When it comes to the promise of a better future for cancer patients, few are in doubt that genomics will lead the way. Ever since the completion of the Human Genome Project, the world has waited for discoveries that promise to move cancer treatment beyond radiation and surgery and into a world where tumors are dealt with deftly at the molecular level. But it seems that the more genomics data there is, the more difficult it becomes to live up to that promise. Locating the chinks in cancer's armor is not, surprisingly, the biggest problem — figuring out exactly how to morph these discoveries into effective therapeutic strategies for the clinic is. "The major challenges to translating genomics into medicine are to figure out which genes are important in which diseases and why," says David Baltimore, a 1975 Nobel laureate and a professor at the California Institute of Technology. "There are countless efforts going on worldwide to associate particular genes with particular diseases where the underlying cause of those diseases is not clear. … That is going slowly because we're finding too many genes — many diseases are associated with 20, 30, or 40 genes — and that makes it difficult to know how to focus therapeutic development."
This has not discouraged Baltimore from serving as director of the newly formed Joint Center for Translational Medicine, a research institute established by Caltech and the University of California, Los Angeles, last December. The first order of business for JCTM is to continue work that began in 2006 as a joint research effort between the two schools. The researchers on that project are investigating a gene therapy-based treatment method for late-stage melanoma. Their technique involves extracting normal lymphocytes from the patients and infecting them with a retrovirus encoding a T-cell-receptor gene, which then activates the lymphocytes into tumor recognition. While for some time, gene therapy, or "gene transfer" as Baltimore prefers to call it, has been relegated to the fringes of translational cancer research, that is something that JCTM investigators are looking to change. "Gene therapy has been a dream — but a not realized dream — for many, many years and is still practiced for only a very small group of diseases where it can be life-saving. It's only done in academic medical centers and has no commercial potential that would take it out of the realm of the orphan diseases into something more mainstream," Baltimore says. "Late-stage melanoma is a very difficult disease to stop, so what we need to do now is follow up on what looks like a promising beginning and add it to other modalities of control, because cancer is such a slippery disease."
More and more, it's becoming clear that it will take concerted efforts from groups specifically focused on moving basic cancer research to the bedside to make translational research deliver on its promise. Research sites such as the Translational Genomics Research Institute, an early bearer of the translational genomics flag, are where the work of cancer investigators can really flourish. Raoul Tibes, an associate investigator in the Clinical Translational Research Division at TGen, is using array CGH and RNAi to characterize leukemias and solid tumors, aiming to pinpoint possible therapeutic targets. Tibes and his colleagues at TGen are among the few teams that have developed an approach using high-throughput, lipid-based RNAi transfection of myeloid suspension cells. "This approach has given us a tremendous platform where we can interpret hundreds of thousands of genes on the functional level. This is exciting because the disease biology that we see in our experiments is amazing," says Tibes, who also leads TGen's Leukemia Research Team. "Based on this work, I'm currently working on initiating a few clinical trials and there's great interest from the various entities. Hopefully within the next couple of months, we will be able to implement one of those clinical trials, a phase one, with leukemia patients."
[ pagebreak ]
Tibes, who is an oncologist by training, says that in order to derive clinical applications for cancer from the massive amount of genomics data that has flooded the bench, more funding is needed for researchers to think creatively about translational approaches. More essential still is the need to increase the number of people who can speak the language of both the lab and the clinic — or have teams in place that can do the translating part of translational research. "What I see from the PhD side is researchers saying to doctors, 'Just give me the patient samples.' But that is the wrong approach," Tibes says. "The physicians say, 'If you just need me just for my samples, forget it! I know what I'm treating, you're not giving me anything yet!' Researchers should sit down with the physicians [and] ask about what they see. Let's work on how we design experiments and become truly equal partners — it seems easy and straightforward, but it's not."
A member of Tibes' team, James Bogenberger, was recently awarded a three-year, $150,000 postdoctoral fellowship by the American Cancer Society to further develop translational approaches to combat acute myeloid leukemia. AML is quite lethal, with only about 10 percent of patients surviving more than three years from date of diagnosis. Bogenberger is attempting to identify targets that increase the anti-leukemic potency of 5-azacytidine and suberoylanilide hydroxamic acid, two drugs that have already been approved by the US Food and Drug Administration. "Then from there, translation in the clinic depends on multiple factors. In a case where identified targets already have inhibitors in the clinic, our data can then result in a design of clinical trials more immediately," Bogenberger says. "If the targets don't have inhibitors available, then our data could facilitate the design of novel therapeutics, potentially. Ultimately, we want to find rational combination therapies that can increase the efficacy of these compounds. In terms of treatment, the goal is always better [to have] outcomes with less toxicity."
A diverse foe
The more researchers bring genomics technology to bear on cancer, the more they discover why the disease is so challenging to treat effectively. An example of this is breast cancer: it is usually regarded as having five subgroups, but there is a considerable amount of heterogeneity within each subgroup, which means that clinicians cannot paint with a broad brush when choosing their treatment approach. In order to better understand how these subgroups will react to treatment, Christina Curtis, a research associate at the Cambridge Cancer Centre in the UK, is taking a focused approach. She is drawing on a set of 1,000 fresh-frozen breast cancer tumor samples, combined with several years of clinical history data for each sample, to sort out which subgroup of patients is more likely to respond to particular therapeutic targets and inhibitors. In one study, Curtis and her colleagues identified a subgroup of basal-like tumors most likely to respond to mitotic inhibitors — an encouraging discovery given that this is an aggressive cancer. "If we can pinpoint the genomic aberrations and the molecular differences that characterize those subgroups, then we have better options for treatment because it's clear that the agents we're using are not always as effective as we would like them to be," Curtis says. "We have multiple tiers of genomic information, copy number expression, and eventually microRNA and methylation, but we're also sequencing these cases and looking at their mutational profiles. It would be good to know which subgroups have a poor outcome and which do not, so we don't put patients through unnecessary treatments."
Curtis points out that there isn't just heterogeneity among tumors but also within the tumors themselves, something that is only now being explored with next-generation sequencing and other approaches. She says that in order to really tackle the intra-tumor heterogeneity and find something that can really inform clinicians, large-scale studies need to be combined with the deep sequencing of tumor samples. "We need to focus on a broad perspective because it's still the case that having a really large study where cancer cells are profiled using the same [microarray] platform gives us new hints for the clinical approaches. However, I don't think we can stop there," Curtis says. "If we want to understand clonal heterogeneity within tumors, we have to probe cancer genomes at single-molecule resolution. ... This will only be practical when single-cell genomics comes of age."
Skipping the valley
The gap between where basic research ends and where pharma picks up is what translational research is supposed to fill. "My view is that I don't like to separate excellent basic from excellent translational research because they're one and the same — translational research is just truly excellent basic research applied in a more focused fashion," says Robert Scheinder, associate director of translational research and co-director of breast cancer research for the New York University Cancer Institute. "Let me put it in perspective: a great deal of the excellent basic research is indeed moving into the clinic, either from an individual's lab or from a company. What we've learned about signaling pathways or receptors on the surface of cells has really hit a tipping point."
Scheinder recently helped lead a team of researchers that demonstrated the potential of using molecular profiling of pretreatment biopsies to identify markers of response. The results of their study, published earlier this year in Clinical Cancer Research, suggest that immune signaling molecules such as DEFA and MAP2 have considerable utility as tumor markers that associate with response to neo-adjuvant taxane-based chemotherapy. For Scheinder, who has first-hand experience of taking basic research to the drug discovery phase, the biggest barriers are the limitations of mouse models and lack of funding. "Most of the cancer models in mice very poorly reflect human disease because they model local disease of primary tumors, so we end up positioning drugs in our models incorrectly because we're doing metastatic cancer generally, and yet we're testing them in mice," Scheinder says. "The other problem is just the staggering cost of moving something into the clinic. Just to do a phase one clinical trial, assuming you already have an agent, takes millions of dollars — the costs rapidly escalate beyond the reach of most institutions' ability ... to do this without a corporate partner."
[ pagebreak ]
According to some researchers, the pharmaceutical industry's reluctance to take more risks has created a no-man's land between where their ideas of what is worth spending capital on begins and the point at which good academic cancer research ends. Lynn Matrisian, a professor at the Vanderbilt-Ingram Cancer Center and former co-chair of the NCI's Translational Working Group, has observed a steady increase during the last few years in the amount of assurances needed by pharmaceutical companies that a drug target is going to work before they will even consider investing. "Now you end up with a gap in the middle because they want to see that it isn't toxic and they want some phase one clinical trial data [before they conduct their own trials]. Then you have to get all the way through all those toxicology studies and making clinical-grade material — all the things that academics haven't traditionally done," Matrisian says. "It's this valley of death where the academics take it up so far, but the pharmaceutical companies — in order to do development and the really large clinical trials that need to be done — want to see data that's further than where the academics can take it, so that's where things are falling down."
But rather than spinning one's wheels trying to find new drug targets, Matrisian says that translational cancer research can more immediately benefit from genomics technology if researchers focus on pathways targeted by existing drugs. "We're getting all kinds of information on the mutations that are common in certain kinds of cancer that lead to target identification. For some of those there are drugs and that's fabulous," she says. "If that happens, then you have shortened that whole process because the drug has already been approved, so it's already gone through that hard step of making a good drug, testing it in humans, and making sure it's safe — that's where genomics can be a very powerful tool."
An effort aimed at transforming academic research into real-world treatments for cancer may just provide some light at the end of the translational tunnel. The National Cancer Institute's Division of Cancer Treatment and Diagnosis together with the Center for Cancer Research recently launched a joint effort, called the NCI Experimental Therapeutics program, which will serve as a drug discovery and development pipeline focusing on therapeutics that are not being pursued by either the pharmaceutical industry or the biotechnology sector. Part of the goal of the NExT program is to shorten the drug development cycle, which usually takes about 10 to 12 years, by at least one year. The idea is not simply to replicate the current drug discovery process of pharmaceutical companies but to focus on bringing to fruition research that would otherwise fall through the cracks — such as projects that focus on rare or pediatric cancers.
"We can actually look at orphan diseases and fill a niche that pharma or biotech isn't addressing by taking that high risk. We want to look at interesting therapeutic targets that pharma is not looking at and try to push it forward with new technologies," saysBarbara Mroczkowski, special assistant to the director of the Division of Cancer Treatment and Diagnosis at NCI. This new program will take tumor data coming out of the academic sector and aim to translate it into the clinic by conducting early target validation using genomics to move it all the way into clinical trials. The NExT program is all--encompassing, with a complete drug development pipeline that involves not only early target validation, but new imaging modalities and biomarkers to aid in the understanding of clinical outcomes as well. "I think it's the government's job to lead the way because we can come in and not just develop new drugs but get a better understanding of what combinations we should be advancing into the clinic," Mroczkowski says.
The new program, which issued its first call for proposals last fall, is also unique for a government consortium in that it has external and internal panels for evaluating applications from industry, non-profit organizations, and academic institutions. "There has been overwhelming interest from the outside community indicating that NCI is filling a critical void by bridging the gap from the lab to the clinic," Mroczkowski says. "There is definitely a need in the academic sector for helping to translate novel therapeutics."