Skip to main content

Immunotherapy Response Data Spurs International Effort to Uncover Germline Polygenic Predictors


NEW YORK – Based on emerging data in melanoma, an international consortium has begun work to characterize, as exhaustively as possible, the influence that germline genetic changes have on immunotherapy response in these, and hopefully other cancer patients.

According to Tomas Kirchhoff, the consortium's principal investigator, the project, dubbed IO GEM, intends to then use this data to develop predictive clinical tools to determine which patients are likely to respond to immune checkpoint inhibitors, which have become increasingly ubiquitous in treatment for advanced melanoma despite only working in a subset of patients.

A researcher at NYU's Perlmitter Cancer Center, Kirchhoff and his colleagues have previously published smaller studies that suggests that germline genetic factors related to things like autoimmunity could help explain differences in drug response and overall outcomes for immunotherapy-treated melanoma patients.

In a presentation at the annual meeting of the American Association for Cancer Research earlier this month, he outlined how his and other groups' findings have hinted that this germline component to immunotherapy response could help capture certain responders or non-responders missed by existing tumor-oriented biomarkers like PD-L1 expression and tumor mutational burden.

Some of the members of the new consortium, which includes more than 45 centers, had already started collaborating in recent years to build out this data, but critical funding for the effort only emerged last year, with an award from the Melanoma Research Alliance.

"We started with an initial focus on the candidate pathways, like autoimmune susceptibility, and immunomodulatory proteins, and then with the generation of these individual collaborations we started to create a bigger, community-based effort," Kirchhoff said.

"Of course, there's a component of discovery, but … definitely the goal here is to develop clinically actionable targets. That's number one," he added "And that's the whole purpose also of the funding mechanism, to develop and improve the portfolio of the current biomarkers that are available for the patients undergoing this treatment."

The IO GEM effort reflects a longstanding hypothesis in this field that a combination strategy, incorporating both tumor and innate host immune factors might be necessary for ideal prediction of immunotherapy response. A group from the Netherlands Cancer Institute and the University of California, Los Angeles, for example, proposed a multi-pronged approach they called a "cancer immunogram" for assessing a patient's likelihood of responding to treatment, years ago as these drugs were first clinically implemented.

Despite this, immune-oncology drugs have advanced with mainly tumor-informed biomarkers, or, in melanoma, without any companion diagnostic mandate.

In recent years, gene expression signatures and other algorithms that assess aspects of the tumor microenvironment have sought to bring non-tumor signals into play. But existing research has only touched the tip of the iceberg in terms of the role of germline genetics, Kirchhoff argued. "Germline genetics impacts the host immunity, and [so] we need, with a capital N and maybe even a capital EE and a capital D, to consider this," he said.

Currently advanced melanoma patients are treated in the first line by anti-PD1 drugs, but about 50 percent do not respond to this treatment. "What is also deeply concerning is that not all the patients respond initially to the treatment [end up] with a durable clinical benefit," Kirchhoff added.

"If we can actually predict which patients would not really benefit from [anti-PD1 treatment] ... they could be positioned to other treatments, either a single line CTLA-4 … or a combined regimen," he said. Although more toxic, this would save the patient the burden of a treatment round to which they would would most likely not respond.

The first task of IO GEM under its new grant is to complete a microarray-based genome-wide association study, boosted by whole-genome and exome sequencing to prioritize loci for further validation, and followed up by informatic analyses and functional studies.

Overall, Kirchhoff said the group intends to pool genomic and harmonized clinical data from 20,000 melanoma patients treated with either anti-PD1 or anti-CTLA4 drugs. The team will examine the association between germline genetic factors, immunotherapy efficacy, patient survival, and toxicity.

According to the program description, investigators will further test the resulting targets using platforms that help map these loci to specific aspects of immune and peripheral T-cell activity.

The hoped-for result is that this will yield polygenic risk scores that, in combination with other aspects of the overall "immunogram," accurately stratify responders and non-responders,

According to Kirchhoff, it's unlikely that there are just one or two variants that explain the fullness of how innate immune function affects cancer immunotherapy response. "Different pathways controlled by these germline genetic factors can also manifest in the same way when the patient is treated by these drugs, so for example I could have higher autoimmune susceptibility and you could perhaps have some kind of cytokine genetic variants that are higher, but both of these would reach the same effect," he said. "The effect also multiplies … so that's where the future is, to develop polygenic scores."

With the discovery portion of the effort complete, the group intends to work, either with academic collaborators, or with industry, to prospectively test these scores, and Kirchhoff said the team is already seeing interest from pharma companies interested in supporting these efforts, though he declined to disclose potential partners at this stage.

According to Kirchhoff, although the immediate focus of the project is melanoma-specific, the group also believes that the predictors they hope to glean would be translatable to other cancer types as well.

One challenge to that will be the fact that the current GWAS effort is limited to white individuals with European ancestry, due to the overrepresentation of this population in the development of cutaneous melanoma.

"In lung cancer, it's a different story, so it's very important to consider this [going forward], and we are already seeing differences among subpopulations, for example in [patients with Ashkenazi Jewish ancestry] in our cohort in terms of toxicities … so other ancestress will be critically important when we start [translating our findings] to other cancers that are treated by these drugs," Kirchhoff said.

Finally, although predictive biomarker discovery is driving the consortium, the group is also hopeful that these same findings could also lead to new therapeutic targets for drug development, or insight into novel combination strategies.