Background Cellular experiments revealed that a decreased histone H3 lysine 9 trimethylation (H3K9me3) level was associated with the upregulation of oncogenes in breast cancer cells. Moreover, the role of H3K9me3 in breast cancer was closely associated with estrogen receptor (ER) status. Therefore, we aimed to examine the prognostic value of H3K9me3 on breast cancer by ER status. The level of H3K9me3 in tumors were evaluated with tissue microarrays by immunohistochemistry for 917 women diagnosed with primary invasive breast cancer. Hazard ratios (HRs) and their 95% confidence intervals (CIs) for overall survival (OS) and progression-free survival (PFS) were estimated using Cox regression models. Interaction between H3K9me3 and ER on the prognosis was assessed on multiplicative scale. Results The level of H3K9me3 in tumor tissues was lower than that in adjacent tissues. The high level of H3K9me3 was associated with a better OS (HR = 0.43, 95% CI: 0.21–0.86) and PFS (HR = 0.49, 95% CI: 0.29–0.81) among only ER-positive but not ER-negative tumors. Moreover, the interaction between the level of H3K9me3 and ER status (negative and positive) on the prognosis was significant (Pinteraction = 0.011 for OS; Pinteraction = 0.022 for PFS). Furthermore, the ER-positive tumors were stratified by ER-low and ER-high positive tumors, and the prognostic role of H3K9me3 was significant among only ER-high positive patients (HR = 0.34, 95% CI: 0.13–0.85 for OS; HR = 0.47, 95% CI: 0.26–0.86 for PFS). Conclusions Our study showed that the prognostic value of H3K9me3 on breast cancer was related to ER status and expression level, and the high level of H3K9me3 was associated with a better prognosis among ER-positive tumors, particularly ER-high positive tumors.

Background The prognostic role of either forkhead box A1 (FOXA1) or anterior gradient 2 (AGR2) in breast cancer has been found separately. Considering that there were interplays between them depending on ER status, we aimed to assess the statistical interaction between AGR2 and FOXA1 on breast cancer prognosis and examine the prognostic role of the combination of them by ER status. Methods AGR2 and FOXA1 expression in tumor tissues were evaluated with tissue microarrays by immunohistochemistry in 915 breast cancer patients with follow up data. The expression levels of these two markers were treated as binary variables, and many different cutoff values were tried for each marker. Survival and Cox proportional hazard analyses were used to evaluate the relationship between AGR2, FOXA1 and prognosis, and the statistical interaction between them on the prognosis was assessed on multiplicative scale. Results Statistical interaction between AGR2 and FOXA1 on the PFS was significant with all the cutoff points in ER-positive breast cancer patients but not ER-negative ones. Among ER-positive patients, the poor prognostic role of the high level of FOXA1 was significant only in patients with the low level of AGR2, and vice versa. When AGR2 and FOXA1 were considered together, patients with low levels of both markers had significantly longer PFS compared with all other groups. Conclusions There was a statistical interaction between AGR2 and FOXA1 on the prognosis of ER-positive breast cancer. The combination of AGR2 and FOXA1 was a more useful marker for the prognosis of ER-positive breast cancer patients.