Serum metabolites predict response to angiotensin II receptor blockers in patients with diabetes mellitus

• DOI: 10.1186/s12967-016-0960-3
• Publication Date: 2016-07-04T22:38:01Z
• Authors: Michelle J. Pena; Andreas Heinzel; Peter Rossing; Hans-Henrik Parving; Guido Dallmann; Kasper Rossing; Steen Andersen; Bernd Mayer; Hiddo J. L. Heers...
• Keywords: CHRONIC KIDNEY-DISEASE; Adult; Male; Models, Molecular; 0301 basic medicine; ALBUMINURIA; NEPHROPATHY; Tetrazoles; BLOOD-PRESSURE; Losartan; Angiotensin Receptor Antagonists; 03 medical and health sciences; Journal Article; Metabolomics; Albuminuria; Humans; RISK; Medicine(all); POST-HOC ANALYSIS; Biochemistry, Genetics and Molecular Biology(all); Research; Biphenyl Compounds; MICROALBUMINURIA; Irbesartan; Middle Aged; 3. Good health; RENAL-DISEASE; TARGET; Diabetes Mellitus, Type 2; Metabolome; Female; LOSARTAN; ARB response

Individual patients show a large variability in albuminuria response to angiotensin receptor blockers (ARB). Identifying novel biomarkers that predict ARB response may help tailor therapy. We aimed to discover and validate a serum metabolite classifier that predicts albuminuria response to ARBs in patients with diabetes mellitus and micro- or macroalbuminuria.Liquid chromatography-tandem mass spectrometry metabolomics was performed on serum samples. Data from patients with type 2 diabetes and microalbuminuria (n = 49) treated with irbesartan 300 mg/day were used for discovery. LASSO and ridge regression were performed to develop the classifier. Improvement in albuminuria response prediction was assessed by calculating differences in R(2) between a reference model of clinical parameters and a model with clinical parameters and the classifier. The classifier was externally validated in patients with type 1 diabetes and macroalbuminuria (n = 50) treated with losartan 100 mg/day. Molecular process analysis was performed to link metabolites to molecular mechanisms contributing to ARB response.In discovery, median change in urinary albumin excretion (UAE) was -42 % [Q1-Q3: -69 to -8]. The classifier, consisting of 21 metabolites, was significantly associated with UAE response to irbesartan (p < 0.001) and improved prediction of UAE response on top of the clinical reference model (R(2) increase from 0.10 to 0.70; p < 0.001). In external validation, median change in UAE was -43 % [Q1-Q35: -63 to -23]. The classifier improved prediction of UAE response to losartan (R(2) increase from 0.20 to 0.59; p < 0.001). Specifically ADMA impacting eNOS activity appears to be a relevant factor in ARB response.A serum metabolite classifier was discovered and externally validated to significantly improve prediction of albuminuria response to ARBs in diabetes mellitus.