MGD013 Demonstrates Encouraging Activity Across Tumor Types

Jason J. Luke, MD

MGD013, an investigational DART protein targeting PD-1 and LAG-3, demonstrated encouraging monotherapy activity as well as in combination with margetuximab in multiple tumor types, according to analyses presented at the 2020 ASCO Virtual Scientific Program.

In findings from the MGD013 dose-escalation phase 1 trial, partial response was observed in 3 patients, including 1 patient with gastric cancer and 1 patient with mesothelioma, both of whom had prior PD-1 therapy, as well as 1 patient with triple-negative breast cancer (TNBC) who was on therapy for more than 5 months. Additionally, 18 patients had stable disease as best overall response for a disease control rate of 48.8%.

Further, preliminary results in patients with relapsed/refractory HER2-positive solid tumors treated with MGD013 and margetuximab in the combination cohort had an overall response rate of 42.9% (6/14).

“The rationale for dual targeting of PD-1 and LAG-3 stems from the observation that checkpoint molecules are leveraged by tumors or antigen presenting cells to evade the immune system,” explained Jason J. Luke, MD, FACP, in a presentation. “PD-1 and LAG-3 receptors are expressed on exhausted T-cells and interact with corresponding ligands in the gate anti-tumor activity.”

Synergy of anti-PD-1 and anti-LAG-3 has been observed in preclinical models, explained Luke who is an associate professor of medicine in the Division of Hematology/Oncology and the director of the Cancer Immunotherapeutics Center within the University of Pittsburg Medical Center Hillman Cancer Immunology and Immunotherapy Program.

MGD013 is an investigational bispecific DART molecule that engages PD-1 and LAG-3 with a single molecule, resulting in greater T-cell activation than either a monoclonal antibody alone.

The phase 1 trial open-label, dose escalation and cohort expansion study (NCT03219268) was designed to characterize the safety, tolerability, pharmacokinetics, immunogenicity, pharmacodynamics, and preliminary antitumor activity of MGD013 administered by IV infusion every 2 weeks. Another hypothesis investigators tested in the study was whether or not tumors can be made more responsive to PD-1 and LAG-3 intervention with the addition of an Fc-engineered monoclonal antibody, such as margetuximab.

In the initial dose-escalation cohort, patients were treated with MGD013 in a 3 + 3 design from 1 mg up to 1200 mg. Upon establishment of 600 mgs every 2 weeks as the recommended dose, monotherapy cohort expansions in ovarian, TNBC and non—small cell lung cancer (NSCLC), as well as multiple other tumor types including squamous cell carcinoma of the head and neck, small cell lung cancer, hepatocellular carcinoma, cholangiocarcinoma, cervical cancer, gastric/gastroesophageal junction carcinoma, and diffuse large B-cell lymphoma began enrolling patients.

In total, 53 patients participated in dose-escalation cohort; 205 patients participated in monotherapy cohort expansion; and 21 patients were treated in combination with margetuximab.

The 3 confirmed partial responses in the dose-escalation cohort each occurred at a different dose. The patient with TNBC had a response after treatment with a 10-mg dose; the patient with mesothelioma had a response with the 800-mg dose; and the patient with gastric cancer had a response with the 1200-mg dose.

The all-grade immune-related adverse events (AEs) of special interest observed in the dose-escalation cohort included rash (13.2%), hypothyroidism (11.3%), and immune-related hepatitis (3.8%). Grade 3 or higher AEs included immune-related hepatitis (3.8%), and 1 patient each reported respective cases of rash, pancreatitis, colitis, and adrenal insufficiency (1.9% each).

“You’ll note that broadly, [MGD013] was well tolerated with a manageable immune related adverse event profile and safety consistent with PD-1 or PD-L1 monotherapy in the treatment value setting,” said Luke. “Importantly, on-mechanism toxicities with immune-mediated hepatitis and lipase increase with radiographic evidence of pancreatitis were observed, though both resolved easily with steroids.”

Anti-tumor activity was also observed in the MGD013 monotherapy cohort expansion across multiple tumor types including epithelial ovarian cancer (n = 23), TNBC (n = 23), and NSCLC, including both checkpoint inhibitor-naïve (n = 14) and post PD-1 (n = 15). Confirmed partial response was observed in 2, 1, 2, and 0 patients, respectively. The overall response rates of both confirmed and unconfirmed partial responses were 8.7%, 4.3%, 21.4%, and 13.3%, respectively. Further, stable disease rates were 43.5%, 34.8%, 50.0%, and 53.3%.

The responses translated to a disease control rate of 52.2% in patients with epithelial ovarian cancer, 39.1% in patients with TNBC, 64.3% in patients with checkpoint inhibitor-naïve NSCLC, and 53.3% in patients with NSCLC following PD-1 treatment.

Several cohorts are still ongoing, including a lymphoma cohort, Luke said. He cited a case from this cohort of a 27-year-old man with diffuse large B cell lymphoma who had progression of disease after chimeric antigen receptor (CAR) T-cell therapy who went on to receive 600-mg dose of MGD013. The patient’s pretreatment biopsy was noted for high expression of PD-1 and LAG-3 co-currently. Following the single dose of MGD013, the patient went on to be admitted on day 11 due to grade 2 cytokine release syndrome, but had a complete response on PET scanning at day 24 per Lugano criteria, Luke said. The patient went on to have allogeneic stem cell transplant and as of the presentation, was in remission 11 months post-MGD013 and 9 months post-transplant.

“Given this early activity of MGD013 and these biomarker results, we are interested to more broadly analyze LAG-3 expression and more overall interferon expression as associated with treatment outcome,” Luke said.

Luke JJ, Patel MR, Hamilton EP, et al. A phase I, first-in-human, open-label, dose-escalation study of MGD013, a bispecific DART molecule binding PD‑1 and LAG‑3, in patients with unresectable or metastatic neoplasms. J Clin Oncol. 2020;38(suppl 15; abstr 3004). doi:10.1200/JCO.2020.38.15_suppl.3004