Circulating Mast Cell Progenitors Are Depleted by Benralizumab in Severe Asthma

Inappropriate mast cell (MC) accumulation and activation in the airways promote symptoms and are associated with eosinophilia/T2 inflammation in severe asthma (SA) [1]. MC-targeted therapeutics are in use (anti-IgE omalizumab) or in development. The origin of increased airway MC numbers in SA is poorly understood. Circulating MC progenitors (MCPs) seed increased airway MCs in mouse models of airway inflammation and infection; however, the clinical associations and therapeutic targeting of MCPs in SA are unstudied [2, 3]. MCPs are IgE-responsive; thus, they may be impacted by omalizumab treatment, but the effect of T2 cytokine biologics on MCP number and phenotype in SA is unknown.

We performed an observational study, titled “Effect of mepolizumab on alternative functions of eosinophils in severe eosinophilic asthma (ALTEOS)”, to determine the impact of mepolizumab and benralizumab treatment of SA on MCPs. Study design, endpoints and participant characteristics of ALTEOS are published [4]. MCPs were measured in PBMCs using flow cytometry (Table S1 and Figure S1: CD45^lo, CD4⁻, CD8⁻, CD19⁻, CD14⁻, CD34⁺, CD117⁺, FcεR1A⁺ progenitors [2, 3]) in a cross-sectional study of 80 SA participants (eosinophilic [EA] n = 32, non-eosinophilic [NEA] n = 23, mepolizumab-treated n = 25). EA was defined as any peripheral blood eosinophil (PBE) count ≥ 300/μL at baseline visit or in the 12 months prior, or ≥ 150/μL while receiving maintenance OCS treatment. Proportion and number of MCPs did not differ between NEA, EA and mepolizumab-treated SA (Figures 1A, S2A). Two of 23 NEA, 9 of 32 EA and 3 of 25 mepolizumab-treated SA were treated with maintenance OCS [4]. MCP did not differ between OCS and non-OCS treated participants (Figure S2B). Of the 32 EA cross-sectional (v1) participants, 20 subsequently commenced mepolizumab and 10 benralizumab, and MCPs were measured at v2 (8–24 weeks) and v3 (> 24 weeks) post-treatment. MCPs were unaltered at v2 and v3 compared to baseline in mepolizumab-treated patients (Figures 1B, S2C,D). In contrast, MCPs were significantly decreased at follow-up in benralizumab-treated patients (Figures 1B, S2E–G). A greater proportion of MCPs expressed detectable IL5 receptor subunit alpha/CD125 (the target of benralizumab) compared to CD34⁺, CD117⁺, FcεR1A⁻ progenitors (termed non-MCP), while IL4R/CD124 (the target of dupilumab) expression was undetectable (Figure 1C,D). IL5R⁺ MCPs expressed increased surface CD117 and FcεR1A compared to IL5R⁻ MCPs, consistent with acquisition of IL5R expression as MCPs progress toward MC commitment (Figure 1E,F) [5]. Neither IL5R⁺ nor IL5R⁻ proportion, number or viability were altered by mepolizumab, suggesting MCPs are not IL5-dependent for survival or maturation in SA. (Figures 1G,H, S2H–P). Complete depletion of IL5R⁺ MCPs occurred with benralizumab, while IL5R⁻ MCPs were relatively unaffected, consistent with known antibody-dependent cell cytotoxicity-mediated depletion of IL5R-expressing targets by benralizumab (Figures 1G,H, S2Q–V) [4]. Lower MCP number at v3 was associated with greater decrease in ACQ5 score in pooled mepolizumab/benralizumab patients, supporting clinical significance of benralizumab-mediated MCP depletion (Figure 1I). The relationship of delta ACQ5 with v3 MCP as reported in Figure 1I remained significant when IL5R⁺ or IL5R⁻ MCPs were examined in all anti IL-5/IL-5R-treated patients.

MC-restricted gene expression measures relate to MC abundance in complex samples [1]. We identified MCP-restricted gene expression using CELLxGENE analysis of 836,148 single blood cell transcriptomes, including MCPs identified by protein expression of CD34, CD117, and FcεR1A [6]. We examined the expression of 8 genes published as highly expressed in MCPs [2, 5], identifying HDC, MS4A2, TPSB2, TPSAB1, CPA3, and HPGDS as MCP-restricted (Figure 2A–I). These genes were then analyzed in bulk blood transcriptomes as MCP markers. In contrast, KIT and SGRN were not MCP-restricted and were excluded from further analyses. MCP gene expression was unaltered between healthy controls, moderate, and SA in U-BIOPRED (GSE69683) and IMSA (GSE207751) datasets (Tables S2–S5). HDC, MS4A2, and CPA3 were decreased in whole blood following 4 months of benralizumab in 41 SA patients (Figure 2J–L) [7], confirming our flow cytometry finding. MCP gene expression was unaltered following 6, 14, and 26 weeks of omalizumab in n = 45 SA patients (GSE134544, Table S6), and was not differentially expressed in blood or PBMC bulk transcriptomes following 3 or 4 months of dupilumab treatment in two adult atopic dermatitis cohorts with high asthma co-occurrence [8, 9].

In summary, we demonstrate benralizumab but not mepolizumab, dupilumab, or omalizumab deplete IL5R-expressing MCPs in SA. While under review, benralizumab depletion of MCPs was independently validated, strengthening our findings [3]. We show MCPs can express IL5R but lack IL4/13R, and their abundance and survival are independent of IL4/5/13 signaling in SA. Our analyses suggest MCP depletion may be clinically significant; future studies should test this in asthma and other MC-related disorders.

M.F. co-designed and conceived the study, led study conduct, performed laboratory analyses, performed data analyses, and drafted the manuscript. J.H. was involved in study conduct and edited the manuscript. S.A.H. designed statistical analyses, performed data analyses, and edited the manuscript. P.G.G. co-designed and conceived the study, was involved in study conduct, and edited the manuscript.

M.F. reports research funding from GSK. J.H. has received honoraria from GSK, AstraZeneca, and Novartis. S.A.H. declares no conflicts of interest. P.G.G. reports personal fees from AstraZeneca, Chiesi, GSK, Novartis, and Sanofi and grants from AstraZeneca and GSK.