The study was conducted to determine the effect of physical activity on DNA methylation and to predict the consequence of this effect concerning gene expression and breast cancer survival. Blood samples, collected from 12 breast cancer patients who participated in a randomized clinical trial of exercise, were examined for exercise-related changes in DNA methylation using a methylation microarray. Tumor samples of 348 breast cancer patients were analyzed with qRT-PCR and qMSP to determine gene expression and methylation identified in the microarray analysis. Cox regression models were developed to predict survival outcomes in association with gene expression and methylation. After 6 months of moderate-intensity aerobic exercise, changes in DNA methylation (P < 5 × 10(-5)) in peripheral blood leukocytes were detected in 43 genes from a panel of 14 495. Based on the list, we analyzed gene expression in association with overall survival in breast tumors and found three genes whose methylation was reduced after exercise were favorably in association with overall survival, i.e., higher expression associated with better survival. Of the three genes, L3MBTL1 was a putative tumor suppressor gene with known function to repress chromatin for transcription, which is activated mainly in germline stem cells. Further analyses of tumor features among patients indicated that high expression of L3MBTL1 was associated with low grade and hormone receptor-positive tumors, as well as low risk of disease recurrence and breast cancer death. In conclusion, the study suggests that increasing physical activity after a breast cancer diagnosis may affect epigenetic regulation of tumor suppressor genes, which have favorable impacts on survival outcomes of breast cancer patients.