Obesity-associated methylation in breast tumors: a possible link to disparate outcomes?

PURPOSE

As a primary risk factor and modifier of breast cancer incidence and prognosis, obesity may contribute to race disparities in breast cancer outcomes. This study examined association between obesity and DNA methylation in non-Hispanic Black and White women diagnosed with breast cancer.

METHODS

Genome-wide DNA methylation was measured in the breast cancer tumor tissue of 96 women using the EPIC array. To examine the association between obesity and tumor methylation, linear regression models were used-regressing methylation β value for each cytosine and guanine (CpG) site on body mass index adjusting for covariates. Significance was set at false discovery rate (FDR) < 0.05. In the top 20 CpG sites, we explored the interactions with race and estrogen receptor (ER) status. We used multivariable Cox-proportional hazard models to examine whether methylation in the top 20 sites was associated with all-cause mortality.

RESULTS

While none of the CpG sites passed the FDR threshold for significance, among the top 20 CpG sites, we observed interactions with race (TOMM20) and ER status (PSMB1, QSOX1 and PHF1). The same CpG sites in TOMM20, PSMB1, and QSOX1 were associated with all-cause mortality.

CONCLUSIONS

We identified novel interactions between obesity-associated methylation and both race and ER status in genes that have been associated with tumor regulation. Our data suggest that dysregulation in two sites may associate with all-cause mortality.