Proteomic and Transcriptomic Signatures of Poor Asthma Symptom Control in the U-BIOPRED Cohort

Background Controlling asthma symptoms remains challenging. Understanding its molecular mechanisms may provide new insights into asthma pathophysiology. We explored the associations between the transcriptome and proteome in blood and sputum and asthma symptom control.Methods This cross-sectional study included asthmatic and healthy adults from the U-BIOPRED cohort. Uncontrolled symptoms were defined as a mean >= 1.5 points on the 5-item asthma control questionnaire. Stability selection using LASSO logistic regression identified genes/proteins associated with symptom control, followed by logistic regression. Analyses were adjusted for age, sex, age of asthma onset, smoking status, body mass index (BMI), FEV1% predicted, exacerbation history, anti-IgE therapy and urinary steroid levels. The selected variables were compared with healthy controls using linear regression, adjusted for age, sex, BMI, smoking status and urinary steroid levels.Results Four serum proteins were selected based on data from 415 asthmatics (median age 52 [41, 61] years; 59% female; 64% uncontrolled). Higher TWEAKR/TNFRSF12A [OR 2.25 (95% CI 1.15, 4.77)] and MBL/MBP-C [1.6 (1.07, 2.42)] levels increased the odds of uncontrolled symptoms, whereas higher MK08/MAPK8 [0.48 (0.29, 0.76)] and CD5L [0.59 (0.41, 0.82)] levels decreased the odds. CD5L levels were significantly lower in severe asthma than in healthy controls [estimate -0.23 (95% CI -0.41, -0.04)]. No associations were found between symptom control and the sputum proteome (N = 90) or the sputum (N = 96) and blood (N = 360) transcriptomes.Conclusion Four serum proteins distinguished asthmatics with uncontrolled from controlled symptoms. CD5L levels were also lower in asthmatics with severe disease than in healthy controls, warranting further investigation into its potential therapeutic value.