Journal of Neuropathology and Experimental Neurology

Copyright q 2003 by the American Association of Neuropathologists

Vol. 62, No. 5

May, 2003
pp. 441 450

Nitrosative and Oxidative Injury to Premyelinating Oligodendrocytes in

Periventricular Leukomalacia
ROBIN L. HAYNES, PHD, REBECCA D. FOLKERTH, MD, RACHAEL J. KEEFE, BA, IYUE SUNG, PHD,
LUKE I. SWZEDA, PHD, PAUL A. ROSENBERG, MD, JOSEPH J. VOLPE, MD, AND HANNAH C. KINNEY, MD

Key Words:
cursors.

Astrocytes; 4-hydroxynonenal-protein adducts; Microglia; Nitric oxide; Nitrotyrosine; Oligodendrocyte pre-

INTRODUCTION
In the United States alone, approximately 55,000 live
newborns are born very prematurely (,1500 gram birth
weight) each year (1). Because of modern intensive care,
nearly 90% of these infants survive. Approximately 10%,
however, develop cerebral palsy (motor deficits) and
25%50% develop cognitive and behavioral deficits (2).
Thus, nearly 5,000 cases of cerebral palsy and 10,000 to
20,000 children with serious learning disabilities result
each year. Periventricular leukomalacia (PVL) is the most
common substrate of this enormous amount of neurologic
disability. The pathogenesis of PVL likely involves multiple factors, particularly ischemia/reperfusion in the critically ill premature infant with impaired regulation of cerebral blood flow (3). Cytokine-induced brain injury
associated with maternal or fetal infection may also play
an important contributory role (3).
Our over-riding hypothesis is that PVL is due to cerebral ischemia with the production of excessive reactive
nitrogen and/or oxygen species in the reperfused tissue,
From the Departments of Neurology (RLH, RJK, PAR, JJV, HCK)
and Neuropathology (RDF, HCK), Childrens Hospital and Harvard
Medical School, Boston, Massachusetts; New England Research Institutes (IS), Watertown, Massachusetts; Department of Physiology and
Biophysics (LIS), Case Western Reserve University School of Medicine, Cleveland, Ohio.
Correspondence to: Robin L. Haynes, PhD, Department of Neurology
(Neuroscience), Childrens Hospital, Boston, MA 02115. E-mail: Robin.
Haynes@TCH.Harvard.edu
This work was supported by grants from the NINDS (PO1 NS38475)
and NICHD (Childrens Hospital Mental Retardation Research Center
[P30-HD18655]).

resulting in nitrosative and/or oxidative damage to immature cerebral white matter; these reactive species preferentially injure vulnerable premyelinating (O4 and O1)
oligodendrocytes, leading to their loss, subsequent decreased numbers of mature oligodendrocytes, and hypomyelination in long-term survivors. During the peak period of PVL (i.e. 2432 gestational weeks), O4 and O1
oligodendrocytes dominate human cerebral white matter,
with the first appearance of mature, myelin-producing
(myelin basic protein [MBP]1) cells around 30 to 35
gestational weeks (4). Recent studies in perinatal rats (5)
and rodent cell culture systems (6) indicate that O4 and
O1 oligodendrocytes are vulnerable to oxidative stress,
suggesting the possibility that the prevention of oxidative
injury may be an important treatment modality in PVL.
The lack of direct evidence of nitrosative and/or oxidative
stress in O4 and O1 cells in human PVL, however, is a
critical missing link of information that is urgently
needed to confirm the relevance of the experimental models to the human disease, and to guide preventive interventions in the premature infant.
The pathology of PVL consists of 2 components, a
focal, necrotic component in the periventricular region, and a diffuse component characterized by reactive gliosis in the surrounding white matter. In this study,
we focused more on the diffuse than the focal, necrotic
component of PVL because the former is thought to result in widespread impairment in myelination, the likely
basis for the global neurological deficits in long-term survivors. Moreover, in modern neonatal intensive care units
the diffuse component of PVL is much more common
than the focal component by neuroimaging (7). While

441

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

Abstract. Periventricular leukomalacia (PVL), the major substrate of cerebral palsy in survivors of prematurity, is defined
as focal periventricular necrosis and diffuse gliosis in immature cerebral white matter. We propose that nitrosative and/or
oxidative stress to premyelinating oligodendrocytes complicating cerebral ischemia in the sick premature infant is a key
mechanism of injury interfering with maturation of these cells to myelin-producing oligodendrocytes and subsequent myelination. Using immunocytochemical markers in autopsy brain tissue from 17 PVL cases and 28 non-PVL controls, we found
in the PVL cases: 1) selective regionalization of white matter injury, including preferential involvement of the deep compared
to intragyral white matter; 2) prominent activation of microglia diffusely throughout the white matter; 3) protein nitration and
lipid peroxidation in premyelinating oligodendrocytes in the diffuse component; 4) preferential death of premyelinating oligodendrocytes diffusely; and 5) virtual sparing of the overlying cerebral cortex, as demonstrated by markers of activated
astrocytes and microglia. These data establish that PVL is primarily a white matter disease that involves injury to premyelinating oligodendrocytes, potentially through activation of microglia and release of reactive oxygen and nitrogen species.
Agents that prevent nitrosative and oxidative stress may play a key role in ameliorating PVL in premature infants in the
intensive care nursery.

442

HAYNES ET AL

TABLE 1A
Summary of Cerebral White Matter Pathology in PVL Cases

Case

GA

PNA

PCA

Focal
acute
necrosis

1
2
3
4
5
6
7
8
9
10
11
12
13
14

35*
40
30*
27*
34*
28*
36*
38
35*
36*
40
29*
36*
31*

0
4
0
4
0
7
0
1
5
6
4
20
15
64

35
44
30
31
34
35
36
39
40
42
44
49
51
95

1
1
1
1
2
1
2
1
2
2
1
2
2
2

Focal
organizing
necrosis

Focal
scar
w/mineral

Diffuse
gliosis

WM
and/or
Vents

Hypomyelin

1
1
2
1
1
1
2
2
1
1
2
2
1
1

1
1
2
2
1
1
1
2
2
1
2
1
1
2

1
1
1
1
1
1
1
1
1
1
1
1
1
1

1
2
2
2
2
1
1
2
2
2
2
1
1
1

2
2
2
2
2
2
2
2
2
1
2
1
1
2

necrotic foci appear pathologically to be micro-infarcts,

thought to be secondary to severe hypoperfusion in the
distal end-zones of the long penetrating arteries (8), the
less severe but more widespread diffuse component of
PVL may reflect other previously unrecognized pathogenetic mechanisms, including protein nitration and oxidative injury to lipids and proteins in partially perfused
tissue. We reasoned that the demonstration of protein nitration and lipid peroxidation in O4 and/or O1 cells
would carry different implications with respect to the underlying pathophysiology of cerebral white matter damage in PVL, and to potential specific treatment modalities. We sought also to determine whether microglial
activation was associated with markers of nitrosative and/
or oxidative stress, given recent evidence that microglia
are sources of nitric oxide (NO) and peroxynitrite (9).
We further tested the hypothesis that O4 and/or O1 cells
in the diffuse component of PVL undergo cell death and
loss by double-labeling these cells for TUNEL product.
MATERIALS AND METHODS
Clinicopathologic Database Information
Cases were collected from the autopsy services of the Departments of Pathology, Childrens Hospital and Brigham and
Womens Hospital, Boston, with permission according to hospital protocols. All cases were classified by the systematic examination of standardized microscopic sections (maximum of
19 sections/case), stained with conventional H&E/Luxol fast
blue, from each brain and spinal cord (8). Eleven PVL cases
(Cases 1, 514; 3495 postconceptional weeks) and 23 control
cases (i.e. lacking PVL) (20183 postconceptional weeks) had
brain tissue fixed in formalin and embedded in paraffin and
were used for single-labeling immunocytochemistry (Table 1).
Four PVL cases (Cases 14; 3044 postconceptional weeks)

J Neuropathol Exp Neurol, Vol 62, May, 2003

and 4 control cases (2340 postconceptional weeks) had tissue

fixed in 4% paraformaldehyde for double-labeling studies. Of
note, Case 1 had both formalin- and paraformaldehyde-fixed
tissue. Ten of the 14 (71%) PVL cases were born prematurely,
and 3 of these 10 infants survived beyond the neonatal period
(.44 postconceptional weeks). In the control group, none of
the cases analyzed with a postconceptional age .37 weeks were
born prematurely.

Single-Labeling Immunocytochemistry in
Formalin-Fixed, Paraffin-Embedded Tissue
Standard methods in deparaffinized tissue sections (4 mm)
were applied (10). Antibodies specific for the following markers
were used: 4-hydroxy-2-nonenal-protein adducts (HNE) (1:100;
laboratory of Dr. Luke I. Szweda, or purchased from Calbiochem, San Diego, CA); malondialdehydeprotein adducts
(MDA) (1:100; Abcam, Cambridge, United Kingdom); nitrotyrosine protein adducts (NT) (1:100; Upstate, Lake Placid,
NY); heme-oxygenase-1 (HO-1) and 22 (HO-2) (both 1:500;
Stressgen, Victoria, Canada), glial fibrillary acidic protein
(GFAP) (1:9,000, Sternberger Monoclonals, Inc., Lutherville,
MD), and CD68 (1:50, Cell Marque, Austin, TX). Stains were
optimized using known Alzheimer and ALS tissue as positive
controls. CD68-positive control was tonsil. Negative controls
omitted the primary antibodies.

Double-Labeling Immunofluorescence for the

Demonstration of Oxidative Damage in
Specific Cell Types
Freshly collected tissue from 4 PVL and 4 control cases was
fixed immediately in 4% paraformaldehyde, and sectioned with
a Leica VTS1000 vibratome (Leica Microsystems, Inc., Bannock, IL) into 40- to 50-mm free-floating sections. Double-labeling experiments were performed sequentially, first with
monoclonal antibodies O4 or O1 for detection of oligodendrocytes (1:1,000) (4) or GFAP for detection of astrocytes. The

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

Legend: 1, present; 2, absent; GA 5 gestational age in weeks; PNA 5 postnatal age in weeks; PCA 5 postconceptional age
in weeks; mineral 5 mineralization; WM 5 white matter, Vents 5 ventricles. * Prematurity is defined as a gestational age less
than 37 weeks.

443

NITROSATIVE AND OXIDATIVE STRESS IN PVL

TABLE 1B
Continued Summary of the Neuropathology of PVL cases
Case

CB

BG

Thal

Hipp

CCH WMH SAH GMH

A, B

2
3
4
5
6

A
2
2
2
A

A
2
F
2
2

C
E
NA
G
E, H

A, B
E
2
G
E, H

A, B
2
2
2
2

D
2
2
2
2

1
2
2
2
1

2
2
2
1
2

2
2
1
2
2

9
10

NA
A

J
L

A
A

A
A

A
2

K
A

1
2

2
2

2
2

11
12

A
2

A
2

NA
A

A, M
2

A
2

2
2

2
2

2
2

2
2

13

14

Diagnosis
Potter syndrome; abnormal temporal lobe gyration;
malrotation of hippocampi; duplication of dentate
gyrus
Congenital diaphragmatic hernia; ECMO treatment
Placental abruption
Breech birth; sepsis; necrotizing enterocolitis
Osteogenesis imperfecta
Necrotizing enterocolitis; bowel perforation; coagulopathy; sepsis
CMV infection in utero; ABO incompatibility;
dwarfism/skeletal dysplasia; fetal heart failure
Urea cycle defect (ornithine transcarbamylase deficiency); hyperammonemia
Breech birth; oligohydramnios; renal sclerosis
Intrauterine growth retardation; breech birth; oligohydramnios; dysplastic kidneys; malformed basal
ganglia and hippocampus
Congenital cardiac malformation; ECMO treatment
Pulmonary hypertension; pneumonia; ECMO treatment
Breech birth; oligohydramnios; pneumothorax; respiratory failure
Congenital cardiac malformation; ECMO treatment

Legend: 1, pathology present; 2, pathology absent; A 5 reactive gliosis; B 5 neuronal loss; C 5 ischemic necrosis; D 5
hemorrhagic infarct; E 5 apoptosis; F 5 focal microhemorrhage; G 5 mineralization; H 5 perivascular hemorrhage; I 5 microglial
nodules; J 5 vermis cortical scar; K 5 organizing infarct; L 5 Purkinje cell loss; M 5 microinfarct; N 5 necrotic focus in
hypothalamus; EMCO 5 extracorporeal membrane oxygenation; NA 5 not available; Bst, brainstem; CB, cerebellum; BG, basal
ganglia; Thal, thalamus; Hipp, hippocampus; CCH, cerebral cortex hemorrhage; WMH, white matter hemorrhage; GMH, germinal
matrix hemorrhage.

same sections were subsequently incubated with antibodies to

HNE (1:250), MDA (1:100), and NT (1:250). Immunostained
sections were visualized with Nikon Eclipse E800 microscope
(Nikon, Melville, NY) with image capture using Spot Software
(Diagnostics Instruments, Incorporated, Sterling Heights, MI).

Tdt-Mediated Dutp Nick End Labeling (TUNEL) Staining

and Double-Labeling Immunofluorescence for
Dying O4 and O1
For O4 and TUNEL double labeling, tissue fixed in 4% paraformaldehyde was cut at 40 to 50 mm. Free-floating sections
were stained for O4 (see above). Sections were then mounted,
allowed to dry, and TUNEL-positive cells were stained using
an in situ cell death detection kit from Roche (Indianapolis, IN).
For GFAP and TUNEL double labeling, formalin-fixed paraffinembedded sections were used.

Grading Method and Statistical Analysis

Grading of the single-labeled tissue sections was performed
by counting positive cells per high power field (hpf) at 3400
magnification (0.173 mm2), which was selected on the basis of
the most intense immunostaining after survey of all fields: 0,
no cell staining; 1/2, 0 to 1 immunopositive cell/hpf; 1, 2 to 10
cells/hpf; 2, 11 to 20 immunopositive cells/hpf; and 3, .20
cells/hpf. Two observers scored each case without knowledge

of case diagnosis and age. The age-adjusted difference was assessed with age as the covariant.

RESULTS
In the PVL and control tissue sections, we assessed
activated microglia/macrophages with an antibody to
CD68, and reactive astrocytes with an antibody to GFAP.
We used an antibody to nitrotyrosine (NT) to indicate
protein nitration, and antibodies to 4-hydroxynoneal
(HNE)-protein adducts and malondialdehyde (MDA)protein adducts to indicate lipid peroxidation, in singleand double-labeling studies in the diffuse component of
PVL. In addition, the enzymes heme oxygenase (HO)-1
and 22 were examined to indicate cellular response to
oxidative stress. We performed single-labeling immunocytochemistry using markers of nitrosative and oxidative
stress and cellular responses in 11 of the 14 PVL cases
(Cases 1, 514) and 23 controls (without PVL), adjusted
for age, in formalin-fixed, paraffin-embedded tissue sections. We performed double labeling studies in brain tissues from 4 PVL and 4 control cases, fixed in 4% paraformaldehyde. We used a semiquantitative grading
system (03) of the density of immunopositive cells/high
J Neuropathol Exp Neurol, Vol 62, May, 2003

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

Bst

444

HAYNES ET AL

TABLE 2
Comparison of Age-Adjusted Means (SEM) for Markers
of Oxidative Stress and Cellular Damage between PVL
and Control Cases
Age*

PVL

Control

p value

GFAP

PCA
GA
PNA

2.64 (0.17)
2.73 (0.22)
2.68 (0.17)

1.73 (0.14)
1.66 (0.18)
1.70 (0.14)

,0.001
0.002
0.002

CD68

PCA
GA
PNA

2.24 (0.26)
2.50 (0.30)
2.27 (0.26)

1.18 (0.21)
1.53 (0.23)
1.15 (0.21)

0.005
0.017
0.003

HNE

PCA
GA
PNA

2.57 (0.13)
2.66 (0.14)
2.58 (0.13)

0.98 (0.11)
0.92 (0.12)
0.98 (0.11)

,0.001
,0.001
,0.001

MDA

PCA
GA
PNA

2.78 (0.27)
3.04 (0.29)
2.80 (0.27)

1.43 (0.23)
1.23 (0.24)
1.42 (0.23)

0.002
0.001
0.001

NT

PCA
GA
PNA

1.98 (0.26)
2.00 (0.29)
1.99 (0.26)

0.38 (0.22)
0.37 (0.24)
0.38 (0.22)

,0.001
,0.001
,0.001

HO-1

PCA
GA
PNA

1.99 (0.31)
1.80 (0.34)
2.04 (0.31)

1.51 (0.26)
1.64 (0.28)
1.47 (0.26)

0.252
0.741
0.176

HO-2

PCA
GA
PNA

2.43 (0.28)
2.48 (0.31)
2.44 (0.27)

1.05 (0.24)
1.01 (0.26)
1.04 (0.23)

0.001
0.003
,0.001

* The age-adjusted mean (6 standard error [SEM]) of the

degree of oxidative damage (scale: 03) for the selected markers of oxidative stress in PVL (n 5 11) and control (n 5 23)
cases using single labeling of glial cells with immunocytochemical techniques in sections of cerebral white matter. The p values are from t tests of the difference between the least-square
means for PVL and Control cases. A p value ,0.05 is considered significant. Abbreviations: PCA, postconceptional age
(gestational plus postnatal age); GA, gestational age; PNA,
postnatal age; SEM, standard error of the mean.

power (3400) microscopic field (hpf, 0.173 mm2) to

compare the degree of cellular injury between PVL and
control cases in comparable single-labeled tissue sections
(Table 2). There was no statistically significant relationship between the selected markers of nitrosative and oxidative stress and postmortem interval (range: 2132
hours; median: 14 hours) (data not shown), indicating
that the differential expression of cellular immunostaining for the selected markers in this study did not represent
an artifact of postmortem autolysis.
The Neuropathology of PVL with Cellular Markers for
Microglial and Astrocytic Activation
The PVL cases (n 5 14) exhibited a spectrum of severity of cerebral white matter pathology. The age at
birth, length of survival, and histopathologic characteristics of the PVL lesions of each PVL case are summarized in Table 1. All of the PVL cases, with the exception
of three (Cases 2, 8, and 11) were born prematurely. The
J Neuropathol Exp Neurol, Vol 62, May, 2003

Fig. 1. Macroscopic and cellular pathology of PVL. A: The

typical appearance of PVL, with focal necrosis causing a cystic
cavity (arrow) in the periventricular white matter of the parietal
lobe in a premature infant born at 35 gestational weeks, and
dying at 5 postnatal (40 postconceptional) weeks (Case 9). B
E: Activated microglia/infiltrating macrophages, as demonstrated by CD68 immunostaining, are differentially distributed
throughout the cerebral white matter in PVL. Macrophages (arrow) are concentrated around and in the focus of necrosis, indicated by asterisk, 3100 (B), where they are noted as large,
round cells with distinct cytoplasmic borders at high power
(3400) (C). D: Activated (ramified) microglia (arrow) are also
present diffusely throughout the deep white matter surrounding
the periventriculuar focus of necrosis, where they are visible as
elongated cells with processes (3200). E: Activated microglia/
infiltrating macrophages are not detected in the cerebral cortex
overlying the white matter damage (3200). F: Immunostaining
for GFAP shows a high density of reactive astrocytes in the
deep white matter (DWM) (grade 3), with striking sparing of
the overlying cerebral cortex (CC); the gray-white matter junction is indicated by asterisks (340). PVL Case 9 is illustrated
with each of the above markers.

term deliveries were complicated by congenital diaphragmatic hernia (Case 2), urea cycle defect (Case 8), and
congenital heart malformation (Case 11) (Table 1). In 9
cases, focal cystic cavities were grossly visible in the
periventricular white matter (Fig. 1A). Microscopically,
7 cases had acute coagulative necrosis, characterized by
fragmentation of all tissue elements and pyknotic or absent nuclei, indicative of an acute insult of 24 to 48 hours
duration; 9 cases had necrotic foci with subacute changes

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

Marker

NITROSATIVE AND OXIDATIVE STRESS IN PVL

with cavitation and infiltration of macrophages, indicative

of an insult at or beyond 3 to 5 days duration; and 8
cases had chronic changes (glial scarring or mineralization) indicating the site of prior focal necrosis (8). Seven
cases had focal lesions of more than one histopathologic
age. Of note, the overlying cortical neurons did not appear acutely necrotic, nor was there appreciable chronic
neuronal cell loss, with the exception of 2 cases with
focal cortical infarctions (Cases 2 and 9) (Table 1). All
cases had the diffusely gliotic component of PVL associated with the focally necrotic component, as per the
study criteria of PVL. Although macrophages in the subacute necrotic foci were CD68-positive (Fig. 1B, C), as
expected, we also found marked microglial/macrophagocytic infiltration, as detected with CD68 immunostaining, throughout the periventricular and central regions of
the gliotic white matter in the PVL cases (Fig. 1D; Table
2), with relative sparing of the overlying cerebral cortex

Fig. 3. Nitrosative damage in white matter in PVL. A: A

high density of cells (Grade 3), including larger (arrowhead)
and smaller (arrow) cells morphologically consistent with reactive astrocytes and oligodendrocytes, respectively, have nitrotyrosine detected by immunocytochemistry, here illustrated in
the diffuse component of PVL in an infant at 44 postconceptional weeks (3400). B: No staining is seen in age-matched
control (3400). C, D: Nitrotyrosine (NT) colocalizes (yellow
signal) with the marker for premyelinating oligodendrocytes
(O4, red; C), and with GFAP (green; D), indicating that both
types of glia undergo protein nitration (3600). PVL Case 11 is
illustrated.

(Fig. 1E). The difference in CD68 immunostaining between PVL and control white matter, and between PVL
white matter and cortex, was statistically significant (p ,
0.005). In the deep white matter surrounding the foci of
necrosis there was marked (Grades $2) gliosis with reactive (GFAP1) astrocytes, again with striking sparing of
the overlying cerebral cortex (Fig. 1F; Table 2). There
was marked regional heterogeneity in the density of reactive astrocytes, and of glial cells with positive immunostaining for the selected markers of nitrosative and oxidative stress in the different components of the white
matter (Fig. 2AC).
Immunocytochemistry for Protein Nitration in PVL
Morphologically identified macrophages within the
subacute necrotic foci immunostained positively with NT
(data not shown). Nitrotyrosine immunoreactivity was
also localized to cells morphologically consistent with
glial cells in the diffuse component (Fig. 3A) and was
increased in PVL when compared to controls (Fig. 3B;
J Neuropathol Exp Neurol, Vol 62, May, 2003

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

Fig. 2. Regional heterogeneity of white matter injury in

PVL. A: There is a heterogeneous distribution of reactive glial
cells with positive immunostaining for selected markers of cellular injury (MDA shown) in the different components of the
cerebral white matter, as demonstrated in a typical PVL case at
35 postconceptional weeks (28 gestational weeks) in the temporal-parietal white matter (level of the temporal horn of the
lateral ventricle [Ven]; white matter regions, periventricular
[PV], central [Cen], and intragyral [IG]; hippocampus [Hipp];
choroid plexus [CP] (with focal intraplexal hemorrhage) (H&E/
Luxol fast blue; 31). Periventricular necrosis is demonstrated
at this level by the rarefaction (pallor) of the periventricular
white matter (arrow). B: The IG white matter distant from the
periventricular foucs of necrosis lacks MDA-immunopositive
cells (Grade 0) (3200); central white matter (Cen) is likewise
uninvolved (not shown). C: PVL Case 6. In contrast, the MDAimmunopositive glial cells are heavily concentrated (Grade 3)
in the periventricular white matter surrounding the focal necrosis (3200).

445

446

HAYNES ET AL

TUNEL staining was performed in 4 PVL cases (Cases

14). Although TUNEL-positive oligodendrocytes were
identified in other PVL cases, there was an increased
amount of TUNEL-positive oligodendrocytes identified
in Case 4, shown in Figure 5A. This finding was consistent with a notable amount of pyknotic nuclei seen in
H&E sections (data not shown). In 1 case of chronic PVL
(Case 1; characterized by glial scarring of and around the
necrotic foci), we found a striking qualitative loss of oligodendrocytes (O1 shown) in the diffuse component of
PVL, compared to control (Fig. 5C). Interestingly, few
TUNEL-positive (i.e. acutely dying) oligodendrocytes
were identified in the chronic PVL. Of note, O4 and O1
immunolabeling is technically suboptimal in formalinfixed, paraffin-embedded tissue sections (4), thereby precluding the evaluation of O4 and O1 cell loss in the extensive archival database studied above with single-label
immunocytochemistry.

Immunocytochemistry for Oxidative Stress-Related

Markers in PVL

DISCUSSION

Antibodies to HNE, MDA, HO-1, and HO-2 were localized to cells morphologically consistent with macrophages within the necrotic foci of PVL (data not shown).
In the diffuse white matter lesion, reactive astrocytes
were significantly increased in number compared to control cases, adjusted for PCA, GA, or PNA (Table 2). Glial
cells, some of which were morphologically consistent
with reactive astrocytes (Fig. 4A), within the diffuse lesion were immunopositive for all oxidative stress-related
markers analyzed, including HNE (Fig. 4B), MDA (Fig.
2C), HO-1, and HO-2 (Table 2). There was a statistically
significant difference in the age-adjusted means of the
grade of immunopositive white matter cells between the
PVL and control groups for HNE, MDA, and HO-2, regardless of the particular age measure, i.e. whether the
cases were adjusted for PCA, GA, or PNA (Table 2).
With double labeling, the oxidative stress markers HNE
and MDA were present in O4-positive oligodendrocytes
(Fig. 4C; HNE shown) and O1-positive oligodendrocytes
(data not shown), as well as in astrocytes (Fig. 4D). In
control cases, a slight but detectable amount of HNE and
MDA colocalized primarily with GFAP-immunopositive
reactive astrocytes, and occasionally with O4- and/or O1positive cells (data not shown).
TUNEL Staining and Loss of Oligodendrocytes in PVL
Dying oligodendrocytes were identified in acute PVL
by colocalization of O4 immunoreactive cells with TUNEL-positive nuclei (Fig. 5A, Case 4). In contrast, the
GFAP-immunopositive reactive astrocytes, which separately labeled with markers of nitrosative and oxidative
damage (Figs. 3D, 4D), were not dying, given the lack
of the colocalization of GFAP and TUNEL-positivity
(Fig. 5B) and their increase in number, i.e. proliferation.
J Neuropathol Exp Neurol, Vol 62, May, 2003

Based upon the use of specific cellular markers with

single- and double-labeling immunocytochemistry, this
study provides major new insights into the cellular pathology of PVL, with important implications for developing new drugs in patients and in testing hypotheses in
animal and cell culture models. The use of markers for
reactive astrocytes (GFAP) and activated microglia/macrophages (CD68) in this study establishes that PVL is a
primary white matter disease characterized by significant,
diffuse astro- and microgliosis, in contrast to a virtually
spared overlying cerebral cortex. Thus, the recent demonstration of cortical volume loss by sophisticated neuroimaging in premature infants with PVL (11) is not explained by histopathologic evidence for primary cortical
damage in the acute and subacute phases of PVL, using
the astrocytic and microglial markers analyzed in this
study. Explanations for cortical volume loss must therefore be sought in a more subtle and likely chronic disturbance of cortical neurons that is secondary to injury
to axons coursing through cerebral white matter (12),
and/or to direct injury to the subplate neurons dispersed
throughout the white matter that are critical for development of cortical neurons (13).
The use of markers for glial injury in this study demonstrates clearly that the involvement of the cerebral
white matter in the diffuse component of PVL is not uniform. Rather, the deep and central white matter is preferentially involved, with relative sparing of the intragyral
(subcortical) white matter, and the white matter in the
vicinity of periventricular foci of necrosis is involved to
a much greater degree than is the white matter distant
from these foci. While heterogeneity of white matter
damage in PVL has previously been suggested using
standard tissue stains (12), the use of injury-specific
markers allows for more complete characterization of the

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

Table 2). The degree of NT immunostaining was statistically significant in PVL cases compared to control cases, adjusted for postconceptional age (PCA), gestational
age (GA), or postnatal age (PNA) (Table 2), with increased positive cell density in areas immediately surrounding the focal PVL lesion. To define precisely the
specific cell types susceptible to nitrosative stress in PVL,
double labeling with markers to O4, O1, or GFAP and
to NT was done. In the PVL cases, NT was present in
O4-positive oligodendrocytes (Fig. 3C) and O1-positive
oligodendrocytes (data not shown) in the diffuse component. Nitrotyrosine also colocalized with GFAP (Fig.
3D), indicating that nitrosative damage occurs in reactive
astrocytes, as well as in oligodendrocytes. In control cases, a slight but detectable amount of NT colocalized primarily with GFAP-immunopositive reactive astrocytes
and occasionally with O4 and/or O1 cells (data not
shown).

NITROSATIVE AND OXIDATIVE STRESS IN PVL

distribution of damage. This regionalization of white matter damage reinforces the idea that PVL is secondary to
ischemia in distal fields of arterial irrigation, i.e. vulnerable watershed territories in the deeper (and not superficial) white matter.
The use of the CD68 marker in this study reveals a
widespread activation of microglia/macrophages in the
diffuse component of PVL, indicating a need for new
emphasis on the role of microglia/macrophages in the
pathogenesis of PVL. CNS microglia/macrophages are
known to be activated by ischemia/hypoxia (14, 15), including up to 1 month after the initial insult (16). Of
particular relevance in PVL is the ability of activated microglia/macrophages to release reactive nitrogen species
(RNS) (17) and reactive oxygen species (ROS) (18) into
the surrounding environment, thus putting oligodendrocytes at risk for nitrosative and oxidative damage (19).
Nitric oxide (NO), in particular, is fundamental to nitrosative injury. While NO has normal functions in the
brain, including synaptic transmission and the regulation

Fig. 5. Oligodendroglial cell death and depopulation in

PVL. A, B: Cell death, as determined by TUNEL staining
(green), preferentially affects (yellow signal) premyelinating oligodendrocytes (O4, red; A), while sparing astrocytes (GFAP,
green; B), here shown in the acute phase of injury in a premature infant dying at 31 postconceptional weeks (3600) (Case
4). C: Compared to control at 31 postconceptional weeks, there
is a dramatic loss of O1 cells (3600) and O4 cells (data not
shown) in the gliotic white matter (chronic, diffuse PVL) of an
infant at 35 postconceptional weeks (Case 1).

of vascular tone, it also plays a causal role in stroke (20).

There are multiple isoforms of nitric oxide synthase
(NOS), the synthetic enzyme for NO, including an inducible form in microglia and astrocytes that is a major
source of excess NO in pathological conditions (20). Nitric oxide reacts with superoxide anion to form peroxynitrite, a highly reactive molecule that targets tyrosine
residues of proteins to form nitrotyrosine residues (21).
Excess NO release and subsequent peroxynitrite and nitrotyrosine formation occur in experimental models of
ischemia (22), while treatment during reperfusion with
inhibitors of inducible NOS reduces damage (23), suggesting a direct role for NO in ischemic injury. In cell
culture, premyelinating oligodendrocytes are sensitive to
NO-mediated injury and cell death (24, 25). Of interest,
the activation of microglia by interferon-g, b-amyloid,
and lipopolysaccharide is neurotoxic to rat neurons
through the release of peroxynitrite (9). Our data showing
evidence of protein nitration in O4 and O1 cells in the
presence of diffuse microglial activation suggest that microglial activation with NO production and formation of
J Neuropathol Exp Neurol, Vol 62, May, 2003

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

Fig. 4. Oxidative damage in white matter of PVL. A, B:

Glial cells, including many with the histopathologic features of
reactive astrocytes (arrowhead), shown on GFAP immunostain
for reference (A), as well as smaller cells suggestive of oligodendrocytes (B, arrow), demonstrate intense HNE immunostaining (Grade 3), here shown in the diffuse component of
PVL in an infant at 44 postconceptional weeks (3400). C, D:
Hydroxynonenal (HNE) adducts colocalize (yellow signal) with
markers for premyelinating oligodendrocytes (O4, red; C) and
astrocytes (GFAP, green; D). PVL Case 11 is illustrated. Double
labeling also identified MDA adducts in oligodendrocytes and
astrocytes in PVL (data not shown).

447

448

HAYNES ET AL

J Neuropathol Exp Neurol, Vol 62, May, 2003

including PVL. Astrocytes may protect surrounding neuronal and oligodendroglial cells under conditions of such
stress due to their capacity to release antioxidants into
the extracellular fluid (36, 37), and to clear extracellular
glutamate (38), which in excessive levels results in excitotoxicity to premyelinating oligodendrocytes (39, 40).
We speculate that reactive astrocytes in diffuse PVL proliferate to act, at least in part, as scavengers of excess
oxidants in the environment, thereby protecting O4 and
O1 cells from reactive nitrogen and oxygen species. Alternatively, reactive astrocytes may not be protective, but
instead contribute to tissue damage in PVL. In this regard, astrocytes upregulate inducible NOS in experimental ischemic injury (41), suggesting that, in addition to
activated microglia, they are another source of the free
radical NO.
In addition to markers of cellular damage, we showed
a statistically significant increase in the density of HO-2immunopositive glial cells in PVL. The HO family functions as a defense mechanism against free radical injury
through degradation of heme, a pro-oxidant, and subsequent production of carbon monoxide, free iron, and biliverdin, which is subsequently converted to bilirubin, a
potent antioxidant (42). In contrast to HO-2, HO-1 was
not significantly increased in the PVL cases compared to
the control. Heme oxygenase-1, shown to be induced by
ischemic injury, is also induced by cytokines and prostaglandins, and in response to hemorrhage, sepsis, and
pregnancy (43). The lack of a significant difference in
HO-1 between PVL and controls suggests that, unlike in
other reported settings, HO-1 may not be a specific marker of oxidative stress in developing human brain.
In conclusion, we show a region-specific vulnerability
to nitrosative and oxidative damage within the deep periventricular white matter, and a relative sparing of the
overlying cerebral cortex. We also show widespread activation of microglia, leading us to speculate that these
cells may play a direct role in the white matter damage
via release of NO. It is unknown whether nitrosative or
oxidative damage in PVL is the primary insult to premyelinating oligodendrocytes, or is secondary to other
insults, e.g. excitotoxicity from excessive glutamate release (44), or cytokine toxicity (45) complicating diffuse
microglial/macrophagocytic activation. Irrespective of the
precise sequence of events, the burden of nitrosative and
oxidative damage to the white matter appears enormous
in PVL. Our findings indicate that continued research into
the role of reactive nitrogen and oxygen species in the
pathogenesis of PVL is urgently needed, with the ultimate goal of the development of drug strategies to prevent this devastating disorder.
ACKNOWLEDGMENT
We thank Drs. Geraldine Pinkus, Neil Kowall, and Pamela L. Follett
for assistance in immunocytochemistry, Ms. Diane Brown and Ms. Lena

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

peroxynitrite may play a key role in O4 and O1 cell injury in PVL.

In addition to nitrosative stress, we demonstrate PVLrelated oxidative damage to oligodendrocytes, most of
which are premyelinating in the ages affected by PVL
(4). Like RNS, ROS are released by activated microglia
and upon release, interact with the lipid component of
cellular membranes, initiating lipid peroxidation that results in the breakdown of lipid constituents into highly
reactive byproducts, including lipid aldehydes, HNE, and
MDA (26). These reactive aldehydes then bind to and
modify protein creating protein adducts. Lipid peroxidation, as indicated by HNE-protein adducts, occurs in
models of ischemia (27). In this regard, a recent study of
premature infants found elevated levels of markers of lipid peroxidation and oxidative protein products in cerebrospinal fluid of infants who had PVL documented by
magnetic resonance imaging (28). In ischemia, lipid peroxidation and adduct formation may play direct roles in
cell toxicity, given that HNE, like NO, is toxic to premyelinating oligodendrocytes in a dose-dependent manner (27, 29). Although RNS and ROS represent different
pathophysiological mechanisms of cellular injury and
may play a different role in the overall damage in PVL,
they may share a similar mode of action with respect to
oligodendrocyte injury, i.e. both have been shown to inhibit mitochondrial respiration (20, 30) and induce mitochondrial membrane permeabilization (31, 32).
In this study we demonstrated, via the TUNEL method,
that O4 and O1 cells undergo cell death in the diffuse
component of PVL. Although we identified dying (likely
premyelinating) oligodendrocytes in an acute case of
PVL, a striking loss of these cells was found in an advanced case of PVL with secondary scarring of the periventricular necrotic foci, suggesting that O4 and O1 cells
die immediately in the acute phase following the ischemic insult and disappear almost completely by the chronic phase. Premyelinating oligodendrocytes are likely
damaged at the time of the initial ischemia/reperfusion,
but also for an extended period thereafter; the latter injury
due to subacute tissue reactions (i.e. diffuse microglial
activation), which may last weeks. Consequently, this
possibility of ongoing inflammatory-related injury extends the window of opportunity for drug therapy to
protect against O4 and O1 cell damage.
In addition to oligodendrocytes, reactive astrocytes in
the diffuse component of PVL immunostain for markers
of both nitrosative and oxidative damage. The markers
used in this study colocalize to reactive astrocytes in several human neurologic diseases not generally considered
ischemic, i.e. amyotrophic lateral sclerosis (33), Alzheimer disease (34), and multiple sclerosis (35). Thus, nitrositive and oxidative damage in reactive astrocytes in
human neuropathology suggests RNS and ROS injury as
a final common pathway in multiple disease processes,

NITROSATIVE AND OXIDATIVE STRESS IN PVL

Liu for assistance in the preparation of the database, Richard Belliveau
for assistance with manuscript preparation, and Mr. Mark Drinkwater
and the Fellows in the Combined Brigham and Womens Hospital and
Childrens Hospital Training Program in Neuropathology for assistance
in tissue collection.

REFERENCES

20. Bolanos JP, Almeida A, Stewart V, et al. Nitric oxide-mediated

mitochondrial damage in the brain: Mechanisms and implications
for neurodegenerative diseases. J Neurochem 1997;69:222740
21. Beckman J, Viera L, Estevez A, Teng R. Nitric oxide and peroxynitrite in the perinatal period. Sem Perinatology 2000;24:3741
22. Sakurai M, Fukuyama N, Takizawa S, et al. Inductions of 3-Lnitrotyrosine in motor neurons after transient spinal cord ischemia
in rabbits. J Cereb Blood Flow Metab 1998;18:123338
23. Meguro M, Katsuramaki T, Nagayama M, et al. A novel inhibitor
of inducible nitric oxide synthase (ONO-1714) prevents critical
warm ischemia-reperfusion injury in the pig liver. Transplantation
2002;73:143946
24. Mackenzie-Graham AJ, Mitrovic B, Smoll A, Merrill JE. Differential sensitivity to nitric oxide in immortalized, cloned murine oligodendrocyte cell lines. Dev Neurosci 1994;16:16271
25. Merrill JE, Ignarro LJ, Sherman MP, Melinek J, Lane TE. Microglial cell cytotoxicity of oligodendrocytes is mediated through nitric
oxide. J Immunol 1993;151:213241
26. Esterbauer H, Schaur RJ, Zollner H. Chemistry and biochemistry
of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free
Radic Biol Med 1991;11:8128
27. McCracken E, Valeriani V, Simpson C, Jover T, McCulloch J, Dewar D. The lipid peroxidation by-product 4-hydroxynonenal is toxic
to axons and oligodendrocytes. J Cereb Blood Flow Metab 2000;
20:152936
28. Inder T, Mocatta T, Darlow B, Spencer C, Volpe JJ, Winterbourn
C. Elevated free radical products in the cerebrospinal fluid of
VLBW infants with cerebral white matter injury. Pediatr Res 2002;
52:21318
29. Gard AL, Solodushko VG, Waeg G, Majic T. 4-Hydroxynonenal, a
lipid peroxidation byproduct of spinal cord injury, is cytotoxic for
oligodendrocyte progenitors and inhibits their responsiveness to
PDGF. Micros Res Tech 2001;52:70918
30. Picklo MJ, Amarnath V, McIntyre JO, Graham DG, Montine TJ.
4-Hydroxy-2(E)-nonenal inhibits CNS mitochondrial respiration at
multiple sites. J Neurochem 1999;72:161724
31. Brookes PS, Salinas EP, Darley-Usmar K, et al. Concentration-dependent effects of nitric oxide on mitochondrial permeability transition and cytochrome c release. J Biol Chem 2000;275:2047479
32. Kristal BS, Elizondo-Fournier M. 4-Hydroxyhexenal is a potent
inducer of the mitochondrial permeability transition. J Neurosci
1995;15:707178
33. Sasaki S, Shibata N, Komori T, Iwata M. iNOS and nitrotyrosine
immunoreactivity in amyotrophic lateral sclerosis. Neurosci Lett
2000;291:4448
34. Dei R, Takeda A, Niwa H, et al. Lipid peroxidation and advanced
glycation end products in the brain in normal aging and in Alzheimers disease. Acta Neuropathol 2002;104:11322
35. Oleszak EL, Zaczynska E, Bhattacharjee M, Butunoi C, Legido A,
Katsetos CD. Inducible nitric oxide synthase and nitrotyrosine are
found in monocytes/macrophages and/or astrocytes in acute, but not
in chronic, multiple sclerosis. Clin Diagn Lan Immunol 1998;5:
43845
36. Loeffler DA, LeWitt PA, Juneau PL, et al. Increased regional brain
concentrations of ceruloplasmin in neurodegenerative disorders.
Brain Research 1996;738:26574
37. Siushansian R, Wilson JX. Ascorbate transport and intracellular
concentration in cerebral astrocytes. J Neurochem 1995;65:4149
38. Anderson CM, Swanson RA. Astrocyte glutamate transport: Review of properties, regulation, and physiological functions. Glia
2000;32:114
39. McDonald JW, Levine JM, Qu Y. Multiple classes of the oligodendrocyte lineage are highly vulnerable to excitotoxicity. Neuroreport
1998;9:275762

J Neuropathol Exp Neurol, Vol 62, May, 2003

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

1. Hoyert DL, Freedman MA, Strobino DM, Guyer B. Annual summary of vital statistics: 2000. Pediatrics 2001;108:124155
2. Volpe J. Neurology of the newborn. Philadelphia: W. B. Saunders
Company, 2001
3. Volpe JJ. Neurobiology of periventricular leukomalacia in the premature infant. Pediatr Res 2001;50:55362
4. Back SA, Luo NL, Broenstein NS, Levine JM, Volpe JJ, Kinney
HC. Late oligodendrocyte progenitors coincide with the development window of vulnerability for human perinatal white matter
injury. J Neurosci 2001;21:130212
5. Levison SW, Rothstein RP, Romanko MJ, Snyder MJ, Meyers RL,
Vannucci SJ. Hypoxia/ischemia depletes the rat perinatal subventricular zone of oligodendrocyte progenitors and neural stem cells.
Dev Neurosci 2001;23:23447
6. Back SA, Gan X, Li Y, Rosenberg PA, Volpe JJ. Maturation-dependent vulnerability of oligodendrocytes to oxidative stress-induced death caused by glutathione depletion. J Neurosci 1998;18:
624153
7. Maalouf EF, Duggan PJ, Counsell SJ, et al. Comparison of findings
on cranial ultrasound and magnetic resonance imaging in preterm
infants. Pediatrics 2001;107:71927
8. Kinney H, Armstrong D. Perinatal neuropathology. In: Graham D,
Lantos P, eds. Greenfields neuropathology, Vol. 1. London: Arnold,
1997:55255
9. Xie Z, Wei M, Morgan TE, et al. Peroxynitrite mediates neurotoxicity of amyloid beta-peptide1-42- and lipopolysaccharide-activated
microglia. J Neurosci 2002;22:348492
10. Lee BH, Hecht JL, Pinkus JL, Pinkus GS. WT1, estrogen receptor,
and progesterone receptor as markers for breast or ovarian primary
sites in metastatic adenocarcinoma to body fluids. Am J Clin Pathol
2002;117:74550
11. Inder TE, Huppi PS, Warfield S, et al. Periventriculuar white matter
injury in the premature infant is followed by reduced cerebral cortical gray matter volume at term. Ann Neurol 1999;46:75560
12. Deguchi K, Oguchi K, Matsuura N, Armstrong DD, Takashima S.
Periventricular leukomalacia: Relation to gestational age and axonal
injury. Pediatr Neurol 1999;20:37074
13. Ghosh A, Antonini A, McConnell SK, Shatz CJ. Requirement for
subplate neurons in the formation of thalamocortical connections.
Nature 1990;347:17981
14. Cai Z, Pang Y, Xiao F, Rhodes PG. Chronic ischemia preferentially
causes white matter injury in the neonatal rat brain. Brain Res 2001;
898:12635
15. Ivacko JA, Sun R, Silverstein FS. Hypoxic-ischemic brain injury
induces an acute microglial reaction in perinatal rats. Pediatr Res
1996;39:3947
16. Bona E, Andersson AL, Blomgren K, et al. Chemokine and inflammatory cell response to hypoxia-ischemia in immature rats. Pediatr
Res 1999;45:500509
17. Chao CC, Hu S, Molitor TW, Shaskan EG, Peterson PK. Activated
microglia mediate neuronal cell injury via a nitric oxide mechanism. J Immunol 1992;149:273641
18. Colton CA, Gilbert DL. Microglia, an in vivo source of reactive
oxygen species in the brain. Adv Neurol 1993;59:32126
19. Thorburne SK, Juurlink BH. Low glutathione and high iron govern
the susceptibility of oligodendroglial precursors to oxidative stress.
J Neurochem 1996;67:101422

449

450

HAYNES ET AL

40. Follett PL, Rosenberg PA, Volpe JJ, Jensen FE. NBQX attenuates
excitotoxic injury in developing white matter. J Neurosci 2000;20:
923541
41. Endoh M, Maiese K, Wagner J. Expression of the inducible form
of nitric oxide synthase by reacitve astrocytes after transient global
ischemia. Brain Research 1994;651:92100
42. Otterbein LE, Choi AM. Heme oxygenase: Colors of defense
against cellular stress. Am J Physiol Lung Cell Mol Physiol 2000;
279:L102937
43. Bidmon HJ, Emde B, Oermann E, Kubitz R, Witte OW, Zilles K.
Heme oxygenase-1 (HSP-32) and heme oxygenase-2 induction in
neurons and glial cells of cerebral regions and its relation to iron

accumulation after focal cortical photothrombosis. Exp Neurol

2001;168:122
44. Yang CS, Lin NN, Tsai PJ, Liu L, Kuo JS. In vivo evidence of
hydroxyl radical formation induced by elevation of extracellular
glutamate after cerebral ischemia in the cortex of anesthetized rats.
Free Radic Biol Med 1996;20:24550
45. Casaccia-Bonnefil P. Cell death in the oligodendrocyte lineage: A
molecular perspective of life/death decisions in development and
disease. Glia 2000;29:12435
Received October 3, 2002
Revision received December 16, 2002
Accepted January 3, 2003

Downloaded from http://jnen.oxfordjournals.org/ by guest on May 22, 2016

J Neuropathol Exp Neurol, Vol 62, May, 2003