Wu, Dianna

Ciltacabtagene autoleucel (cilta-cel) consists of autologous T cells genetically modified to express a two-binding domain chimeric antigen receptor (CAR). The target antigen of the CAR is B-cell maturation antigen (BCMA), which is highly expressed on malignant plasma cells (PC) from multiple myeloma subjects. In the Phase 1b/2 CARTITUDE-1 study, cilta-cel led to early, deep, and durable responses and was approved for the treatment of patients with relapsed or refractory multiple myeloma (RRMM) after four or more prior lines of therapy by FDA. Despite cilta-cel's outstanding overall response rate (97.9%) and duration of response (DOR) in RRMM patients, relapse still occurs. Biomarkers associated with this durable response and acquired resistance remain to be elucidated. We present updated correlative data from the CARTITUDE-1 study. At the ~3 year follow-up (October 2022 data cutoff), 97 patients had received a single infusion of cilta-cel (median 0.71 x10 6 cells/kg [range 0.52 x10 6-0.94 x10 6]); the median follow up was 33.4 months (range 1.5-45.2), median DOR was 33.9 months (95% CI, 25.5-not estimable [NE]) and median PFS was 34.9 months (95% CI, 25.2-NE) (Lin Y et al, ASCO 2023). Drug product (DP), baseline and on-treatment whole blood and bone marrow samples from enrolled patients were analyzed by methods including flow cytometry, MSD immunoassays, Cellular Indexing of Transcriptomes and Epitopes by Sequencing (CITE-seq) and TCR sequencing. In CARTITUDE-1, the DP contained a mixture of transduced and non-transduced T cells; the median transduction efficiency was 16% (range 5-32%), with a balanced distribution of CAR+CD4+ and CAR+CD8+ cells, median frequency 12% (range 2-28%) and 6% (range 2-20%), respectively. Further, the DP T cell subset composition was variable but balanced between central (Tcm) and effector memory (Tem) phenotypes. CAR+ T cells expanded reaching a median peak concentration (C max of 730 cells/µL; range 3-13805 cells/µL) in blood between days 12-14 post-infusion and persisted in circulation for a median of 100 days (range 20-912 days). High efficacy (best response) and DOR were achieved despite variable CAR-T cell expansion and lack of detectable CAR-T cell persistence over time. At C max, the CD4:CD8 ratio (based on % of CAR+ cells) was 0.29, demonstrating a CAR+CD8+ T cell preferential expansion, and both CAR+CD4+ and CAR+CD8+ T cells were predominantly of a Tcm phenotype, median 95% (range 62-99.5%) and 96% (range 33-99.7%), respectively. Higher frequency of early-memory T cell phenotypes has been associated with clinical responses to CAR-T therapies in MM (Dancy et al., Blood 2018). CITE-seq analyses showed that DP characteristics such as a high CD8+ stem-like phenotype and a low CD4+ Treg-like phenotype in the CAR+ T cell compartment associated with longer DOR. Some of the most abundant clones in the DP single cell TCR data were CAR+CD8+ T cells expressing granzyme B, suggesting high cytotoxic potential. Additional bulk TCR sequencing data suggest both diversity and individual clonal expansion longitudinally after cilta-cel infusion. We also evaluated tumor and patient intrinsic characteristics at baseline. BCMA expression on bone marrow PC was prevalent and highly variable among RRMM patients; however, expression did not associate with best response or DOR. Further, high ORR was observed in patients with high-risk cytogenetics, high tumor burden (>= 60% bone marrow PC), or baseline plasmacytomas, although with a non-statistically significant shorter PFS. Sustained MRD negativity strongly associated with longer DOR. Conversely, patient-intrinsic characteristics associated with inflammation (CRP, ferritin, neutrophils, pro-inflammatory cytokines) correlated with shorter DOR. Treatment benefit of the magnitude seen with cilta-cel had not been reported with any other approved therapy in this setting. Correlative analyses suggest an emerging profile to help understand such outstanding clinical activity: a balanced CD4/CD8 ratio and Tcm/Tem phenotype of the manufactured DP and preferential expansion of CAR+CD8+ Tcm cells. Further, certain baseline patient-intrinsic and tumor characteristics may explain the longer DOR in the RRMM setting. These investigations help identify markers of response to cilta-cel and may lead to CAR-T cell design or manufacturing strategies that enhance drug product characteristics and thus clinical efficacy.

Abstract Background: Previous data on the relationship between PD-L1 expression and activity of anti-PD-1 and anti-PD-L1 antibodies have been conflicting. MK-3475, a humanized monoclonal IgG4 antibody against PD-1, has demonstrated durable antitumor activity in MEL and NSCLC. We evaluated tumor PD-L1 expression and its relationship with outcomes in a phase I clinical trial of MK-3475. We also evaluated T cell activation as a pharmacodynamic marker of MK-3475 activity. Methods: 135 MEL pts received MK-3475 10 mg/kg Q2W, 10 mg/kg Q3W, or 2 mg/kg Q3W. Tumor response was assessed by RECIST 1.1 per independent central review, including requirement for a confirmatory CT scan. Pretreatment biopsy was required. Serial blood samples were collected before infusion at the start of cycles 1-5 (Q2W dosing) or 1-4 (Q3W dosing). Tumor PD-L1 expression was assessed by IHC. A preliminary cutoff of 1% of stained cells was used to define PD-L1 positivity. Absolute CD4+ and CD8+ T cell counts and the percentage of activated (HLA-DR as marker) CD4+ and CD8+ T cells in peripheral blood were assessed by multiplex flow cytometry (n = 101). Results: Median PFS was 36 weeks, 6-month overall survival (OS) rate was 89%, and 12-month OS rate was 81%. Median duration of response and OS were not reached. In the 116 pts with measurable disease, ORR was 41%. PFS and response rate were significantly associated with tumor PD-L1 expression (Table). At week 6, a statistically significant percent increase from baseline in the percentage of activated (HLA-DR+) CD4+ and CD8+ T cells was observed at all doses (pooled mean percent change [SE], +24.0% [4.7%] for HLA-DR+/CD8+, +17.5% [2.7%] for HLA-DR+/CD4+). Conclusions: Tumor PD-L1 expression levels were associated with tumor response and PFS in MEL pts treated with MK-3475; activity was also observed in pts with low PD-L1 expression. As assessed by HLA-DR expression, post-treatment T cell activation in the circulating pool was increased at all doses tested. Patients WithMeasurable Disease and Interpretable PD-L1 IHC ResultsPatients With InterpretablePD-L1 IHC Results and PFS and OS Data PD-L1 TumorExpressionNORR, n (%)NPFS at 6 months, % (95% CI)Median PFSOS at 6 months, % (95% CI)Median OSPositive (membrane staining in ≥1% of cells)5529 (53)6058 (47-72)10.6 mo93(87-100)Not reachedNegative (membrane staining in <1% of cells)161* (6)2232 (16-64)2.9 mo75(58-97)Not reachedHR (95% CI) for PD-L1-positive vs negative____0.54(0.28-1.05)_0.67(0.25-1.83)One-sidedP value(PD-L1 association test)_<0.004 (logistic regression)__0.034(Cox regression)_0.220(Cox regression)*IPI-pretreated patient who received MK-3475 10 mg/kg Q2W. Citation Format: Adil I. Daud, Omid Hamid, Antoni Ribas, F. Stephen Hodi, Wen-Jen Hwu, Richard Kefford, Jedd Wolchok, Peter Hersey, Jeffrey S. Weber, Richard Joseph, Tara C. Gangadhar, Roxana S. Dronca, Amita Patnaik, Hassane Zarour, Anthony M. Joshua, Kevin Gergich, Dianna Wu, Jared K. Lunceford, Kenneth Emancipator, Marisa Dolled-Filhart, Nicole Li, Scot Ebbinghaus, S. Peter Kang, Caroline Robert. Antitumor activity of the anti-PD-1 monoclonal antibody MK-3475 in melanoma(MEL): Correlation of tumor PD-L1 expression with outcome. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr CT104. doi:10.1158/1538-7445.AM2014-CT104