FLICKR, ED UTHMANCancer is a devious foe, revealing new complexities just as scientists find new ways to tackle them. A recent hope has been a new generation of “targeted therapeutics” that home in on specific molecular defects in cancer cells, promising more effective and less toxic therapy than imprecise chemotherapeutic agents. However, researchers are now realizing that they may have previously underestimated one of cancer’s oldest and best-known complexities: tumor heterogeneity. This may help explain the successes and disappointments with targeted therapeutics. It should also motivate a broader re-examination of research strategies.
Cancer can occur in any organ, although it occurs most commonly in the breast, lung, colon-rectum, and prostate. Researchers have long known about two types of microscopic heterogeneity lying beneath these macroscopic distinctions: inter-patient heterogeneity, involving differences among patients (e.g., different types of breast cancer), and intra-patient heterogeneity within a single individual’s cancer—including heterogeneity within a primary tumor, between the primary tumor and metastases, and among metastases. While there have been both success, and failures with so-called “targeted therapeutics,” our evolving understanding of intra-patient heterogeneity has added urgency to an old question: Can we treat complex tumors by attacking a limited number of molecular targets (and resulting pathways), or does intra-patient heterogeneity severely limit this approach?
Current theory suggests that cancer forms when a single cell acquires a set of genetic mutations that cause it to divide uncontrollably and—often—spread to other parts of the body. Cancer cells then “evolve” in response to selective pressures, including those caused by diversity in the tumor’s microscopic environment or even from therapies. Intra-patient heterogeneity occurs when cancer cells evolve in different ways.
Consider intra-patient heterogeneity as the branch structure of a tree, with its trunk representing the original malignant cell type. The branches represent evolved subpopulations, varying by genetic mutations, gene expression profiles, protein expression profiles, or tumor microenvironment. The further from the trunk, the greater the differences from the original malignancy. The scope and complexity of the branch structure can be a measure of heterogeneity within a single patient.
Finding the drivers
The hope for targeted therapeutics is that all cells in a complex tumor depend on specific genetic mutations in the trunk of the tree, and—despite the tumor evolution—that those mutations remain critical in all the branches. These are the so-called “driver” mutations. Disabling any one driver might stop the growth. The tree trunks might vary from patient to patient, but the downstream branch variations aren’t as critical.
Advances in molecular biology have given us a better understanding of such tree trunks, and how they vary among patients. Breast cancer patients can now be classified into at least 10 subgroups. Targeted therapeutics have significantly improved treatment for Her2-positive and estrogen receptor-positive breast cancers. Even more dramatic are the enduring remissions from the drug imatinib in select leukemias and gastrointestinal stromal tumors.
Unfortunately, these successes have been exceptions. Most targeted therapies produce only short-lived benefits, and at high costs. Vermurafenib for BRAF-mutated melanoma prolongs life for a few months at annual costs of $100,000 to $300,000 per patient. This brings us back to those tree branches. Have “passenger” mutations morphed into driver mutations under the selective pressure of treatment? Have resistance mutations developed that limit the effectiveness of therapy, perhaps by impeding binding of the therapeutic to its receptor, activating alternative bypass pathways, or upregulating downstream pathways?
More generally, does intra-patient heterogeneity imply that a single targeted therapeutic, while producing short-term responses, brings us no closer to enduring disease control? Or do we simply need a better—or a second or third—targeted therapeutic?
The jury is still out, but recent measurements are sobering. Sequencing of tumor samples shows a surprising genetic, transcriptional, and proteomic diversity in the branches. It also is creating confusion about which mutations are drivers versus those that fulfill some other cancer-sustaining function, such as supporting therapeutic resistance. New single-cell analysis technologies are showing similar results. Both techniques suggest significant heterogeneity between—and among—primary tumors and metastatic sites.
This creates questions for basic science, clinical practice, and the translational research that bridges the two. More basic research is needed because our knowledge of intra-patient heterogeneity is still limited. Another basic science issue is the relevance of the cancer stem cell hypothesis, which argues that specifically targeting only a subset of key tumor initiating cells holds the key to sustained response.
Translational research strategies must also be reviewed. Despite disappointments, perhaps we should re-emphasize targeting the blood vessels that feed cancers rather than the heterogeneous tumors? Some immunotherapies become more effective when exposed to heterogeneous antigenic signals. Can augmenting or adjusting the patient’s immune response help overcome heterogeneity? We must also probe deeper into the biology of targeted therapeutics. Trastuzumab for Her2-positive breast cancer, in combination with chemotherapy, produces dramatic remissions, even though the tumors are highly diverse. Is the chemotherapy effective against all but the Her2-positive cells, or is some other biological process at work? Should targeted therapy generally be combined with broad-based, conventional chemotherapies? There is also growing interest in so-called “liquid biopsies” that monitor circulating tumor cells and free tumor DNA in the blood. Technologies are moving forward but heterogeneity complicates knowing what to measure.
And there are clinical implications. Perhaps combinations of targeted therapeutics can overcome heterogeneity. However, any drug could have off-target effects, so toxicity might limit this approach, as could cost. Is a single biopsy from a primary tumor sufficient for diagnosis, or should we be doing multiple biopsies that include metastatic sites? What can imaging or other diagnostic technologies reveal about heterogeneity?
These are only some of the issues arising from increased concern about intra-patient heterogeneity. This fall, the Stanford Cancer Institute will convene an international symposium to discuss these and related issues. While uncertainties remain, there is sufficient data for every major cancer research and treatment center to assess how intra-patient heterogeneity will affect research priorities, clinical trial design, and the choice of treatment for patients.
Beverly Mitchell is the director of the Stanford Cancer Institute, where David Rubenson is the associate director for administration and strategic planning. Daniel S. Kapp is a professor emeritus of radiation oncology at Stanford University.