AUTHORS
Danny V. Jeyaraju, PhD, Manuel Ugidos, PhD2, Keyur Desai3, Alberto Risueño4, Francesco Passamonti, MD5, Haifa Kathrin Al-Ali, MD6, Julie Vienne Buerki7, Fabian Sanabria8, Rajasekhar Suragani, PhD9
AFFILIATIONS
1 Bristol Myers Squibb, Princeton, NJ, United States
2 Bristol Myers Squibb, Princeton, NJ, United States
3 Bristol Myers Squibb, Princeton, NJ, United States
4 Bristol Myers Squibb, Princeton, NJ, United States
5 Dipartimento di Oncologia ed Onco-Ematologia, Università degli Studi di Milano, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Milan, Italy
6 Universitätsklinikum Halle (Saale), Halle, Germany
7 Bristol Myers Squibb, Boudry, Switzerland
8 Bristol Myers Squibb, Boudry, Switzerland
9 Bristol Myers Squibb, Princeton, NJ, United States
ABSTRACT
Background: MF is characterized by ineffective erythropoiesis, leading to anemia. LUSPA, an erythroid maturation agent, binds to TGF-β superfamily ligands to suppress SMAD2/3 signaling and enhance early- and late-stage erythroid differentiation. INDEPENDENCE (NCT04717414) is a phase 3, randomized, placebo-controlled study evaluating LUSPA in TD MF pts receiving JAKi. To understand LUSPA’s biology in MF, we integrated serum pharmacodynamics, single-cell RNA-sequencing (scRNA-seq), and mutational profiling.
Methods: Serum biomarkers were measured at Day 1 and Day 169 in responders within the biomarker evaluable population. scRNA-seq analysis was performed in peripheral blood mononuclear cells (PBMCs) in a subset of pts at baseline (BL) only or BL and Wk 24. Targeted next-generation sequencing (NGS) was used to assess BL mutational status (74 gene panel). The above analyses were correlated against the primary endpoint of red blood cell (RBC) transfusion independence for ≥12 consecutive wks starting within the first 24 wks (RBC-TI ≥12) or other clinical parameters.
Results: Biomarker analyses confirmed target engagement and preservation of LUSPA’s mechanism of action in responders. Compared to Day 1, at Day 169, LUSPA RBC-TI ≥12 responders demonstrated significant changes in serum biomarkers, including GDF11, a negative regulator of erythropoiesis (decreased from 113 to 92 pg/mL; P=0.015), hemoglobin (increased from 8.58 to 9.40 g/dL; P=0.021), reticulocytes (increased from 65 to 106; P<0.001), and serum EPO levels (increased from 163 to 402 U/L; P=0.010), indicating improved erythropoiesis. Notably, the index of ineffective erythropoiesis (sTfR1/reticulocyte ratio) decreased (459 to 296; P=0.006). Hepcidin, a marker of iron availability, also decreased (233 to 178; P=0.039), suggesting improved iron homeostasis and sustained JAKi effect. None of these changes reached significance in placebo responders (all P≥0.15), confirming a LUSPA-specific signal.
Longitudinal scRNA-seq mirrored observed biological changes in the study. In RBC-TI ≥12 responders, TGF-β/SMAD signaling, MYC, IL6-JAK-STAT, and G2/M checkpoint pathways, along with heme metabolism, declined by Wk 24. In the pool of megakaryocyte-erythroid progenitor (MEP) signature-positive erythroid cells, the MEP erythrocyte fraction that represents the proportion of mature RBCs, decreased from ~75% to 30% in responders and remained stable in non-responders. As scRNA-seq of PBMCs preferentially captures immature erythroid precursors, this reduction is interpreted as a shift toward terminal erythroid differentiation, consistent with the observed reticulocyte dynamics. In contrast, responders showed reduced HSC/MPP (myeloid stem cell/megakaryocyte progenitor), megakaryocyte, and common myeloid progenitor/monocytic fractions, alongside increased CD8+/NK (cytotoxic T cell/natural killer cell) fractions. Interestingly, JAK/STAT and interferon-γ activity remained correlated at both timepoints (R=0.52/0.65, P<0.05), suggesting LUSPA did not disrupt JAKi-mediated JAK-STAT suppression.
BL mutational profiling of 217 pts revealed a diverse and balanced mutational landscape. JAK2 V617F was most prevalent (68% of cases), followed by ASXL1, TET2, CALR, and other genes. A random-effects meta-analysis of RBC-TI ≥12 by mutation showed a favorable effect of LUSPA across mutational subgroups. The pooled risk difference was 0.17 (95% CI, 0.11–0.22) and significant benefit was observed in pts harboring ASXL1, TET2, and HMR mutations. Importantly, independent of treatment (tx) arm and disease risk (MIPSS70+ balanced; P=0.84), pts who subsequently transformed to AML harbored a pre-tx CD34+ hematopoietic progenitor cell state at BL, characterized by enhanced stemness signatures (elevated LSC17 score, notably within MPP-GMP compartment), impaired myeloid maturation programs, and suppressed interferon-γ responses. These features associated with AML transformation were detectable before starting therapy.
Conclusion: LUSPA engages GDF11 and activates erythroid maturation in responders without disrupting concomitant JAKi activity. Benefit was observed across mutational backgrounds, and AML transformation appears to be driven by pre-existing disease-intrinsic progenitor biology rather than tx exposure. Collectively, these mechanistic findings provide a coherent biological rationale supporting the clinical benefit of LUSPA in TD MF pts receiving JAKi.
DISCLOSURES
Danny V. Jeyaraju, PhD: Bristol Myers Squibb - Current equity holder in publicly-traded company; Bristol Myers Squibb - Current Employment