Biomarker Analysis in a Phase Ib Study Combining Ibrutinib with R-CHOP in Patients with CD20-Positive B-Cell Non-Hodgkin Lymphoma (NHL)

Introduction

The primary objective of the DBL1002 study was to determine the recommended phase 2 dose and assess dose limiting toxicities of ibrutinib in combination with R-CHOP in subjects with CD20-positive B-cell NHL (diffuse large B-cell lymphoma [DLBCL], mantle cell lymphoma [MCL], and follicular lymphoma [FL]). Secondary objectives included assessment of pharmacodynamic markers of ibrutinib in peripheral blood mononuclear cells, as well as identification of biomarkers predictive of response to the combination regimen.

The clinical study is described in detail elsewhere (Younes, et al. 2013, 2014). Briefly, 33 subjects were enrolled in the study, of which 29 completed 6 cycles. The ibrutinib + R-CHOP combination demonstrated clinical activity across all dose cohorts in the all treated population (all subjects were treatment-naïve DLBCL, MCL, and FL) with an ORR of 90.9%. This abstract presents the results from evaluation of subject samples with SomaLogic SOMAscan proteomic analysis, flow cytometry (phospho-flow and immunotyping), Foundation Medicine (FM) sequencing, and gene expression profiling. Patients’ plasma samples were evaluated with SomaLogic SOMAscan to measure secreted levels of 1124 protein markers at baseline, cycle 2 day 1 (C2D1), and cycle 4 day 1 (C4D1). As a part of pharmacodynamics biomarker assessment, PBMCs were evaluated at cycle 1 day 1 (C1D1), cycle 1 day 3 (C1D3), C2D1, C4D1, and end of treatment (EOT) time points by flow cytometry for cell surface markers and phosphorylated kinases in the BCR pathway. The primary tumor samples were evaluated by FM sequencing and gene expression profiling. Some subjects did not have sufficient tumor biopsies (bone marrow or lymph node) available, therefore a subset of PBMC samples enriched for CD19+ cells from time points C1D1 (N = 9) or C1D3 (N = 4) or C2D1 (N = 2) or EOT (N = 3) were used for sequencing.

The overall response rate in evaluable DLBCL patients was 100%. In 13 patients whose lymphoma’s subtype for cell-of-origin was determined by the Hans method of immunohistochemistry at the central laboratory, 2 of the GCB sub-type were not evaluated for post baseline assessment (both subjects prematurely discontinued: 1 due to non-compliance and 1 due to rituximab infusion reaction prior to exposure of ibrutinib). Of the 11 evaluable subjects, 4/4 subjects with non-GCB DLBCL achieved CR (for CR rate of 100%), 5/7 subjects with GCB-DLBCL achieved CR (for CR rate of 71.4%), and 2/7 with GCB-DLBCL achieved PR for CR rate of 81.8% (9/11). Differential expression of proteins revealed that Aurora kinase B and HCK were overexpressed at baseline in GCB subjects, whereas peroxiredoxin-1 and interferon gamma were among those significantly overexpressed in the non-GCB subjects. After one cycle of therapy across all patients, matrilysin and WNT7A were decreased while IL-12 and macrophage colony stimulating factor-1 were increased. Phospho-flow analysis of BCR signaling pathways (p-BTK, p-PLCγ2, p-ZAP70/SYK, p-NFκB, p-AKT, and p-ERK1/2) in CD19+ B-cells revealed that these pathways were active in all treated DLBCL patients. No major changes were seen with treatment in the levels of cell surface proteins analyzed, including CD38 and CD79b. Baseline sequencing data showed that variants in MLL2, CREBBP, BCL2, TNFRSF14, and MYC were observed, consistent with recent reports in DLBCL. Mutational analyses correlating to response cannot be made at this time as the response rate among DLBCL patients was 100% and only one patient had progressive disease at the last data cutoff.

In conclusion, ibrutinib once daily in combination with R-CHOP is active in treatment-naïve subjects with NHL, with an ORR of 100% in DLBCL patients at the recommended dose. Biomarker analyses showed that the BCR signaling pathway was active at baseline in these patients, and revealed differential protein expression patterns between GCB and non-GCB patients. Sequence variants were noted which are consistent with recent reports in DLBCL.