Beyond PD-L1: The search for better predictors of beneficial response
The immuno-oncology field as a whole can agree that there is a desperate need for better predictors of not only initial response to immunotherapies, but also of the likelihood of durable response.
Status does not guarantee treatment success
While much of the focus of the immuno-oncology field to date has been on the development of PD-1/PD-L1 checkpoint inhibitors, PD-L1 status has not yet proven to be an effective predictor of clinical response to immunotherapy. In fact, studies show that positive expression of PD-L1 does not guarantee a response, and conversely, the lack of PD-L1 expression does not definitively mean that no response will be observed.
Why do some patients experience a prolonged respond to anti-PD-1 therapy, while others respond initially before relapsing shortly thereafter?
This question was recently addressed by a group at UCLA who investigated the mutations that were found in metastatic melanoma patients with acquired resistance to pembrolizumab (Merck) therapy (Zaretsky et al. 2016 NEJM).
It goes without saying that a variety of factors are in play in this complex scenario. The results of this study showed that clonal evolution resulted in the silencing of pathways that are involved in interferon signaling (via loss-of-function mutations in JAK1 and JAK2 genes) and the altering of antigen presentation (truncating mutation in the B2M gene). It is clear that clonal evolution of tumors may play a key part in the development of resistance.
A better understanding of the process of acquired resistance to PD-1 blockade will help to further tailor treatment and will lead to more rational design of combination therapies. The findings from studies such as this will ultimately help us to combine inhibitor drugs that block multiple resistance routes, and/or to identify new predictive biomarkers.
Mutational Burden and Patient Response
Besides PD-L1 status, it has been suggested that mutational burden evaluation can also be used to provide guidance on whether or not a patient will respond to immunotherapy. It makes sense that the higher the mutational load, the higher the number of “foreign markers” that can be flagged by the immune system. But will mutational load prove to be a better predictor for response to therapy than PD-L1 has been?
Neoantigen load as a predictor of immune response
Quantifying the load of neoantigens (the tumor-specific mutations that form neoepitopes and are presented by the MHC complex) could be a more reliable indicator of response. Given the inherent complexity of cancer, it is likely that we will rely on a combination of different genomic and immune signatures to provide insights and guidance as to who will benefit and have a lasting clinical response to these promising immunotherapies – and who will not.
There are still many questions to answer: What is the cut off for the number of mutations that could be considered a high, intermediate, or low mutational burden for determining likelihood of response? Does the mutational burden spectrum vary across different cancer types? What role will neoantigen load quantification play?
In recent years, there has been a renaissance in immuno-oncology as our understanding of the underlying molecular mechanisms has improved and deepened at a pace that is unprecedented in the cancer treatment space. With the technological advances in next-generation sequencing (NGS) and downstream bioinformatics analysis, we’re able to profile tumors at a much higher throughput more accurately than ever before. As we look towards the future, it’s important for us to understand that at the patient-level, it may be a variety of factors that interplay to determine patient response, and we collectively need to look towards how best to leverage the molecular clues to develop safer and more effective immunotherapies.