February 1, 2006

Seeing the shimmer of biology in action

Creatures that glow lend their proteins to biomedical research

Field of fireflies near New London, Conn.

Courtesy of James E. Lloyd, Ph.D., University of Florida

It’s a scene that says summertime—sparks of light in the gloaming and children darting after them, Mason jars at the ready. The captured fireflies might be released or they might be tortured. They are sure to be admired.

The glow of fireflies—called lightning bugs in some regions of the country—is a source of wonder to children and adults alike. Over the past decade, scientists have discovered how to harness this biological glow, called bioluminescence, to reveal secrets from inside living animals. The chemical reaction that produces light can be used to follow cancer cell metastasis, stem cell migration, gene expression, and protein activity, all as they are happening in vivo.

Bioluminescence imaging is part of the burgeoning field of molecular imaging, which aims to “see” not just anatomy, but specific molecular or cellular processes, says Sanjiv Sam Gambhir, M.D., Ph.D., director of the Molecular Imaging Program at Stanford University.

“The goal is to do this as non-invasively as possible so that one can interrogate a living subject repeatedly over time,” Gambhir says. “Ultimately, we want to fundamentally change the way in which we diagnose and manage disease by really looking at molecular information.”

Let there be light

Fireflies are not alone in their ability to generate light—marine organisms including jellyfish, sea pansies and squid, along with various worms, fungi and bacteria all possess the biochemistry to shine. They glow to signal interest in courtship and mating, lure prey, defend, camouflage, and respond to stress.

Light production depends on the presence of a protein enzyme called a luciferase, from the Latin “lucem ferre”—bringer, or bearer of light. The luciferase performs a biochemical reaction on its substrate—luciferin for the firefly protein—usually requiring energy, oxygen and other co-factors, with the end products including the release of a single photon of light.

Of the wide variety of luciferases, the protein from the firefly has been most commonly used in biological research. It was first purified and characterized 30 years ago, and it gained widespread exposure as an “optical reporter gene” for cells in culture beginning in the late 1980s.

"The greatest contribution to human medicine that probably all molecular imaging approaches, including bioluminescence, will have is in refining and accelerating our animal models of disease, such that we can test and develop drugs more efficiently."
Christopher H. Contag, Ph.D., co-director of the Molecular Imaging Program at Stanford University

Such luciferase assays to study gene regulation in cultured cells were in full swing when Christopher H. Contag, Ph.D., got frustrated with the methods for studying infectious diseases. Contag, a virologist by training, was following mother-to-infant transmission of HIV by comparing viral genetic sequences.

“It turned out to be a very complicated analysis, and I said at the time, ‘wouldn’t this be so much easier if we could just watch the whole process?’” recalls Contag, now co-director of the Molecular Imaging Program at Stanford. “It occurred to us that we should be developing tools for watching complex biological processes in the context of living animals, and at some point in time, living humans.”

A search of the available imaging modalities and a fortuitously timed lecture by an environmental microbiologist about luminescent bacterial enzymes pointed Contag and his wife Pamela R. Contag, Ph.D., in the direction of bioluminescence.

“We figured that since animals and people don’t glow in the dark, if you put something in the body that does glow in the dark, you should get great signal-to-noise ratios since there should be relatively no background noise,” Contag says. We now know there are background signals—from biological processes in animals that produce small amounts of light—but for bioluminescence imaging, he notes, “the signal-to-noise ratio is really extraordinary.”

For their first studies, the Contags and their colleagues followed bioluminescent bacteria in a mouse.

“When we saw the first images of glowing bacteria in the intestines of a mouse, I said, ‘Every biology lab in the world will want to use this; this will be fantastic,’” Contag recalls. The team published its findings in 1995, and Contag anticipated that use of the technology “would explode.” It took a little longer than he expected, with adoption by many laboratories occurring only in the last five years.

New software is improving the spatial resolution of bioluminescence imaging.  In these test images, a luminescent bead has been implanted inside a silicon mouse model.  On the left, a "standard" planar bioluminescence image shows light scattered as it moves through the model mouse.  The image on the right shows the result of bioluminescence tomography using Xenogen's Living Image Software 3D Analysis Package.  The red pixel indicates the reconstructed light source location.

Courtesy of Jack Virostko, graduate student in the lab of E. Duco Jansen, Ph.D.

To propel what they saw as a powerful technology, the Contags and David A. Baneron, M.D., founded a company called Xenogen to market the technology, along with unique instrumentation and biological reagents that utilize luminescent signals for studying biology in animals. Pamela Contag has served as president of Xenogen since the company’s inception in 1995.

Christopher Contag remained at Stanford. “I decided to take a very broad approach and demonstrate the breadth of this technology,” he says. “So we’ve been tracking viruses and tumor cells and stem cells, and in the early days attempted to show how versatile this technology is. Now we’re focusing on stem cells and cancer biology.”

Light bulbs inside cells

Investigators at Vanderbilt embraced bioluminescence imaging early on to follow cells and gene expression in living animals. Watching cells as they migrate through a living animal, take up residence, multiply, and in the case of tumor cells, metastasize to new sites, has been the most popular application of bioluminescence to date.

“What bioluminescence gives you is a level of sensitivity of detection that is not attainable by any other current method,” says E. Duco Jansen, Ph.D., associate professor of Biomedical Engineering at Vanderbilt.

The way it works is conceptually quite simple, Jansen explains. Cells of any sort can be infected with viruses or engineered to incorporate a luciferase gene. After being injected into small animals, usually mice, the cells begin to produce the luciferase protein. Investigators then inject the substrate molecule—such as luciferin—into the animals, and the luciferase acts on it, releasing photons of light.

“So we have effectively a light bulb inside the cell,” Jansen says.

That light bulb is really quite weak—the mice do not actually glow like fireflies. But some of the photons of light do make their way out of the animal, and sophisticated charge couple device (CCD) cameras, cooled with liquid nitrogen to minimize noise, can capture them. Imaging systems such as those produced by Xenogen, which are available to Vanderbilt scientists via the new Institute of Imaging Science, make the process relatively straightforward. The systems manage everything from administration of the inhaled anesthetic to quantitation of the detected light.

Tumor cells were the early front-runners in the “cells-to-watch” category.

“It’s a pretty well-established paradigm now to incorporate a luciferase into a tumor cell line, implant those modified cells into animals, and then monitor the luciferase activity to find where the tumor cells become established and to follow the growth of the tumor,” says J. Oliver McIntyre, Ph.D., research professor of Cancer Biology at Vanderbilt.

And because of the high sensitivity of bioluminescence, the very early stages of tumorigenesis and of metastasis are open for study.

“Bioluminescence imaging lets us detect very small numbers of cells—in the hundreds—from the internal organs of a small animal,” says P. Charles Lin, Ph.D., associate professor of Radiation Oncology at Vanderbilt. “There is no other way right now to detect those cells.”

Watching tumor growth and metastasis in real time gives investigators a window to a tumor’s molecular environment and to its susceptibility to therapeutic interventions. McIntyre, who works with Lynn M. Matrisian, Ph.D., professor and chair of Cancer Biology at Vanderbilt, describes how the group has used bioluminescence to study the effect of an enzyme called MMP-9, which “chews up” the matrix material between cells, on tumor growth.

Heath Acuff, Ph.D., at the time a graduate student in the group, compared the establishment of lung tumors in mice with and without MMP-9. Luciferase-expressing tumor cells were injected into the tail vein of mice; they then homed to the lung and grew, and the investigators followed their growth by looking at the light being produced.

Mice lacking MMP-9 had fewer tumors at the end of the study, and by following the mice over time, the investigators knew this difference occurred very early, within the first 24 hours.

“The imaging really provides this temporal information from individual animals or groups of animals that is not so easy to obtain by other methods,” McIntyre says.

Shining light on a diabetes therapy

The high sensitivity of bioluminescence imaging was just what Alvin C. Powers, M.D., director of the Vanderbilt Diabetes Center, needed to track his favorite cells—those that populate pancreatic islets. Islets—so-named because they appear to be small cellular “islands” within the pancreas—are home to the insulin-producing beta cells and several other hormone-releasing cell types.

Islet transplantation is an emerging experimental therapy for type 1 diabetes and has shown promise in multiple small clinical trials. One difficulty in moving the field forward, Powers explains, is that investigators have no way to follow the islets after transplantation.

Human pancreatic islets glow after infection with a virus carrying the firefly's light-producing luciferase gene.  Number of islets in each well:  (top row, from left) 0, 1,000, 50; (bottom row) 100, 500, 1,000.  The 1,000 islets in the middle well of the top row were not infected, do not produce the luciferase enzyme, and therefore do not emit light.

Image courtesy of Alvin C. Powers, M.D.

“We really need a way to assess where the islets go after transplantation, how many survive, (and) what kinds of therapies promote islet survival,” Powers says.

In collaboration with Jansen, Powers and colleagues have “tagged” islets with luciferase. They have used primarily a strategy of infection: first the investigators harvest islets, both from mice and humans, then they infect the islets with a virus carrying the luciferase gene. A certain percentage of the islet cells incorporate the luciferase, and after transplantation into a mouse the surviving cells can be followed with bioluminescence imaging.

The team is also beginning to use islets from genetically modified mice that have luciferase in all of the beta cells of the islet. These light-emitting islets offer the advantage that all cells permanently express the luciferase.

In both models, the investigators are attempting to optimize transplantation parameters, Powers says. What is the best site for survival? Which growth factors best promote survival? Is it best to treat the islets with growth factors before transplantation, to treat the animals after transplantation, or both?

“Bioluminescence is really the only way to non-invasively assess these islets over time,” Jansen says. He notes that it would be possible to sacrifice animals to get single time-point snapshots, but that methodology would require a very large number of animals and retain the problem of biological variation between individuals. Non-invasive imaging of any sort, in the same animal over time, gives the best statistical results by avoiding interindividual variability, he says.

"Being able to image the biological event is critical in studies of drug development.  The advantage here is that bioluminescence imaging can be done in vivo and repeatedly in the same animal, which gets you better data."
E. Duco Jansen, Ph.D., associate professor of Biomedical Engineering at Vanderbilt.

Photo by Anne Rayner

And as with the luminescent tumor cells, the glowing islets offer a model system for evaluating new pharmaceutical interventions. Suppose, Jansen says, that a drug company has a new immunosuppressant drug candidate. That company can use the transplanted islet model and bioluminescence imaging to quickly assess the drug candidate’s efficacy.

“Being able to screen compounds in live animals, with a relatively high throughput readout, is critical for pharmaceutical companies,” Jansen says. “If a drug candidate can be eliminated from the pipeline earlier because of in vivo molecular imaging, that translates into huge cost savings. And likewise, if a candidate can make it to the market sooner, that means millions of dollars of added revenue. Drug companies are very interested in these small animal in vivo molecular imaging approaches.”

Watching the blaze of inflammation

Bioluminescence imaging changed the way Timothy S. Blackwell, M.D., associate professor of Medicine at Vanderbilt, thinks about lung inflammation and injury. In addition to tracking cells and bacterial pathogens, Blackwell’s group has been following gene expression in the context of inflammation.

To do this, the team engineered a “transgenic” mouse that contains the firefly luciferase gene, linked to a stretch of DNA that responds to a transcription factor—a protein that controls the expression of other genes—called NF-kappa-B. When NF-kappa-B is active in cells, it turns on the production of luciferase and the cells light up. NF-kappa-B is an important mediator of inflammatory processes.

In some of their first studies with these transgenic reporter mice, the investigators studied short-term versus long-term infusions of endotoxin, “with the idea of trying to figure out the parameters of inflammatory signaling that lead to lung injury—was it dose, timing, duration, cellular distribution—those sorts of things,” Blackwell says.

The bioluminescence imaging revealed that the duration of the inflammatory stimulus affected the final outcome, he says. A single bolus injection of endotoxin caused a peak of NF-kappa-B activation, as measured by light output, but no lung injury. The same dose of endotoxin given as an infusion over 24 hours caused a progressive and sustained activation of NF-kappa-B in the lung, and resulted in lung injury.

“That was something that would have been very time consuming to try to figure out without the use of bioluminescence,” Blackwell says, “and it has led us to other studies now trying to understand which cells are activated over this period of time and identify specific injury-provoking gene products.

“The ability to look non-invasively at NF-kappa-B activity over time in a relatively quantitative way is helping us to define the balance of factors that cause either lung injury or effective host defense against infection,” Blackwell says. “Ultimately we might be able to come up with ways to prevent injury and still maintain adequate defenses.”

These inflammation-reporting mice have been useful for Vanderbilt’s Lin as well. Lin is interested in blood vessel formation—angiogenesis—in the context of diseases including cancer, arthritis, and cardiovascular disease.

In the case of arthritis, inflammation appears to trigger excessive angiogenesis, which facilitates tissue growth and eventually causes bone damage, Lin explains. Anti-angiogenic therapies may be effective in preventing the tissue growth. Lin’s group is using multiple imaging technologies—an increasingly common approach known as multi-modality imaging—to probe the arthritic joint: bioluminescence to see the inflammation, x-ray to look at bone damage, and fluorescence techniques to visualize the blood vessels.

“We think this is a very powerful way to study how the blood vessel affects disease progression as well as what kind of therapy we can use to stop this process,” Lin says. “Imaging has really moved from traditional modes of looking at structure into functional imaging, from static into kinetic. We’re no longer satisfied looking at single point snapshots of dynamic processes.”

A cornerstone technology

While bioluminescence imaging offers excellent sensitivity for tracking cells and seeing gene expression in living animals, it suffers from poor spatial resolution. Because light is absorbed and scattered by tissues as it makes its way out of the animal, images become “fuzzy.” Jansen likens it to having a pencil in a glass of water and adding a few drops of milk—you can still make out an image, but it’s no longer clear that it’s a pencil.

“All of the imaging modalities have strengths, and they all have weaknesses,” Jansen says. “Bioluminescence is great at sensitivity, but it’s lousy at resolution. So in many cases we combine it with something like CT or MR, or even fluorescence.”

There are current attempts to improve the spatial resolution of bioluminescence imaging by collecting data at different wavelengths of light, taking a surface map of the animal and then computing the three-dimensional source distribution, Jansen says. Other attempts include using a rotating stage to image the animal from different planes.

Absorption of light by tissues poses another problem for bioluminescence. Although it is a very useful imaging technique for small animals, where it has to travel only a short distance to reach the surface, it’s not likely to translate to humans except in niche areas. One of those areas could be encapsulated cell therapies—therapeutic cells, like glucose-sensing, insulin-producing beta cells that are “contained” within a membrane of some sort and implanted just under the skin.

“Developing encapsulated cell therapies will be greatly enhanced by building into those cells markers that tell us if the cells are alive and doing what they’re supposed to be doing,” says Stanford’s Contag. “Since the cells would be under the skin, bioluminescence imaging should work. I think that’s a perfect scenario for its first clinical application.”

Another place bioluminescence imaging might find clinical use is in breast cancer detection, Contag says. Proteins tagged with luminescent markers would have the advantage of great signal to noise for detecting very small numbers of cells.

“The question is, is it going to be better than anything else out there, and we won’t really know until someone tries it,” Contag says.

Whether or not bioluminescence imaging makes it to the clinic, “my guess is that optical imaging in small animals will become the mainstay for many laboratories using animal models,” Contag says, “and bioluminescence will be one of the cornerstone technologies for developing new ways to treat disease and to visualize biology as it occurs in the living body.”

That’s bright stuff for those flashing summertime lights.

Fighting infection—the glow from within
Bioluminescence imaging reveals inflammation-related gene expression in mice infected with the pathogen Pseudomonas aeruginosa.  The images show increasing intensities of bioluminescence, from deep blue to bright white, in mice before injection of the pathogen (A), and 24 hours after injection of increasing doses of P. aeruginosa: (B), (C) and (D).  Studies like this could lead to new treatments aimed at augmenting the host defense, particularly in critically ill patients.

From Sadikot RT, et. al.; The Jornal of Immunology, 2004, 172:1801-1808.  © 2004, American Association of Immunologists, Inc.