SITC 2017 Scientific Highlights - Nov. 11

The Society for Immunotherapy of Cancer (SITC) is pleased to present highlights from the Saturday programs (Nov. 11, 2017) of the 32nd Annual Meeting in National Harbor, Md. (Scroll to the bottom of this blog post to view the Glossary).

Dendritic cell acquisition of MHC I controls CD8+ T cell priming

Brandon MacNabb, BS (University of Chicago) presented data supporting the concept of MHC I antigen presentation by Batf3-lineage DCs as a critical component of CD8+-mediated anti-tumor response. Researchers initially generated H2-KbAB (MHC I+) and Kb-/-(MHC I-) C1498.SIY acute myeloid leukemia cell lines and subsequently engrafted C57BL/6 mice to determine the contribution of tumor cell MHC I presentation in CD8+ T cell priming. Initial assessment revealed reduced tumor growth in C1498.SIY KbAB mice compared to C1498.SIY Kb-/- mice. CD8+ T cell proliferation was also increased in KbAB mice compared to Kb-/- mice (p < 0.05). The observed C1498.SIY KbAB-dependent CD8+ proliferation was abolished in Batf3-/- C57BL/6 mice (p < 0.01). IFN-ϒ secretion was also decreased in Batf3-/- KbAB mice, suggesting that Batf3 initiates CD8+ priming. Transfer of autologous T cells from KbAB tumor-bearing mice offered complete protection from tumor growth in tumor-free mice. Conversely, autologous T cells from Kb-/- mice offered no such protection. Interestingly, DCs isolated from the TME and the tumor-draining lymph node in tumor-positive mice had increased KbAB MHC I expression (p < 0.001), suggesting DC acquisition of tumor-derived MHC I. Ex vivo experiments confirmed that migratory KbAB DCs are capable of CD8+ priming. These data reveal the importance of Batf3-lineage DCs and tumor-derived MHC I presentation in CD8+ T cell activation of anti-tumor response, providing insight into potential development of DC-oriented therapies. 

Risk/benefit ratio differs between checkpoint inhibitor monotherapy and combination therapy in the neoadjuvant setting in patients with advanced melanoma.

Sangeetha Reddy, MD (University of Texas MD Anderson Cancer Center, Houston, TX) discussed preliminary clinical trial results examining the efficacy of neoadjuvant nivolumab as monotherapy, or in combination with ipilimumab, in patients with high-risk advanced melanoma. Patients with resectable stage III or oligometastatic stage IV melanoma received either nivolumab monotherapy (n=12, 3mg/kg IV Q2W for 4 doses) or combination nivo/ipi (n=11, 1mg/kg and 3mg/kg, respectively, Q3W for 3 doses) prior to surgical resection, and then nivolumab (3 mg/kg Q2W) for an additional 6 months. ORR in the nivolumab and nivo/ipi groups was 25% (3 CR) and 73% (5 CR) respectively (RECIST 1.1). All patients receiving nivo/ipi were able to undergo surgical resection; conversely, 2 patients from the nivolumab cohort were not eligible for resection. Patients with pathological CR had 100% RFS (p = 0.05 versus non-responders). Incidence of grade 3/4 AEs, however, was higher in the nivo/ipi than the nivolumab monotherapy cohort (73% vs. 8%). These data resulted in trial stoppage after DSMB review. Researchers subsequently used IHC to investigate the TME to determine if specific biomarkers could predict therapeutic response and resistance. During treatment, responders had increased CD8+ T cell counts (p = 0.002), PD-L1 expression (p = 0.01) and TIL counts compared to non-responders. Additionally, T cell clonality was increased in responders during treatment (p = 0.006 vs. non-responders). RNAseq analyses of tumor samples are ongoing. These data reveal that both neoadjuvant nivolumab monotherapy and combination nivo/ipi have activity in patients with advanced, resectable melanoma, but the risk/benefit ratio differs.

PIVOT-02: Promising early efficacy and safety with CD-122-biased agonist NKTR-214 plus nivolumab in patients with locally advanced/metastatic solid tumors

Adi Diab, MD (MD Anderson Cancer Center, Houston, TX) presented preliminary results from the phase 1/2 Pivot-02 dose escalation study of NKTR-214, a novel CD-122-biased agonist, in combination with anti-PD-1 agent nivolumab, in patients with advanced/metastatic solid tumors (NCT02983045). NKTR-214 is metabolized to an active cytokine that harnesses the IL-2 pathway to preferentially activate and expand effector T cells and NK cells in the tumor microenvironment, resulting in increased proliferation of TILs and greater T cell PD-1 expression. Based on anticipated synergism between drug classes, NKTR-214 (0.003 – 0.009mg/kg) and nivolumab (240mg Q2W or 360mg Q3W) were administered to 38 checkpoint inhibitor-naive patients (mean age 61 yrs, 79% male, 68% no prior treatment for metastatic disease) with BRAF +ve metastatic melanoma (mM, n=11; 9/18/73% stage M1a/b/c; 46% PD-L1 ≥5%), renal cell carcinoma (RCC, n=22; 29% first-line PD-L1 ≥ 1%; 63% second-line PD-L1 ≥ 1%) or non-small cell lung carcinoma (NSCLC, n=5; 0% PD-L1 ≥1%). At data cutoff on November 1, 2017, 26/36 (72%) evaluable patients experienced target lesion tumor shrinkage, the majority of whom (n=19) received NKTR-214 0.006 mg/kg Q3W. ORR (RECIST v1.1 and iRECIST) was 64% (2 CR, 5 PR) and 46% (1 CR, 5 PR), in patients with mM and RCC, respectively, treated first-line. ORR in second-line NSCLC (PD-L1-negative) patients was 75% (1 CR, 2 PR). DCR in the three groups was 91%, 85% and 75%, respectively, and mTTR approx. 1.8 months across groups. Treatment is ongoing in all responders. Grade 3 DLTs (hypotension, metabolic acidosis, diarrhea, hyperglycemia) occurred in 4 patients in the highest dose cohort. Grade 1/2 TRAEs occurred in ≥50% of patients, most commonly fatigue, flu-like symptoms, and rash. There were no discontinuations due to TRAEs, no treatment-related deaths, and NKTR-214 did not increase the risk for irAEs with nivolumab. The trial established NKTR-214 0.006 mg/kg plus nivolumab 360 mg IV Q3W as the maximum tolerated dose. Enrollment to 13 expansion cohorts is underway. 

Early FDG-PET serves as a pharmacodynamic biomarker of treatment response in patients with microsatellite stable mCRC treated with carcinoembryonic antigen T-cell bispecific (CEA-TCB) antibody plus atezolizumab

Jose Saro, MD (Roche Innovation Center Zurich, Zurich, Switzerland) reported preliminary results from a phase 1b trial of CEA-TCB, a novel T-cell bispecific antibody targeting CEA on tumor cells, and CD3 on T cells (5-300mg IV, QW), in combination with the anti-PD-L1 antibody atezolizumab (1200mg IV, Q3W), in patients with microsatellite stable metastatic colorectal cancer (mCRC) (NCT02650713). mCRC is a disease with high unmet medical need in which >90% of patients have CEA-high tumors. In the first trial to employ F-fluorodeoxyglucose-positron emission tomography (FDG-PET) imaging as a pharmacodynamic biomarker of response, Dr. Saro and colleagues evaluated change in maximum standardized uptake value (SUVmax ; a measure of tumor metabolic activity), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) before, and 3-7 weeks after, treatment in the tumors of 25 patients with locally advanced/metastatic CEA +ve disease and poor response to standard therapy. FDG uptake 4 weeks after starting treatment was noticeably reduced compared to baseline, and metabolic response decreased in 17/25 (68%) patients, mostly those receiving a CEA-TCB dose ≥ 80mg. Changes in on-treatment metabolic response appeared to be CEA-TCB dose-dependent (p=0.08), and correlated with tumor shrinkage (p = 0.0002), which was observed in 9/25 (36%) patients, and with longer PFS [HR 0.08 (95%CI:0.16-0.35); p<0.0001]. These data suggest that early on-treatment change in FDG-PET may have value as an efficacy biomarker and to guide dose selection in patients with MSS mCRC. Analyses to validate FDG-PET are ongoing. 

Novel single cell analyses offer unique insight into acquired immunotherapy resistance in patients with Merkel cell carcinoma. 

Kelly Paulson, MD, PhD (Fred Hutchinson Cancer Research Center, Seattle, WA) discussed research using Merkel cell carcinoma (MCC) as a model to better understand acquired immunotherapy resistance mechanisms. A 59-year old male with widely metastatic, heavily pre-treated MCC received autologous HLA B*3502-restricted MCPyV-specific CD8+ T cell therapy, which targets the immunogenic MCPuY MCC-specific oncoprotein, and despite initial response (90% reduction in tumor burden) experienced metastasis after ~630 days. Using novel single-cell RNA sequencing technologies (scRNAseq) to assess mixed tumor cell populations across the patient timeline, researchers observed that CD8+ T cells overexpressed genes related to cell division, activation, and glycolysis during the initial immune response and tumor regression. Genes indicating CD8+ activation were most highly expressed at day 370 during patient response (day 376, p < 0.01 versus pre-treatment), and reduced during early relapse (day 614, p < 0.01 versus time of response). CD8+ T cells were observed within the tumor samples at time of relapse (acquired resistance), eliminating stromal exclusion as a resistance mechanism. Altered tumor cell gene expression was observed during acquired resistance, resulting in downregulation of tumor HLA-B and upregulation of HLA-E. Reduced tumor HLA-B expression mediated therapeutic resistance via loss of HLA-B-restricted T cell recognition and was reversible in vitro following 5-azacitidine or IFNγ treatment. HLA-E upregulation suppressed NK cell activity. Only achievable using scRNAseq, these results elucidate mechanisms that dictate long-term tumor progression. Additionally, researchers provided a foundation for using scRNAseq to better understand acquired immunotherapy resistance in other malignancies.

Unique qualities, but not quantity, of neoantigens predict long-term survival in patients with pancreatic adenocarcinoma

Vinod Balachandran, MD (Memorial Sloan Kettering Cancer Center, New York, NY) presented a series of studies designed to increase mechanistic understanding of the relationship between T cell immunity and rare cases of long-term survival in patients with pancreatic adenocarcinoma (PDAC). A combined analysis of whole exome sequencing, T cell receptor (TCR) Vβ-chain sequencing, 9-color multiplexed immunohistochemistry, and transcriptomic profiling was conducted on tumors collected from stage-matched short (n=68; median OS, 0.8 yrs) and long-term PDAC survivors (n=82; median OS, 6 yrs). All patients had surgically resected, non-metastatic PDAC and had not received neoadjuvant chemotherapy. Antigen immunogenicity was evaluated using a computational fitness model that integrated clonal genealogy, epitope homology and TCR affinity, while in vivo reactivity between T cells and neoantigens was assessed using functional assays in a small (n=7) subgroup of long-term survivors. Results showed that tumors from long-term survivors had 12-fold greater density of cytolytic CD3+ CD8+ granzyme-B + cells (p <0.0001 versus short-term survivors) and >94% unique intratumoral T cell clonality. Long-term survival was predicted by the combination of high predicted number of neoantigens and density of CD3+ CD8+ tumor infiltrate (median OS not reached vs. 0.8 yrs, P=0.004); neither factor alone was sufficient. High-quality neoantigens independently predicted long-term survival (median OS 8.6 vs. 0.8 yrs; p = 0.002); conversely, quantity of neoantigens did not. This held true in a cohort of patients not stratified by survival. Interestingly, mutation in MUC16 was highly correlated with long-term survival (p=0.03) and mutations were selectively lost during metastasis. T cells from two long-term survivors were capable of recognizing MUC16 neopeptides to a statistically significant degree. This study, which indicates that neoantigen quality is a biomarker of immunogenic tumors in long-term PDAC survivors, may contribute towards rational application of neoantigen-based therapies in future.

Tumor and T cell metabolic adaptations contribute to acquired checkpoint inhibitor resistance

Ashvin Jaiswal, Dpharm, MS, PhD (UT MD Anderson Cancer Center, Houston, TX) reported data investigating the role of metabolism in acquired resistance to checkpoint inhibitors in melanoma. Researchers introduced B16/F10 melanoma cells into BL6 mice and treated animals with combination anti-PD-1, anti-CTLA-4, anti-PD-L1, and FVAX. Tumor cells from non-responding animals were subsequently re-introduced into naïve mice, and this process was repeated until a CPI-resistant tumor cell line (3I-F4) - characterized by increased tumor growth in naïve mice while receiving CPI therapies - was generated. CPI-resistance in 3I-F4 tumors was verified to be immune driven. Gene expression analyses of TME cells from 3I-F4 subjected mice revealed signatures indicating a hypoxic environment. CPI-resistant tumor cells, however, had gene expression signatures suggesting hypoxic adaptation and oxidative stress tolerance. 3I-F4 tumors had increased ECAR and OCR compared to parental B16/F10 tumors, confirming gene expression analyses. 3I-F4 tumors displayed increased glycolytic activity compared to parental tumors (p < 0.05). It was hypothesized that these environmental alterations in resistant tumors would negatively affect TIL survival, but interestingly TIL counts were similar in parental and resistant tumors both off and during combination CPI treatment. TILs had reduced glycolytic capacity (p < 0.01) and effector function (p < 0.05), and gene expression analyses identified significant alterations in ADH7 and PGAM2 expression. Mice subsequently challenged with B16/F10 cells overexpressing ADH7 or PGAM2 had increased resistance to combination CPI therapy compared to control mice (p < 0005 and p < 0.0001, respectively). Biopsy analyses reveal that melanoma patients resistant to anti-CTLA-4 and anti-PD-1 have similar alterations to glycolytic activity. These data suggest that tumor and T cell metabolic adaptations can facilitate CPI resistance in a pre-clinical model, and that CPI-resistant tumors from melanoma patients have similar metabolic characteristics.

Glossary

AE = adverse event

CI = confidence interval
CPI = checkpoint inhibitor

CR = complete response

DC = dendritic cell

DCR = disease control rate

DLT – dose limiting toxicity
ECAR = extracellular acidification rate

HR = hazard ratio

IHC = immunohistochemistry

irAE – immune-related adverse event

MHC = major histocompatibility complex

mTTR = mean time to response

NK = natural killer
OCR = oxygen consumption rate

OS = overall survival

ORR = objective response rate

PFS = progression-free survival

PR = partial response

Q2W = every 2 weeks

Q3W = every 3 weeks

TCR = T cell receptor

TIL = tumor infiltrating lymphocyte

TRAE = treatment-related adverse event