Untargeted metabolomic analysis reveals different metabolites associated with response to mepolizumab and omalizumab in asthma

Background There is limited evidence on biomarkers associated with response to the monoclonal antibodies currently approved for asthma treatment. We sought to identify circulatory metabolites associated with response to treatment with mepolizumab or omalizumab. Methods We conducted global metabolomic profiling of pre-treatment plasma samples from 100 patients with moderate-to-severe asthma who initiated mepolizumab (n=31) or omalizumab (n=69). The primary outcome was the change in exacerbations within 12 months of therapy. Negative binomial models were used to assess the association between each metabolite and exacerbations adjusting for age, sex, body mass index, baseline exacerbations, and inhaled corticosteroid use. Chemical Similarity Enrichment Analysis (ChemRICH) was conducted to identify chemical subclasses associated with treatment response. Results The mean age of the mepolizumab group was 58.7 years and 2.9 exacerbations over the year prior to initiation of biologic therapy. The mean age in the omalizumab group was 48.8 years with 1.5 exacerbations in the preceding year. Patients with higher levels of two tocopherol metabolites were associated with more exacerbations on mepolizumab (delta-carboxyethyl hydroxychroman (CEHC) (p=2.65E-05, false discovery rate (FDR=0.01) and delta-CEHC glucuronide (p=2.47E-06, FDR=0.003)). Higher levels of six androgenic steroids, three carnitine metabolites and two bile acid metabolites were associated with decreased exacerbations in the omalizumab group. In enrichment analyses, xanthine metabolites (cluster FDR=0.0006), and tocopherol metabolites (cluster FDR=0.02) were associated with worse mepolizumab response while androgenic steroids (cluster FDR=1.9E-18), pregnenolone steroids (cluster p=3.2E-07, FDR=1.4E-05), and secondary bile acid metabolites (cluster p=0.0003, FDR=0.006) were the top subclasses associated with better omalizumab response. Conclusion This study identifies distinct metabolites associated with response to mepolizumab and omalizumab, with androgenic steroids associated with response to both mepolizumab and omalizumab.