Main links:
Home | About us | News and publications | Centre-led programmes | Funding schemes | Information portal
Site map | Help | Contact us | Media Centre | Accessibility | Disclaimer | Acknowledgements | Glossary
Dr Kevin P Francis, Xenogen Corporation
Non-invasive, whole body imaging of small animals using techniques such as X-ray, CT, SPECT, PET and MRI has become commonplace in both academic and commercial research settings and, to some degree, has helped to reduce the number of animals used in basic research and preclinical drug development (1-3). However, because the majority of these techniques involve the use of radiation and/or low throughput, expensive imaging equipment, the extent of this reduction is somewhat limited, making conventional animal modelling techniques more attractive from both a practical and an economic standpoint.
Biophotonic imaging (BPI) of small animals is a technique that does not involve radiation, is relatively high throughput due to short imaging times (1s - 5 min with up to five animals per session), and is comparatively low cost. BPI allows biological processes, including gene expression that is both temporal and spatially defined, to be monitored longitudinally in live animals in real-time and non-invasively (1-3). The principle of BPI is that light (especially that above 600 nm, red light) passes through tissue at depth (up to several centimetres) and can be detected from outside of a live animal (4-6). To utilize this process, genes encoding specific photon emitting proteins (luciferase or fluorescent proteins) are engineered into viruses, cells (bacteria, fungi and cancer cell lines), protozoa and animals (mice and rats) enabling them to emit light that can be visualised through the tissues of a live subject using specialized imaging equipment (Figure 1). To date, this technology has been used predominantly to develop small animal models (predominantly rodents) that can facilitate drug discovery and basic research in areas such as oncology, infectious disease, inflammation, endocrinology, cardiology, metabolic disease, neurology, toxicology, gene therapy and stem cell research (7-27).
Figure 1. Cells, infectious agents or genes can be
labelled with bioluminescent or fluorescent markers, such as
firefly luciferase or GFP (A). Small animals (e.g., rodents)
into which these labelled cells or genes are introduced (B) can
be imaged non-invasively using a highly sensitive CCD based
imaging system (C) to allow the photonic signature of these
labelled entities to be seen and quantified using data analysis
software (D). Printed with permission from Xenogen
Corporation. All rights reserved.
Bioluminescence is a natural biological process by which
certain organisms can generate light through an enzyme-mediated
reaction. Firefly and glow worm are probably the most familiar
examples of this phenomenon, with both of these insects
producing visible light. The application of these
bioluminescent systems to monitor gene expression in cells is
now routine in molecular and cellular biology, as is the use of
fluorescent proteins such as GFP. Typically, the gene(s)
encoding the luciferase or fluorescent protein is cloned into a
cell adjacent to the region of a gene controlling expression
(the promoter), such that the reporter protein is produced in a
fashion similar to that of the native protein. BPI can be
monitored from cells containing these proteins using a light
sensitive detector, such as a luminometer or fluorometer.
Alternatively, the light source (e.g. bioluminescently
engineered cell) can be placed inside a live, intact animal and
monitored; a technique termed in vivo biophotonic imaging. In
the same way that bioluminescent light is transmitted from
cells within the firefly tail, so light emitted from
bioluminescently engineered cells (e.g. cancer cells or
transgenic tissue) placed or generated within a small animal
can be detected at the surface of the subject using a suitably
sensitive detector (e.g. CCD camera) and image processing
software (4). Fluorescent cells can be detected in a similar
manner by using an additional excitation source at the surface
of the animal (5).
Typically, the bioluminescent light generated by genetically engineered cells can penetrate 2 - 3 cm of tissue (at sufficient cell densities) making rodents ideal subjects in which to monitor such activity (4). The location and number of such cells can then be tracked in the live animal. Moreover, the same animal may be imaged multiple times, not only allowing the expansion or regression of the disease to be followed more accurately, but allowing a significant reduction in the overall number of animals used in the study (in some cases by 80% or more) (Figure 2).
Figure 2. BPI allows the same small group of animals to be anaesthetised and imaged at consecutive time points throughout the experiment (lower panel), as opposed to conventional methodologies that require the sacrificing of multiple groups of animals (upper panel). Printed with permission from Xenogen Corporation. All rights reserved.
Biophotonic imaging is unique in that it can be applied to monitor virtually any biological process in real-time in a live animal. Furthermore, because different luciferases and fluorescent proteins emit light at different wavelengths (and in the case of luciferases, use different substrates), it is possible to monitor two biological events in the same animal at the same time (5,12,20).
The pharmaceutical industry currently invests significant time, money and animals lives into testing the efficacy and safety of its drugs and compounds under development (1). On average it takes 10 to 15 years to bring a drug to market, with one of the major bottlenecks in this drug development process being the requirement for preclinical animal studies. These essential in vivo experiments are not only laborious and expensive, but necessitate the sacrifice of large numbers of animals, usually rodents, in experimental procedures that are becoming increasingly more controversial. Being able to reduce the number of animals used in these studies without reducing the quality of the data obtained would be extremely advantageous, especially from an animal welfare perspective.
Using BPI it is possible to develop predictive assays with the capacity to replace conventional animal studies that use death as an endpoint. The idea behind this approach is relatively straightforward. If an animal has a disease that is rapidly expanding in an uncontrolled manner, it is likely that that disease will eventually lead to the death of the animal. Thus, labeling the disease with a BPI marker will allow the changing photon signal to be monitored from outside of the animal so that an accurate prediction of the speed and magnitude of the diseases expansion or regression can be recorded and graphed. The data obtained can then be used to draw a best fitting straight line and a slope, which in turn can be used to estimate whether an animal will survive or die. We have used such an approach in a murine sepsis model employing bioluminescently engineered bacteria (e.g. Pseudomonas aeruginosa and Staphylococcus aureus ) (Figure 3). Using data from only the first 12 hours, we were able to predict the death or survival of individual animals. Additionally, we were able to estimate the hour of death for those mice predicted to die.
Figure 3. BPI of a single mouse infected inter-peritoneally with a bioluminescently engineered strain of Pseudomonas aeroginosa . The rapidly increasing biophotonic signal can be graphed to give an accurate prediction of the animals survival or death (including the hour of death), thus allowing it to be humanely euthanised rather than relying upon its death to mark the experimental endpoint (as in the case of antibacterial drug efficacy studies). Printed with permission from Xenogen Corporation. All rights reserved.
As with most BPI studies, fewer animals are used because of the ability to obtain quantitative data from an individual animal at multiple times throughout the experiment. This capacity negates the need to repeatedly sacrifice groups of animals for data collection at each time point. Moreover, the predictive nature of the assay allows for humane euthanasia after as little as twelve hours, rather than requiring death as an endpoint. Another less obvious advantage of this methodology is the ability to improve on the quality of the data obtained over conventional methodologies. One reason for this improvement is that the bioluminescent signal can be seen at the initial time point (t = 0), allowing mice that are improperly infected to be removed from the study, thus increasing both the reliability and reproducibility of the data.
Using BPI it is possible to observe disease processes as, when and where they occur in a live intact animal, in some cases allowing unidentified pathologies to be seen for the first time (e.g. 15). Moreover, because whole live animals can be observed repeatedly over long periods of time (several months) it is possible to tell whether the treatment of a particular disease is successful or whether disease relapse will occur. This ability has been shown to be particularly useful for monitoring and treating infectious diseases and cancers in animals, including micrometastases that would otherwise require histopathology for an accurate evaluation. BPI allows therapies to be assessed before, during and after treatment, offering an excellent preclinical strategy to assess disease response and relapse (9) (Figure 4).
Figure 4. Bioluminescently engineered prostate cancer cells (PC-3M) were introduced orthotopically into the prostate gland of a small group of mice. Subsets of these mice were then treated with saline (control group), 5-FU or mitomycin C. A single representative animal from each of these three sets of mice is show in the above figure. The animal treated with mitomycin C shows an initial response to the drug at week 5, indicating this compound to be efficacious. However, because there is no requirement to sacrifice the test subject to gain this data, the same mouse imaged at week 7 shows the cancer to have once again expanded in the animal, indicating disease relapse to have occurred. Printed with permission from Xenogen Corporation. All rights reserved.
To date, most BPI models have relied upon the use of bioluminescent or fluorescent labelled cell lines to monitor a disease process or its treatment in vivo. However, biophotonic reporter genes can also be incorporated directly into an animals genome (e.g. using pronuclear microinjection) to allow a specific BPI transgenic animal to be generated (e.g. 17,23,24,26). Such transgenic animals have proven to be particularly useful in allowing a distinct disease process to be visualised from the hosts perspective. For example, rather than simply monitoring the arbitrary growth of a cancer cell line in vivo, a BPI transgenic animal can be generated to allow the cross-talk between the developing tumour and the animal to be seen more clearly. This approach can allow for a particular gene that is being up or down regulated in the animal during oncogenesis (e.g. a gene involved in angiogenesis, apoptosis, or spontaneous tumor formation) to be tracked over time, and for compounds with the potential to alter this gene expression to be screened accordingly (17) (Figure 5). Moreover, by differentially labelling the cell line and the transgenic animal with dissimilar BPI reporters (e.g. renilla luciferase and firefly luciferase), both oncogenic events can be monitored simultaneously (e.g. tumour growth and angiogenesis).
Figure 5. Lewis lung cancer cells were injected subcutaneously into the thighs of a transgenic mouse containing a firefly luciferase gene downstream of the VEGFR2 promoter. Bioluminescent signals are indicative of angiogenesis occurring around the periphery of the expanding dark tumour. Printed with permission from Xenogen Corporation. All rights reserved.
In addition to oncology applications, transgenic animal models containing luciferase and/or a fluorescent reporter gene have been developed and utilized for drug screening and basic research in a range of different disease areas, and have been shown to be particularly appealing for monitoring disease processes in inflammation, endocrinology, toxicology and neurology (23,24,26), including damage to the CNS caused by infectious agents (Figure 6).
Figure 6. Bioluminescent signals recorded from transgenic mice in which the glial fribrillary acidic protein (GFAP) promoter driving firefly luciferase is upregulated in response to astrocyte damage caused by the administration of Kainic acid (A) or Streptococcus pneumoniae (B). Printed with permission from Xenogen Corporation. All rights reserved.
Over the past decade, the majority of BPI studies conducted in animals have been to monitor the expansion/regression of a disease or the expression of a target gene during a disease process. Monitoring light (photons) at the surface of an animal using a CCD based imaging system has enabled researchers to gain valuable information regarding the disease processes occurring within the animal. However, identifying the exact location of the source of the disease or gene expression has been somewhat more challenging, usually necessitating the sacrifice of the animal and ex vivo analysis of organs and tissues. Thus, in order to facilitate the spatial visualisation of a disease within a living animal, spectral and multidimensional BPI systems have been developed (Figure 7). These 3D BPI systems have the capacity to determine the location of multiple sites of infection/gene expression to within a few millimetres, approach the special resolution achievable by PET and MRI (1-3). Moreover, in comparison to these other whole body imaging modalities, BPI can much more readily determine when, where and to what degree a particular gene is expressed. Furthermore, although 3D BPI requires longer imaging times than 2D BPI, it is still significantly faster and less expensive than either PET or MRI.
Figure 7. 3D reconstruction of a
phantom mouse containing a single internal point-source of
light near its centre. Phantom mouse was imaged lengthwise in
eight sequential fields of view, each 45 degrees apart from one
another. Shown is a composite image (using 3D software)
allowing the location of the point-source to be estimated to
within 1 - 2 mm. Printed with permission from Xenogen
Corporation. All rights reserved.
Further information can be obtained by contacting:
Kevin P. Francis Ph.D.
Senior Director of Technical Applications
Xenogen Corporation
860 Atlantic Avenue
Alameda, CA 94501
Tel: 510 282 1110
Email: kevin.francis@xenogen.com
Website: www.xenogen.com
All views and opinions expressed in this article are those of the author and do not necessarily reflect the views and opinions of the NC3Rs.
In this section:
Banners:
NC3Rs | 20 Park Crescent, London W1B 1AL | Tel: 020 7670 5331 | Fax: 020 7670 5178 | Email: enquiries@nc3rs.org.uk