Prophylactic Ruxolitinib for Cytokine Release Syndrome (CRS) in Relapse/Refractory (R/R) AML Patients Treated with Flotetuzumab

Abstract

Introduction: CRS is a potentially life-threatening toxicity observed following T cell-redirecting therapies. CRS is associated with elevated cytokines, including IL6, IFNγ, TNFα, IL2 and GM-CSF. Glucocorticosteroids (GC) and the IL6 receptor blocking antibody tocilizumab (TCZ) can reduce CRS severity; however, CRS may still occur and limit the therapeutic window of novel immunotherapeutic agents. Disruption of cytokine signaling via Janus kinase (JAK) pathway interference may represent a complementary approach to blocking CRS. Ruxolitinib (RUX), an oral JAK1/2 inhibitor approved for the treatment of myelofibrosis and polycythemia vera, interferes with signaling of several cytokines, including IFNγ and IL6, via blockade of the JAK/STAT pathway. We hypothesized that RUX may reduce the frequency and severity of CRS in R/R AML patients (pts) undergoing treatment with flotetuzumab (FLZ), an investigational CD123 x CD3 bispecific DART® molecule. Methods: Relapse/refractory (including primary induction failure, early relapse and late relapse) AML pts were included in this study. RUX pts were treated at a single site, Washington University, St. Louis, MO. RUX was dosed at 10 mg or 20mg BID days -1 through 14. Comparator (non-RUX) pts (n=23) were treated at other clinical sites. FLZ was administered at 500 ng/kg/day continuously in 28-day cycles following multi-step lead-in dosing in week 1 of cycle 1. CRS was graded per Lee criteria1. Results: As of July 1st, 2020, 10 R/R AML pts, median age 65 (range 40-82) years, have been enrolled and treated in the RUX cohort (6 at 10mg, 4 at 20 mg of RUX). All pts had non-favorable risk by ELN 2017 criteria (8 adverse and 2 intermediate); 1 (10.0%) pt had secondary AML; pt characteristics in the RUX and non-RUX cohorts were balanced, except for median baseline BM blasts which was higher in non-RUX pts: 15% (range 5-72) vs (40% (range 7-84), RUX and non-RUX pts respectively. Cytokine analysis showed statistically significant (p<0.05) lower levels of IL4, IL12p70, IL13, IL15, IL17A, IFNα2, but higher levels of GM-CSF were measured in RUX vs non-RUX pts, specifically during co-administration with FLZ (Fig. 1). However, incidence and severity of CRS events were similar. In the RUX cohort, 9 (90%) pts experienced mild to moderate (grade ≤ 2; 48.6% of events were grade 1) CRS events whereas no grade ≥ 3 CRS were reported; in the non-RUX cohort, 23 (100%) pts experienced mild to moderate (grade ≤ 2; 73.1% of events were grade 1) CRS events, 1 (4.3%) grade ≥ 3 CRS was reported. Most CRS events occurred in the first 2 weeks of FLZ administration (75% and 92%, respectively). No differences in duration of CRS events were noted. However, more CRS-directed treatment was used in the RUX cohort. Five (50%) pts received a total of 12 doses of TCZ, 1 (10%) pt received GC and 1 (10%) pts received vasopressors in the RUX cohort. In the non-RUX cohort, 5 (21.7%) pts received 8 doses of TCZ, 3 (13.0%) pts received GC and 1 (3.7%) pt received vasopressors. Dose intensity (DI) at FLZ dose of 500 ng/kg/day was comparable, with median DI of 97.6% and 98.0% in RUX and non-RUX cohorts, respectively. Time to first response (TTFR; BM < 5% blasts) and time on treatment (ToT) were similar between both groups. Median TTFR was 1 cycle for both groups (range 1-2 cycles), and median ToT was 1.4 (range 0.9-5.1) and 1.8 (range 1.3-5.1) months, for RUX and non-RUX pts, respectively. Complete response rate (BM < 5% blasts) was similar: 4 (40%) in RUX pts, and 8 (34.8%) in non-RUX pts; 2 RUX (50%) and 5 non-RUX (62.5%) responders transitioned to stem cell transplant. Conclusion: Prophylactic RUX produced a clear difference in cytokine profiles but no discernable improvement in clinical CRS or response rates in FLZ treated patients. A larger study may be required to determine the prophylactic role of RUX in CRS.