Kidney Week

Abstract: FR-PO438

Human Kidney Tubule Cytosine Methylation Changes Can Improve Models for CKD Progression

Session Information

• CKD: Risk Factors for Incidence and Progression - II
  November 03, 2017 | Location: Hall H, Morial Convention Center
  Abstract Time: 10:00 AM - 10:00 AM

Category: Chronic Kidney Disease (Non-Dialysis)

• 301 CKD: Risk Factors for Incidence and Progression

Authors

• Gluck, Caroline A., The Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States

Caroline A. Gluck,

• Qiu, Chengxiang, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Chengxiang Qiu,

• Han, Sang Youb, Inje University, GoYang, Korea (the Republic of)

Sang Youb Han,

• Huang, Jing, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Jing Huang,

• Park, Ae Seo Deok, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Ae Seo Deok Park,

• Ko, Yi-An, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Yi-An Ko,

• Mantzaris, Ioannis, Montefiore Medical Center, Bronx, New York, United States

Ioannis Mantzaris,

• Chen, Yong, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Yong Chen,

• Verma, Amit K., Albert Einstein College of Medicine, Bronx, New York, United States

Amit K. Verma,

• Palmer, Matthew, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Matthew Palmer,

• Susztak, Katalin, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Katalin Susztak,

Background

Chronic Kidney Disease (CKD) progresses at variable rates. Patients who progress rapidly are more likely to reach end stage renal disease (ESRD). Current models to predict CKD progression are centered on baseline GFR, age, and albuminuria. These clinical phenotypes do not explain the pathophysiology of progression and cannot account for environmental influence on progression. Cytosine methylation is a stable, cell type specific and environmentally responsive epigenetic signal that affects gene expression patterns. The aim of this project was to determine if genome wide cytosine methylation changes can improve baseline models for CKD progression and identify novel pathways underlying CKD progression.

Methods

Biobanked human kidney tissue was microdissected to isolate kidney tubules. The data set included 69 human kidney tubule samples with associated cross-sectional and longitudinal clinical data. Histopathology for samples were graded on 20 independent parameters. Genome wide cytosine methylation was analyzed using the Illumina Infinium 450K chip. Transcript level changes were determined using the Affymetrix RNA microarray. Subject-specific adjusted GFR slopes were determined using best linear unbiased prediction to account for random variation. Variables for CKD progression models were selected by a machine learning regression analysis method, "LASSO", to improve model accuracy and reduce model overfitting. Methylation and transcript levels were added to baseline linear regression models and both R² and akaike information criterion (AIC) values were ranked to determine model fitness.

Results

The final model (M1) for CKD progression based on LASSO-selected variables included: baseline GFR, age, albuminuria (dipstick), CKD stage, diabetes, height, and vessel intimal fibrosis. Adding the top gene transcript level to M1 (M2) improved model fitness. Finally, addition of the methylation level at the top genome loci to M2 (M3) further improved model fitness.

Conclusion

Human kidney tubule cytosine methylation levels may improve CKD progression models even after gene expression levels are included. Methylation changes in kidney tissue may be useful biomarkers for CKD progression and may be functionally important in CKD progression.

Funding

• NIDDK Support