Immunological Effects and Predictive Gene Signatures in Patients with Relapsed Follicular Lymphoma Treated with CT-011, a Humanized Anti-PD-1 Monoclonal Antibody

Abstract 162 Background: The coinhibitory receptor programmed death (PD)-1 is markedly increased on peripheral blood (PB) and intratumoral T cells in follicular lymphoma (FL) and is associated with impaired T-cell function. CT-011 (pidilizumab), a humanized anti PD-1 monoclonal antibody, blocks the PD-1/PD-ligand pathway and enhances the function of antitumor T and NK cells. We conducted a phase II trial to determine the safety and efficacy of CT-011 and rituximab in patients (pts) with relapsed FL and the clinical results are reported separately. Here, we evaluated the effects of CT-011 on immune cells both in the PB and tumor microenvironment. We also determined predictors of clinical outcome based on PB immune cell subsets and molecular features of tumor-infiltrating immune cells at baseline. Methods: CT-011 was given every 4 wks × 4 and rituximab weekly × 4 starting on day 17 after the first infusion of CT-011. Pts with response or stable disease received 8 additional optional infusions of CT-011 every 4 wks. PB and core needle biopsies from involved lymph nodes were collected prior to and on day 14 after the first infusion of CT-011. PB mononuclear cells (PBMC) were analyzed by multiparametric flow cytometry to determine various immune cell subsets. Whole genome gene expression profiling (GEP) was performed on core needle biopsies. Results: Of 29 pts eligible for efficacy analysis, 19 had an objective response for an ORR of 66%. CR was observed in 15 (52%) and PR in 4 (14%). After a median follow up of 14 mo, median PFS was 21.1 mo, and was not reached for the responders. We observed a significant increase in the absolute number of PB immune cells in day 14 samples compared with baseline including total lymphocyte count (p<0.01), CD3+ T cells (p=0.01), CD4+ T cells (p<0.01), and CD4+ naive (p=0.01), effector memory (p=0.02), and central memory T cells (p<0.05). We also observed increase in the expression of the activating receptor NKG2D on NK cells (p=0.01). In contrast, CD8+ terminally differentiated T cells were decreased (p=0.02). Comparison of GEP from core needle biopsies obtained pretreatment and day 14 (n=8 pairs) showed up regulation of several genes associated with T cell activation in day 14 samples including CD58, CTLA4, NFATC1, CCR5, PBK, TSLP, and ZBTB32. We analyzed GEP of baseline tumor biopsies from 18 pts to find gene signatures that correlated with clinical outcome, as measured by PFS and/or quantitative change in tumor size (%change). We tested a “PD-1hi_Down” signature of 41 genes previously reported by us (Chu et al, Blood, Nov 2011; 118:2648) to be less expressed in CD4+ T cells strongly positive for PD-1, likely to be follicular helper T cells (Tfh, PD-1hi), than in CD4+ T cells with intermediate or low levels of PD-1 surface staining, likely to be exhausted effector T cells (Teffs, PD-1int) or activated effector or naïve T cells (PD-1lo). The PD-1hi_Down gene signature correlated significantly with PFS by univariate Cox regression and was also significant when examined by gene set enrichment analysis based on ranking all genes by correlation with %change. Low expression of this signature, suggesting more Tfh and fewer PD-1+ Teffs within the tumor, predicted shorter PFS duration and less tumor shrinkage. Combined with our in vitro findings that anti-PD-1 Ab enhances the function of Tfh and PD-1+ Teffs but has no effect on PD-1lo T cells in FL, these results suggest that CT-011 therapy enhances the respective pro- and anti-tumor effects of one or both of these cell types. In support of our conclusion that the effect of the PD-1hi_Down signature on outcome depends on CT-011 therapy, we found that this signature did not correlate with overall survival in an external dataset of FL pts treated largely with chemotherapy alone (Dave et al., NEJM 2004; 351: 2109). Conclusions: Administration of CT-011 (pidilizumab) was associated with increase in the numbers of naïve, effector memory, and central memory CD4+ T cells and resulted in activation of T and NK cells in the PB and the tumor microenvironment in FL. A high expression of PD-1hi_Down gene signature, consistent with relatively increased numbers of antitumor Teffs compared with protumor Tfh was predictive of good response and improved PFS suggesting that CT-011 restores function of exhausted Teffs. These results provide insight into the mechanism of action of CT-011 and offer a predictive biomarker for selection of pts for future clinical trials with this class of agents in FL. Disclosures: Rotem-Yehudar: CureTech Ltd: Employment, Research Funding. Neelapu:Cure Tech, Ltd.: Research Funding.