Paper Details

ABSTRACT

Introduction:

Development of effective low-risk oral therapies for immune-mediated bone marrow failure (iBMF) remains an unmet need. The JAK-STAT pathway plays an important role in immune cell development and function. Ruxolitinib (RUX) improved blood counts, suppressed activated T-cells, mitigated Fas-mediated apoptotic destruction of HSPCs, and improved long-term survival in a pre-clinical mouse model of iBMF, prompting development of this clinical trial.

Methods:

In this prospective investigator initiated phase I/II study (NCT05998408), we enrolled 13 patients (median 67 yrs, range 54-75) with relapsed/refractory iBMF in 3 disease cohorts; SAA (n=5), PRCA (n=1) and T-LGL (n=7). The study closed early due to withdrawal of drug support from Incyte Co.

Co-primary endpoints were: 1) Safety; rate of RUX induced hematologic toxicity, and 2) Efficacy; overall hematologic response (HR) at 6 months post-treatment. Secondary endpoints: tolerability of max RUX dose, HR at 3 and 12 months,  depth of response (partial [PR] versus complete response [CR]), transfusion independence, relapse, response to re-initiation, evolution to secondary myeloid neoplasm (SMN), overall survival.

Response criteria were disease specific. For SAA, blood counts no longer meeting Camitta criteria. For PRCA, Hb increase by >1.5g/dL. For T-LGL, CR was ANC >1.5x10⁹/L, platelets >150x10⁹/L, and lymphocyte count <4x10⁹ /L with no clonal T cells detected while for PR, lineage specific responses were used.

Patients commenced on RUX 5mg BID, titrated to a maximum 20mg BID by weekly escalation. Hematologic toxicity, either increased transfusion dependence (TD) or significantly worsened cytopenia, was assessed weekly with protocol defined dose adjustments.

Luminex panel with iBMF-related cytokines and chemokines, and flow cytometry for T, B, NK, DC, and monocytes were performed. Full-spectrum phospho-flow cytometry assay quantified intracellular phosphorylation of STATs 1, 3, 4 and 5 in major PBMC subsets following stimulation with different cytokines. Assays were performed at baseline and after 3 and 6 months on RUX.

Results:

All patients were refractory, with 10/13 having received >2 lines of therapy. Eleven were red blood cell (RBC) (n=5; SAA, n=5; T-LGL, n=1; PRCA) and five platelet (n=5; SAA) TD. HR was achieved in one patient with T-LGL and isolated neutropenia, who achieved PR with persistent T-cell clonality. Eight patients met criteria for hematologic toxicity necessitating dose reduction, due to progressive neutropenia (n=5), thrombocytopenia (n=2), or new RBC-TD (n=1). One permanently stopped RUX due to hematologic toxicity.

Notably, in three non-responders at 6 months, we observed progressive hematologic improvement after drug discontinuation. UPN-1 (T-LGL/PRCA) had RBC-TD, and decreasing TD while on study; Hb improved to >11g/dL a month after drug discontinuation but relapsed 4 months later. UPN-2 (T-LGL/AA) was RBC-TD at enrollment; a month after stopping RUX, Hb rose to >10g/dL with sustained response 8 months after drug discontinuation. UPN-3 (T-LGL/PRCA) was non-RBC-TD prior to ruxolitinib, but had low baseline Hb (6-7g/dL). New RBC-TD developed on RUX; after discontinuation, Hb rose to >9g/dL but relapsed after 3 months. Two patients with relapse re-initiated RUX off study.

Activated T-cells (CD8+/CD25+) were reduced in 4 T-LGL patients at 3 months on RUX compared to baseline, but not in 1 assessed SAA patient. There was no significant difference in cytokine levels on/off therapy. In one T-LGL responder, we noted reduction from baseline to 3 months on RUX in CXCL9 (117.85 v 0), IFN-g (2.26 vs 0.70), and TNF-a (9.16 vs 4.56). In the three late responders, we saw reductions at 3 months in CXCL9 (2/3) and TNF-a (3/3) but not in IFN-g. There was overall reduction of STAT1 and STAT3 phosphorylation and frequency of pSTAT1+ and pSTAT3+ total T cells and activated CD25+ T cells at 3 months on RUX compared to baseline (n=4/5), both unstimulated (physiological) and when stimulated with IFN-a, IFN-g, IL-6, IL-10, and TNF-a. STAT5 phosphorylation response to inflammatory cytokines was not affected by RUX.