Integrated single cell network profiling data of ERK signaling and mutations of SF3B1 gene refine prognosis in chronic lymphocytic leukemia

Introduction: Extracellular signal-regulated kinase (ERK) is a major pathway downstream of the B-cell receptor (BCR) and its activation is a significant, independent predictor of shorter clinical progression in chronic lymphocytic leukemia (CLL). Recently identified driver genetic lesions, most likely favored by BCR signals, have reshaped the genetic landscape and added a further level of complexity in CLL. Despite the established driving role of BCR signaling and gene mutations in pathobiology and clinical behavior of CLL, little is known about the clinical influence of integrated BCR responses and genetic alterations. In this study, we investigated the clinical impact of integrated BCR response of ERK and gene mutations in CLL. Methods: Peripheral blood cell samples at diagnosis from 152 CLL patients were analyzed in this study. ERK phosphorylation was analyzed using Single Cell Network Profiling (SCNP), a flow cytometry-based assay that allows signaling analysis at the single cell level. NOTCH1, SF3B1, TP53, MYD88, and BIRC3 gene mutations were analyzed by direct DNA sequencing. Univariate and multivariate models for time to first treatment (TTFT) were generated using Cox proportional hazards regression. TTFT curves estimated using the Kaplan-Meier method for the respective groups of patients were compared using the log-rank test. Results: BCR stimulation with anti-IgM induced a signaling response of ERK (anti-IgM→p-ERK) that was significantly higher in UM subset (P=0.0020), in CD38-positive CLL (P=0.0059), in treated patients (P=0.0003), and in mutated- SF3B1 cells (P=0.0098). Univariate analysis identified increased anti-IgM→p-ERK (P=0.001), UM-IGHV (P<0.0001), positive CD38 (P=0.001), mutated SF3B1 (P<0.0001), as significant, independent predictors of shorter TTFT. In multivariate analysis, only increased anti-IgM→p-ERK (P=0.03) and SF3B1 mutation (P=0.0001) were independent significant parameters of prognosis. Integrating anti-IgM→p-ERK data and SF3B1 mutations stratified patients in three independent prognostic categories (P<0.0001) and identified an intermediate-risk group that included patients with low p-ERK and mutated SF3B1 or high p-ERK and wild-type SF3B1. Consistent with the results in the whole patient set, integrating anti-IgM→p-ERK and SF3B1 mutation in an unselected group of Binet stage A patients (n=110) identified an intermediate-risk group including patients with low p-ERK and mutated SF3B1 or high p-ERK and wild-type SF3B1 (P=0.0002). Conclusions: These data reveal that integrated dynamic ERK signaling and SF3B1 mutations independently predict disease progression risk in CLL and identify a novel risk-group of patients, thus providing complementary prognostic information and suggesting a functional synergy between BCR-induced ERK signaling and altered SF3B1.