Abstract 3290: Evaluation of immune phenotypes and gene expression profiles in tumors from patients treated with bintrafusp alfa

Abstract

Abstract Background: Bintrafusp alfa (M7824) is a first-in-class bifunctional fusion protein composed of the extracellular domain of the TGF-βRII receptor (a TGF-β “trap”) fused to a human IgG1 mAb blocking PD-L1. In expansion cohorts of global, phase 1, open-label trials (NCT02517398 and NCT02699515), extensive analysis was conducted to identify potential tumor biomarkers correlated with bintrafusp alfa efficacy. Methods: Immunohistochemistry was performed on archival or fresh formalin-fixed and paraffin-embedded tumor samples using anti-PD-L1 antibody clone 73-10. Based on H&E and PD-L1 staining with negative controls, tumors were scored as having an immune desert (<1% of the tumor microenvironment area populated by immune cells [ICs]), immune excluded (≥1% of the tumor stroma area populated by ICs that do not infiltrate the tumor), or inflamed (abundance of tumor-infiltrating ICs) phenotype by a pathologist blinded to clinical responses. Whole transcriptome messenger RNA sequencing was performed on tumor samples to determine expression levels of various genes/genes signatures associated with immune and TGF-β pathways. Results: The frequency of immune phenotypes varied by tumor type. Triple negative breast cancer (TNBC) and biliary tract cancer (BTC) had no/few inflamed tumors (0/30 and 2/28, respectively) while gastric cancer (GC) and esophageal squamous and adenocarcinoma cancers (ESCC and ESAD) had few immune desert tumors (GC, 1/30; ESCC, 1/30; ESAD, 1/27). Response to bintrafusp alfa varied by immune phenotype. The majority of responders to bintrafusp alfa in GC and squamous cell carcinoma of the head and neck (SCCHN) had inflamed phenotypes (4/5 and 3/4, respectively) while non-small cell lung cancer (NSCLC) responders had similar numbers of inflamed (7/17) and immune excluded (9/17) phenotypes. Most responders to bintrafusp alfa in ESCC, ESAD, TNBC, and BTC had immune excluded tumors (3/3, 5/6, 2/3, 5/6, respectively). Responses rarely occurred in immune desert tumors (n=1 each in NSCLC, BTC, and TNBC). Evaluation of genes/gene signatures found inflamed tumors had a significantly higher level of immune genes/gene signatures compared to immune excluded or immune desert tumors. In TNBC and ESCC there were significantly higher levels of some genes related to TGF-β biology in immune excluded tumors compared to the other two phenotypes. This potential link to TGF-β biology was less clear in other tumor types. Immune excluded and inflamed tumors could not be distinguished by molecular methods across all indications. Conclusion: The data suggest that both pathology evaluation of immune phenotypes and molecular methods are useful to understand response to bintrafusp alfa. Moreover, several complementary biomarkers are likely needed to more precisely predict response to bintrafusp alfa, a drug designed to have a bifunctional mechanism of action, across cancer indications. Citation Format: George Locke, Christian Ihling, Christoph Helwig, Laureen S. Ojalvo, Olaf Christensen, Isabelle Dussault. Evaluation of immune phenotypes and gene expression profiles in tumors from patients treated with bintrafusp alfa [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3290.