Comparative Medicine Vol 54, No 6

Copyright 2004 December 2004

by the American Association for Laboratory Animal Science Pages 631-634

Overview
In Vivo Bioluminescence Imaging

Akiko Sato, VMD,1 Brenda Klaunberg, VMD,2 and Ravi Tolwani, DVM, PhD1,*

In vivo bioluminescent imaging (BLI) is a versatile and sensitive tool that is based on detection of light emission
from cells or tissues. Bioluminescence, the biochemical generation of light by a living organism, is a naturally occur-
ring phenomenon. Luciferase enzymes, such as that from the North American firefly (Photinus pyralis), catalyze the
oxidation of a substrate (luciferin), and photons of light are a product of the reaction. Optical imaging by biolumines-
cence allows a low-cost, noninvasive, and real-time analysis of disease processes at the molecular level in living
organisms. Bioluminescence has been used to track tumor cells, bacterial and viral infections, gene expression, and
treatment response. Bioluminescence in vivo imaging allows longitudinal monitoring of a disease course in the same
animal, a desirable alternative to analyzing a number of animals at many time points during the course of the dis-
ease. We provide a brief introduction to BLI technology, specific examples of in vivo BLI studies investigating bacte-
rial/viral pathogenesis and tumor growth in animal models, and highlight some future perspectives of BLI as a
molecular imaging tool.

BLI: An Overview
Advances in molecular and cell biology techniques have led to
the development of new in vivo imaging strategies. In vivo biolu-
minescent imaging (BLI) is a sensitive tool that is based on de-
tection of light emission from cells or tissues. The utility of
reporter gene technology makes it possible to analyze specific
cellular and biological processes in a living animal through in
vivo imaging methods.
Bioluminescence, the enzymatic generation of visible light by
a living organism, is a naturally occurring phenomenon in many
non-mammalian species (4, 19). Luciferases are enzymes that
catalyze the oxidation of a substrate to release photons of light
(8). Bioluminescence from the North American firefly (Photinus
pyralis) is the most widely studied. The firefly luciferase gene
(luc) expression produces the enzyme luciferase which converts
the substrate D-luciferin to non-reactive oxyluciferin, resulting
in green light emission at 562 nm. Another example of biolumi-
nescence is from the sea pansy (Renilla reniformis). The Renilla
luciferase gene (ruc) uses the substrate coelenterazine to pro-
duce a blue light at 482 nm. Because mammalian tissues do not
naturally emit bioluminescence, in vivo BLI has considerable
appeal because images can be generated with very little back-
ground signal.
BLI requires genetic engineering of cells or tissues with an ex- Figure 1. Bioluminescence imaging. (A) Bioluminescence expression
cassette containing the luciferase gene and a promoter is transfected
pression cassette consisting of the bioluminescent reporter gene into the cell of choice. (B) The transfected expression cassette, when
under the control of a selected gene promoter constitutively driv- present in cells or tissues, produces luciferase enzyme inside the cell.
ing the light reporter (Fig. 1). When these engineered cells are in- When the luciferin substrate is added, the luciferase enzyme catalyzes
jected into the mouse, their dissemination can be tracked by luciferin substrates to emit photons of light. The emitted light then is
detected by a charge-coupled device camera.
detecting the location and intensity of the light signal. In order to
induce light production, the substrates luciferin or coelenterazine
must be provided. These substrates usually are administered by
intravascular or intraperitoneal injection. To date, there have been
Received: 10/04/04. Revision requested: 11/16/04. Accepted: 11/22/04.
1
Department of Comparative Medicine, Stanford University School of Medicine, no reports of toxicity related to repeated dosing of substrates. In
Stanford, California 94305-5410; 2Mouse Imaging Facility, National Institutes of addition to constitutive promoters, inducible promoters can be en-
Health, 10 Center Drive, Room B1D-69 Bethesda, Maryland 20892-1060.
*
Corresponding author. gineered into the expression construct, making it possible to ma-

631
Vol 54, No 6
Comparative Medicine
December 2004

nipulate light reporter gene expression conditionally (7).

The choice of reporter is dependent on the goals of the re-
search, but there are some biomechanical obstacles associated
with the use of Renilla luciferase, mostly related to the stability
of the coelenterazine substrate. The firefly luciferase is a larger
molecule (61 kDa), compared with Renilla luciferase (36 kDa)
(1). Firefly luciferase requires the presence of ATP, oxygen, and
magnesium, whereas the Renilla luciferase needs no cofactors.
These two luciferases emit photons of different wavelengths:
firefly luciferase emits green light at 562 nm, whereas Renilla
luciferase emits blue light at 482 nm. Bhaumik and Gambhir
examined the use of dual reporters in vivo and found that it was
possible to use both in a single animal (1). They found that sepa-
rate signals were distinguishable, but differences in light kinet-
ics governed the imaging techniques. Firefly luciferase produced
a stronger signal than ruc, and the signal persisted longer. Less
coelenterazine is needed for substrate injection compared to lu-
ciferin, but coelenterazine was found to be unstable in plasma,
and target site delivery was not as efficient as for luciferin. The
luc bioassay is well defined, but potential advantages to ruc in-
clude its rapid light kinetics and lack of need for cofactors.
The light emitted by luciferase is able to penetrate tissue
depths of several millimeters to centimeters; however, photon in-
tensity decreases 10-fold for each centimeter of tissue depth (3). Figure 2. An imaging system consists of a light-tight box where ani-
Sensitive light-detecting instruments must be used to detect bi- mals are placed, the cooled charge-coupled device camera to detect light,
oluminescence in vivo. The detectors measure the number of and image processing software.
photons emitted per unit area. Low levels of light at wavelengths
between 400 and 1000 nm can be detected with charge-coupled availability of cofactors (luciferase only); d) the time between
device cameras that convert the light photons that strike silicon substrate injection and maximal signal for the assay (10); e) tis-
wafers into electrons (21). The software is able to convert elec- sue depth; and f) potential signal impedance, such as from pig-
tron signals into a two-dimensional image. The software is also mented skin or fur. Clearly, sensitivity of BLI is difficult to define
able to quantify the intensity of the emitted light (number of and must be established for each biological assay; however, it
emitted photons striking the detectors) and convert these nu- may be possible to image as few as several hundred to a thou-
merical values into a pseudocolor graphic. The actual data is sand reporter cells, depending on the biological system.
measured in photons, but the pseudocolor graphic enables rapid Compared with other in vivo imaging techniques, BLI is rela-
visual interpretation. Quantitative measurements within a re- tively inexpensive. It is possible for ambitious investigators to con-
gion of interest may be necessary for more subtle differences. struct an imaging device of their own design; however, several
The use of cooled CCD cameras reduces the thermal noise, and a commercial devices are available. Some of the advantages of pur-
light-tight box allows luciferase-produced light to be optimally chasing a commercial product include the acquisition and analy-
visualized and quantified (2) (Fig. 2). sis software, as well as technical support. BLI devices vary in
It is useful to have the luciferase image superimposed on an- cost and features and range in price from approximately $65,000
other type of image, such as a photograph or radiograph, for ana- to $200,000. The substrates represent additional cost. The lu-
tomical location of the emission signal. Most commercial ciferin dose is 150 mg/kg, at a cost of approximately $5/mg (1).
imaging devices are equipped to generate an anatomical image, Coelenterazine is more expensive at $190/mg, with a mouse dose
as well as the optical emission image. The software superim- ranging up to 100 µg/mouse (1). With the popularity of this imag-
poses the images for visualization and interpretation. Although ing technique increasing, the availability of these substrates is
the detectors can measure any light source, it is necessary to improving, and more suppliers are decreasing the costs.
have appropriate wavelength filters and a fluorescence light The imaging times in BLI are short compared with those of
source if one desires the combined ability of bioluminescence and other in vivo imaging techniques. Typically, a diagnostic image can
fluorescence imaging. Commercial devices are available with be generated in a time frame of a few seconds to several minutes.
combined capabilities (22). In order to generate the best possible image, it is important that
One of the challenges of in vivo small animal imaging is the subject be immobilized, and this is best accomplished with an-
achieving image resolution that is fine enough to distinguish esthesia. Animals may be anesthetized with injectable or inhalant
subtle details between tissues. The physical sensitivity of the anesthetics, depending on the set up of the imaging device.
optical imaging detector is one consideration, and higher sensi-
tivity usually translates into higher equipment costs. Additional Infection Models to Study Host–Pathogen
important influences on the resolution of optical imaging are the Interactions
organic aspects. The sensitivity of BLI is dependent on the fol- Luciferase imaging has been used to trace bacterial and viral
lowing organic factors: a) the number of cells expressing the re- infection in vivo. In vivo bioluminescence imaging was first de-
porter gene; b) the efficiency of the gene promoter; c) the veloped using a Salmonella typhimurium infection model (2, 3).

632
 In vivo bioluminescence imaging

Three strains of Salmonella, each expressing lux genes (bacte-

ria-specific luciferase genes), were marked with bioluminescence
through transformation with a plasmid conferring constitutive
expression of bacterial luciferase (3, 6). Integration of the genes
into the chromosomes of the bacteria increases the stability of
light production from the labeled bacteria and represents a pro-
nounced improvement over plasmid-based expression (2). Bacte-
ria labeled in this manner can be detected in the tissue of mice
and can reveal the location and extent of infection (2).
Labeled bacterial pathogens can be detected in the infected host
in various applications. Jawhara and colleagues produced a biolu-
minescent Escherichia coli and evaluated the effects of antibiotics
for the treatment of acute infections in rats (12). To identify cells
that would enhance survival after stem cell transplantation,
Brown and colleagues transplanted irradiated mice with hemato-
poietic precursor cells and then challenged them with Pseudomo-
nas bacteria that constitutively expressed a bacterial luciferase
(2). The bone marrow reconstitution was enhanced by myeloid pro-
genitor pools that were capable of protection against an otherwise
lethal bacterial challenge (2). BLI allowed the assessment of the
location and extent of infection and thus enabled a better under-
standing of the host–pathogen interaction in the context of living
animal models (2). Hardy and colleagues recently employed in vivo
BLI to determine the location of Listeria monocytogenes infection Figure 3. Examples of in vivo bioluminescence imaging (BLI) in mice
in mice and reported that the organism could replicate in the mu- with luciferase-expressing tumor cell lines injected intra-renally (left
rine gall bladder (9). The authors concluded that even though it is mouse) and subcutaneously (right mouse). Tumor growth can be tracked
by the intensity and location of the bioluminescence. Figure courtesy of
unknown whether L. monocytogenes replicates in the gall bladder Drs. James Vasselli and W. Marston Linehan, National Cancer Institute.
of asymptomatic humans, the organism may be present there (9).
In vivo visualization of viral infection is also feasible using lu- burden (20). BLI often has better sensitivity in detecting bone
ciferase imaging. The first demonstration of a virus-mediated gene marrow micrometastasis when compared with other noninvasive
transfer to host cells using BLI was reported by Lipshutz and col- imaging methods, such as radiography. In one study, Wetterwald
leagues using an adeno-associated viral (AAV) vector in a therapy and colleagues found that BLI was sensitive enough to detect
model (12, 15). In this study, the approach of detecting viral gene small foci of 0.5 mm3 of bone marrow metastasis of luciferase-ex-
delivery efficiency using bioluminescence was verified as prenatal pressing human mammary carcinoma cells in mice (24).
administration of luciferase expressing AAV in the mouse resulted The capability to longitudinally track tumor growth and me-
in stable integration and bioluminescence from many tissues in tastasis is useful not only to investigate specific tumor character-
these progeny mice. Luker and colleagues used BLI in living mice istics but also can be used to better characterize genetically
to monitor the herpes simplex virus type 1 (HSV-1) virus that ex- engineered mice with altered tumor suppressor genes. Lu-
presses luciferase. The effects of valacyclovir treatment were ciferase-expressing tumor cell lines inoculated into these engi-
monitored using BLI, and it was determined that both viral titers neered mouse models allow the study of various aspects of the
and imaging data showed similar dose-dependent inhibition of the oncology disease process that are most relevant to the functions
virus by valacyclovir in living mice (16). Cook and Griffin used BLI of the specific tumor genes altered and being studied in these
to monitor Sindbis virus (SV) infection in a murine model of viral mice (11, 17, 23). Furthermore, mice have been engineered in
encephalitis (5, 6). Here, virulent and avirulent strains of SV were which the luciferase reporter is expressed on tumors spontane-
engineered to express firefly luciferase. Mice were infected subcu- ously arising from engineered mice, allowing investigation of dis-
taneously in the footpad, and viral replication was detected by bi- ease onset and progression (17).
oluminescence before spread to the central nervous system (5, 6). BLI also has been utilized to noninvasively assess therapeutic
efficacy and gene delivery. Adenovirus-induced ovarian tumor
Tumor Models to Study Growth, Metastasis, killing in mice was demonstrated effectively by using biolumi-
and Therapeutic Efficacy nescence (13). In a study utilizing rats injected with luciferase-
The ability to track a small number of tumor cells expressing transfected colon carcinoma cells, the antineoplastic effects of
luciferase has allowed the study of tumor growth, metastasis, and cisplatin were effectively tracked by bioluminescence as con-
therapeutic responses in vivo using BLI (Fig. 3). Many tumor cell cluded by decreased number of light emitting tumor sites and
lines that constitutively express luciferase have been developed. In light signal intensity (25).
a study characterizing the utility of bioluminescence to track tu-
mor burden, Paroo and colleagues monitored luciferase expres- Assessment of Real-Time Gene Expression
sion in subcutaneous tumors of mice in a longitudinal study. BLI also has been used to study in vivo gene expression. To
They found that the bioluminescent tumor growth profile was accomplish this, mice are engineered to express luciferase under
similar to that observed with caliper measurements, thus fur- the transcriptional control of promoters from genes of interest.
ther validating the utility of bioluminescence to assess tumor Events leading to activation of the promoter, therefore, result in

633
Vol 54, No 6
Comparative Medicine
December 2004

the production of luciferase, which can be monitored by the level 7. Gossen, M. and H. Bujard. 2002. Studying gene function in eu-
of light produced. karyotes by conditional gene inactivation. Annu. Rev. Genet. 36:153-
173.
In one example, transgenic mice engineered with the luciferase
8. Greer, L. F., III and A. A. Szalay. 2002. Imaging of light emission
gene under the control of the estrogen-responsive element pro- from the expression of luciferases in living cells and organisms: a
moter were used to detect gene activation of estrogen receptors. review. Luminescence 17:43-74.
Here, administration of estrogen compounds led to a dose- and 9. Hardy, J., K. P. Francis, M. DeBoer, P. Chu, K. Gibbs, and C.
time-dependent activation of luciferase activity in vivo allowing H. Contag. 2004. Extracellular replication of Listeria
monocytogenes in the murine gall bladder. Science 303:851-853.
the non-invasive study estrogen induction (14). In another ex- 10. Hollingshead, M. G., C. A. Bonomi, S. D. Borgel, J. P. Carter,
ample, mice engineered with the luciferase gene under the control R. Shoemaker, G. Melillo, and E. A. Sausville. 2004. A poten-
of an androgen-dependent promoter were used to detect androgen- tial role for imaging technology in anticancer efficacy evaluations.
dependent gene regulation. In these mice, testosterone treatment Eur. J. Cancer 40:890-898.
11. Jacks, T., L. Remington, B. O. Williams, E. M. Schmitt, S.
increased luciferase expression, whereas anti-androgenic com-
Halachmi, R. T. Bronson, and R. A. Weinberg. 1994. Tumor
pounds decreased the bioluminescent signal (18). Bioluminescence spectrum analysis in p53-mutant mice. Curr. Biol. 4:1-7.
also was used to study in vivo angiogenesis. Transgenic mice with 12. Jawhara, S. and S. Mordon. 2004. In vivo imaging of biolumi-
the vascular endothelial growth factor-2 (VEGFR2) promoter nescent Escherichia coli in a cutaneous wound infection model for
fused to the luciferase gene were used to study VEGFR2 gene ex- evaluation of an antibiotic therapy. Antimicrob. Agents Chemother.
48:3436-3441.
pression, an indicator of angiogenesis, during wound healing (26). 13. Kanerva, A., K. R. Zinn, K. W. Peng, T. Ranki, L. Kangasniemi,
T. R. Chaudhuri, R. A. Desmond, M. Wang, K. Takayama, T.
Future Directions Hakkarainen, H. Alfthan, U. H. Stenman, D. T. Curiel, and
BLI allows investigation of many biological and disease pro- A. Hemminki. 2004. Noninvasive dual modality in vivo monitor-
ing of the persistence and potency of a tumor targeted condition-
cesses in vivo. Bioluminescence has been utilized successfully for ally replicating adenovirus. Gene Ther. Sept. 23.
noninvasive monitoring of bacterial or viral infection and repre- 14. Lemmen, J. G., R. J. Arends, A. L. van Boxtel, P. T. van der
sents an attractive approach to investigating the host–pathogen Saag, and B. B. van der. 2004. Tissue- and time-dependent es-
interaction in vivo. The ability to track tumor burden and meta- trogen receptor activation in estrogen reporter mice. J. Mol.
Endocrinol. 32:689-701.
static disease and to assess anticancer therapy represents an
15. Lipshutz, G. S., C. A. Gruber, Y. Cao, J. Hardy, C. H. Contag,
important application of this technology, but new applications and K. M. Gaensler. 2001. In utero delivery of adeno-associated
are rapidly evolving. For example, we are investigating viral vectors: intraperitoneal gene transfer produces long-term
neurodegeneration in the central nervous system by using BLI. expression. Mol. Ther. 3:284-292.
The rapid diffusion of luciferin across the blood–brain barrier (4) 16. Luker, G. D., J. P. Bardill, J. L. Prior, C. M. Pica, D. Piwnica-
Worms, and D. A. Leib. 2002. Noninvasive bioluminescence im-
makes possible in vivo BLI studies of neurodegeneration. Ge- aging of herpes simplex virus type 1 infection and therapy in liv-
netically engineered mice have been developed in which the lu- ing mice. J. Virol. 76:12149-12161.
ciferase gene is under the control of promoter elements of genes 17. Macleod, K. F. and T. Jacks. 1999. Insights into cancer from
activated upon neuronal injury. Bioluminescence can be used to transgenic mouse models. J. Pathol. 187:43-60.
18. Malstrom, S. E., O. Tornavaca, A. Meseguer, A. F. Purchio,
identify and track degenerating neurons in these engineered
and D. B. West. 2004. The characterization and hormonal regula-
mice, whose neurons express luciferase when injured. tion of kidney androgen-regulated protein (Kap)-luciferase
The capability to monitor a biological or disease process longi- transgenic mice. Toxicol. Sci. 79:266-277.
tudinally in the same mouse is a pronounced advantage of BLI 19. McElroy, W. D., H. H. Seliger, and E. H. White. 1969. Mecha-
technology. Sequential chronological analysis not only reduces nism of bioluminescence, chemiluminescence and enzyme function
in the oxidation of firefly luciferin. Photochem. Photobiol. 10:153-
the number of animals needed but allows for collection of robust 170.
scientific data. In addition, the pathophysiology of disease may 20. Paroo, Z., R. A. Bollinger, D. A. Braasch, E. Richer, D. R. Corey,
be better understood through the analysis of disease progression P. P. Antich, and R. P. Mason. 2004. Validating bioluminescence
or regression in individual animals. imaging as a high-throughput, quantitative modality for assessing
tumor burden. Mol. Imaging 3:117-124.
21. Spibey, C., P. Jackson, and K. Herick. 2001. A unique charge-
coupled device/xenon arc lamp based imagining system for the ac-
References curate detection and quantitation of multicolour fluorescence. Elec-
1. Bhaumik, S. and S. S. Gambhir. 2002. Optical imaging of Renilla trophoresis 22:829-836.
luciferase reporter gene expression in living mice. Proc. Natl. Acad. 22. Troy, T., D. Jekic-McMullen, L. Sambucetti, and B. Rice. 2004.
Sci. USA 99:377-382. Quantitative comparison of the sensitivity of detection of fluores-
2. Contag, C. H. and M. H. Bachmann. 2002. Advances in in vivo cent and bioluminescent reporters in animal models. Mol. Imaging
bioluminescence imaging of gene expression. Annu. Rev. Biomed. 3:9-23.
Eng. 4:235-260. 23. Van Dyke, T. and T. Jacks. 2002. Cancer modeling in the modern
3. Contag, C. H., P. R. Contag, J. I. Mullins, S. D. Spilman, D. K. era: progress and challenges. Cell 108:135-144.
Stevenson, and D. A. Benaron. 1995. Photonic detection of bac- 24. Wetterwald, A., G. van der Pluijm, and I. Que. 2002. Optical
terial pathogens in living hosts. Mol. Microbiol. 18:593-603. imaging of cancer metastasis to bone marrow: a mouse model of
4. Contag, C. H., S. D. Spilman, P. R. Contag, M. Oshiro, B. minimal residual disease. Am. J. Pathol. 160:1143-1153.
Eames, P. Dennery, D. K. Stevenson, and D. A. Benaron. 1997. 25. Zeamari, S., G. Rumping, B. Floot, S. Lyons, and F. A. Stewart.
Visualizing gene expression in living mammals using a biolumi- 2004. In vivo bioluminescence imaging of locally disseminated co-
nescent reporter. Photochem. Photobiol. 66:523-531. lon carcinoma in rats. Br. J. Cancer 90:1259-1264.
5. Cook, S. H. and D. E. Griffin. 2003. Luciferase imaging of a neu- 26. Zhang, N., Z. Fang, P. R. Contag, A. F. Purchio, and D. B.
rotropic viral infection in intact animals. J. Virol. 77:5333-5338. West. 2004. Tracking angiogenesis induced by skin wounding and
6. Doyle, T. C., S. M. Burns, and C. H. Contag. 2004. In vivo biolu- contact hypersensitivity using a Vegfr2–luciferase transgenic
minescence imaging for integrated studies of infection. Cell mouse. Blood 103:617-626.
Microbiol. 6:303-317.

634