KLRG1 Depletion Is a Novel Therapeutic Strategy for Patients with Mature T and NK/T-Cell Lymphomas

Abstract

Outcomes for patients with aggressive T-cell lymphomas (TCLs) remain dismal. Killer cell lectin-like receptor G1 (KLRG1) is a co-inhibitory immune checkpoint receptor expressed on late differentiated CD8+ T effector memory (TEM) and TEMRA cells. Monotherapy with KLRG1 neutralizing antibodiesextended survival in solid tumor models. Here, we report the therapeutic potential of a novel and distinct strategy characterized by depletion of KLRG1+ malignant cells in hematologic malignancies and specifically in TCLs. We defined the molecular epidemiology of KLRG1 by determining tumor cell-specific RNA and protein expression on T cell subsets from 8 healthy donors, 26 TCL lines, 15 patient-derived xenografts (PDXs), and 165 tumor and blood specimens representing various subtypes of T and NK/TCLs. Microarray andRNA-seqdata demonstrated highertranscript levels of KLRG1 in patients with T-LGLL and PDXs of T-PLL and HSTCL relative to other subtypes, which were corroborated by protein expression. We quantified tumor cell-specific protein expression of KLRG1 by IHC by generating H scores (% of KLRG1+ tumor cells × measure of staining intensity) ranging from 0-300 across 141 primary TCL biopsies. PTCL-NOS and AITL had heterogeneous KLRG1 expression, whereas ALCL, CTCL, and EATL demonstrated none. The majority of patients with T-PLL (median 180), δ/γ TCLs (median 140), and NK/TCLs (median 100) exhibited higher H scores relative to other subtypes. Up to 99% of the CD3+/CD8+/CD57+ and up to 97% of the CD3-/CD56+/CD57+/CD16+/CD94+ leukemic cells in the blood of patients with T-LGLL (n=12) and CLPD-NK (n=12) respectively expressed KLRG1 protein (Fig.1). None of the TCL cell lines expressed surface KLRG1. Therefore, SMZ1 cells (TP63 rearranged PTCL-NOS line) were stably transduced to express KLRG1. We utilized two human KLRG1 binding antibodies, 13F12F2 (a commercial mIgG2a Ab) and mAb208 (a humanized afucosylated mIgG2a Ab) for in vitro and in vivo studies. Antibody binding capacity assays confirmed a higher number of KLRG1 receptors per cell on SMZ1 KLRG1+ (175,800), HSTCL (23,586), and CD8+ T cells (16,326) compared with B-cell lymphoma cells (242). Competition binding assays demonstrated that concentrations of 0.1 ug/ml of mAb208 saturated receptors on SMZ1 KLRG1+ cells. mAb208 induced a 2-5-fold dose-dependent increase in antibody-dependent cellular cytotoxicity (ADCC), antibody-dependent cellular phagocytosis (ADCP), and complement-dependent cytotoxicity (CDC), but not apoptosis of SMZ1 KLRG1+ cells relative to SMZ1 KLRG1- WT and isotype control (CTL). This was exhibited across a variety of human and murine effectors such as human PBMCs, human and murine macrophages, and human and murine complements. Reporter-based murine FcγRIV ADCC bioassay demonstrated a dose-dependent 2-6 fold induction of bioluminescence by mAb208 in SMZ1 KLRG1+ cells relative to WT (p=<0.0001). In 4/4 tested T-LGLL patient samples, mAb208 induced ADCC of CD8+/CD57+ LGLs with autologous PBMCs (p=0.03) relative to CTL. Finally, mAb208 induced selective ADCC of KLRG1+ CD8+ TEM and TEMRA cells while sparing KLRG1- naive and central memory T cells in 3 healthy donors (p=0.06). Duvelisib (Duv), a dual PI3K-δ/γ inhibitor (PI3Ki), has demonstrated a high ORR of 50% in phase II trials in TCLs and its ability to reprogram macrophages to an M1 phenotype in TCL PDXs. Thus, we hypothesized that the combination of PI3Ki with anti-KLRG1 mAbs would enhance the clearance of apoptotic and opsonized KLRG1+ TCL cells by proinflammatory macrophages in several PTCL subtypes. We tested this combination in vivo using SMZ1 KLRG1+ subcutaneous xenografts in NSG mice. After 17 days of treatment, combination (mAb208 + Duv)- treated mice had a marked reduction in their mean tumor volumes (159 mm3) relative to mAb208 (1003 mm3 ; p=0.03), Duv (465.6 mm3 ; p=0.03) and CTL (2104 mm3 ; p=0.03) mice resulting in prolonged survival (Fig. 2). Studies in PDX models of T-PLL and HSTCL with varying levels of KLRG1 expression (30% and 80% respectively) have been initiated and will be reported at the meeting. Selective depletion of tumor cells with an anti-KLRG1 depleting mAb in specific subtypes of TCLs, T-LGLL, and CLPD-NK is effective in vitro and in vivo, and its combination with Duv has marked activity in an aggressive PTCL-NOS xenograft model. A multicenter phase I/II trial of the KLRG1 depleting antibody ABC008 (Abcuro) in patients with R/R T-LGLL has been initiated. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal