Shuwei Liu - Final Dissertation Defense - HUGEN PhD Candidate

Department of Human Genetics Doctoral Candidate, Shuwei Liu, will defend the following dissertation on “Deciphering Cognitive Function through DNA Methylation: Insights from the Exercise Program in Cancer and Cognition Study.”

COMMITTEE CHAIR: Daniel E. Weeks, PhD 

Committee Members:

• John R. Shaffer, PhD
• Yvette Conely, PhD, FAAN
• Wei Chen, PhD

ABSTRACT:

Breast cancer is one of the most frequently diagnosed cancer types among women worldwide and many patients experience cancer-related cognitive decline, which can negatively affect their quality of life. Exploring DNA methylation signatures has the potential to shed light on the underlying biological mechanisms of cognitive function (CF) variation in postmenopausal women with early-stage hormone receptor-positive breast cancer and may provide novel insights on interventions that mitigate CF decline for these patients.

This study leverages a randomized clinical trial, Exercise Program in Cancer and Cognition (EPICC), which collected CF and genome-wide DNA methylation data for whole blood before and after aerobic exercise intervention for early-stage breast cancer patients. Here in Aim 1, epigenome-wide association studies of eight CF phenotypes for patients before randomization (n=109) identified one genome-wide significant differentially methylated CpG for processing speed (cg10331779 in CTNND2, p-value=9.65×10^(-9)) and subjective CF (cg25906741 in MLIP, p-value=2.01×10^(-8)), respectively, along with differentially methylated regions for various CF domains. Downstream functional analysis demonstrated the role of beta-estradiol in regulating CF for breast cancer patients. In Aim 2, I used high-dimensional mediation analysis to test for mediation effects of DNA methylation between exercise and CF as well as endocrine therapy and CF in 74 breast cancer patients - several suggestive DNA methylation mediators were identified. Further pathway analyses suggested the role of oxytocin signaling pathway in understanding the impact of exercise on CF. Finally, in Aim 3, I developed a DNA methylation-based predictor of general CF based on random forest algorithm in the Canadian Longitudinal Study on Aging cohort (n=904) and validated it on the EPICC data. The predictor accounted for about 3.5% of the total variance. The CpGs selected in this predictor offered additional insights on biomarkers that inform CF status.

This work enhanced our understanding of biological mechanisms underpinning CF variation in breast cancer patients and provided novel perspectives on the interplay of DNA methylation, treatment, exercise, and CF. In the long term, this research could pave the way for risk assessment and targeted interventions to effectively address cancer-related cognitive decline.