Predictive Biomarkers of Response to Luspatercept in Patients with Myelofibrosis- (MF) Associated Anemia: Biomarker Analysis from the ACE-536-MF-001 Study

Abstract

Introduction: MF is a clonal myeloproliferative neoplasm that results in fibrosis of the bone marrow and impaired hematopoiesis. Among other clinical manifestations, ineffective erythropoiesis leads to anemia resulting in poor quality of life and reduced overall survival (Passamonti et al., Blood 2010). Current treatment options for MF-associated anemia with most agents lack optimal efficacy and durability of response (Cervantes et al., Expert Rev Hematol., 2016). Luspatercept is a transforming growth factor-β (TBF-β) family ligand trap that inhibits the TGF-β-Smad signaling axis and promotes late-stage erythropoiesis. Luspatercept is approved in low-risk MDS with ring sideroblasts and β-thalassemia requiring red blood cell (RBC) transfusion. The ACE-536-MF-001 study investigates the safety and efficacy of luspatercept in the treatment of MF-related anemia. Here, through mutational and cytokine analysis of ACE-536-MF-001 biomarker samples, we sought to identify mutational and cytokine profiles associated with response. Methods: Eligible patients were divided into 4 cohorts based on transfusion dependence (TD) and treatment with ruxolitinib (RUX). Cohorts 1 and 3A were non transfusion dependent (NTD) while cohorts 2 and 3B were TD patients receiving 4-12 RBC units/84 days up to C1D1. Cohorts 3A and 3B received a stable dose of RUX. All patients received 1.0 mg/kg of luspatercept (1.33 mg/kg for cohort 3B) with titration up to 1.75 mg/kg once every 21 days. Primary endpoint was anemia response (defined as an increase of ≥ 1.5 g/dL hemoglobin (Hgb) from baseline for cohorts 1/3A or RBC transfusion free for ≥ 12 consecutive weeks for cohorts 2/3B). Secondary endpoint was mean Hgb increase ≥ 1.5 g/dL from baseline for ≥ 12 consecutive weeks for cohorts 1/3A, or ≥ 50% decrease in RBC transfusions from baseline for ≥ 12 consecutive weeks and duration of RBC transfusion-free period for cohorts 2 and 3B. Mutational analysis was performed in DNA from peripheral blood mononuclear cells using a panel targeting 74 genes. Rules-Based Medicine's Human Inflammation panel was used to measure changes in 54 serum cytokines. Nexelis custom panel was used for detection of GDF8, GDF15, erythroferrone, soluble transferrin receptor 1 (sTfR1) and hepcidin. Changes in soluble factors from cycle 1 to end of week 24 (day 169) were examined for correlation with clinical response. Results: All 4 cohorts (N = 95) displayed improvement in anemia and TD patients receiving RUX had the greatest benefit (Gerds et al. EHA 2023 S167). In the biomarker cohort (n = 63 across all 4 cohorts), mutational analysis revealed that most patients carried the JAK2 V617F mutation, and patients with SF3B1 or SF3B1 with co-occurring DTA mutations ( DNMT3A, TET2 or ASXL1) had a favorable prognosis related to anemia response (p= 0.0260, 0.0495; OR= 3.54, 3.86). In addition, JAK2 V617F patients in cohort 3B showed clinical benefit with luspatercept (OR= 6.75, p= 0.0445). Across cohorts and endpoints, the general trend of responders was low baseline (2-3-fold lower relative to nonresponders) levels of key inflammatory cytokines. Low levels of IL-10 (TD patients) and VCAM-1, TNFR2, CRP and IL-18 (NTD patients) were significant ( P < 0.01) predictors of response. Paired analysis between cycle 1 and week 24 revealed increases in alpha-2-macroglobulin, haptoglobin, TIMP1, and decrease in C3, vitamin D-binding protein, hepcidin and sTfR1 associated with anemia response. When transfusion burden reduction was used as a response criterion, ferritin, hepcidin, thrombopoietin, von Willebrand Factor, sTfR1, and beta-2-microglobulin were significantly ( P < 0.05) associated with response (Figure). In addition, Hgb increase was associated with increased TGF-β1 and VEGF levels and decreased GDF-15 and ERFE. These cytokines are implicated in iron metabolism (ferritin, hepcidin, alpha-2-macroglobulin), Hgb catabolism (haptoglobin), anemia (vitamin D-binding protein, C3), and luspatercept-target engagement (hepcidin, sTfR1). Conclusions: Our analysis identifies unique predictors of response to luspatercept treatment in patients with MF. While the data describe our findings across cohorts, cohort 3B (TD, receiving RUX), which had the highest response rate, also had higher levels of RANTES and TPO at baseline in responders. The ongoing phase 3 study, INDEPENDENCE (NCT04717414), would validate these observations in a larger cohort of patients