A Phase 1, Open-Label Study of MGD013, a Bispecific DART® Molecule Binding PD-1 and LAG-3 in Patients with Relapsed or Refractory Diffuse Large B-Cell Lymphoma

Abstract

Background: MGD013 is an investigational, first-in-class, Fc-bearing bispecific tetravalent DART molecule designed to bind PD-1 and LAG-3 and sustain/restore the function of exhausted T cells (1). MGD013 demonstrates in vitro ligand blocking properties and improved T-cell responses beyond that observed with anti-PD-1 and anti-LAG-3 benchmark antibodies alone or in combination. PD-1 targeted therapy with nivolumab in patients (pts) with relapsed or refractory (R/R) diffuse large B cell lymphoma (DLBCL) has yielded modest efficacy (2). LAG-3, highly expressed in DLBCL (3), has emerged as another therapeutic target of interest in this population with continued unmet need. Methods: This study characterizes the safety, tolerability, dose-limiting toxicities, maximum tolerated dose, PK/PD, and antitumor activity of MGD013 in pts with advanced solid and hematologic malignancies. R/R DLBCL pts are being treated at the recommended Phase 2 dose of 600 mg every two weeks in Cohort Expansion. Results: At data-cutoff, 17 DLBCL pts received MGD013 (2.5 median prior lines of therapy, 41.2% with prior CAR-T therapy). Treatment-related adverse events (TRAEs) occurred in 11/17 (64.7%) pts, with no same TRAE occurring in > 1 patient except pyrexia (n=3). One grade ≥ 3 TRAE occurred (pneumonia), and no events of tumor lysis syndrome were observed. Among 11 response-evaluable pts (i.e. received at least one on-treatment scan), 1 complete response (CR) and 3 partial responses (PRs) per the Lugano Classification were observed. Analyses of available pre-treatment tumor biopsy samples corresponding to responding pts demonstrated relatively high levels of LAG-3, PD-1, and PD-L1 by IHC. More comprehensive translational analyses were undertaken for the pt with CR, observed after a single MGD013 infusion in a 28-year-old male in relapse 6 months after CD19-directed CAR T-cell therapy. In contrast to a pre-CAR T biopsy specimen demonstrating no LAG-3 or PD-1 expression, IHC analysis of a lymph node specimen biopsied post-CAR T and prior to MGD013 treatment revealed high levels of both LAG-3 and PD-1 on both infiltrating T-cells and malignant B-cells. Consistent with CD19 CAR T resistance, no CD19 expression was evident. MHC class II and PD-L1 expression were observed, which when bound to LAG-3 and PD-1, respectively, form immune checkpoints that can decrease T cell function. Following MGD013 treatment, serum IFN-γ markedly increased to >140-fold above baseline. No significant changes were observed for IL-6 or IL-2. Expansion of circulating CD3+CD8+ and CD3+CD4-CD8- T-cell subsets and associated cytolytic markers (i.e., perforin, granzyme B) were observed following MGD013 treatment. The patient underwent allogeneic stem cell transplant (allo-SCT) and remains in remission 14 months post-MGD013 and 12 months post-allo-SCT. Further correlative biomarker analyses are underway in the ongoing clinical trial. Conclusion: MGD013, a novel molecule designed to coordinately block PD-1 and LAG-3, has preliminarily demonstrated an acceptable safety profile and encouraging early evidence of anti-tumor activity in R/R DLBCL pts with and without prior treatment with CAR T. Biomarker analyses confirmed expression of both PD-1 and LAG-3 axes in responding pts with evidence of pharmacodynamic responses consistent with the ability of MGD013 to enhance T-cell function.