MI explores tumor microenvironment

June 4, 2005
Charles Bankhead

Solving the mysteries of the hostile microenvironment of tumors serves as a research focal point in the laboratory of Robert Gillies, Ph.D., a professor of biochemistry and molecular physics and radiology at the University of Arizona.

Cancer is often described as a genetic disease, as it invariably contains genomic alterations or mutations. These mutations are selected during cancer progression, which occurs against a backdrop of environmental alterations, Gillies said.

The microenvironment within tumors differs significantly from that of normal tissue. One major difference is tumors' chaotic vasculature, which results in an unbalanced blood supply and significant perfusion heterogeneities. Many regions within tumors are transiently or chronically hypoxic, and this exacerbates tumor cells' natural tendency to overproduce acids, resulting in very acidic pH values.

"We know that the environment is acidic and hypoxic," Gillies said. "We're trying to mechanistically understand the interrelationships that exist between vascular perfusion, low pH, and low oxygenation. We're trying to understand the molecular controls that are involved."

The research at the University of Arizona makes extensive use of molecular imaging techniques in an effort to develop methods for imaging pH and hypoxia. Imaging includes contrast-enhanced MRI for measurement of pH and vasculogenesis, and optical imaging, using luciferase and fluorescent protein reporter genes driven by hypoxia, for hypoxia imaging.

"A secondary aspect of the work is to understand the consequences of the low pH and hypoxia," he said. "We're using imaging to set boundary conditions for what the environments might be in tumors. Then we recapitulate these in vitro to investigate the effects on the biology of cells."

The list of potential consequences of low pH and hypoxia is long because the two conditions have such ubiquitous and far-reaching effects, Gillies said. One key effect is resistance to therapy. Gillies and others believe that low pH and hypoxia also have an impact on cancer progression; specifically, that the hostile microenvironment creates environmental selection pressure for cells that have a particular phenotype (Nat Rev Cancer 2004;4:891-899).

Gillies and his colleagues have demonstrated that therapeutic resistance can be reversed by raising the environmental pH. Other laboratories have tried to reverse tumor hypoxia, with mixed results.

Low pH and hypoxia lead to more aggressive tumors that are more likely to metastasize, according to Gillies. One line of investigation in his laboratory involves monitoring metastasis in vivo and ex vivo. Reporter genes are used for the in vivo work and automated image analysis for the ex vivo studies.


A second major area of investigation involves use of noninvasive imaging to monitor response to chemotherapeutic agents. The rationale for the work is improved efficiency of drug development.

"Currently, a new anticancer drug is estimated to cost $1.7 billion," Gillies said. "It is expected that this cost can be reduced if we can discover valid biomarkers for response. Imaging has a lot of promise as a response biomarker."

The laboratory has developed a model that includes a fixed number of imaging end points in animal tumor xenografts, some of which can be immediately translated into the clinic, and some of which cannot. The end points are compared for their reproducibility, the magnitude of change induced by the drug, and the relationship of the change to the ultimate response to therapy.

"We have an animal model that is responsive to a drug, and we'll also use a model that has been shown to be resistant to the drug," Gillies said. "We treat the animals with the drug and then monitor the time course and dose response with imaging."

The work involves diffusion MRI, dynamic contrast MRI, MR spectroscopy, blood oxygen level-dependent (BOLD) contrast MRI, and PET imaging with FDG.

The first phase of the investigation is concerned with evaluating change in a tumor as a function of time after a single dose of a drug. The different end points are compared, and the one that produces the highest magnitude of change on imaging will be proposed when that investigative drug enters clinical trials. The second phase of the work involves relating imaging changes to clinical response, typically measured as a delay in tumor growth. Currently, only clinical evaluation using diffusion MRI is being done.

"This is a fairly active program that is primarily interested in drugs under development in preparation for phase I and II clinical trials," Gillies said. "During phase I and II, we take pretherapy images of patients, followed by post-therapy imaging, using the end point determined in the animal models. The expectation is that analysis of images before therapy might be a predictor of which patients will ultimately respond to the drug."

If imaging can provide a predictive biomarker, it would be carried into phase III and would greatly reduce the cost of phase III clinical trials, he said.


The third major research program in Gillies' laboratory involves what he calls reverse engineering of targeted ligands. Target identification uses genomic and proteomic analyses to determine the epitope expression pattern of target cells. These are compared with the epitope expression patterns in normal human tissues derived from autopsy.

Ligands that bind to these targets are identified though a novel high-throughput approach. They are then modified to contain chelating groups that are appropriate for attaching either imaging or therapeutic cargo. The potential clinical application of this work would be targeted or molecular surgery.

"If you have a targeted cell and you want to deliver a payload to it, such as a chemotherapeutic or radiotherapeutic agent, you use an imaging dose to identify whether you have appropriate targeting," Gillies said. "If you do, you just add a therapeutic dose to the same patient."

Other researchers have asked to join the ligand work with their own particular targets. Gillies and his colleagues have been working with pancreatic cancer and melanoma, and they will expand the program into ovarian cancer and gliomas through new collaborations.

Zaver Bhujwalla, Ph.D., director of the molecular imaging lab at Johns Hopkins University, has collaborated with Gillies for several years. Bhujwalla has a particular interest in tumor pH and hypoxia and their role in cancer cell invasion and in the vascular characteristics of tumors. Gillies' research has been key to understanding the tumor microenvironment, she said.

"A critically important observation he has made is that pH in cancer cells is very well regulated," Bhujwalla said. "For many years, people thought that tumor pH is purely acidic. His work has clearly illustrated that tumor cells have a normal to alkaline intracellular pH and an acidic extracellular pH. (Gillies) is pursuing this therapeutically to see if this difference in pH can be utilized so that more of the drug gets into the cell."