Kidney Week

Abstract: SA-PO136

Urine Metabolites Report on Dysregulated Kidney Metabolism in Diabetes and Hypertension

Session Information

• Diabetic Kidney Disease: Basic - III
  October 27, 2018 | Location: Exhibit Hall, San Diego Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Diabetic Kidney Disease

• 601 Diabetic Kidney Disease: Basic

Authors

• Scialla, Julia J., Duke University Medical Center, Durham, North Carolina, United States

Julia J. Scialla,

• Bain, James R., Duke University Medical Center, Durham, North Carolina, United States

James R. Bain,

• Muehlbauer, Michael, Duke University Medical Center, Durham, North Carolina, United States

Michael Muehlbauer,

• Ilkayeva, Olga, Duke Molecular Physiology Institute, Durham, North Carolina, United States

Olga Ilkayeva,

• Kovalik, Jean-Paul, Duke-NUS Medical School, Singapore, Singapore

Jean-Paul Kovalik,

• Newgard, Christopher B., Duke University Medical Center, Durham, North Carolina, United States

Christopher B. Newgard,

• Coffman, Thomas M., Duke University Medical Center, Durham, United States

Thomas M. Coffman,

• Gurley, Susan B., Oregon Health and Science University, Durham, North Carolina, United States

Susan B. Gurley,

Background

Interrogating the molecular pathophysiology of kidney disease is challenged by limited clinical biopsy specimens. We evaluated concordance of kidney and urine metabolites in mouse models of diabetes mellitus (DM) and hypertension (HTN) to explore urine metabolomics as a tool to investigate kidney diseases in patients.

Methods

We profiled kidney and urine metabolites in the Akita model of type 1 DM (Akita^+/-), the Akita Renin transgene model of DKD (Akita^+/-Rtg^+/-), the Rtg^+/- model of HTN, and wild type mice (WT; n=14 each). 24h urines were collected by metabolic cage and organs harvested at 12 wks. Kidneys were frozen and homogenized in acetonitrile:formic acid and urines concentrations were standardized. Nontargeted GC/MS was performed and annotated with spectral libraries. Unidentifiable metabolites or those missing >50% were removed, with other values imputed. Comparisons utilized principal components (PC), partial least squares discriminant analysis (PLS-DA), and sparse generalized canonical correlation.

Results

Using PLS-DA, higher levels of multiple metabolites discriminated DM (Akita^+/- and Akita^+/-Rtg^+/-) from non-DM (Rtg^+/- or WT) including: branched chain amino acids, simple sugars, and ketones (β-hydroxybutyrate; BHB) in kidneys; and lactate, pyruvate, and ketones (BHB and acetoacetate) and related compounds (2-hydroxybutyrate) in urine. After data reduction by PC, similar kidney and urine PCs were correlated within phenotypes. In DM, kidney PC1 contained valine, pyruvate and simple sugars. This correlated with urine PC1 (r=0.65) characterized by branched fatty acids, keto acids and acyl glycines. Within non-DM, urine PC1 composed primarily of tricarboxylic acid (TCA) cycle-related organic anions (OAs) correlated directly with a similar TCA-cycle related PC in kidney lysates (r=0.71) with levels higher in Rtg^+/-.

Conclusion

Similar kidney and urine metabolite PCs correlate in models with and without DM. The predominance of sugars and ketones in profiles from uncontrolled DM models may obscure discovery relevant to treated DM in patients. Studies in non-DM highlight urine TCA cycle-related OAs as potential biomarkers of kidney metabolism. Whether TCA cycle-related OAs may be useful in human studies involving treated DM requires confirmation.

Funding

• NIDDK Support