A Novel Approach To Identify Genetic Signatures Of Clinical Outcome To Ipilimumab

Immune therapies focused on unleashing the body’s own immune system against cancer have heralded a revolution in clinical oncology. Dramatic responses are now being observed in several intractable tumor types like metastatic melanomas which were previously associated with very short survival times.

At the forefront of this revolution are cancer immunotherapy drugs known as “checkpoint inhibitors.” In a healthy individual, small tumors that arise in the body are continually detected and eliminated by the immune system. In cancer patients, however, the tumor often evades the immune attack by rendering itself invisible to immune killer cells. Checkpoint inhibitors are powerful new tools that can reverse this process, thereby enabling the immune system to successfully launch an attack against cancer.

Although checkpoint inhibitors can be powerful, they are also associated with severe, sometimes fatal, side effects. In metastatic melanoma, one of the most well-studied tumors in the context of immune therapies, physicians have struggled to predict which patients will benefit from a given drug and which may suffer unnecessary toxicity. Scientists at Dana-Farber Cancer Institute have now discovered genetic markers for resistance to ipilimumab, the first checkpoint inhibitor approved by the Food & Drug Administration for the disease, according to a new report in the journal Cell.

Checkpoint inhibitors target surface proteins on T cells that can act as switches preventing these immune-system cells from attacking tumors. Two predominant classes of checkpoint inhibitors currently are ones that target either the CTLA-4 or PD-1 proteins on T cells. Ipilimumab, an anti-CTLA-4 therapy, was the first checkpoint inhibitor approved by the FDA for metastatic melanoma on the basis of work led by Dr. F. Stephen Hodi, a Dana-Farber physician scientist who was closely associated with this project. Since CTLA-4 and PD-1 work through significantly different pathways,  drugs that target PD-1 are also given for metastatic melanoma, sometimes in combination with ipilimumab. This, unfortunately, results in increased toxicities, and a burning question in the field is to identify biomarkers that can predict responders and non-responders to these drugs which can lead to the effective pairing of patients with the right treatments.

Sachet A. Shukla, PhD, who leads the computational group in the Translational Immunogenomics Lab in the Center for Immuno-Oncology at Dana-Farber, and medical oncologist Pavan Bachireddy, MD, a Damon-Runyon Fellow completing his post-doctoral research in the laboratory of Dr. Catherine Wu, MD, devised a novel approach to identify genetic signatures of clinical outcome to ipilimumab by comprehensively analyzing multiple clinical cohorts and publicly available data sets. In all, they studied expression of thousands of genes in tumors from 146 melanoma biopsies collected by four clinical studies of checkpoint inhibitors, and further validated the results using a separate melanoma cohort of 465 samples and advanced imaging techniques Interestingly, they found  that expression of a cluster of genes located on chromosome X was associated with resistance to ipilimumab but  not with resistance or response to nivolumab, a PD-1 inhibitor.

More surprisingly, these genes that were most highly expressed genes in patients that did not benefit from ipilimumab were from a sub-family of genes known as MAGE-A. The MAGE-A genes were first discovered in 1990 and are known as cancer-germline or cancer-testis antigen genes. Although several trials have been conducted to harness these genes in the clinic, these efforts have largely been unsuccessful. The results from this work suggest that MAGE-A genes themselves can suppress critical factors needed to stimulate effective responses to therapy, and researchers believe they may be good targets for drug development efforts.

The team validated their findings by analyzing tumor samples from a prospective, randomized clinical trial that studied whether or not the order in which nivolumab and ipilimumab are given is clinically important to these patients. Additionally, researchers examined a set of patients treated with anti-PD-1 therapy alone as well as genetic data from 465 melanoma samples collected by The Cancer Genome Atlas program. The scientists again discovered that high expression of MAGE-A genes correlated with resistance to ipilumumab, but they didn’t see a similar correlation for nivolumab. These results are consistent with the notion that the CTLA-4 and PD-1 pathways occupy biologically and clinically distinct niches.

If the findings about MAGE-A genes are validated in larger studies, they may point toward the development of clinical tools that help to select which patients should receive ipilimumab rather than other treatments and may help reinstate anti-CTLA-4 therapy as a predominant checkpoint inhibition strategy.

Previous work by other groups had revealed a connection between MAGE-A genes and the cell housekeeping process called autophagy, in which cells degrade a certain amount of their own internal contents. Some cancer chemotherapies end up boosting autophagy, while other drug candidates designed specifically to heighten the process are in clinical trials. If these drug candidates prove useful, it may make sense to try combining them with ipilumumab to treat melanoma. Most existing drugs are designed to act on proteins that are also present in normal tissues of the body, which is the root cause of serious toxicities. In contrast, the MAGE proteins are expressed only in tumors and testis, and a MAGE-directed drug would likely have very limited side effects. Discoveries around MAGEs as oncogenes and key players in autophagy, combined with their role in modulating the immune response provides a solid opportunity to build therapeutic platform initially targeting solid tumors and ultimately targeting broader aspects of immune system dysregulation.

The wealth of immunotherapy clinical trials underway for melanoma will offer many further opportunities to mine genomic data to better understand which drugs work for which patients and ultimately lead to a rational clinical decision-making framework. While drug discovery efforts against the identified MAGE-A genes are still in their infancy, success in this area will lead to a novel first-in-class therapeutic that could have ramifications beyond metastatic melanoma.

These findings are described in the article entitled Cancer-Germline Antigen Expression Discriminates Clinical Outcome to CTLA-4 Blockade, recently published in the journal Cell. This work was conducted by Sachet A. Shukla (Center for Immuno-Oncology at Dana-Farber), and Pavan Bachireddy and Catherine Wu (Broad Institute, Brigham & Women’s Hospital, and the Dana-Farber Cancer Institute).