
edema. These two events are controlled by vasodilation and the inflammatory process.
Inflammation is often a blood-borne event that is transported by the circulatory sys-
tem. The cornea is devoid of blood vessels which may explain why the cornea typically
will not demonstrate irritation although the conjunctiva presents with inflammation and
hyperemia. The conjunctiva is very rich in blood vessels and even the manipulation of the
lids can cause an increase in flushing of these blood vessels.
The redness and chemosis are regulated by neurogenic and non-neurogenic factors.
The neurogenic effects will be to direct regulation of blood vessel dilation and pain. The
non-neurogenic effects will be from blood-borne chemicals such as vasoactive substances,
histamines, leukotrienes, serum proteases, prostaglandins and cytokines. Arachidonic acid
breaks down into prostaglandins, 12-R-HETE and leukotrienes that affect vessel dilation
which causes conjunctival redness. The cytokines are produced in the epithelial cells, stro-
mal cells, etc. The chemical transmitters regulate leukotrienes and interferon. The system
gets more and more complex as one recognizes the underlying biochemical events. The in-
flammatory cells involved in this system are platelets, polymorphonuclear leukocytes,
macrophages and lymphocytes. White blood cells, that are either granular or agranular
leukocytes will be the site of further chemical cascade into other chemical transmitters.
Granular leukocytes become neutrophils which clean up debris and will add to the dis-
charge seen on the eye. When an inflammatory ocular response occurs, many factors are
involved and many chemical compounds are subsequently being produced. So if you are
undertaking a test that evaluates one of these factors, you may get deceptive information.
The list of possible events resulting from an initial insult is overwhelming. The tis-
sue takes up a foreign chemical, and a whole series of events subsequently follows (Fig-
ure I). When a Draize test is performed, however, one only looks at the end products of
the responses, namely corneal opacity, fluorescein staining and neovascularization after
this underlying and unseen cascade of events has taken place. The iris is examined for red-
ness or vasodilation. The conjunctiva may show redness, chemosis and discharge. What is
assumed to be a simple test turns out to be very complex and the events underlying the fi-
nal outcome are long and arduous.
Edelhauser. Thank you, Bernie. I think that we have a very complex issue as we
break down the individual processes underlying Draize testing.
Professor Dr. Otto Hockwin, St. Augustin, Germany. Why do you still recommend
the use of albino animals? It is well known that albino animals have sick eyes. Does it
mean that when you test solutions and contact lenses on the cornea of an albino rabbit that
you can recommend the use of these solutions and lenses only in albino human eyes?
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 213
xenobiotic agent
+
locular tissue uptake I
~
ocular cellular chemical release
(cytokines, prostaglandins, leukotrienes)
~
CQRNEA
opacity
lBlS
!
redness
-----
CQNJUNCTIVA
redness
fluorescein staining chemosis chemosis
neovascularization discharge
McCarey. The albino rabbit probably isn't the ideal animal model, but it is a practi-
cal animal with respect to handling and literature acceptability. Giving topical eye drops
to, or applying contact lenses on, a cat is a challenge that I do not volunteer to do.
Fitting a rabbit with contact lenses has its problems but is manageable. If you go
through the literature, however, you will notice people are making conclusions about lens
fit on rabbit eyes after using various procedures such as tarsorrhaphy (sewing the lids
closed). It is totally unphysiological and means nothing unless they are out to cause an in-
flamed eye. In the best of circumstances, contact lens fitting and post-fit inflammation in
the rabbit model may not mimic the response in the patient.
Hockwin. I am not so much against rabbit as a model for contact lens fitting research
or tests; I am against using albino animals in ophthalmic research or any toxicological
studies.
Hockwin. During the meeting we have heard several times about observations on the
iris; this is a major problem in albino animals where no pigment is present. In my opinion,
eye research should not be performed any more with "sick" albino eyes.
Keith Green, Ph.D., D.Sc., Medical College of Georgia, Augusta, Georgia. I just wanted
to point out that Holden has been very successful at fitting cats with contact lenses with a high
degree of retention. I have seen these animals and they appear quite content even with dif-
ferent contact lens materials being used.
On another topic that you mentioned concerning ocular tissue responses to stimuli;
these can also vary depending upon the location of the stimulus. All tests do not involve
solutions or contact lenses. For example, we have examined tissue adhesives derived from
mussel glue (this is the adhesive that keeps the mussel attached to rocks in the tidal zone).
The underlying concept for the proposed use of these compounds was that since the mate-
rials worked well as adhesives for tissue cultured cells, by allowing faster cell attachment
to different substrates, then the reasoning was that they should work well in vivo to close
214 H. F. Edelhauser and K. Green
penetrating wounds of the globe, to help corneal epithelial healing in diabetics, to assist in
re-epithelialization in corneal ulcers, and many other uses as a glue or sealant on the eye.
While these compounds promoted corneal endothelial and other ocular cell attachment
in vitro without the overt induction of any tissue toxicity, the in vivo toxic effects were site de-
pendent. The compounds, that varied depending primarily upon the presence or absence of
different amino acids or the number of L-DOPA residues in the chemical backbone, showed
different responses when placed sub-epithelially or intracamerally. The response was very
muted when the compounds were injected sub-epithelially relative to placement of many
compounds in the anterior chamber. Even the most quiescent of the glues in the corneal site
caused a small inflammatory reaction of short duration when placed in the anterior chamber.
The glues polymerize very rapidly (i.e., in seconds) and in the cornea presumably
this process is so fast that no toxic substances are released that can reach the anterior
chamber. Release of toxic substances must occur, however, when many of the compounds
are placed in the anterior chamber. The result is that one has an exceptionally good adhe-
sive with little or no toxicity to many of the derivatives if confined to the cornea or exter-
nal to the intraocular fluids. If these chemicals accidentally reach the anterior chamber or
intraocular space then toxic inflammatory reactions of varying degrees may occur depend-
ing upon the specific compound under use. Although further testing is required in order to
delineate the structural requirements for minimizing toxicity, it is already evident that
these adhesives show potential for use as an extraocular and intraocular adhesive.
This is an example of different tissue responses using the revised Draize test where
the results depend upon the specific location of the test agent in the eye and differs from
the "usual" Draize test to assess solution irritation potential. This Draize test used more
frequent observation periods, slit-lamp observation only, and longer times to allow recov-
ery. This was published in early 1996 in a Congress Proceedings on Tissue Glues and Ad-
hesives (Surgical Sealants and Adhesives, Eds. Sierra DH, Saltz R, Technomic Publishing
Company, Lancaster PA).
David E. Potter, Ph.D., Morehouse School ofMedicine, Atlanta, Georgia. For the most
part you've been talking about organic chemicals but what about recombinant DNA prod-
ucts and/or biotechnologically-derived agents? Would a system such as the one you de-
scribed really work or would you have to go to another model?
Edelhauser. David, the Draize test still has to be used in terms of evaluating inflam-
mation. The other possibility is to use intravitreal or AC injections to evaluate recombi-
nant DNA products. I think that's a good question but I really don't know if we have an
answer at this time.
Janice B. Allen, Ph.D., North Carolina State University, Raleigh, North Carolina.
There have been recent reports in IOVS and other eye journals concerning gene therapy
where they have either applied material topically or injected intravitreally in rabbit, mice
and rat eyes. This may also address questions about where liposomal complexes have been
used. Of course, the end-product or the termination of the experiment is going to be differ-
ent and the design of the experiment is different, but there are studies where they've used
these kind of reagents in the eye.
Allen. Most studies used a reporter gene such as beta-galactosidase to see if a gene
can be inserted into certain parts of the eye and they have been successful. Borise did an
experiment with rabbits and I believe did an intravitreal injection and got some inflamma-
tion. A Japanese group administered these liposomal DNA complexes topically onto the
eye. Actually they compared four routes of injection and maybe three or four different
liposomal preparations and didn't report any inflammation. But they were able to actually
transduce those cells to express beta-galactosidase.
Edelhauser. The article you're referring to is Masuda et aI., Gene transfer with
liposomes to the intraocular tissues by different routes of administration. Investigative
Ophthalmology and Visual Science 1996;37: 1914-1920. The best route was an intravitreal
injection.
Allen. Right, the best results, but they did a topical administration as well.
McCarey. No, and we're probably living with history and convenience rather than
reality.
Ramesh Tripathi. Well, I can understand the convenience and cheapness but there
are other animal models such as miniature pigs that are more akin to human in their eye
function. Why keep on using rabbit and especially albinos?
Green. Because, for the Draize test, it's dictated by the law. Regulatory agencies ask
for albino rabbits. This is a requirement of the Consumer Product Safety Regulations and
of different Federal groups that require submission of toxicity testing data
Green. Yes, it's written in the laws and regulations. Other aspects of the test protocol
appear to have more latitude if there is a scientific basis and justification for their adop-
tion, e.g., the amount of test product added to the eye, observation times (as long as the
minimum is performed), etc., but the preferred species seems to be the albino rabbit. Pre-
sumably this is because of precedent.
Javier Ava/os, Ph.D., Food and Drug Administration, Rockville, Mary/and. It is ref-
erenced in the Regulations. However, the reviewer and supervisor as well as the Agency
would be willing to review a method which uses another test species if a good scientific
argument is provided. Most of the historical data for irritants, both dermal and ocular, has
216 H. F. Edelhauser and K. Green
been obtained from albino rabbits. Certainly, this species is the preferable one, but the
FDA will consider other species if scientifically warranted. Certainly, the albino rabbit is
the preferable use but FDA will take any model.
Edelhauser. The classic Draize test has examination at 1 hour, 24 hours and other
specified times. But many effects, when dealing with a contact lens wetting solution or
maybe a drug preparation that has a preservative in it, may be evoked at 2, 3, or 4 hours.
One of the things that needs to be addressed is slit-lamp evaluation as well as the time
frame in which a compound or drug is evaluated.
McCarey. There are drugs that penetrate deep into the eye before a tissue response can
be noticed. For example, dilation of the iris. So if events occur transiently following drug ap-
plication, it is possible that evaluations at 1 hour and 24 hours may completely miss the tissue
response. The investigator must be aware of the possible tissue reactions and design the test
relative to the drug being tested rather than being inflexible to a set of rules.
McCarey. You can apply 100 III but it's not going to stay there.
McCarey. I personally use 30 III test volume. I try to get an amount that is small
enough that it will stay on the eye surface but large enough to apply a reasonable amount
of test substance. Ten III will stay but that's a very small application of the test chemical.
Whereas 30 III will increase the volume of fluid on the lower lid margin but will not in-
duce excess blinking of the eye.
Green. Years ago we did studies with the Xerox Company and we obtained informa-
tion on solid toners which was great because you could actually collect the material left
behind. We put in 50 mg and after the first blink we washed the eye surface, collected,
dried and weighed the powder and measured 5 mg remaining in the conjunctival sac. Be-
cause these substances are black, complete recovery from all parts of the eye surface could
be guaranteed. If you put 100 mg of a solid test substance on the eye you're only going to
finish up with 5 or 6 mg, certainly less than 10 mg, remaining on the ocular surface and I
think that's a very important consideration. The test, therefore, calls for much more mate-
rial to be applied than can possibly remain on the surface of the eye tissues. We published
that in 1985 (J Tox Cut Ocular Tox 1985;4: 13-26).
From a pharmacokinetic standpoint, the addition of 100 III of a liquid test solution
on the eye increases the volume from 7 to 107 III which, theoretically, is reduced to 7 III at
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 217
the first or second blink. This represents the loss of over 93% of the originally added drug.
Using these high volumes, and even assuming no loss of fluid from the lids to the cheek,
only a small fraction of the applied material is under test. Reducing the applied volume to
25 III allows testing of 78% of the added drug mass - a much better test system. Joe Ro-
binson and his co-workers have performed an excellent series of definitive studies on ap-
plied volume/drug delivery relationships and the eye.
FT. Fraunfelder, M.D., Casey Eye Institute, Portland, Oregon. A month ago in Lon-
don I was with someone who I respect as probably one of the finest ocular toxicologists in
the world. He made the comment that in his 30 years of ocular toxicology he felt that
he had done a great deal of disservice to animals because it had very limited predictive
value to the human. Now, that comment was made for oral medications. Would you
comment as far as topical ocular medications? Contact lenses are a different subject
but what is your opinion of the validity of ocular toxicology in the animal as it relates
to the human?
McCarey. I am not able to respond as to how the data that we collect in the rabbit
model relates to clinical experiences with the test substance. This may sound naive on my
part but as an academic person hired by a company to do a certain test, we generally do
not even know what we are testing. What I present to you are experiences that I have had
in the realities of dealing with the animal model and trying to evaluate it rather than the
added experience of how the test correlated to the human ocular response. I would like to
ask Keith Green to share his many more years of experience in this area.
Green. There are papers by Griffith and co-workers (J Tox Cut Ocular Tox 1984;1 :5~;
J Tox Cut Ocular Tox 1986;5:115-123; Fundam Appl ToxicoI1986;7:626-{i34) that relate anec-
dotal findings in man, particularly reports back to companies of events that occurred when
the popUlation at large was exposed to a variety of materials such as contact lens solu-
tions, soap powders, cosmetics, etc. compared with reports in rabbit. Comparisons have
been made between these limited number of reports available from human and those in the
literature with the same chemicals in Draize rabbit toxicity studies. Generally the rab-
bit responds much more vigorously than human by about IO-fold to the same stimulus.
That was the reason for the recommendation by the National Academy of Sciences in
1977 to go from 100 III (liquid) or 100 Ilg (solid) down to 10 ml or 10 mg as the
amount delivered to the rabbit eye in a testing paradigm (see Ophthalmic Toxicology,
Ed. Chiou GCY, Raven Press, NY. Pp 1-16). Much of the commercial information is
unfortunately not available for widespread comparisons since it lies in the proprietary
archives of companies.
Fraunfelder. What I suggest for topical ocular medications is that this type of testing
is more for irritancy than it is for finding side effects of drugs. The human is probably the
only species where you are going to find true ocular side effects and drug-induced effects.
Pierre Duprat, D. V.M., Merck Laboratories, Riom, France. I still have trouble using
the Draize test for ophthalmic preparations because we know the Draize score will be very
low. It is probably in the lowest part of the scale where we see inappropriate data. Usually
a drug that you put onto the human eye has a total score of 15 or below (out of 110). We
are not using the Draize test properly which was initially designed for human safety,
something like safety of chemicals splashed onto the eye. It was never designed to test
ophthalmic drugs.
McCarey. I agree completely. I tried to make the point that the original test was de-
signed for substances that are only accidentally applied to the eye. The Draize test does
not have a fine enough sensitivity or resolution to evaluate the potentially toxic effects of
an ophthalmic product. Recently, I have been using non-invasive fluorophotometry to
measure corneal epithelial permeability and other tissue functions in order to increase the
sensitivity of ocular toxicity testing. You get a much more sensitive read of what is going
on in the tissue from these tests as compared to the very crude scoring system of the
Draize test.
Duprat. You can push the test a little bit further by using a more appropriate volume
which mimics the human delivered volume, 30 /-ll, and use the scale (i.e., the Draize scor-
ing system) especially for conjunctiva, during repeated administration. If you go for
longer than a week dosing, one month, three months or six months, the cornea will not be
affected if you have a new drug but the Draize scoring scale applies to conjunctival pa-
rameters.
McCarey. When you say a longer duration do you mean beyond what would be nor-
mally requested for the use of that drug? For instance, ifit says apply it once a day, you're
going to apply it eight times a day?
Duprat. No, the application regimen could be for several days but it doesn't mean
you're looking for a magic number. We have to use common sense; something that we are
going to use once a day, for instance, such as an anesthetic or something like that, you can
accept a little bit of irritation or "some irritant effect" on the eye. But we have to keep in
mind that, depending on the duration of the drug being applied onto the eye, we have to
specifically design a drug administration regimen in animals. We have to consider the goal
we are seeking, we cannot have a single way to assess any particular situation; we have to
sit down and work on a case-by-case basis. Otherwise, we don't have any uniform system
that is working worldwide.
McCarey. Keith Green mentioned that the rabbit is up to ten times more sensitive.
As far as I'm concerned as an eventual end-product user, I like the idea of a test model
that is more responsive to a product being tested than is the human. I'd rather see the ani-
mal be far more sensitive than equally sensitive than the human response. Maybe the rab-
bit is a safety margin indicator.
Green. I agree with you, Bernie. We published information a long time ago on the
comparison between the Draize and corneal epithelial fluorescein permeability (J Tox Cut
Ocular Tox 1985;4: 13-26). The fluorescein epithelial permeability measurement using the
fluorophotometer is far more sensitive. I've gone to using that as one of my prime indica-
tors over the last fifteen years; we actually used the test, and validated it over several
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 219
years, before publishing the technique and results. As you said, you can measure corneal
epithelial permeability, wait a few hours and then measure corneal endothelial permeabil-
ity and aqueous humor flow rate. All of these measurements provide for a much more sen-
sitive model in which changes can be measured at far lower test substance concentrations.
One can, therefore, avoid any massive overt changes that may occur with large amounts of
chemical placed on the eye. Use of this methodology helps to avoid problems associated
with the induction of severe responses which may complicate the data interpretation. The
method also allows the measurement of far more subtle changes that can be induced by
prospective ophthalmic compounds, especially when overt responses may be negligible.
Nevertheless, as Dr. Fraunfelder said, these tests are for irritation and short-term
safety evaluation. Often, side effects are not seen until after long-term use of a product,
then we go back to an animal model in an attempt to elucidate the mechanism of action.
The shortcomings of the Draize test in evaluating ophthalmic products, and espe-
cially the multiple attempts to create alternative models, are becoming more evident with
the passage of time. The complexity of the interactions of chemicals with different ocular
tissues and the generation of multiple biochemical, pharmacological and immunological
responses from each eye tissue has become more obvious as we uncover more of these re-
actions. Data is available that reveals a large differential response of identical tissues in
different species, this is especially true in terms of the release of various arachidonic acid
metabolites either inherently or in response to the same stimulus. Rabbit is one of those
species that produces vast amounts of arachidonic acid metabolites from iris and conjunc-
tiva when stimulated. This was initially inferred from the eye responses (intraocular pres-
sure, uveitis, conjunctival hyperemia [Exp Eye Res 1984;39:807-829]) but was later
quantitated by direct analysis of the inflammatory products from individual tissues (Curr
Eye Res 1984;3:447-452; J Pharm Exp Therap 1988;747:1064-1072).
The development of alternatives is impeded by the very presence of the growing
scope of our knowledge concerning these tissue/chemical interactions. It is blatantly obvi-
ous that no one test can simulate the wide variety of responses seen in the eye. What origi-
nated as relatively simple observations of tissue reactions is being revealed as an
extremely complex series of events.
In addition, as pointed out by others in this workshop, the Draize test was originally
invoked to test the behavior of the eye towards soaps, cosmetics, agents found around the
home and any other extraneous chemical that might reach the eye. The adaptation to test-
ing ophthalmic products represents a leap that was not in the conceptual processes under-
lying the origination of the test and is a misplaced use of the test.
Another aspect of the Draize test concerns the use of a slit-lamp to provide a de-
tailed, magnified examination of the ocular responses. The slit-lamp offers a great instru-
ment with which to look at the eye, and I have used it almost without exception over a
considerable number of years. Without the use of this apparatus one relies upon what
could be relatively poor illumination and ones' eyesight. Slit-lamp examination is a must,
especially for potential ophthalmics. With the tissue glues, mentioned earlier, the slit-lamp
allowed one to see a thin, wrinkled, veil-like area lying sub-epithelially following injec-
tion of some of the compounds at that site, that was not visible with the naked eye.
Edelhauser. I think we all recognize the faults of the Draize test. At the same time
we have regulatory guidelines that we have to follow. We also have the European concern
that eventually programs are going to move forward to in vitro-type of tissue culture as-
says. I hope we can address some of these issues as to what would be acceptable. We are
going to review a number of these issues such as re-epithelialization in a tissue culture
220 H. F. Edelhauser and K. Green
model and organ culture. Should we use inbred strains of animals for some tests? Do we
use dogs? Obviously we're not going to come up with complete answers. It will be left to
the individual drug houses to develop their own protocols but hopefully some of the
guidelines obtained from this discussion will narrow the scope of the specific tests that
have to be run.
I was at a meeting at International Life Sciences Institute in which Procter & Gam-
ble, Amway, Gillette and other companies were represented. The companies were con-
cerned that they are spending a lot of money on in vitro testing, but where is the testing
going and how can the data be used?
It was unanimously felt by the committee that the Draize rabbit eye model test
should not be considered the gold standard. The Draize test has a number of serious flaws
in it's ability to predict the human response. It has historical precedence and that's one of
the things that came out this morning. There is no scientific evidence to support use of the
Draize test. Major anatomical differences exist between rabbit and man, including the nic-
titating membrane in rabbit; thus, if one is evaluating a powder the nictitating membrane
may trap some of the powder so a greater exposure can occur.
Tear formation, the blink rate, and no method to measure pain or ocular discomfort
are other problem areas. One of the points that this workshop emphasized was the failure
to use the biomicroscope in ocular examination. It is a wonderful tool and by failure to use
the biomicroscope there are many severe limitations to the evaluation process.
Ocular irritation should be based on, 1) Human experience. One of the things that
was obvious is that the companies were, in many cases, unaware of the ophthalmic litera-
ture. There are companies developing a bovine corneal irritation technique with minimal
reference to published data in ophthalmology. 2) In vitro testing. Where is in vitro testing
going to be important? Hopefully, one of the points that we might be able to come to some
agreement on is, are there in vitro tests that are important? For example, what about the
human corneal epithelial cell line in tissue culture? Animal testing should be used to either
help corroborate the in vitro test and to fill in unknown responses in a case where a new
family of chemicals has never been tested.
One of the issues upon which there was agreement, was a need to evaluate corrosive
and severe irritation compounds in tissue culture. Obviously, this is not a major concern of
the audience consideration today, but has to be considered in an overall scheme. Does a
chemical cause loss of greater than 70% of the limbal area? Can it be quantitated with
staining? If there's a loss in visual acuity, does cataract form? Is there a persistent stromal
ulcer and is there uveitis? If these categories are present, this test substance would fall into
some type of corrosive classification. What would classify something that severe? Again,
uveitis. Well, how do you document uveitis other than with a slit-lamp? What about all of
the cytokines· and other inflammatory mediators that are being released? Is there stromal
damage and can it recover? Do we have an in vitro model to evaluate recoverability? One
can measure stromal edema with the slit-lamp by doing optical or ultrasonic pachymetry
to determine if recoverability occurs.
The goal of this whole meeting was to provide a direction for future national and in-
ternational research efforts to develop and replace the test for assessing acute chemical or
induced eye injury. The companies are looking for a white paper to help them evaluate
where to invest money in in vitro testing. There has to be a payoff and utility at the end.
Because there's a lot of money invested in alternative ocular toxicity testing, how can aca-
demia and industry work together to bring this to fruition? What are the future directions?
There should be standardization of ocular slit-lamp exams. If you're evaluating in-
flammatory reactions, it may be important to use Jackson Lab mice strains, for example,
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 221
which are predictable. If you're evaluating dry eye, such models exist in dogs. Impression
cytology; is this useful when considering limbal area damage? When one prepares to
evaluate inflammation and uveitis, what about studies for mediator release? IL I,lL 6,
TNFa. Where do adhesion molecules fit in terms of re-epithelialization if recovery is be-
ing evaluated? Can we then take this one step further, can you correlate corneal or con-
junctival damage with the use of the confocal microscope as well as scanning and
transmission electron microscopy? The various companies and the ophthalmic community
need to work together to evaluate tissue culture for assessing ophthalmic products.
We initially tested known disinfectant substances that are known to set standards of
operations offered by GLP. The pattern and protocol for drug testing already shown via
this included several well-known agents. When benzalkonium chloride (BAK) was added
to the medium, at concentrations used in ophthalmic preparations, it caused marked cell
retraction. Mitotic activity ceased and the cells died within 2 hours of exposure.
Thimerosal caused cells to gradually retract. Mitotic activity terminated and cell destruc-
tion, due to inhibition of cell membrane processes, was apparent within 9 hours. Sorbic
acid suppressed cell movement and mitotic activity, but no cell retraction or cell death oc-
curred during 24 hours of exposure. When cultures were returned to normal medium with-
out sorbic acid, cell movement and mitotic activity gradually returned to normal.
Chlorobutanol at 0.5% caused an instantaneous cessation of cell movement as well as mi-
totic activity and large membranous blebs developed within 2 hours of culture exposure.
After 24 hours of continuous exposure to medium containing chlorobutanol, cells had de-
generated and morphologically were unlike the normal monolayer.
We investigated the effect of hydrogen peroxide at concentrations ranging from 30
to 60 parts per million and find that a dose and time-dependent response in the cells. This
agent is also employed as a preservative. There was marked cell retraction; mitosis, cell
movement and cytokinesis ceased and, depending on the concentration, cell death oc-
curred within 4 to 8 hours (8 hours at 30 parts per million and 4 hours at higher concentra-
tions). A conspicuous feature of HPz toxicity was the development of numerous optically
dense spots on the cell surface; these evolved over a period of time into optically lucent
membranous vesicles which measured up to 10 !lm in diameter. These probably were a re-
sult of oxidative damage to plasma membrane proteins and lipids. Scanning electron mi-
croscopy clearly revealed the presence of little vesicles covering the surface of the cells,
as well as marked retraction, after I Y2 hours exposure to 50 parts per million of hydrogen
peroxide.
Some of the newer preservatives were also studied. Polyquad, a high molecular
weight polymeric quaternary ammonium compound, caused no suppression of cell move-
ment. No adverse effect occurred on mitosis and normal cell proliferation during 24 hours
of continuous exposure. Dymed, which is a polyaminopropyl biguanide, also had no effect
on cytokinesis; no significant reduction in mitosis occurred during 24 hours of exposure
and nor was there any cell death or degeneration.
I would like to summarize some of the major advantages in using this model. First,
time-lapse video micrography can document continuously the response of the cells. Sec-
ond, one can use human cells which is a major advantage. And thirdly, different agents do
have different mechanisms of action. Different times exist at which cellular toxic adverse
reactions occur. Cell culture can be a very valuable and very sensitive model.
Edelhauser. Are there questions for Brenda? The data that Dr. Tripathi presented is
one model system, an epithelial monolayer to classify preservative toxicity.
McCarey. The one question I have is basically your tests are 24 hour exposure; is
that correct?
McCarey. In the real life clinical world these agents may be there for minutes. I was
questioned earlier about the efficacy of the rabbit model and here we're talking about 24
hours of exposure. How do you relate back to a clinical experience?
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 223
Brenda Tripathi. That's a very good question. I wanted to emphasize that the experi-
mental paradigm could, in fact, be changed depending on what you wanted to look at spe-
cifically. We have changed the concentration of serum in the medium so it is akin to tear
protein concentration. One can expose the cells to the agent for any amount of time. You
can calculate how much time you anticipate the compound remains in the conjunctival sac
before it's diluted out. So those types of modifications can be readily made and incorpo-
rated.
McCarey. Let's say you've flushed BAK solution over for two minutes or five min-
utes or so.
Brenda Tripathi. Those experiments are currently in progress. But I can tell you, for
example, that in a comparison of compound X and compound Y, we are noticing a differ-
ence.
Ramesh Tripathi. To answer Bernie's question, the changes are recorded from 0 sec-
onds to any time you want. We took it 24 hours maximum to see what happened to the fate
of the cell. But our findings are from 0 seconds, five minutes, an hour, to whatever you
want because with continuous time-lapse video micrography you can analyze the data and
can say at any given time what changes have occurred.
McCarey. Surely you can titrate it down to 1 minute or one whatever and I just
wanted to point out the issue that when you push it so far, 24 hours isn't a little bit of a
push. We're not talking 5 minutes, we're talking 24 hours. Almost any of these things are
probably going to kill the cells in that time. We should stick to something a little more re-
alistic in exposure.
Fraunfelder. With contact lenses it can be 24 hours because soft lenses can act as
slow release delivery systems. The amount of chemical released into the tear film will be a
function of the binding of the chemical to the lens material but a release can occur over a
long time period,
Ramesh Tripathi. We did not want to necessarily emphasize 24 hours but there are
drops and medications that you put in the eye that last for 24 hours. For example, take ti-
moptic XE. One drop is effective for 24 hours. But analysis can be done at ant time; this is
just an example of what you can do with this technique.
Ocular surface drug kinetics are such that dilution is instant and drug is lost from the tear
film rapidly, usually within a minute or so. This has been well established for many drugs
and tracers added to the tear film to simulate topical drop administration.
Exceptions occur, however, since BAK is retained in the corneal epithelium as are other
chemicals, too, for over 96 hours after a single 30 III drop application in vivo of a 0.03% solu-
tion. Exposure of the cornea to solution was normal, with no effort made to enhance contact
time. Based upon the kinetic analysis of tissue content of the radioactive material it was evi-
dent that the BAK diffused back into the tear film at a fairly steady rate over a long time pe-
riod rather than penetrating into underlying corneal stromal tissue (J Tox Cut Ocular Tox
1986;5: 133-142; 1987;6:89-107). In this case, therefore, the agent is present within the cor-
neal epithelial cells with subsequent loss to the tears despite a very brief exposure of the cells
to BAK. One should, therefore, have as much information on the pharmacokinetics of the
drugs under test in order to create a realistic in vitro environment and paradigm.
In contrast to the disposition of the cationic surfactant after topical application to the
eye, an anionic surfactant showed penetration into all ocular, and most systemic, tissues.
The duration of retention of this compound, sodium lauryl sulfate, was almost equivalent
to that of BAK, but the distribution was completely different. In both cases, however,
more uptake occurred in younger (and hence smaller) animals.
On another topic, there seems to be some good correlation between your tissue cul-
ture results with different disinfectants/preservatives and the effects of these chemicals on
in vivo corneal epithelial wound healing. Several years ago we presented the effects of
various compounds on mechanically-created epithelial lesions and the rate of re-epi-
thelialization. The overall results mirror your findings regarding toxicity; this includes
Polyquad and chlorhexidine digluconate, as well as cationic and anionic surfactants. In
our tests, sodium lauryl sulfate (an anionic surfactant found in shampoos) caused expan-
sion of the original lesion with a long delay in healing; many other compounds had no ef-
fect on healing rate (J Tox Cut Ocular Tox 1989;8:253-269).
These in vivo:in vitro relationships are important since they can reveal whether the phe-
notype of the cultured cells remains the same in terms of the response to a particular stimulus.
It is probably important only to use cells in vitro that have only been through three to six or
seven passages in order to insure that the phenotype expression is not significantly altered.
Nobuo Takahashi, M.D., Kanazawa Medical University, Kanazawa, Japan. The cy-
totoxicity of chlorobutanol is different depending upon the temperature of the bathing so-
lution. At what temperature did you treat the cells?
Edelhauser. Brenda, I have one other point, are these mono layers of epithelial cells?
Edelhauser. Sherry Ward at Gillette, I remember her poster at ARVO, has been able
to get epithelial cells to grow in layers.
Edelhauser. How does a monolayer tissue culture system that you have compare to
hers? Are we going to see differences or do you think we'll see similarities?
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 225
Brenda Tripathi. Well, you probably will see some differences. Clearly, the human
corneal epithelium in vivo has five to six cell layers. If one is interested in obtaining
meaningful data about cell morphology from time-lapse video micrography and also phase
contrast micrography, one really has use cell mono layers for resolution.
Walter H. Bee, Ph.D., Corning-Hazleton, Munster, Germany. Brenda, have you ever
tested other cell types or has it been exclusively epithelial cells?
Brenda Tripathi. For these toxicity studies we have concentrated on human corneal
epithelium because this seems to be more pertinent when you're using topically prepared
ophthalmic preparations or contact lens solutions. Other studies have been done with en-
dothelial cells. That is an important criterion; when undertaking in vitro studies you have
to choose the right cell type, the one that you're interested in and the one that is going to
be exposed.
John N. Norton, D. V.M, Ph.D., D.A.B. T, Alcon Laboratories, Ft Worth, Texas. I will
give you a perspective of how I, as an in vivo toxicologist, use or will use in vitro assays
in risk assessment to examine the toxic potential of a compound or a chemical device. Re-
lated to that, I will extend more into the mechanistic area and actually go over a few ex-
amples using an alternative tissue system and show some pharmacological approaches that
are used in deciphering mechanisms.
What are the objectives of an in vitro toxicity study? First, it can identify toxic po-
tential. Secondly, the tests are commonly used to rank order toxicity of compounds within
the same class and between classes. Thirdly, species differences related to the pharmacol-
ogy or pharmacotoxicity interpretation can be examined. For example, Hank Edelhauser
with his in vitro corneal endothelial perfusion model typically uses rabbit corneas and
then substantiates and confirms his results in human corneas. In addition, in vitro studies
can delineate different mechanisms due to different abundance of receptors and different
cycling pathways.
How do we use the information from in vitro toxicity studies? First, we could use
the data to screen compounds for toxic potential and perform rank ordering of toxicity. It
is also possible to screen raw materials that will be used to make the final product. Formu-
lations can be screened. For example, a compound in different formulations can affect ef-
ficacy as well as impact on toxicity. Finished product lots of different compounds can also
be evaluated for product specifications or release testing. Another approach is as an ad-
junct to help reduce the number of animals that are utilized. Second, the mechanism of ac-
tion is another approach. For example, if toxicity is observed in vivo, you may take a step
backwards and try to delineate why toxicity is being elicited. The third example is replace-
ment of animal studies with in vitro testing and any time that an in vivo study is replaced
the in vitro model used as a replacement must be validated as truly predictive.
The following is a summary of what I consider are the most important points for
correlating in vitro observations to in vivo relevance. Ideally you want an end point that is
quantifiable, not only in the in vitro system but from animal to human. This is further
226 H. F. Edelhauser and K. Green
Edelhauser. These studies illustrate tissue culture techniques that can be used to
evaluate human safety and possibly even efficacy. How many people in the audience have
used tissue culture? Let's see a show of hands; maybe 10% or so. Do you feel that these
tissue culture models are going to be useful as we try and assess metabolism, gene ther-
apy, toxicity and efficacy? Any thoughts along those lines? I think there's some potential,
but obviously we were all wondering where tissue culture fits into routine toxicity testing.
Alfred Wegenel; Ph.D .. University of Bonn. Bonn. Germany. I would like to make a
comment on both talks that we've heard because I'm not working in tissue culture. My
feeling is that there are a lot of options that we can have with tissue culture in comparison
to animal work and to human data but we do not use enough of those options.
I've seen a lot of biochemical parameters that are of interest that you have pre-
sented, Dr, Norton. My question would be, have you found parameters that have already
been shown to be predictive?
I would like to ask you, Dr. Tripathi, if you look at the morphology of your cells and
compare it to a whole cornea that has been exposed in the same way then what are the re-
lationships? Have you any experience with how predictive this is?
Norton. Related to the reproductive system, several investigators have been able to
delineate earlier changes using second messengers and reporter genes prior to actually
seeing cell death in cell culture. There has been some work with therapeutic agents using
receptor binding of competitive-type nature. This has become more prevalent in other tis-
sues such as reproduction, liver toxicity, those areas where you have further charac-
terization of the tissues and what's going on. PCR is a valuable tool because the sensitivity
of detection increases dramatically. Related to that there's a company in Rockville, Mary-
land that is determining the human genetic make-up and a lot of pharmaceutical compa-
nies are using over-expression of secreted proteins and seeing their pharmacological
effects. So I think in those regards PCR is a valuable tool. People in other areas have used
228 H. F. Edelhauser and K. Green
it more readily but I think, just like I mentioned with the antisense in phase III for CMV
infections in retinitis, it's coming and so its importance will grow.
Brenda Tripathi. We have not used our model in the strict sense of predictability be-
cause we have only investigated compounds that have been available on the market. Looking
at long-term clinical effects in terms of toxicity is extremely difficult in humans because you
don't know whether it's due to the drug per se, the preservative, the vehicle, and so on and it
largely relies on anecdotal accounts. But I would like to add that at least some of the effects
that we have documented with BAK and HP2' do correlate very well with reports of toxic
side effects in humans. Cell culture hasn't been used as a predictive-type of test as yet, but it
could be done very easily to screen products which are going to come, or be considered,
for marketability, either as the entire formulation or as individual components.
Wegener. I must say I am glad to hear that and I think this can be one of the advan-
tages of this workshop, namely to find out what direct comparisons can be made between
results from your tests and, for example, from my animal tests. I am going to come to that
later on but I think we should begin cooperation between investigators and compare and
contrast in vitro test results and in vivo results on the same compounds.
Norton. One of the things I did not discuss was that I did some work with oramu-
cosal cultures to evaluate dentifrices. There you have a lot of surfactants. I was able to
take a similar biochemical marker that I could assay in my cell culture. This marker could
be captured from irritated tissues in a dog model and quantitated. I was also able to cor-
roborate results from consumer product test panels in humans. By designing studies care-
fully I was able to actually identify markers in in vitro systems and extrapolate results to
higher order testing. I think that was a good way to approach use of this technique.
Wegenel: To conclude let me just mention that you should not forget that there is
also the possibility of doing work in a two-sided chamber where you have either a pig or
even a human or bovine cornea as a membrane in between. We could directly try to corre-
late the results from that system to corneal responses in vivo.
Brenda Tripathi. Yes, I would concur that it is extremely important to correlate and
compare results from in vitro cell culture to what we know is happening in animal models
and also in human patients.
Edelhauser. Thank you. Bob Hackett from Alcon Laboratories has agreed to summa-
rize Alcon's program for ophthalmic drug development. Obviously he's not going to re-
veal company secrets but Alcon's program is one way of how you bring a product from
the drawing board to become an active product and on the market.
About five years ago when the FDA listed what was important to them, pharmacokinetics
was number 33 on the list, it is now number two. It is a very important aspect of toxico-
logical evaluation and clinical risk assessment. So, depending on what is projected in
terms of metabolism, the pharmacological class of the compound, what's known about
that class, etc., determines what Alcon will do in terms of conducting systemic toxicity
studies on novel chemicals.
For example, a non-steroidal anti-inflammatory which is only going to be used for
two weeks in duration clinically is going to have a different program than a glaucoma
drug which could be used for 45 to 50 years of somebody's life. In these studies we do
what we need to do, we don't do everything for every specific compound. But what they
are positioned to do is determine the target organ toxicity and give us an indication as to
what to look for in subsequent clinical studies. Is a compound a hepatotoxin? That gives
the clinical people some index of what to look for. The data establishes the no-effect level,
to estimate the margin of safety and to determine a carcinogenic potential if that's neces-
sary and warranted for that compound. These consist of short-term studies, four weeks in
duration, which are dose-finding studies used to set the doses for the 13 week studies
which are used to set the dose levels for longer-term studies which may be six months to
one year in duration. That's under discussion with the ICH right now in terms of what is
the appropriate length. Is anything gained by carrying a study to 12 months or is mankind
better benefitted by shortening the drug development process and getting drugs on the
market and to the patient sooner?
I mentioned carcinogenicity studies; in my opinion they mayor may not be war-
ranted. There is a lot of controversy about the bioassay model. We have seen a large ge-
netic drift in the rats that are used in terms of body weight gain and decrease in longevity.
Generally these bioassays are conducted in two species, similarly that's the case with our
ophthalmic studies. We use the albino rabbit currently; we are developing a pigmented
strain. We use the primate which provides us with information on any interaction such as
was seen with latanaprost in terms of iridial color change.
Melanin interaction is becoming a very big concern with the FDA. We know that
certain drugs bind to melanin and increase local concentration within the eye. What are
the ramifications of drug binding in terms of toxicity? So, for an intermediate used drug
like an antibiotic or a non-steroidal, we run through subchronic studies. They cost about
a quarter million dollars. For longer-term compounds, adding to the chronic tox, that
would be an additional half million dollars. If the carcinogenicity studies are deemed
necessary, you're looking at approximately two million dollars. More importantly you're
looking at a three year time commitment which is preventing your drug from reaching
market.
Reproduction and fertility impairment induced by a drug are critical. In Europe and
Japan typically the government agencies want to see if your compound affects male fertil-
ity, spermatogenesis, etc., prior to going into humans. In the United States, until recently,
it was a concern that you not put your drug into women of childbearing potential because
of unknown birth effects. The FDA has indicated that they would prefer to see women in
phase I safety studies in order to find out up front what's going on. There are three seg-
ments to reproductive fertility development studies. Segment I, which evaluates whether
or not your compound impairs fertility, both in terms of male and female reproductive ca-
pacity. Segment II, teratology studies using two species, rats and rabbits, that came about
because of thalidomide in the 60's. The drug is metabolized differently between rat, rabbit
and man. Segment III studies which are peri- and post-natal development. That's looking
at the effect of your compound through multi generations. The animal is exposed in utero
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 231
and through lactation and then followed for two generations to see if there is an effect
down the line. These are pretty expensive studies, they are about $350,000 and about a
year in terms of time commitment.
There are several specialized studies which are necessary to conduct to assure
yourself that your drug is not a contact sensitizer. Generally this is the guinea pig maxi-
mization test (GPMT). Some can argue it's relevance to the eye. But if we get something
that's a hot sensitizer, typically then we've got a real problem. We know from our devel-
opment program that a compound that is positive in the GPMT will likely produce sensi-
tization in some patients. So there is at least some correlation; how strong, I'm not sure.
Quite often, depending upon the class of the compound we'll look at cell proliferation
assays, looking at whether or not a compound can cause cancer through an epigenetic
mechanism.
Lastly the SHE-cell transformation assay which is pretty new on the horizon. It has
yet to be fully accepted by the FDA but so far it appears to have very good correlation in
terms of predicting epigenetic carcinogens. If we're looking at trying to get rid of animals
for testing, we have got a long way to go. There are a myriad of things we're looking at in
terms of assessing whether or not a compound is safe for clinical use.
Edelhauser. Thank you, Bob, I think this really gives a very nice overview. I had not
realized the degree to which a compound has to be tested. Are all the tests necessary? We
have representatives from Japan and Europe here, what's the international market require?
Hockwin. Bob, I very much enjoyed your presentation and I think you should make
an asterisk on slit-lamp biomicroscopy. Alcon is equipped with two Scheimpflug cameras
and they perform Scheimpflug photography and image analysis where special problems
with the lens occur.
Hackett. Well, it's funny you say that, Otto, because we are involved in a rather
sticky legal issue with a company who sold us some technology for determining lens den-
sity and we're assessing a newer system. So that will be part of the ocular parameters that
we'll be looking at, certainly for long-term drugs for glaucoma or something like that. I
think it's important, I'm glad you pointed that out.
Edelhauser. Has there been a test that's been overlooked that may be used in Europe
or in Japan?
Peter Walter, M.D., University of Cologne. Cologne, Germany. That was a really
complete program, I think. Two years ago in Annecy we discussed the pros and cons of
functional electrophysiology testing. Do you consider this procedure?
Norton. Dealing with devices, many of these assays and testing that Dr. Hackett
mentioned are not conducted. It's not quite as labor intensive. The tests are used for look-
ing for potentialleachables from a solid device like a viscoelastic or an IOL. The first test
that needs to be evaluated is actually a cytotoxicity test. It is not usually a corneal-type
cell although for some European submissions they did require us to go back and evaluate
corneal effects. Usually the cytotoxicity study uses a modified cell line such as the L929
mouse fibroblast line. Another difference that relates to genotoxicity, only a mutagenicity
assay such as the Ames is typically conducted. Chromosomal aberrations and other geno-
toxicity assays are rarely done. The overall cost associated with the development of a de-
vice, as far as toxicity testing is much less.
Edelhauser. I have one comment concerning the cell flare meter. A company devel-
ops an instrument and says we want you to use this clinically but has that instrument been
validated? This particular apparatus probably has been validated now but what we're deal-
ing with is a case not only where an instrument is developed but how is it validated to be
used for ophthalmic testing?
A good example is specular microscopy. The "gold standard" for specular micros-
copy has been the Keeler-Konan Contact specular microscope. But we have three new
non-contact specular microscopes. How do they compare? There have been no validation
studies. We get good cell numbers but the companies who develop these instruments ha-
ven't done the validation to be able to give toxicologists an instrument that is ready for
use, it's incumbent upon yourself to undertake a study to validate that company's instru-
ment.
Hackett. We have done extensive validation of our non-contact versus the Keeler
versus your work as well, so I think we've got two or three publications on the validation.
But you're right, it is incumbent upon the company acquiring instrumentation to validate
it for use in-house and that is becoming a bigger issue with the FDA and rightly so. You
can't just pull something out of a box and start using it and make judgements of clinical
safety on that.
Frances Kane, Ph.D .. Ciba Vision Ophthalmics. Duluth. Georgia. I would like to
hear a discussion of other company's and academic's views of the utility of doing carcino-
genicity testing for ophthalmic products.
Hackett. Hackett says yes. What you are probably referring to is in the ICH which
came out with a statement that carcinogenicity studies may not be applicable for ophthal-
mic drugs. We also agree with that position, but it depends. They have got four parameters
to look at. What is the extent of systemic exposure to your population? That is going to be
key. Pharmacokinetics comes into play there. Are there structural alerts? What is the phar-
macological class? What is the genotoxicity? A myriad of things require evaluation but we
are of that opinion and we work very closely with the FDA and I'd also like to stress and
reiterate what Dr. Avalos said. Customize your program and work with the FDA, they are
very willing to do so.
I cannot say that carcinogenicity studies are unanimously not necessary because I think
there are instances where they might be appropriate. Latanaprost and it's effect on iridial color
change, for example, where a recent evaluation of a patient showed an iris growth which
some pathologists thought was a pre-melanoma and some did not; it's still in discussion. In
that case if you're looking at prostaglandins, which affect iris color, you may want to consider
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 233
a carcinogenicity study. If you're looking at an antibiotic which has limited use, I don't think
you do but you've got to build a case for it. If you're looking at drugs for CMV for AIDS pa-
tients, does it really matter if you do carcinogenetic studies? So, a lot of common sense and a
lot of good logic and working with the FDA is warranted in this case.
Avalos. Dr. Hackett has made some very good comments. Our policy is to discuss
your development program with you at the earliest time possible.
I want to add to what we consider chronic use. Usually we define six months of con-
tinuous use or six months of intermittent use over a ten year period as chronic. So if your
product or the indication for your product is something that fits that definition, most likely
the Agency will recommend that a carcinogeneticity study be conducted.
For dermal products, a dermal carcinogenicity study may be recommended in addi-
tion to systemic carcinogeneticity studies. The systemic carcinogenicity studies may be
recommended if substantial absorption is observed after topical administration. But all this
information may be obtained from the toxicologists at the FDA. We are more than willing
to give that information.
Hockwin. I agree that carcinogenicity studies for ophthalmic drugs are often not nec-
essary. On the other hand I would like to stress the situation for many pharmaceutical
companies for drugs other than specifically for ophthalmic field, who have to perform a
two year carcinogenicity study using systemic drug administration. They should include
ophthalmological examinations in such protocols because this is the best that can be done
in terms of continuous systemic supply of drug in high concentrations to eye tissues.
Fraunfelder. In fairness, there is some data to show that the antivirals can cause epi-
thelial dysplasia which may be pre-malignant. So maybe local studies may be indicated.
Edelhauser. Are there any comments from our international representatives about
the battery of tests that Alcon has used? Are there tests that are overlooked or are there
tests that are required by your country that should be incorporated into this program?
mation by evaluation of the results from systemic and local toxicity studies with respec-
tive toxicokinetic data as well as from kinetic and metabolic studies after local applica-
tion.
Hackett. My only comment to that is look at latanaprost. The color changes were
seen in primate studies but not until after approximately three to four months. We have an-
other compound that is on the market, and it has been associated with sub-epithelial cor-
neal haziness in some patients. That was also seen in-house but it took eight months in a
primate study to see that; so granted you get greater exposure systemically but I think that
with chronic drugs or novel delivery systems that we don't know anything about, I believe
it is warranted to run out longer term ocular studies to look for these subtle effects.
Edelhauser. Thank you, Bob. Dave Potter, who has had experience both in industry
and academia with the development of glaucoma drugs, will discuss various animal mod-
els used in drug testing.
Potter. I will give you examples of toxicity-related situations that I've encountered
during a variety of activities, both in academia and in industry to reinforce some of the
things that you've heard.
For example, one of the compounds we became interested in as a potential anti-glau-
coma drug are a group of compounds that fall broadly into a class of aminotetralins. We
examined the drugs in three species, cat, rabbit and primate. Interestingly enough, the only
species that showed toxicity was primates. There was virtually no toxicity either in rabbits
or cats. One of the things that a drug designated N0437 would do is that it produced a
dose-related change in monkey pupil diameter causing a modest degree of miosis. The dif-
ference in pupil diameter between eyes in the same animal was thought to be an acute ef-
fect. However, the ocular pressure changes occurred in animals treated in June and results
were still seen in July. We got a prompt decrease in intraocular pressure in these animals
followed by a more sustained effect.
We began to look at these animals a little more closely and noticed that the drug had
produced a Horner's Syndrome where we observed a significant miosis in the treated com-
pared to the untreated eye. The miosis was accompanied by a narrowing of the palpebral
fissure and these effects were accompanied by a sustained lowering of intraocular pres-
sure. The responses persisted for weeks and weeks, with 13 versus 18 mm Hg in the
treated versus experimental eyes, respectively. We applied fluorescein to the treated eye
of these animals and noted almost completely denuded corneas. The corneal endothelium
may have been damaged as well. So that's my example of how important it is, as the FDA
has pointed out, to use more than one species in testing a drug because there can be very
species-specific effects.
I will now talk about proliferative diseases of the eye and how important it is to con-
sider the species involved. There are a number of proliferative diseases in the eye that lend
themselves to novel approaches. Specifically I'm talking about situations where you use
biotechnologically derived products in an effort to control a disease that is an example of
selective toxicity.
I will be talking about an approach to secondary cataract. When performing extra-
capsular cataract surgery, after removal of the cataract what is left is a layer of epithelial
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 235
cells or even only a few cells on the posterior capsule that may later cause a problem. The
cells left behind can begin to grow, dedifferentiate and become myoepithelial-like in na-
ture. As a result of getting multiple layers, and contraction of the remaining lens capsule,
you can get a significant opacity which is known as a secondary or after-cataract. The sig-
nificance of the problem is that there are a number of people that have cataract surgery
and after a while a significant proportion will develop an opacity on the posterior capsule.
A simple solution to the problem is to use a laser to blow a hole in the posterior capsule
and restore vision. One can treat with a YAG laser but in this case there can be sequelae,
that is a rise in intraocular pressure can occur. It is also possible to rupture the anterior
hyaloid face and displace vitreous with the possibility of a retinal detachment. There may
or may not be corneal endothelial cell damage due to laser use.
The novel idea was to use selective toxicity in the form of an immunotoxin, and per-
haps prevent the formation of a secondary cataract. This is an example of how to test tox-
icity related to the species in question. In this case referring to human. The approach was
to produce specific monoclonal antibodies to a specific epitope on the surface of lens epi-
thelial cells. The epitope on a human lens epithelial cell is very characteristic to the hu-
man species. That monoclonal antibody will not cross-react with lens epithelial cells from
rabbit, cat, etc. Thus, it makes it very difficult to use animal models to test for possible
toxicities. The general approach is to take a monoclonal antibody that's specific for an
epitope on a particular cell type and have a linking agent like SPDP which links the mono-
clonal antibody to the cytotoxic agent. In this case, we're dealing with part of a ricin
molecule, the chain known as ricin-A.
When you're dealing with a biotechnologic product it is not one agent but, in fact,
three agents that have toxicity potential. The idea is that the immunotoxin will be taken up
by the cell. Once the ricin gets released it blocks protein synthesis resulting in cell demise.
In this particular circumstance, the idea was that the antibody or the immunotoxin would
be cell-specific. That is the ricin released from the antibody would kill the lens epithelial
cell without damaging other cells in the eye.
Because the compound was effective only in humans it had to be tested on other
types of cells. Assays were done on human lens epithelial cells conducting dose-related
responses at various exposure times using a number of end-points to measure not only in-
hibition of protein synthesis but also to some extent cell viability. This agent worked in vi-
tro causing the death of human lens epithelial cells (HLE); usually a little better on a
confluent cell culture than in subconfluent cultures, but it did work fairly well in vitro.
The question at that point became what is this particular product going to do to a popula-
tion of ocular cells that may not have the ability to regenerate, namely human corneal en-
dothelial cells? We used methionine labeling as an index of what might be happening to
these cells. The immunotoxin had very little or no effect, at least at the doses tested, in
terms of preventing the protein synthesis or survival of human corneal endothelial cells
(HCE). Comparing the effect on HLE versus HCE-type of cells it was more toxic to HLE
than HCE cells. In many cases effects can be species-specific and tests must be designed
accordingly.
Basil Worgul, Ph.D., Columbia University, New York, New York. It strikes me that
not only do you have to pick your species and your model correctly but also the cell type.
A subset of posterior capsular opacifications are due to iridial seeding; Fred lakobiec and
our laboratory have shown that. One would expect that the system wouldn't be too effec-
tive if you have a ricin attached to immunoglobulin against lens proteins if the cell is not
lenticular in origin.
236 H. F. Edelhauser and K. Green
Potter. All I can tell you is that I no longer am associated with the company in ques-
tion but apparently this product is now in phase II clinical trials getting ready to go into
phase III.
Wegener. I might have missed it but how do you get your active agent to the site
where you want to have it active in the eye?
Wegener. Yes.
Potter. In this case, there are one of several mechanisms you can use. This agent can
be deposited in the eye by injection at the time of surgery. It can also be coated on an in-
traocular lens, as well. But I think that the idea was that it would just be deposited into the
surgical site.
Wegener. You know that George Duncan's group in Norwich also works on the prob-
lem of secondary cataract and they coated their lenses with thapsigargin, which inhibits
calcium metabolism, and found this drug offers a quite effective way of preventing secon-
dary cataract.
Potter. Yes, I think there are a number of approaches, this was just one that was util-
ized.
Bret Jessee. Ph.D .• Bausch & Lomb. Rochester. New York. Bausch & Lomb's perspec-
tive on ocular toxicology is that of a manufacturer of generic drugs, over the counter
ophthalmics and novel contact lens and contact lens care products. While direct ocular in-
stillation products, such as eye lubricating and moisturizing eye drops, are relatively
straight-forward to evaluate by means already discussed, contact lenses and related prod-
ucts present a more complicated challenge.
Contact lenses can act as delivery devices for most chemicals that a lens is soaked in
prior to insertion onto the eye, including the packaging solutions or drugs. Bausch &
Lomb does not indicate or recommend the use of contact lenses as delivery devices, as
prior regulatory approval is required for such use. But lenses, especially soft contact
lenses, can retain agents both in the interstitial water volume and bound to the lens poly-
meric matrix itself.
As soft contact lenses vary in water content from low « 40%) to high (> 55%), in
ionic nature from negatively charged to little or no charge, and also vary in specific
chemical functionality groups, different lenses can uptake and release different exogenous
compounds to different extents and at different rates. Aside from regulatory hurdles to
overcome, these differing lens/compound interactions have contributed to the lack of pro-
gress in formally approving contact lenses as delivery devices. Bausch & Lomb strives to
develop products to be used with all contact lenses. This goal can be technically translated
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 237
into attempts to develop multi-functional products that are biocompatible with all lenses,
which can be viewed as different ion exchange resins with different affinities to solution
ingredients. Thus we approach ocular toxicology not only in the realm of contact lens for-
mulation and design, but also in lens care product formulation and design.
People have come to consider contact lenses as purely cosmetic devices and would
like to be able to buy them in grocery stores or in vending machines. But although contact
lenses appear outwardly simple, only eye care professionals have the skill and knowledge
to provide patients with appropriate choices of lenses and care systems. The professional
literature provides numerous and frequent examples of health, comfort and visual acuity
problems caused by mixing and matching lenses and lens care products without proper
training. There have been a number of lens and lens care product incompatibilities, even
with some globally marketed products.
Just as Dr. Hackett has described a complete program for ophthalmic drug develop-
ment, there are also numerous tests required for lens and lens care product regulatory sub-
missions. Contact lens disinfecting solutions are a medical device in the US, a
quasi pharmaceutical in Japan and cosmetics in other countries. Thus, the range of studies
required for global lens and lens care product submissions includes many of the same
acute and subacute general toxicity, sensitization and genotoxicity protocols, with accom-
panying PKlTK work, as required for novel pharmaceuticals.
As you get into new discovery agents for such simple actions as cleaning and disin-
fecting contact lenses, one man's excipient or vehicle turns out to be another man's active.
So although an ingredient may not be listed in a pharmaceutical monograph, a lot of the
compounds that my laboratory has screened do have biological activity. The methods that
we would view as compendial or standard guideline methods are probably satisfactory for
proving to a regulatory body that you have developed a safe drug, contact lens disinfectant
or cleaner but they are not sufficient to discover one.
I think that new product research and development is really the greatest application
we have in in vitro toxicology - sorting through the dozens of potential actives to find
those with the greatest potential efficacy and safety to then take forward to human clinical
trials. When presented with a final product formulation, in vitro toxicology becomes a
very small part of my repertoire. While good for ranking potential product formulation
choices such in vitro methods have less power to predict anything less than a significant
ocular irritant.
Edelhauser. Dr. Mary McKee, do you have any comments you would like to make as
an industry representative?
Edelhauser. That's a good point, 21 days for contact lens solutions compared to the
5 days; what is the thought of the audience?
Rodeheaver. It does and in some of the cases we do not have enough experience
with the chemicals and their structure/activity relationships, as we would classically from
ophthalmics and drug perspective how they are going to interact with some of the other
things that we see in cleaners. This occurs simply because they are not normally used in
those contexts.
Jessee. I guess I have to pick sides and support a move more toward a shorter rabbit
test procedure. If, as Alcon's studies have shown, some problems can be detected only af-
ter months of animal testing, then even a three week animal test may alone be too short to
assess product success 100% of the time. Certainly, no one would advocate a six month
rabbit test for a contact lens product. Lens retention and animal welfare become major
barriers to conducting studies with lens wear longer than about 4 weeks. The required
chemical tests, such as lens compatibility and chemical uptake and release studies, should
provide manufacturers with enough data to assess whether long term chemical interactions
with lenses is a significant issue.
Because conduct of human clinical trials for contact lenses and care products does
not require approval by FDA in the US, it is incumbent on the manufacturer and the insti-
tutional review board (IRB), otherwise known as the human assurance committee, to make
sure appropriate information is available to safely conduct human studies. Approval to
market a product comes only after extensive 3 to 6 month human clinical trials designed to
elicit any adverse reaction potential of a product. No animal model can yet predict prob-
lems arising beyond 6 months of human studies. The rabbit ocular irritation model was de-
veloped to assess acute irritation. Beyond about a week, the model is no longer useful, as
now one is reasonably sure that any problems that might arise in human trials would be
slow in onset and represent no significant risk to health in monitored studies.
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 239
It is difficult to understand how a 3 week rabbit study can add much value to a PMA
submission, when accompanied by a presumably successful extended human clinical trial.
Obviously, if human studies show potential problems, additional chemical, animal and hu-
man studies may be necessary to assess safety. Given that contact lenses have little or no
general therapeutic benefit, no significant health risk can or will be tolerated by regulatory
bodies or the consumer.
Edelhauser. Let's hear what Dr. Wegener has to say about animal models.
Wegener. I will use my short talk as an update on our experience with animal mod-
els. We have animal cataract models based on hereditary defects and those due to acquired
or induced defects. Originally we thought that hereditary models were not very useful be-
cause it seemed difficult, a priori, to interfere with genetically-based defects. In addition,
in several models it still is unknown what the underlying genetic effect really is.
We have been able to show in some models, however, that it is possible to interfere
with progress of the lesion. We refer to a mouse model from a large collection of heredi-
tary cataract models that are bred in a national research facility near Munich in Germany
(GSF). They had a mouse model with an anterior suture cataract. We could demonstrate
that if we irradiated these lenses with UV-B we enhanced the expression of the genetic de-
fect. This was the first case where the possibility of interaction with an inherited cataract
model was demonstrated. Maybe one of the major advantages of this model is that the in-
duction triggers changes in whole eye development in a rodent from it's origin at day 10
or I I post-conception to senile changes. There are a lot of options although so far there
have not been many interferences with genetically triggered processes.
Let us focus more on cataract models that are based on induced diabetes or on a
compound that is given in specific doses or on radiation applied to the animal. There are
some prerequisites or requirements that are necessary before starting with these models.
First, the model has to have clinical relevance. It does not make sense to look to a cataract
model where the mechanism has never been found in humans. We have to require that the
mechanism is more or less known, and that we can set the starting point. These are impor-
tant features that led us primarily to the usage of cataract models where we induce a proc-
ess.
One of the models that we studied during the last two years was a UV -cataract
model. If we apply UV-A on it's own, it causes a subliminal effect. Over months it will
not cause lenticular or corneal changes but we could show that if we make the animal vita-
min E deficient via diet, we enhance the UV -A effect so that the lens density increases,
over a period of eight to ten weeks, without a visible cataract. This is a good example of
working in the phase where there is no visible cataract but we definitely can say that the
scattering in the superficial cortex increases all the time. UV-B is a direct cataractogenetic
factor which causes an anterior polar cataract.
If you look at the suture, not only the center but also the branches of the suture be-
come opaque. How can this model be used for drug testing? We have found a drug that
slows down the UV effect, so we could show that it is possible to interfere with lens den-
sity development. It was a drug that not only protected the eyes as a whole or prevented
UV from entering the eye while it also depressed the effect of UV in the corneal epithe-
lium.
240 H. F. Edelhauser and K. Green
Sugar cataracts are a model widely used in drug testing. One finds drugs that act
against these cataracts, like a number of the aldose reductase inhibitors, but we also found
that there are drugs that speed up their development. Among various compounds from the
group of quinolones, one speeds up diabetic cataract development and thus gave a positive
interaction. Based upon this finding, one must be careful when prescribing this drug for a
patient.
Another chemically-induced model is the naphthalene cataract, where there are in-
teractions in both directions, prevention and enforcement. Prevention was achieved by ap-
plying certain types of aldose reductase inhibitors, but other compounds also enforced this
model, thus the density of the cortical zonular cataract was higher at the end of the experi-
ment.
In the rabbit, the naphthalene cataract model is not useful because the homogeneity
of cataract development is low. Concomitant with the cataract appearance is a band-
shaped keratitis formed by naphthaquinone in this model. In my experience this is an in-
teresting model for corneal changes. In a tolerance study of a compound that was not
marketed, we could show that this compound enforced the formation of a band-shaped
cataract in the rabbit.
Cataracts occurred in animals with tryptophane-deficient diets. We have not used
this model for testing of cataractogenic effects of drugs, however, but we tried zinc defi-
ciency in animals. Zinc deficiency in the cornea and lens can change a drug effect. In the
cornea, zinc deficiency induced via diet suppressed the sensibility to UV damage.
We also used various models in combination with each other which enhanced one of
either cataract forms but there never has been a formation of both cataract types in paral-
lel. I would like point out that when we conduct these studies we always do examinations
with slit-lamp microscopy on all eyes. We also do Scheimpflug photography with con-
secutive image analysis. When I indicated that we see an enforcement of the naphthalene
cataract, for example, this is based on the fact that we have measured density increases.
We always do biochemistry after the lenses have been dissected. For example, if we
have an effect on lens growth, which months later could lead or contribute to cataract forma-
tion, it will be observed early based on the lens fresh weight. Of course, there are many other
parameters including protein content, enzyme activity, co-enzyme concentration, etc. that can
tell you something about how a particular drug interfered with this cataract model.
In our experience with a large variety of test drugs such as antibiotics, anti-inflam-
matory drugs, non-steroidal anti-inflammatory drugs and many drugs of internal systemic
medical application, these cataract models proved to be a very valuable tool. Although
you need certain equipment for evaluation, especially a slit-lamp and if a lens effect is ob-
served we are convinced that Scheimpflug photography is an extremely valuable tool. I
hope this overview shows that these models, that primarily focus on the lens, are impor-
tant tools for drug side-effect evaluation.
9. ELECTRORETINOGRAPHY
Bee. I will tell you about ERG in primates. We will start with toxicology but I will
finalize my presentation with a few words on pharmacokinetic testing.
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 241
change in the ERG. I will give you one short example. If we have a 50% reduction of the
amplitude, the question is, are 50% of your cells no longer functional or are all cells work-
ing at a 50% level? Now, if 50% of your cells are dead you expect the others to work nor-
mally. That means latency will not change. If all cells work at the 50% level probably also
the latency will be changed.
This is a model which can also be used in pharmacokinetic testing so you can do this
test pre-dose, dose the animal and do further tests in the following hours or days. Much
depends on the pharmacokinetics of the test drug. But fast results can be obtained because
the original curve, as far.as the individual is concerned, is stable. So you have a valuable
model for pharmacokinetic testing.
The procedure for photopic stimulations is as follows: from the first flicker-imme-
diately after the background illumination-and the second one to minutes later, there
should be an increase in amplitude with the response occurring a little bit earlier.
Beyond the requirement of the SeE we also do pure cone testing. We do that by
eliminating all rod responses by using red flash. We record three intensities with the red
flash before we do the white flash cone response. Under background illumination condi-
tions we can insure that we have a pure cone response. Data transfer to a graph shows an
entirely different curve than in the scotopic measurements especially as far as latency is
concerned. The curve is very similar to that seen in humans.
This was the procedure I wanted to point out; we are now installing this for pharma-
cokinetic tests.
Dennis Carson, Ph.D., Alcon Laboratories, Ft. Worth, Texas. Are there compounds
that affect ERG in primates that do not affect rabbit or another species; are you aware of
any?
Bee. I am not aware of that. Do you mean using the rabbit or another animal might
be cheaper?
Carson. Exactly. I know the response is similar to the human and primate.
Bee. Bob Hackett and I have talked about the cost of the development of such a drug
in the pre-clinical phases. It is about three and a half million dollars and an ERG study
with rabbits would not cost very much less than a study in primates. The savings would
only be $50,000 of that total, it is ridiculous to select that option on the basis of cost; you
have a much better model using the primate which is much closer to humans. Why should
you save $50,000 if you do not know whether you will get good results or not?
Hackett. The only time I could see that there may be a difference is with melanin
binding. Unless you are using a drug that binds to the retinal pigment epithelium like
chlorpromazine, which could cause an effect which you may not see in an albino rabbit,
then responses could be reasonably similar. Differences in metabolism could playa role in
causing different responses which may be generating something which has an effect in a
primate because it is metabolizing it one way versus a rabbit that may not produce that
metabolite.
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 243
Bee. There is also another advantage of the primate, with respect to dogs that are
also occasionally used in ERG studies. The dog has a tapetum which can act as site for
drug action itself and change responses to everything. Also the dog does not have a mac-
ula, which means that a macula effect cannot be checked. Even if dogs are a little bit
cheaper, the results obtained-unlike those from primates--are not applicable to man.
Walter. Some comments to that. First of all, the ERG components show essential dif-
ferences between cold and warm-blooded species. In warm-blooded species, the ERG's
are about the same.
Another thing I'd like to mention concerns the story with latencies and amplitudes in
the ERG; it is not that easy to say that an amplitude reduction of 50% with a maintenance
of a latency means that 50% of b-wave generating neurons are down or if the latency is
prolonged they are functioning at a 50% level. The problem is which species of neurons
are involved in the pathological process. There are different kinds of retinal neurons with
different implicit times contributing to the generation of the b-wave. So you cannot put it
all together.
Bee. It is, of course, a simplification but it just shows how this model could work.
Bee. Probably the word value was not a correct use, I confess that.
10.-------------------------------.
Tear/Cornea
10 20 30 40
Relative Location
Figure 2. The natural fluorescence ofthe rabbit eye measured with the non-invasive technique.
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 245
1987 measured epithelial permeability and came up with a difference of 1.6 times less be-
tween the two agents. We found about a 12-fold difference. Regardless of the ratio differ-
ence, it is clear that sodium fluorescein has a greater permeability across epithelial cells
because it can penetrate through the lipid component of the membranes and the carboxy-
fluorescein penetrates across breaks or gaps in the cell layer.
Generally, fluorescein staining of the corneal epithelium is representative of epi-
thelial damage, such as micropunctates or cell dropouts. In contrast, it is assumed if there
is no epithelial staining then there is no epithelial damage. Graham Wilson in 1995 re-
ported that if sodium fluorescein is topically applied, then fluorescein uptake can be ex-
pected even in a normal healthy epithelium. It does not require damage. The epithelium is
a significant barrier to small ions. Specifically, it is the superficial epithelial cell junctions
that creates the barrier. For small ions it has a resistance of 80,000 times less permeable
than the stroma. So molecules struggle to penetrate across the epithelial barrier and then
can easily diffuse through stroma before being restricted by another barrier at the endothe-
lium. The endothelial barrier is about 100-fold less restrictive than that of the epithelium.
The low levels of fluorescein penetration into a healthy cornea are too low to be visual-
ized with a blue filter on the slit-lamp. It is at these lower fluorescein concentrations that
the fluorophotometer's value is evident because of its enhanced sensitivity.
Let me return to the fluorophotometry technique that I use in my laboratory. Follow-
ing the test substance application, the corneal surface is bathed in fluorescein for five min-
utes by extending the lower lid, then rinsed with 50 ml of balanced salt solution. The
fluorescein penetrates the cornea and is measured as a peak on the fluorophotometer scan
of the tissue, Figure 3. The aqueous does not show an immediate increase in fluorescein. It
takes 45 to 60 minutes for the fluorescein in the aqueous to reach a maximum and fluo-
rescein loss rate will be equal to the very slow loss of corneal fluorescein across the endo-
thelial barrier into the aqueous.
There are two basic methods of using fluorophotometry in either an unanesthetized
patient or a rabbit. In the five minute bath technique, which is the technique I prefer in my
rabbit model, the corneal surface is bathed for 5 minutes with a solution of fluorescein; the
ocular surface is rinsed free of fluorescein; the corneal fluorescence is measured and fi-
1000".--------------------,
Tear/Cof'!'lea
E 100
01
.s
~
c: 10
~
II)
CIl
2
::::I
u:::
10 20 30 40
nally the permeability calculated in nanometers per second. In a patient the bath technique
could be considered uncomfortable.
There is a second technique that will conveniently provide a relative permeability. A
specific small volume of fluorescein is applied topically to the ocular surface; I use 20 fJ.I,
but it can be as low as 2 fJ.1. The patient is permitted to blink normally without rinsing the
fluorescein from the ocular surface. After 45 minutes the fluorescence concentration in the
cornea is measured. The concept is that by waiting 45 minutes the fluorescein is diluted
and flushed away from the ocular surface by the tears. The tear fluorescein concentration
is considered negligible and the corneal scan value will solely represent the fluorescein
that permeated across the epithelial cell layer. The experimental technique yields a rela-
tive measurement referred to as the F-45. It is not a permeability value. Unfortunately, it
does not take into account the individual's tear turnover rate. Each of the two basic tech-
niques are useful within their specific limitations.
There is a spread in the epithelial permeability data for a normal rabbit cornea after a
five minute bath with carboxyfluorescein (Figure 4). The spread is a skewed bell-shaped
curve with the skew appearing in the larger values. The reason for this is that the rabbit has a
tendency to have natural damage from tear drying in-between blinks and accidental touch
while preening with the fore paws. The epithelial permeability to carboxyfluorescein is about
0.06 nm/sec, while the permeability to sodium fluorescein is 12 times higher. Either one of the
fluorescein molecules can be used to measure epithelial permeability.
Let me present examples of utilizing fluorophotometry to measure the toxicity of
commercially available products through the use of the corneal epithelial permeability
technique. Preservative-free tear lubricating solutions are available in unit dose bottles
and presumably should cause no change in corneal epithelial permeability. The experi-
mental baseline permeability is measured following no test or control solution application.
The control solution is BSS Plus (Alcon Laboratories, Fort Worth, TX). The data for sev-
eral commercial tear lubricating solutions is graphed in Figure 5. AquaSite, a product sold
as a preservative-free solution, causes a statistically larger epithelial permeability as com-
pared to the baseline or control solution. This is because it has sorbic acid and EDT A as
solution stabilizers.
I
Figure 4. The distribution in nonnal rabbit corneal epithelial penneability after a five minute bath with carboxy-
fluorescein.
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 247
Figure 5. Rabbit corneal epithelial permeabilities slightly increased following a five minute exposure to several
commercial. preservative-free tear lubricating solutions.
In Figure 6, the data for preserved tear lubricating solutions is presented following a
5 minute exposure to the rabbit corneal surface. Generally the preservative is benzalk-
onium chloride. HypoTears, Murine and Visine cause increases in the corneal epithelial
permeability with peaks approximately 100 times or more greater than the baseline of 0.06
nm/sec.
In other experimental series, the tear lubricating solution was applied topically as
two drops every 30 minutes for six hours for one day or five days. This regimen approxi-
mated the clinical application for the products. Under these conditions the AquaSite prod-
uct causes epithelial permeability changes that are reduced to near control values.
CelluVis and HypoTears preservative-free are statistically equivalent to control. Even the
preserved tear lubricating solutions that cause a 100-foid increase in permeability when
~
:0
III
Q)
...
E
Q)
0..
Figure 6. Rabbit corneal epithelial permeabilities were significantly increased after a five minute exposure to sev-
eral commercial tear lubricating solutions containing preservatives.
248 H. F. Edelhauser and K. Green
exposed to the cornea for 5 minutes have great reductions in their toxicity to the corneal
epithelium when dosed as multiple drops. Yet, they are not equivalent to the BSS Plus
control. It is very important to notice the different impression one obtains by varying the
topical application regimen.
In conclusion, I hope I have presented to you the concept of using the fluoro-
photometer as an in vivo, non-invasive technique to test the toxicity, specifically corneal
epithelial permeability changes, caused by commercial products. The animal is unanes-
thetized and the normal ocular physiology is maintained. Also, after an appropriate delay,
the animal can be reentered into the experimental studies. This will lessen the number of
animals used for toxicity testing. Furthermore, the data clearly shows that the toxicity of a
product can be accentuated by a five minute exposure versus the clinical treatment dosing.
I believe fluorophotometry is a much more sensitive test than utilizing visual observations
of fluorescein staining of the tissue.
Edelhauser. This is a new test which I think has great possibility for evaluating topi-
cal medications with or without preservatives.
Hackett. Bernie, what is going on? In clinical practice you showed that there is a re-
turn of resistance to breakdown; how is it compensating?
McCarey. Return is not the right word, maybe I spoke too quickly. I have presented
two different exposures to the tissue. One was five minute continuous bath, while another
was a multiple drop application over 6 hours for 1 or 5 days. So I did not measure a re-
versibility phenomenon. That could be tested since the study can be done noninvasively
but that would be another experimental series.
Green. We have been using epithelial permeability measurements over at least the
last decade and a half to assess solution, and other test agent, effects on the cornea. This
approach has proven invaluable since it provides a quantitative measure of what is occur-
ring at the membrane level. We use an application of topical 2% sodium fluorescein in a
10 III drop that is washed off after 5 minutes with excess saline (about 5 ml) and measure-
ments made at 45 or 60 minutes after initial dye application. This is more sensitive than
any Draize assessment of the cornea and was published initially as a comparison of Draize
versus corneal epithelial fluorescein permeability after assessment of cetylpyridinium
chloride toxicity on the eye, either alone or in combination with toner, in 1985 (J Tox Cut
Ocular Tox 1985;4: 13-26).
As an adjunct to Dr. McCarey's comments upon the multiple uses of fluorescein as a
tracer in the eye, there are two other methods that have been described and validated al-
though they do not necessarily apply to the cornea per se. These application methods are:
iontophoresis and direct injection into the vitreous of fluorescein-labeled dextrans (FITC-
dextran). The former method uses a small electric current to drive fluorescein (present as
the only ion) out of an agar gel-filled electrode into ocular tissues. The FITC-dextran can
be directly injected into the vitreous and provides a source of tracer for days or weeks in
order that drug effects on aqueous humor flow rate can be determined.
Rodeheaver. If there is a disruption of the mucin or the tear film does that change
any of your uptake at all?
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 249
McCarey. That is a good question and it worries me a lot because inherently in the
test method there are several aggressive rinses of the ocular surface in order to remove the
fluorescein from the tear film. I do not know how to avoid this problem.
Edelhauser. The next thing we have on the schedule is contact lenses which will be
presented by Dr. Rodeheaver, Dr. Jessee and Dr. McKee. I think fluorescein permeability
is a unique test that might be of great value once we start to look at contact lenses.
Rodeheaver. I would like to make a few comments about some of the in vitro testing
and some difficulties that we encounter with contact lenses. Dr. Jessee and Dr. McKee will
also have some insights on that. I will then concentrate on the screening tests that we rou-
tinely use for some of our compounds.
With the contact lens we are not only testing the solution but we also have to test the
interaction with the contact lens and what that means for the eye, the burning, irritation or
cytotoxicity potential. To date we have looked at some different in vitro tests that try to
incorporate the contact lens. We have had some limited success with those but we have
not found one that is applicable to all the different chemical classes that we routinely look
at.
I think we may find some of the in vitro tests are applicable to specific chemical
classes. For instance, one test may work with surfactants but it may not work with some-
thing that is a straight disinfectant or a cleaning agent. We have various models that we
utilize, including our routine model. We have some other ones that we use once we nar-
row down the number of compounds that we look at in more detail and allow us to go into
detail concerning the test chemicals a little bit more. Dr. Yao has some rabbit corneal epi-
thelial and lens epithelial cell models that allows us some different test possibilities.
One of our first screens is to look at kidney epithelial cells. We have compared these
with some immortalized human corneal epithelial cell lines and have not seen a difference
in response. The canine cells allow us to have many more cells in a lot shorter time and
since we routinely have as many as 12 compounds that we need to screen at one time, it
affords us a little more opportunity to get through it quickly. For the modified neutral red
procedure we do not see a difference so we use the canine cells for the first screen. When
a chemical passes that test and then is tested in rabbits, if we find it to be acceptable then
it is acceptable in the clinic.
We routinely look at lens uptake, because if we have our use range down low and
yet we have a great deal of uptake this could put us, for some compounds, into a cytotoxic
range in vivo. Comparing to some of our standards that we know are acceptable we would
find some not to be acceptable and we would not continue any further with that product.
Another compound may show very little uptake and with low use range then we would go
ahead and test, particularly if that compound shows good efficacy in some of our pre-
clinical models. In some cases we can tell immediately with this test that we would not be
using a specific chemical. We found that some compounds gave some very acute toxi·city.
In most cases we have a predictive model in vitro so that we can cull out of some of the
more toxic and more irritable compounds. We found that with some compounds the use
concentration is very low, the uptake is very minimal and it has been acceptable in the
clinic. Other compounds have very low use range, low uptake and we predicted that we
250 H. F. Edelhauser and K. Green
would not see any adverse effects and yet when we went to the rabbit we saw some cor-
neal toxicity; not irritation but specifically corneal toxicity.
We have to pay close attention to the in vitro results and also be very alert in the in
vivo models to make sure that we pick up any adverse reactions. We are looking at some
of the cytokines to see if there is any correlation between irritation and corneal toxicity.
We are also looking at some epithelial resistance models. Dr. Yao has one in-house that
we have put some of our contact lens solutions through. We have had some success and
we hope to expand on that test in the future. We routinely use in vitro tests to screen out a
large number of compounds and then proceed with some more detailed in vitro and some
in vivo tests.
Jessee. I have to agree with Dr. Rodeheaver that looking at lens uptake and accumu-
lation of cytotoxic residue is probably the most telling way of determining whether you
are going to have an acceptable contact lens product. This conclusion is based on my ex-
perience with the major disinfectant active ingredients on the market today as well as my
new product R&D work. We do find, however, that the more in vitro assays that we de-
velop to screen out acute ocular irritation and inflammation, the more often we experience
failure due to more subtle mechanisms of toxicity.
We have tested formulations that presumably cause minor transient corneal edema.
People see halos when they look at bright lights, but the physiological effects are not suf-
ficient enough to measure by laser pachymetry in humans or in animals with extended
testing with elevated doses. We have tested formulations that directly stimulate the cor-
neal nerve endings and cause excruciating pain but at the same time cause no overt signs
of human or animal toxicity by fluorescein slit-lamp analysis.
I should mention that we have done some work with Dr. McCarey's fluoro-
photometry model. I'll let him describe his work, but I will say that everything that we
have had fail in the clinic due to minor human corneal staining or irritation has not been
detected in a standard 21 day FDA-mandated test. The fluorophotometry method is more
revealing simply because the method of staining and detection is far more sensitive. So
while we have good success explaining clinical failures, we can also use our experience to
increase our odds of successful product development.
I want to show you some of the tests that we have been using for measuring uptake
and accumulation of cytotoxic lens residues, which is a way that one can more readily fail
in making a contact lens product. Accumulation of cytotoxic residues from your product
onto the lens, with subsequent prolonged ocular exposure by slow release or even direct
contact and no release, can cause ocular toxicity. The first example I will show you is ben-
zalkonium chloride (BAK). This is typified by the worse case example ofthe classic inter-
action between BAK and high water content contact lenses. This is illustrated using
Etafilcon A polymer, which is a lens material used in a major selling brand of contact
lens. The lens has a significant negative charge, has a carboxyl functional group and a
high water content. Etafilcon A therefore tends to uptake positively charged compounds,
such as tear lysozyme and BAK.
We use a very simple test where we take a mouse suspension of L929 cells, lay a
small drop in the dish of a treated contact lens, and allow exposure of the cells to the lens
surface for a couple of hours. The cells are then stained with fluorescent probes for viabil-
ity or metabolic function and analyzed using an automated fluorescence microscopy im-
age analysis system so we can do hundreds of lenses at a time. We have validated this
model using several known problematic lens/solution combinations. We have used this
method to reject possible formulations and have had good success in the clinic. When we
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 2St
experience clinical failure, it's usually by mechanisms eliciting other than irritation after
only a month or two of clinical lens wear.
Our procedure tests lenses that have been cycled in 2.5 ml of solution overnight fol-
lowed by soaking in saline or in an artificial tear solution (ATS) mix to mimic possible
chemical release by lens wear. The artificial tear solution is a formulation published by
CibaVision that contains a myriad of proteins, mucin carbohydrates, salts, divalent cations
and lipids, etc. Lenses are soaked in a solution of I ppm BAK, which shows no cytotoxic-
ity by Agar Overlay or I day rabbit ocular irritation studies.
We found that if we cycle an Etafilcon A lens five times alternately in BAK, and
either saline or ATS then cell viability drops markedly after five cycles. If you extend
studies to 25 cycles, cell viability increases again. By doing protein analyses and chemical
uptake studies, what we are finding is the layer of protein, mucin and carbohydrate that
builds up on the lens after that extended period of cycling is sufficient to neutralize or
slough off most of the BAK. Unfortunately, this reversal of cytotoxicity may occur only
after the damage is done, so BAK has no predicted use with Etafilcon A lenses. This ex-
ample illustrates that within a short period of time you can measure chemical uptake by
lenses. If you allow rabbits to wear such BAK-treated lenses rather than soaking them in
artificial tears, you can then use the lenses to demonstrate that as conjunctival irritation
progresses to corneal irritation, increasing levels of BAK can be measured on the lens
through this assay and as well by chemical means.
Domiphen bromide, another quaternary ammonium compound like BAK but with
less irritation potential, shows the same pattern of cytotoxicity over 25 cycles with saline
release soaks, with the critical difference that artificial tear solution reverse soaks result in
complete reversal of cytotoxicity. Domiphen bromide may interact with the lens, or with
protein/mucin/carbohydrate bound on the lens, differently than does BAK.
The present major selling soft contact lenses brands are of the low-water, non-ionic and
high-water, ionic types. There are, however, a number of intermediate lens types classified as
Group II and III lenses. Using the in vitro procedure described, we have recently been testing
some recently marketed products using a spectrum of lens types across the range of FDA
groups I, 2, and 4. We can detect build-up of cytotoxic residues on many different Group 2
polymers, but we do not detect much difference between reversal of cytotoxic build-up with
cycling intervening by saline or artificial tear solution soaks. Only real market experience will
determine if minor transient irritation is seen with select lens polymers.
As a final note, at this and other meetings the FDA is often asked when they will accept
alternatives to animal eye tests. Japan requires conduct of an in vitro cytotoxicity test, called
the V79 colony assay, which is a far more sensitive test than USP methods currently in use. In
this method, plates are seeded with V79 cells to allow attachment of about 100 colony form-
ing units after a 24 hour incubation. The nascent colonies are then exposed to concentrated
extracts of lenses which have been treated using a lens care regimen for 30 consecutive cy-
cles. Any depression in colony growth below control cultures is interpreted as evidence of
possible biocompatibility problems, regardless of animal and human data presented. The
method shows many chemical disinfection and sterilization procedures used successfully for
years with contact lenses and other medical devices to be significantly cytotoxic.
The V79 Colony Assay was driven into the Japanese Pharmacopoeia, pharmaceuti-
cal and medical device guidelines with little international review. Japan is presently urg-
ing adoption of the assay by other worldwide standards promulgating organizations. The
lesson to be learned here is to be careful what you wish for, for you might end up getting
it. I agree with FDA's long-standing comments, that no method has been validated as an
acceptable alternative to animal eye tests, but at the same time I am concerned that inap-
252 H. F. Edelhauser and K. Green
propriate methods will be adopted simply due to political and public pressures to replace
animal testing.
Edelhauser. Dr. McKee, do you want to make any comments at this time?
Mowrey-McKee. I would go along with what Dr. Rodeheaver and Dr. Jessee have indi-
cated as far as in vitro testing for contact lens care products. In our experience at Ciba Vision
the cytotoxicity of cycled lenses is probably the most significant indicator most closely corre-
lated with a toxic response in an animal or human. As far as in vitro testing of the solutions by
themselves, a toxic result in those tests is not always indicative of the human response. How-
ever, if a solution tested is non-cytotoxic in the in vitro test according to our test methods, it is
non-irritating in the rabbit and in humans. I think you have to be careful about throwing out
compounds or formulations that are cytotoxic by these in vitro solution-type assays because
you could be throwing out a useful product. In fact, there are products on the market that
would look very poor in some of these but provide good patient comfort and lack of irritation.
Green. One must remember that contact lenses frequently bind excipients or active
ingredients from solutions, as pointed out by Drs. Rodeheaver and Jessee earlier, and in
many cases the release kinetics can be extremely slow. This means that one cycle of im-
mersion in uptake solution followed by wearing or immersion in a tear-substitute solution
may only allow uptake but no release whereas after several cycles the binding sites are
saturated. At that point one then sees a "sawtooth" uptake and release process wherein the
amounts are more or less equal in each portion of the cycle.
The degree of saturation of binding sites varies depending upon the nature of the
lens material and the physicochemical properties of the compound under examination.
Over the years I have seen a spectrum of lens-chemical interactions in in vitro lens stud-
ies, whether it be between BAK (Arch Ophthalmol 1990; 108:244-246), domiphen bro-
mide, chlorhexidine digluconate and many others with differing lens types. In some
instances the uptake is complete within one day and uptake and release then occurs from
both lens water and totally or partially from the constituent lens material. In other cases
uptake is not complete within 30 or 40 days and uptake continues to exceed release on a
daily basis so that total uptake continues to increase.
With chlorhexidine digluconate, for example, the uptake has been measured from
contact lens immersion volumes (= 2 ml) during 8 hour periods with washout in artificial
tears over 16 hours; this results in a sawtooth uptake by soft contact lenses (with uptake
exceeding release) that reaches a steady state only after many cycles. Even when im-
mersed on a continuous basis in an excess volume of uptake solution per lens, a maximal
uptake value is not reached for over two weeks.
Because our work in this area formed the basis of the industry standard for the inter-
action between chlorhexidine and soft contact lenses (J Pharm Pharmacol
1978;30:678-682), we were the focal point for many of these studies. Although I am not
at liberty to disclose any specifics, there was an entire spectrum of interactions between
lens types and chlorhexidine that reflected the composition of the lens material as well as
the amount of lens hydration. This was also true for BAK which we performed inde-
pendent of industry and published later.
Edelhauser. What do you think would happen if you go back to the standard FDA
protocol using three rabbits? Do you think that three rabbits is enough to show some po-
tential cytotoxicity in the Draize test?
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 253
Mowrey-McKee. I would do the test with six rabbits and I would not do the solution
tests where you use a 100 f..L\ drop. I would use lenses that are cycled and go through at
least one cycle prior to initially being placed in the rabbit eye. The rabbit would wear the
lens for 8 to 10 hours a day for 5 consecutive days. The lenses would be cycled daily us-
ing the care regimen. These are important deviations for testing contact lenses.
Edelhauser. Dr. McCarey is now going to share with us how you can evaluate con-
tact lenses and measure the epithelial permeability with carboxyfluorescein. It is an add-
on to in vitro tests.
McCarey. I would like to share with you my experiences of fitting rabbits with hy-
drophilic contact lenses. It is a required FDA test for new contact lenses and their solu-
tions, but it is not easy.
The first issue is that as a rabbit gets older, the body weight increases, as well as the
eye dimensions, such as corneal diameter increase and keratometry flattens. In order to fit
a rabbit with a contact lens it is necessary to match the rabbit body weight to the contact
lens dimensions. With a young rabbit, the fit of the contact lens will constantly change
with passing weeks. Eventually, the rabbit body weight will plateau. A 10 pound New
Zealand White rabbit is approximately at its maximum body weight. A six pound rabbit
kept for a month will be eight or nine pounds by the end of that month and the eye dimen-
sions will also change.
The second issue is an available source of contact lenses for the rabbits. Will you
have the lenses specifically made for the rabbits? If you are intending to use contact lenses
commercially available for the human eye, then it is necessary to have large body weight
rabbits in order to match the rabbit corneal curvature to human corneal curvature.
I have had success fitting rabbits with Gentle Touch Contact Lenses (PBH Inc.,
Sunnyvale, CA) and Permalens (Cooper Vision Laboratories, CA) with the base curve of
8.2 mm, diameter 14.5 mm and 65% water content. I like a large diameter lens because
there is a better opportunity of getting the eyelid on top of the edge of the lens which
helps lens retention. I do not remove the nictitating membrane since it can slide over the
contact lens and keep the lens hydrated. The FDA regulations say you may remove the
nictitating membrane but I prefer to retain it. The contact lens should be fit so that the
nictitating membrane can slide over the contact lens with each blink of the eyelids. If the
nictitating membrane slides under the lens, then the contact lens is doomed to not stay on
the eye.
I have not had equal success with other commercially available lenses. Each contact
lens material will have different material characteristics, which even with the same lens
dimensions will result incompletely different fitting success. It is amazing to realize FDA
requires all contact lens materials to be tested in the rabbit. I have tried many types of con-
tact lenses and cannot have success with them all. I only have success with certain contact
lens manufacturers.
The fluorophotometer can be used to assess the corneal effects of contact lens mate-
rials and their solutions. A typical rabbit test regime is as follows: the contact lens is worn
for eight hours each day for five days, with overnight care in ReNu (Bausch & Lomb,
Rochester, NY). The corneal epithelial permeability at the end of the contact lens wear re-
254 H. F. Edelhauser and K. Green
gime with +1.0 diopter lenses was 0.037±0.003 nrn/sec (n=36, SEM) which is equivalent
to non-contact lens control eyes (0.046±0.007 nrn/sec, n=12, SEM). Next, I selected to use
a 6 diopter lens because I thought the lens center thickness would be excessive and physi-
ologically challenge the eye. Unfortunately, the Gentle Touch contact lens has a 4 or 5
mm optical zone which minimizes the central lens thickness. The +6.0 diopter lenses were
worn for five days and resulted in insignificant increases in epithelial permeability
(0.041±0.009 nm/sec, n=12, SEM). When I fit the 65% water daily-wear lenses as a +6.0
diopter continuous wear lens, i.e., continuously for 24 hours a day for a week, then the
epithelial permeability was not statistically different from control eyes. Thus, even with an
excessive wear situation, the lens did not destroy the corneal epithelium. The +6.0 diopter
lens caused the same change in epithelial permeability as + 1.0 diopter lenses.
In conclusion, fitting a rabbit with a contact lens is not easy and likely impossible
with all contact lens types. Secondly, noninvasive fluorophotometry of corneal epithelial
permeability may enhance our understanding of the effects of contact lens on corneal
physiology.
EdelhauseI: This brings up an interesting issue; if you are testing contact lenses,
what laboratories do you use? Do you do this in-house or do you send it out to a contract
laboratory? Obviously, experience helps as you fit an animal with a contact lens.
McCarey. There is another issue that I forgot to mention; the rabbit's adaptation to
the contact lens. If you were to do a clinical study with a particular contact lens, then you
would want that patient to be adapted to wearing a contact lens. In most cases when we
are asked to do a contact lens wear test in the rabbit model, there is no adaptation period.
The investigator simply starts the experiment. In my experience, the first three days are
the toughest fitting success days with the rabbit. During the first few days the rabbit eye is
often irritated from the contact lens. After the third day, the rabbit appears to adapt to
wearing the contact lens. I think consideration of adaptation should be part of the test but
it is not at the present time.
Edelhauser. Any comments from anyone who has had experience with this?
Rodeheaver. Dr. McCarey, I think misery loves company. We find that there are
some lenses that just will not stay on the eye but for most of the lens groups we do get bet-
ter than 80 or 90% retention so if we can help you in that respect we will be happy to.
Rodeheaver. We check the lenses every two hours to see if they are in. If they are in
then we count that as a two hour wear, if they are out we have no wear for that period of
time. Then we add up lens presence over the eight hour period of time and create a per-
centage. The question I had about your model is do you have a feel at which point you be-
gin to see an effect that might be clinically unacceptable? Can you draw a line or is it not
that clear yet?
McCarey. Do you mean did I see a unacceptable contact lens wear and then related
that to epithelial permeability changes?
McCarey. A poorly fit lens may cause large epithelial permeability changes but it is
difficult to have sufficiently good lens retention to test the lens effect. All these rabbits in
this series looked clinically quiet. Every day the conjunctival redness was minimal and the
eyes looked good. I had good retention of the contact lenses. A rabbit might lose its lens
every other day. I am very concerned about data derived from animals that are frequently
refit because of lost lenses. I am not sure if we are testing the solution that the lens is be-
ing bathed in, or if we are testing the manipulation of the animal during refitting, or if we
are testing the lag time when you did not have a contact lens in the eye. I think it is really
tough to sometimes interpret what happens if there is frequent refitting of new lenses.
Rodeheaver. That is true. When we have to refit the lens several times during the day
then our scores do go up.
Jessee. I might have one comment. Several ocular toxicology experts this morning
expressed great concern about the inadequacies of the Draize test which in essence tests
the toxicity of a single application of a test substance. The three week contact lens rabbit
test also uses the same Draize eye irritation test, except that it is repeated for 22 consecu-
tive days. Given that it suffers from the same faults as the Draize eye test, I am surprised
that this test is not receiving the same criticism.
We have found that if you boost the level of active ingredients in some marketed
contact lens products, those that have been used with great consumer and professional eye
care specialist satisfaction for many years, by as little as 3-fold, 20% of the people in a hu-
man clinical study will drop out due to ocular complications in the first week of lens wear.
The concept of defining a margin of safety by Draize eye irritation methods may not nec-
essarily be as applicable for products that are designed for use in the eye than it is for acci-
dental splash exposure risk assessment.
Edelhauser. Next we have on the schedule FDA requirements. Dr. Avalos, do you
have a few comments you would like to make?
Avalos. Actually the more important comment I have is first to thank Dr. Tripathi,
Dr. Norton and Dr. Hackett for making my job a lot easier. What I want to do is follow-up
on what Dr. Hackett said. As we heard earlier the advantages and disadvantages were well
presented for a lot of in vitro alternatives.
Since the FDA's mission is to assure that there are safe and effective drug products
for humans, the most relevant data to our mission is in vivo data. At this point we are not
quite ready to accept in vitro alternatives. The more we keep doing these studies the more
information we will have, so that one day, upon re-evaluation, we will accept them. The
main drawbacks are, I) there is no systemic toxicity provided in the in vitro alternatives,
and 2) there is no relationship between the assay and humans. Still, the in vitro assays may
provide valuable information in many different aspects.
Potential changes are under discussion at the FDA. When we look at some of the
special toxicity studies like the dermal irritation assay, the assay design will probably be
changed in the near future. The number of animals will probably drop from six to three.
Another change is the carcinogenicity study. Transgenic animal models are being consid-
ered in lieu of the 2-year carcinogenicity study. We understand the importance of time to
256 H. F. Edelhauser and K. Green
market and transgenic animal models will decrease that time. Concurrently, we are willing
to accept some transgenic animal models, the TAGC model, for instance.
Lastly, a general time frame for drug development. In order to go into the human phase
of drug development we generally ask for both ocular and dermal irritation studies during
phase I of the IND. Genotoxicity and subchronic studies are also recommended in order to
support safety in the Phase I study. The subchronic studies, again, are going to depend on the
indication on the intended length of use in the clinical study. During phase II clinical trials,
what we usually ask for are some additional toxicity and reproductive studies. Because you
are going to use females in the clinical studies we would like to see the reproductive studies or
the segment II study prior to initiation. In some cases we do make exceptions when the con-
sent form contains a statement advising the female volunteer to use some form of contracep-
tion. Finally, the chronic studies as well as the carcinogenicity studies, are usually submitted
during the Phase III. Well into the development program of your drug product.
McCarey. Wound healing can be used to assess the effects of drugs on the rate of
corneal epithelial wound closure. Roswell Pfister published a paper in 1975 on epithelial
wound healing that contained an extensive series of scanning electron micrographs. He
found that 15 hours after producing a wound, a single layer of epithelial cells was ob-
served' to have moved across the basal lamina or Bowman's layer, depending on which
animal species was under study. The epithelial cells had ruffled edges and filopodia reach-
ing outward. He noticed that the filopodia extended in the direction of movement and ap-
peared to have grabbed onto the surface. We now know fibronectin is a part of this
process. The cell moves forward because of cell contractile proteins within the cell. Other
cells on top of the leading edge cell continue the process of expansion and sliding into the
lesion.
Crosson et al in 1986 performed key research in defining the rate of corneal epi-
thelial wound healing. It is important to interpret the data relative to the surface over
which the cells are migrating. If the epithelial cells are removed by scraping leaving the
basement membrane intact, then the migration is more rapid than across stroma following
a keratectomy. They also observed that cells migrate at the same rate regardless of the size
of the wound. This is an important point because it makes the study a lot easier to per-
form. Wounds of2, 4, or 6 millimeter diameter were made and it was observed that for the
first five or six hours there is no wound closure. Afterward there was a 60 /lm per hour
rate of wound closure regardless of initial wound diameter. The analysis is performed by
digitizing the wound area, converting it into a circle and plotting the radius of the wound
versus post-surgery time. The radius in mm per hour has a linear relationship with time.
Workshop on in Vitro versus in Vivo Models for Ocular ToxIcity Testing 257
E 150
E
2.. 120
¥J
II: 90
e
::J
<II
0 60
0
"0
c: 30
::J
::
0
0 ..:
c: c: Q)
e
Q)
.5 :§ Q)
Qi
l:l ~0
Jl!0 .s::
Q)
>
>
~
.c > > (5
>
tj > ii
Figure 7. Corneal epithelial wound closure rate was unaffected by Voltaren treatment.
258 H. F. Edelhauser and K. Green
~c:
~ 0.90
.,
1.20,------------------,
,,
,,
,,
:is ,,
Ol ,,
E
Q)
0.60 ,
,,
Q)
D.. ,,
,,
~ ,,
Qi 0.30 ,
-=
'is.
w ;~~~~~~~~ "".
0.000
10 20 30 40 50
Post Wound Closure Time (hr.)
Figure 8. The corneal epithelial permeability of rabbits treated with Voltaren was measured non-invasively soon
after wound closure and 24 hours later. • Voltaren; 0 Voltaren vehicle.
Cheng Yao, Ph.D., Alcon Laboratories, Ft. Worth, Texas. As you mentioned for
wound healing you know there are two movements. One is a horizontal movement, one is
a vertical movement. In a way I do not see how you are going to quantify the horizontal
plus vertical. Would it be easier to just do the in vitro cell culture and look at one layer
quantification of the artificial wound?
Yao. So what you are saying is you only see one layer?
McCarey. I am only saying I am seeing closure. I do not even know how many cells
are on that layer.
McCarey. Depends on what you are quantifying. If you are quantifying closure, yes.
If you are quantifying how many cells is it thick when it closes, no, I do not have any idea.
McCarey. No, you cannot determine the epithelial thickness with a slit-lamp. A full
thickness epithelial layer is 35 to 50 f..lm thick in the rabbit and can only be identified as a
gray line with the slit-lamp. You could potentially measure epithelial cell layer thickness
with a specular microscope and a very, very narrow slit.
A more interesting question is: Do I know if the epithelial permeability can return to
normal if there is only a mono cell layer? From the literature I would estimate that it takes
15 or 24 hours before tight junctions start to form. Even though the wound is closed it
does not mean the epithelial barrier is normal. Even 24 hours after wound closure, our
Workshop on in Vitro versus in Vivo Models for Ocular Toxicity Testing 259
1.20 , - - - - - - - - - - - - - - - - - - ,
~c:
~ 0.90
:0
«l
Q)
E 0.60
Q)
D..
]!
CD 0.30
-=
'0.
w
o
0.000
10 20 30 40 50
Post Wound Closure Time (hr.)
Figure 9. The corneal epithelial permeability of the control eyes from Figure 8 was measured non-invasively soon
after wound closure and 24 hours later. 0 individuals; • Voltaren; 0 Voltaren vehicle.
data indicates the epithelial permeability was several fold greater than baseline. It must
take greater than 24 hours for it to become a normal epithelial barrier. I did not try to
measure how long it takes to return to normal.
Green. The correlation that I have found between permeability and wound closure is
that a minimum of30 to 36 hours is needed for a 6 mm trephined-then-mechanically scraped
lesion to reach normal, background permeability. There is not a full thickness epithelium at
this time, only two or three cell layers but this is evidently sufficient to establish the full bar-
rier function. A slightly longer time is required for n-heptanol generated lesions presumably
due to some residual chemical interference with initial re-epithelialization.
It is important to note that the measurement ofre-epithelization in vivo requires the identifi-
cation of specific markers in order that successive photographs of the lesion can be accurately
compared. I use the limbus in the horizontal meridian as the reference points and lesion size is de-
termined relative to this value for each eye. Normalization allows direct comparison either be-
tween eyes or between the same eye at different times from creation of the lesion.
faa. Do you create a wound right in the center or do you have a certain location?
Because the stem cell is on the limbus and they will move at different rates if they are not
in the center and it makes the quantification even harder.
McCarey. You bring up an interesting point. I did make them in the center; I made
different shaped wounds and I thought if I had a rectangle I have a completely different
rate of closure than if I had a circle. Not so, it comes out the same. I did not try off-center
and so forth. After seeing how the rectangle and circle responded I said I am not going to
go further on this. But the issue of migration when you calculate like this, I think neutral-
izes the exact shape of the wound. I do not know if it neutralizes if it is closer to the pe-
riphery or center.
Edelhauser. I would like to thank the speakers and thank the audience. I hope that
this workshop has been beneficial to each of your interests and that maybe you can take a
few of the discussed methods for ophthalmic drug testing home with you. Thank you all
for your participation.
24
1. INTRODUCTION
* This statement is an informal communication under 21 CFR 10.90 (b)(9) that represents the best judgment of the
Division of Dermatologic and Dental Drug Products at this time. This document does not necessarily represent
the formal position of the Center for Drug Evaluation and Research or the Food and Drug Administration, and
does not bind or otherwise obligate the Center or Agency to the views expressed.
The focus of this chapter will be ocular drug products. Regulations that address the
safety profile, commonly submitted assays that are part of the safety profile, and potential
alternatives for evaluating ocular irritation for an ocular drug product to be marketed will
be discussed. Additionally, the current role and advantages of these alternatives in the
drug development process will be emphasized.
2. REGULATORY CONSIDERATIONS
temic exposure of ophthalmic drug products. In order to initiate the clinical studies, toxic-
ity studies via the intended route that are of equal or longer duration than the proposed
clinical study are recommended. Reproductive toxicity studies are also recommended for
drug products intended for intermediate-term or chronic use. Chronic non-clinical studies
would include the one year in duration or lifetime carcinogenicity studies. However, sys-
temic absorption of the drug products would be critical in determining the need for the
carcinogenicity studies. The non-clinical studies should include both genders and two spe-
cies (one rodent and non-rodent). In addition, the studies should be conducted in accord-
ance with current Good Laboratory Practices (21 CFR 58)2. The Sponsor is encouraged to
contact the FDA Division as early in the development of the drug product as possible to
obtain guidance and comment on protocols.
As part of the safety profile, ocular and dermal drug products, as well as ocular de-
vices, are evaluated for their potential to cause ocular lesions. A current method for deter-
mining ocular irritation is the modified Draize assay (16 CFR 1500.42)3. This protocol is
very similar to the protocol proposed by Draize et at. (1944t The main differences in the
protocol proposed by the CPSC in 1979 and the original protocol were a decrease in rab-
bits (9 to 6), limited observation periods, inclusion of a rinsing procedure, and modifica-
tion of the scoring system. More importantly, a provision for testing solid materials was
included, as well as a suggestion for the use of a slit lamp, binocular loupe, and fluo-
rescein dye, to facilitate the detection of corneal injury. The original use of this protocol
was to determine if accidental exposure to agents could induce injuries to the eye.
assays and the advances made in biotechnology may provide useful information in the
ocular safety assessment process. Living cells, tissues, or recently enucleated eyes are be-
ing employed in many of these assays. Although several assays have been proposed as
substitutes for the Draize ocular assay, none of the alternatives may be capable of rep lac-
ing the in vivo assay due to the complexity, structurally and functionally, of the eye. Per-
haps the assays would be of greater value as complements to the in vivo assays or assist in
defining a product's potential biological response. In addition, the assays could provide
methodologies for addressing the potential biotransformation end-products or charac-
terizing a mechanism of action for the product. The use of these alternatives as prescreens
or part of tier testing systems has been proposed by several investigators5-9.
In the approach by Hurley et al. (1993), in vitro alternatives were described as part
of a tier testing system for assessing eye irritationS. The tier testing system is illustrated in
Figure 1. The authors, members of the Interagency Regulatory Alternatives Group
(IRAG), proposed as their final report a stepwise progression of screens so that little or no
animal testing would be required for agents meeting the criteria identified in the report.
The first parameter proposed as a screen was measurement of the pH of the test agent.
Murphy et al. (1982) reported the effects of pH on ocular irritation 10. Acids with a pH be-
low 2.5 and alkalies with a pH above 11.5 produce severe damage to the corneal stroma of
rabbit eyes that had been denuded of the corneal epithelium. However, preliminary infor-
mation indicates that the total alkalinity of the test agents may be a better predictor of irri-
tation than pH 'o." . Structure-activity relationship may also be used as a screen.
Comparison of the structural moieties of the test agent to known irritant structural moie-
ties would provide insights into the agent's potential to cause ocular irritation. Similarly,
the requirements for ocular testing may be waived for reformulations which contain
known irritants.
The third step in this tier testing system was the use of in vitro tests. In the following
section, some of the advantages and disadvantages are discussed. As part of the proposed
tier testing system, a battery of these assays may assist in identifying those compounds
that are likely to be severe irritants to the eye. Severe irritants or corrosive agents to the
Drug Product
PhYSicochemical
Properties
(e.g., pH)
SAR
In Vitro
Observations Figure 1. Tier testing system recommended for the as-
sessment of ocular irritation of drug products. Historical
ocular irritation data of drug products and ingredients are
Acute Dermal TOxicity
first obtained. The physicochemical properties of active
Findings
ingredient and other ingredients of the drug product are
determined. If necessary, the material is then evaluated
in a battery of in vitro assays. After the in vitro assays,
either the acute dermal toxicity or primary dermal irrita-
tion assays could be conducted. Then, the modified
Draize assay could be conducted, if necessary. (Modified
Draize Modified Assay
from Hurley et a/,R)
Toxicity Testing for Ocular Drug Products 265
skin are also most often irritating to the eye. The use of rabbits in the primary dermal irri-
tation assay or a dermal toxicity study would provide valuable information in assessing
tbe potential ocular irritation of a drug product. Additionally, these in vivo assays would
provide systemic toxicity data for review in the drug development process. As part of the
tier testing approach, dermal irritation (PDII > 5) or acute toxicity (LD50 <200 mg/kg)
would preclude any ocular irritation testing.
If after using this approach an ocular irritation assay is deemed necessary, a couple
of options for the ocular irritation assay are available. The current modified Draize as de-
scribed in (3.1.) could be employed. A second alternative would be the use of a more re-
cently modified version of the Draize assay. This later method would use a similar method
of exposure, observation periods, and scoring system as described earlier. However, the
number of animals treated with test material would be reduced to three animals. Because
of the lower number of animals, the interpretation of the assay will also differ.
In the report of the IRAG symposia, a reduction in animals used in the ocular irri-
tancy assay was proposed l2 . Two options were given. Both options would decrease the
number of animals originally proposed by CPSC from six to three. In the first option, two
animals would be tested. If positive findings were reported, then no further testing would
be necessary. If a negative response was obtained, a third animal would be treated with
test material. In the second option, three animals would be used. If two of three animals
had a positive response, the material would be considered an irritant. With the first option,
an accuracy of 97% was given with low false positive «3%) and false negative (0%) re-
sponses l2 • For the second option, an accuracy of 94% was noted with a very low false
negative rate of 1% and an acceptable false positive rate of <5% 12.
Another option would be to treat one animal first. If severe reactions are noted, no
further testing would be recommended. However, if mild or no response is reported, then
two more animals are treated and the findings noted. Using this method, two of three ani-
mals would need to have positive findings for designation as an irritant.
Additionally, the dose employed would be 0.01 ml of a liquid or O.OI-ml equivalent
weight dosage of a solid if the material is suspected to be a severe ocular irritant. If not,
then the dose would also be similar as described in (3.1). This low-dose volume was also
proposed at the IRAG symposia \3.
ess and, thus, would be expected to provide a better estimate of the irritant potential of
chemical agents. Even though the identical end-point can be determined in either the cell
culture or organ test systems, the evaluation of chemical agents would be less expensive
and more versatile with the cell culture system.
Use of end-points like cytotoxicity (membrane integrity or cell adhesion), cell func-
tion (cell metabolism), or cell growth with the cell culture systems would provide a less
expensive, but reproducible, robust, and adaptable model when compared to the cytokine
release assay. Even though these assays do not directly assess the irritant reaction, valu-
able information can be attained in regards to the possible mechanism of action of the test
substance. In addition, a prioritization scheme for further analysis can be generated from
these readily adaptable assays. Ranking of chemical-induced in vitro cytotoxicity, cytok-
ine release, and cellular function have correlated well with historical in vivo ocular irri-
tancy data for certain chemical classes l 4-l7. The correlation of the end-points was
dependent on the concentration of test material, duration of exposure, and class of com-
pound. However, the studies have not been extensive, and more data are required before
this correlation can be accepted, especially for different classes of chemicals.
Tissue organ end-points like alteration in smooth muscle contractility of isolated
rabbit ileum or evaluation of electrical conductivity in rat skin slices provide a more direct
evaluation of tissue rather than isolated cells. However, such tissues have little relation-
ship to ocular toxicity. The use of enucleated eyes for evaluating changes in corneal opac-
ity or permeability would be a more direct measure of ocular toxicity.
5. SUMMARY
As part of the safety assessment, the Food and Drug Administration evaluates der-
mal and ocular drug products for their potential ocular toxicity. Several factors (pH, struc-
ture-activity relationships, in vitro alternatives, primary dermal irritation index, and acute
dermal toxicity) should be considered for the drug product prior to conducting an ocular
irritation assay. The drug product should not be tested further if the drug product has a pH
of 2.5 or below or 11.5 or above. The assay should also not be conducted if the active in-
gredient(s) have analogous chemically active moieties that are known irritants or if the re-
formulation contains known irritants. Additionally, the assay should not be conducted if
the drug product has an acute dermal toxicity lethality at or below 200 mg/kg or a primary
dermal irritation index score of 5 or above. The findings from in vitro systems like Cor-
rositex™ with ocular and dermal drug products should also be considered prior to per-
forming an irritation assay. Strongly positive findings in these assays could preclude the
need for a modified Draize assay. Other alternatives could be used to complement the in
vivo studies, assist in defining a mechanism of action, or identify possible biotransforma-
tion metabolites of the drug products. When animal testing is to be conducted, a low vol-
ume dosing (0.0 I ml or g) in one animal may be performed for suspected severe irritants.
If the product is not classified as an irritant or moderate irritant, then 0.1 ml or 0.1 g of
material should be instilled in two or three more rabbits. The responses are scored accord-
ing to the modified Draize scoring system at 24, 48, and 72 hours, and, if necessary, at 7
days to assess the potential recovery from the insult. In summary, the Agency encourages
the use of alternative tests prior to conducting a modified Draize assay for ocular irrita-
tion. Additionally, the Agency encourages the Sponsor to contact the Agency as early as
possible for guidance in the development of the drug product. Such interaction could con-
serve valuable resources and potentially reduce the time to market for the drug product.
6. ACKNOWLEDGMENT
JA would like to thank Dr. Wiley Chambers for his support and critical review of
the manuscript.
7. REFERENCES
I. Code of Federal Regulations Vol 21 parts 312 and 314.
2. Code of Federal Regulations Vol 21 part 58.
3. Codc of Federal Regulations Vol 16 part 1500.42.
4. Draize. JH. Woodard. G. Calvery, HO. Method for the study of irritation and toxicity of substances applied
topically to the skin and mucous membranes. J Pharm Exp Ther 1944;82:377-390.
5. Frazier. JM. Gad. Sc, Goldberg, AM, McCulley, JP. A critical evaluation of alternatives to acute ocular irri-
tation testing. In: Goldberg AM. ed, Alternative Methods in Toxicology Vol 4, 1987. Mary Ann Liebert,
New York.
6. Wilcox, OK, Bruner, LH. AT LA 1990; 18: 117.
7. Bruner, LH, Shadduck, J, Essex-Sorlie, D. Alternative methods for assessing the effects of chemicals. In:
Hobson OW, ed. Dermal and Ocular Toxicology: Fundamentals and Methods, 585-606, 1991. CRC Press,
Inc., Boca Raton, FL.
8. Hurley, PM, Chambers, WA, Green, S, Gupta, KC, Hill, RN, et al. Screening procedures for eye irritation.
Food Chern Toxicol 1993;31 :87-94.
268 J. Avalos et al.
9. Chan, P., Hayes, AW. Acute toxicity and eye irritancy. In: Hayes AW, ed. Principles and Methods of Toxi-
cology 3rd edition, 579-648, 1996. Raven Press, New York.
10. Murphy, JC, Osterberg, RE, Seabaugh, VM, Bierbower, BW. Ocular irritancy responses to various pHs of
acids and bases with and without irrigation. Toxicology 1982;23 :281-291.
II. Booman, KA, DeProsp, J, Demetrulias, J, et al. The SDA alternatives program: comparison of in vitro data
with Draize test data. J Tox Cut Ocular Toxicol 1988;8:35-49.
12. Springer, JA, Chambers, WA, Green, S, Gupta, KC, Hill, RN, et al. Number of animals for sequential test-
ing. Food Chern ToxicoI1993;31:105-109.
13. Lambert, LA, Chambers, WA, Green, S, Gupta, KC, Hill, RN, et al. The use of low-volume dosing in the
eye irritation test. Food Chern ToxicoI1993;31:99-103.
14. Borenfreund, E, Borrero, O. In vitro cytotoxicity assays: potential alternatives to the Draize ocular irritancy
test. Cell Bioi ToxicoI1984;1:33-39.
15. Borenfreund, E, Shopsis, C. Toxicity monitored with a correlated set of cell culture assays. Xenobiotica
1985;15:705-712.
16. Demetrulias, J, North-Root, H. Prediction of the eye irritation potential for surfactant-based household
cleaning products using the SIRC cell toxicity test. In: Goldberg AM, ed, Alternative Methods in Toxicol-
ogy Vol. 6, 1988. Mary Ann Liebert, New York.
17. Shadduck. J, Everitt, J, Bay, PHS. Use of in vitro cytotoxicity to rank ocular irritation of six surfactants. In:
Goldberg AM, ed, Alternative Methods in Toxicology Vol 3, 641-649, 1985. Mary Ann Liebert, New York.
PARTICIPANTS
269
270 Participants
273
274 Index
Thioltransferase, recovery from lens oxidative insult, Vascular hemodynamics. alterations in disease and in-
42 jury, 139
Time-lapse photography Vascular endothelial growth factor, 144
corneal epithelial cells, 221 Vascular penneability, 140
trabecular cells, 82 Vertebrate retina, 107, 113,240
Tissue culture, 63, 81,169,181,193,221,227 Virus transfonned cells
Toxicity corneal epithelium, 181
iris color changes, 203 lens epithelium, 169
ocular viscoelastics, 103 Viscoelastics, testing in rabbit vitreous, 103
retinal implant, 113 Visual evoked response, 99, 113
Trabecular meshwork, barotoxic effects, 83 Voltaren
Transfonned cells, virus, 169, 181 corneal epithelial penneability, 256
corneal reepithelialization, 256
U-V radiation, 63 Volume added, Draize test, 216
corneal epithelial cells, 63
interaction with cataract models, 239 Water insoluble proteins, lens, 27
lens, 239 Water soluble proteins, lens, 27
protective filters, 63 White cell counts
Uveitis, 93, 138 rabbit vitreous, 103
nitric oxide and Iipoxygenase, 121, 151 viscoelastics, 103