Biological and Therapeutic Advances in the Pursuit of Effective Immunotherapy for Prostate Cancer

Anis A. Hamid; Atish D. Choudhury


Curr Opin Urol. 2020;30(1):30-35. 

In This Article

Biological Correlates of Immunotherapy Response

In light of the observation that a subset of patients with prostate cancer respond and durably benefit from immunotherapy, there has been intense focus on better understanding the determinants of immunotherapy response (and resistance) with the goal of improving the precision of patient care. In recent years, large-scale, systematic analyses of the prostate cancer genome have revealed a so-called 'long tail' of mutational drivers[24] as well as distinct molecular and phenotypic disease subtypes associated with response to immunotherapy.

The best characterized prostate cancer subtype linked to CPI response constitute tumors with deficient DNA mismatch repair (dMMR) and/or microsatellite instability (MSI). Following a proof-of-concept study in colorectal cancer, Le et al.[25] demonstrated the efficacy of PD-1 blockade across numerous mismatch repair-deficient cancers, including prostate cancer, with functional studies in responders confirming proliferation of neoantigen-specific T cells driven by the exceptionally high somatic mutation and mutant peptide load present in these tumors. Consequently, in May 2017, the FDA provided tumor-agnostic approval of pembrolizumab for dMMR or MSI-high advanced, resistant solid tumors.

What is the prevalence, clinical features and therapy response to CPI in dMMR/MSI-high prostate cancer? The largest case series to date examined tumors from 1346 patients with prostate cancer by targeted sequencing using a clinical assay,[26] where MSI was quantified using a computational tool called MSIsensor. Additionally, immunohistochemistry (IHC) for MMR protein expression was performed in a subset of tumors. In total, 32/1033 (3.1%) of patients with evaluable tumor for MSIsensor analysis were MSI-high/dMMR of whom 22% had evidence of germline mutations in Lynch syndrome-associated genes. Prior studies indicate that the frequency of dMMR/MSI-high prostate cancer does not appear to differ between primary versus metastatic/resistant tumors.[24,27] Uniquely, 11 patients in the cohort with MSI-high/dMMR CRPC received CPI therapy, of whom more than half achieved a PSA decline of 50% or greater on therapy. Both radiographic and durable clinical responses were observed. These findings were recapitulated in an independent cohort of 13 patients with metastatic prostate cancer and deleterious tumor MMR mutations.[28] Prolonged responses to hormonal therapy were observed despite aggressive clinicopathologic features, and two of four patients who received CPI therapy achieved a PSA response, with a median progression-free survival of 9 months among the CPI-treated.

Although increased neoantigen burden related to mismatch repair deficiency may explain sensitivity to CPI, Rodrigues et al.[29] sought to dissect the immunogenomic complexity of dMMR/MSI-high prostate cancer using an integrated, multiomic approach to refine biomarker-guided selection for immunotherapy. In total, 8.1% of advanced prostate cancer patients exhibited dMMR/MSI (by IHC and PCR, respectively) and these tumors were enriched for tumor-infiltrating T lymphocytes, PD-L1 protein expression and poorer clinical outcomes. Importantly, this study highlighted the significant degree of discordance between conventional and orthogonal approaches of detecting dMMR/MSI in prostate cancer (e.g. IHC; PCR; MMR gene mutation by sequencing) and suggests that a dMMR-associated DNA signature as the phenotypic readout of these alterations may perform better as a predictive biomarker.

Beyond the small subset of patients with dMMR/MSI-high prostate cancers (typically associated with a high rate of somatic mutation), it is currently unclear whether the larger set of patients with mutations in other genes involved in the DNA damage response (DDR) are enriched for responders to CPI, and studies of CPI with or without poly (ADP-ribose) polymerase inhibition are in progress.[30,31] Recent reports have described other somatic genetic alterations that may correlate with CPI responsiveness. Biallalic CDK12 loss is enriched in metastatic CRPC and is associated with a distinct mutational signature from MMR-deficient and homologous recombination-deficient tumors, and such tumors are marked by genomic instability, increased gene fusion-induced neoantigen burden and high inferred immune infiltration.[32] Two of four CDK12-mutant patients treated with a PD-1 inhibitor monotherapy in this study exhibited exceptional PSA responses. Recently, a case report of a patient with metastatic CRPC and an isolated somatic POLE V411L mutation, which is associated with extreme hypermutation despite microsatellite stability, detailed exceptional response to pembrolizumab without evidence of disease progression after 24 cycles of therapy.[33] Collectively, these data suggest that in prostate cancer, DNA-based biomarkers may have an important role in defining discrete molecular/mutational subtypes likely to respond to immunotherapy.

Beyond single gene mutations, global tumor mutational burden (TMB) is also promising as a predictive biomarker of response to CPI therapy. There is a significant correlation between TMB and response to anti-PD-1 or anti-PD-L1 therapy across tumor types[34] as well as improved survival among CPI-treated patients.[35] Responses to the combination of ipilimumab with nivolumab in mCRPC appeared to be enriched in patients with high TMB,[16] but these findings will require validation in a larger cohort. Furthermore, non-DNA-based biomarkers are being studied and will provide additional layers of biological information for integration, in order to improve our understanding of the complexity of tumor-immune interactions and to develop robust, multimodal biomarkers. For example, TMB and a T-cell-inflamed gene expression profile (GEP) exhibited joint predictive utility in identifying responders and nonresponders to the PD-1 antibody pembrolizumab across tumor types.[36] The role of tumor immune infiltration, tumor immune-related transcriptomic signatures (both activating and repressive),[29,37] PD-L1 expression and host immune factors (including the dynamic and modulatory effect of the microbiome) are all active areas of investigation across cancer types.