Abstract: FR-PO221
Identification of Gene Signatures and Molecular Pathways Associated with Urine Albumin Creatinine Ratio Response to Renal Protective Drugs
Session Information
- Diabetic Kidney Disease: Basic - II
November 08, 2019 | Location: Exhibit Hall, Walter E. Washington Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Diabetic Kidney Disease
- 602 Diabetic Kidney Disease: Clinical
Authors
- Ju, Wenjun, University of Michigan, Ann Arbor, Michigan, United States
- Hartman, John R., University of Michigan, Ann Arbor, Michigan, United States
- Nair, Viji, University of Michigan, Ann Arbor, Michigan, United States
- Harlan, Shannon Marie, Eli Lilly and Company, Indianapolis, Indiana, United States
- Burdet, Frédéric, Swiss Institute of Bioinformatics, Lausanne, Switzerland
- Baker, Hana, Eli Lilly and Company, Indianapolis, Indiana, United States
- Heuer, Josef G., Eli Lilly and Company, Indianapolis, Indiana, United States
- Gomez, Maria F., Lund University, Malmö, Sweden
- Duffin, Kevin L., Eli Lilly and Company, Indianapolis, Indiana, United States
- L Heerspink, Hiddo Jan, University Medical Center Groningen, Groningen, Netherlands
- Kretzler, Matthias, U.Michigan, Ann Arbor, Michigan, United States
Group or Team Name
- Biomarker enterprise to attack DKD (BEAt-DKD)
Background
Renal protective drugs reduce the risk of progression and urine albumin/creatinine ratio (UACR) in some but not all patients with type 2 diabetes and CKD. Previous studies showed that the response in each individual is consistent to different drugs. Here we aim to identify gene signatures and molecular pathways that are associated with the individual response to renal protective drugs.
Methods
ReninAAV db/db uNx mice were treated with the angiotensin-converting enzyme inhibitor Lisinopril, the angiotensin receptor blocker Losartan, the Janus-associated kinase inhibitor Ruxolitinib, the sodium glucose transporter 2 inhibitor Canagliflozin or vehicle control for 2 weeks (n=8 per group). ACR was measured at baseline and after treatment. RNAseq profiling of kidney cortex was performed. weighted gene co-expression network analysis and machine learning approaches were used to identify genes associated with UACR; ingenuity pathway analysis was used to identify enriched molecular pathways.
Results
The fraction of mice that responded to treatment, defined as >35% decline in UACR varied depending on the drug, with all mice responding to Lisinopril, 5/8 to Losartan, 4/8 to Ruxolitinib, and none to Canagliflozin. Network analysis identified 35 co-expression modules, several of which were distinctly associated with phenotype variables. Selecting 12 of the modules, a cross-validated logistic lasso regression model was able to predict responders with 83% accuracy (40/48). The cross-validated model was compared to results from 1000 random permutations of its class labels, and found to be significant (p<0.001). Enriched signaling pathways include genes involved in tight junction-, mTOR- and sirtuin signaling, as well as in nicotine degradation.
Conclusion
Our study identified gene signatures and molecular pathways associated with UACR response to renoprotective treatments, as well as pathways previously reported to be associated with human kidney disease progression. Results may increase our understanding of the molecular mechanisms underlying responses to treatment and help stratify patients to predict their response to treatment. Our findings will be compared to clinical trial data sets.
Funding
- Commercial Support –