Safety and Preliminary Antitumor Activity of the Anti-PD-1 Monoclonal Antibody REGN2810 Alone or in Combination with REGN1979, an Anti-CD20 x Anti-CD3 Bispecific Antibody, in Patients with B-Lymphoid Malignancies

Introduction: Immune checkpoint inhibitors and T-cell engaging bi-specific antibodies (Abs) have shown clinical activity in lymphoid malignancies. Combinations of these agents provide promising therapeutic options in hematological diseases. Here we report findings for 69 patients (pts) with B-cell lymphoid neoplasms treated with the REGN2810 anti-PD1 Ab as a single agent or in combination with REGN1979, a CD20xCD3 bispecific full-length T-cell engaging Ab.

Methods: This is an open-label, multi-center 2-arm 3+3 dose escalation trial to evaluate the safety and activity of REGN2810 alone or in combination with REGN1979 in pts with B-lymphoid malignancies. Arm 1 evaluates 1 or 3 mg/kg REGN2810 every 2 weeks (Q2W) in pts with indolent B-NHL, aggressive B-NHL, or HL, followed by expansion cohorts for each group. Arm 2 evaluates Q2W REGN2810 in combination with REGN1979 (weekly x 11, followed by Q2W) in pts with indolent or aggressive B-NHL, followed by expansion cohorts for each group. For pts receiving REGN1979, the full dose is preceded by a lower initial dose. Safety is determined by assessment of pre-specified dose limiting toxicities (DLTs). Responses are assessed by Cheson 2007 criteria and by Lugano 2014 criteria for metabolic response.

Results:

Arm 1 (REGN2810) : 60 pts with HL (n=25) or B-NHL (n=35) were treated. Most HL were nodular sclerosis (n=14) or mixed cellularity (n=5) subtypes, and most B-NHL were DLBCL (n=18) or FL (n=11) subtypes. Median number of prior therapies was 3 (range 1-12). At the data cut-off, 22 pts were on treatment, and 38 pts discontinued (the majority [28] due to PD). Median number of REGN2810 doses was 7 (range 1-24). There were no DLTs. The most common treatment-related adverse events (TR-AEs) were pyrexia (12%), fatigue (12%), infusion-related reaction (IRR) (10%), and chills (10%). The most common Grade ≥ 3 TR-AE was neutropenia (5%). 4 patients died while on treatment or within 30 days of discontinuation due to Lyell syndrome (TEN) and stomatitis, acute liver failure, urosepsis, or PD. Overall response rate (complete response [CR] and partial response [PR]) was 52% for HL (12% CR, 40% PR) and 17% for DLBCL (11% CR, 5.6% PR). 48% of HL patients had a metabolic response (12% CMR, 36% PMR), 17% of DLBCL pts had a metabolic response (11% CMR, 5.6% PMR) and 9.1% of FL pts had a PMR.

Arm 2 (REGN1979+REGN2810) : 9 pts were treated with REGN2810 3mg/kg q2W in combination with REGN1979 at doses ranging from 2-4 mg (initial dose 1 mg). B-NHL subtypes included DLBCL (n=6) and FL, MCL, and DLBCL transformed from FL (1 pt each). Pts received a median of 5 (range 2-8) prior regimens. Pts received a median of 6 (range 1-10) doses of REGN2810 and 11 (range 1-15) of REGN1979. At the data cut-off, 6 pts were on treatment, and 3 pts discontinued (2 PD; 1 death due to unrelated AE of intracranial hemorrhage). There were no DLTs. The most common TR-AEs were cytokine release syndrome (CRS) (89%); increase in ALT (67%); pyrexia (67%); and increase in AST, alkaline phosphatase, LDH, and GGT (56% each). The most common Grade ≥ 3 TR-AEs were lymphopenia (44%); and increase in ALT and AST, and leukopenia (33% each). CRS/IRR events were mostly grade 1 or 2, with the majority occurring after the initial dose. 5 pts were assessable by Cheson criteria (3 SD and 2 PD), and 3 pts by Lugano criteria (2 NMR and 1 PMD). Evaluation of higher dose levels is ongoing.

PK: REGN2810 and REGN1979 concentrations in serum, as a single agent or in combination therapy, were similar to those observed with single agent therapy in the first-in-human studies.

PD/Cytokines: Interim analysis indicates REGN2810 monotherapy (n=44) induces no or modest changes in cytokine levels, while combination therapy (n=6) induces robust cytokine release. No augmented effect of combination therapy was observed on cytokine levels when compared to the effects of REGN1979 monotherapy explored in the first in human study. Transient target occupancy on CD3+ T cells (n=6) assessed by flow cytometry in the peripheral blood reflected the corresponding circulating REGN1979 concentrations in serum.

Conclusion: REGN2810 as a single agent is well tolerated and has preliminary evidence of clinical activity in patients with HL and B-NHL. Combination REGN2810 with REGN1979 has been well tolerated, and evaluation of the combination with higher REGN1979 doses continues. CRS/IRR events were similar in frequency and severity to those previously observed in single agent studies of REGN1979.