Hematological and inflammatory response trajectories in vexas syndrome patients treated with hypomethylating agents

Abstract Background: VEXAS (Vacuoles, E1-enzyme, X-linked, Autoinflammatory, Somatic) syndrome is a severe autoinflammatory syndrome frequently resulting in progressive bone marrow failure, including myelodysplastic syndromes (MDS). Hypomethylating agents (HMA) have demonstrated efficacy in VEXAS, but have high rates of cytopenic and infectious toxicities. We prospectively studied the response trajectory of dysfunctional hematopoiesis at a predefined time point post HMA therapy in a prospective cohort, and correlated outcomes with quality of life (QoL). Methods: Patients enrolled on a natural history study (NCT 0257866) at the National Institutes of Health (NIH) receiving HMA treatment were included. Laboratory and clinical data were collected from electronic medical records. Standardized evaluations, including complete blood counts (CBC), inflammatory markers, cytokine profiling, bone marrow (BM) biopsies and next-generation sequencing were performed pre-treatment and every 6 months (mos) in most patients. Medical Outcomes Survey Short Form (SF)-36 questionnaires from 2018-2024 were offered at first and follow-up visits. Hematological improvement was assessed using revised IWG criteria. CBC and CRP means (95% CI) were estimated at baseline and at 6mos post-HMA using generalized additive mixture models (Turturice et al. Ann Rheum Dis 2025). Complete molecular response (MR) was defined as UBA1 VAF <2% and partial MR as a ≥25% reduction in initial UBA1 VAF. Inflammatory complete response (iCR) was defined as asymptomatic state and prednisone dose <10 mg/day. Results: Twenty-six male patients with median age of 69 years (range, 41-85) were included. Canonical UBA1 variants were confirmed in all but one: M41L (n=12), M41V (n=6), M41T (n=4), splice motif (n=3). Most patients were treated with azacitidine in a 5- (n=18) or 7-day (n=4) regimen. The median number of prior lines of therapy was 4 (range, 0-14); 25 (96%) were on prednisone (median 25.5mg, range 0-70) at HMA initiation, and 15 received a concomitant steroid-sparing agent, mostly tocilizumab (n=8). Thirteen (50%) had a diagnosis of MDS, mostly lower-risk per IPSS-R (n=10/13: low blast count (<5%) (n=11), normal karyotype (n=10)). Cytopenias were present in 20 patients (Hb<10 (n=16), Plt<100K (n=13), ANC <500 (n=1)) and 11 were RBC transfusion-dependent. DNMT3A (n=8) and TET2 (n=8) were commonest co-occurring mutations. First response assessment at NIH occurred after median of 6mos (range, 2-45mos). None achieved iCR; median prednisone dose decreased from 26mg to 20mg (p=0.65). Paired (n=8) comparisons of cytokines showed increased levels of IL-18 (p=0.016) but other cytokines and CRP levels were not significantly different at 6mos. Three (17%) patients achieved complete MR and 9/17 (53%) a partial MR. Hematological improvement (HI) was achieved in 12 (67%) of the 18 evaluable patients (HI-E: n=9/15, HI-P: n=7/12, HI-N: n=0/1). CBC was assessed at 512 time points prior and 322 post-HMA. CBC expected means were largely unchanged between baseline and 6mos without HMA therapy. There was a significant improvement in Plt count post-HMA (mean difference HMA vs no HMA: 38x103/mL [95% CI: 24.89, 51.11]). A trend towards increased Hb (0.91g/dL [-0.01, 1.83]) was observed and 6 patients became RBC transfusion-independent. ANC was significantly reduced (-1.57x103/mL [-2.95, -0.20]) while other blood counts did not significantly change. BM cellularity significantly decreased post HMA from a median value of 85% to 35% (p=0.01). The median proportion of vacuolated cells decreased by 67% (p=0.008). The M:E ratio down-trended from 5:1 to 2.2:1 (p=0.07). QoL SF-36 responses (n=8) did not change significantly. Commonest adverse events were hospitalization for infection, sepsis or VEXAS flare, occurring in about half of patients. Conclusions: In a prospective cohort of heavily pre-treated VEXAS patients who were comprehensively evaluated at pre-defined 6mos post therapy, we show promising HI and MR with blood count recovery, changes in bone marrow morphology and reduction in UBA1 mutational burden, but lack of complete inflammatory control evidenced by ongoing prednisone requirement, inflammatory symptoms, increased IL-18 expression and unchanged QoL. Our iCR rates were inferior compared to data from FRENVEX, perhaps due to their retrospective design, and higher initiation dose of HMA; longer follow-up may be needed as HI has previously been shown to precede inflammatory response.