T-cell Signatures Suggest Existence of Immune-Responsive Microsatellite-Stable mCRPC

November 23, 2020 - Patients with microsatellite instable (MSI) metastatic castration-resistant prostate cancer (mCRPC) have a distinct T-cell signature that is detectable in the peripheral blood. Moreover, this signature may yield insight into immune parameters that suggest which patients with microsatellite stable (MSS) mCRPC are likely to respond to checkpoint inhibitors, according to findings from a study published in JCO Precision Oncology.

Considerably higher blood T-cell richness and diversity was found in MSI patients vs MSS patients. Moreover, patients with MSI disease had more shared T-cell clusters with low generation probability (pGen) compared with those patients with MSS disease.

These findings suggest that patients with MSI mCRPC derive more complex T-cell responses because of a greater neoepitope load vs patients with MSS mCRPC.

Moreover, patients with MSS disease who also had low pGen T-cell clusters responded better to checkpoint inhibitors vs other therapies compared with patients who did not have low pGen T-cell clusters.

The majority of patients with MSI mCRPC derive a response to checkpoint blockade. The goal of the study was to identify biomarkers of definite susceptibility to checkpoint inhibitor benefit in patients with MSS mCRPC.

The study evaluated the biomaterial from 15 patients with mCRPC before and during treatment with anti–PD-1/PD-L1 checkpoint inhibitors (cohort 1), as well as 13 healthy donors. In cohort 1, baseline peripheral blood mononuclear cells were collected before the second or third cycle of checkpoint blockade. Fresh tissue core biopsies were performed before checkpoint inhibitor therapy initiation when possible per institutional protocol.

MSI disease was defined as loss of mismatch repair proteins MSH2, MSH6, MLH1, or PMS2 or identification of instable microsatellites by next-generation sequencing. Additionally, PD-L1 positivity was assessed by clone E1L3N, and CDK12 loss or mutation was assessed by targeted or whole-genome sequencing.

Prior therapies for CRPC included chemotherapy with docetaxel (n = 12; 75%), cabazitaxel (Jevtana; n = 6; 37.5%), and carboplatin (n = 1; 6.3%), second-generation androgen deprivation therapy (ADT) with abiraterone acetate (Zytiga; n = 7; 43.8%) and enzalutamide (Xtandi; n = 11; 68.8%), nuclear therapy with radium-223 dichloride (Xofigo; n = 2; 12.5%) and lutetium-177 PSMA-617 (n = 1; 6.3%), and experimental therapy with dendritic cell vaccination (n = 1; 6.3%), olaparib (Lynparza; n = 5; 31.3%), and nivolumab (Opdivo; n = 1; 6.3%).

Additionally, 1 MSI patient (16.7%) and 2 MSS patients (22.2%) had loss of or mutated CDK12.

Most patients (n = 8; 53.3%) received checkpoint blockade as fourth- or later-line treatment for CRPC, whereas 26.7% (n = 4) received it in the third-line setting and 20% (n = 3) received it in the second-line setting. No patients received checkpoint inhibitors up front. Of the checkpoint inhibitors received, 6 patients (40%) had prior nivolumab, 7 (41.2%) had prior pembrolizumab (Keytruda), and 3 (20%) had prior atezolizumab (Tecentriq). Notably, no patients with MSI disease had prior pembrolizumab, and no patients with MSS disease had prior nivolumab. Patients received therapy with checkpoint blockade until complete response, progressive disease, or unacceptable toxicity.

In the overall population, the median duration of response to checkpoint inhibitor therapy until prostate-specific antigen (PSA) progression was 2.8 months (range, 1.7-17.1). Patients remained on therapy with checkpoint blockade for a median of 3.2 months (range, 1.4-27.6). Seven patients (41.2%) experienced a PSA decline of more than 50% with checkpoint inhibitor therapy (5, MSI patients; 2, MSS patients).

In patients with CDK12 loss, one MSI patient and one MSS patient experienced a PSA decline of more than 50%. The median duration until PSA progression was 2.6 months (range, 1.7-3.5), and patients remained on checkpoint inhibitor therapy for a median of 3.7 months (range, 2.6-3.7).

In cohort 1 (n = 15), 6 patients were MSI and 9 were MSS. The median age of patients was 61.4 years. The majority of patients (n = 11, 68.8%) had a Gleason score of 8 or greater at diagnosis, whereas 18.7% (n = 3) had a Gleason score of less than 8, and 6.3% (n = 1) had an unknown score. At diagnosis, about half (n = 8; 53.3%) of patients were M1 and the remainder (n = 7; 46.7%) were M0.

Tumor mutational burden (TMB) data were available in 9 patients (n = 6, MSI; n = 3, MSS) and revealed that patients with MSI disease had a median of 38.8 mutations/Mb compared with 2.1 mutations/Mb in MSS patients (P = .0065)

In addition, biomaterial from 13 healthy donors and 39 MSS patients with mCRPC who did not previously receive a checkpoint inhibitor was collected (cohort 2).

The results indicated that the most pronounced clonal overlap between blood and tumor samples was identified from the same patient. Although this was expected, it suggests that peripheral blood findings may partially mirror the tumor-infiltrating T-cell composition and that future analyses could evaluate blood samples alone.

Additionally, patients with MSI mCRPC had higher T-cell repertoire richness, diversity, and clonality compared with patients with MSS mCRPC. Patients with MSS disease also showed different levels of T-cell richness and diversity between patients. As such, a positive trend toward T-cell repertoire diversity and clinical outcomes with checkpoint inhibitors was observed in the MSS subgroup (R2 = .39; P = .068).

One patient responded initially to nivolumab and later, they responded again to atezolizumab after being switched because of toxicity. The patient’s TMB was significantly reduced during treatment with atezolizumab. This suggests a patient with high TMB who responds to first-line treatment with checkpoint inhibitors may have cleared some of the tumor subclones that induced TMB high–status. Moreover, T-cell richness also decreased, suggesting that the parameter may be correlated with TMB/neoepitope burden.

The repertoire overlap analysis showed that no patients shared individual T-cell sequences. Per grouping lymphocyte interaction by paratope hotspot (GLIPH) analysis, 246 T-cell clusters were shared in the group of healthy donors, and 1307 clusters were shared exclusively among patients with mCRPC. Most of the clusters in the patients with mCRPC were enriched in the MSI subgroup.

Additionally, 232 low pGen T-cell clusters were shared between at least 2 patients with mCRPC. These clusters are expected to be shared if selected for functionally, suggesting the presence of T-cell receptors directed against common prostate cancer antigens or neoepitopes in patients with overlapping genetic loci.

Five mCRPC T-cell clusters were identified as being shared among at least 3 patients and excluding contamination bias. Notably, these clusters were found in pre-checkpoint inhibitor samples compared with peripheral blood drawn during treatment with checkpoint inhibitors.

Five patients with MSI disease and 5 patients with MSS disease, including 2 with CDK12 loss, demonstrated CDR3 sequences belonging to at least 1 of the 5 shared low pGen clusters. Patients with MSS mCRPC had a significantly longer time-to-progression (TTP) with checkpoint inhibitors vs patients without these clusters (mean TTP, 3.9 vs 2.1 months; P = .0025).

The sample size of the study was limited in the second cohort of patients. As such, prospective validation of these data are needed to determine the clinical relevance of using blood T-cell profiles as predictive biomarkers for checkpoint inhibition in patients with MSS mCRPC.

Reference

Simnica D, Smits M, Willscher E, et al. Responsiveness to immune checkpoint inhibitors is associated with a peripheral blood T-cell signature in metastatic castration-resistant prostate cancer. JCO Precis Oncol. 2020;4:1374-1385. doi:10.1200/PO.20.00209