Since ESCC patients with the same TNM stage and/or CRT response often display considerable variability in disease progression and survival, the traditional grading system may have reached its limits in providing critical information influencing patient prognosis and treatment strategies. Thus, novel diagnostic and risk assessment are urgently needed. Recently, histone modification, which is important for normal cell growth, has become an increasingly important aspect of cancer biology [6, 19–21]. One such modification, the trimethylation of lysine 27 on histone H3 (H3K27me3) is required for Polycomb Repressive Complex 2 (PRC2) mediated repression of a large number of genes essential for cell proliferation, cell differentiation and tumor development [22, 23]. It is believed that maintenance of the H3K27me3 mark during cell division is essential for normal embryogenesis and for preserving cell identity [24, 25]. In human cancer, H3K27me3 has been evaluated as a prognostic factor in prostate, breast, ovarian, pancreatic and esophageal cancers [8, 14, 15], however, some of the results are totally contradictory. In the present study, we examined the expression of H3K27me3, in 30 samples of normal esophageal mucosa and 98 biopsy specimens of primary ESCCs treated with definitive CRT.
Our results showed that a significant percentage of the normal esophageal mucosa was observed positive staining of H3K27me3. However, the frequency of positive expression and levels of H3K27me3 in our ESCC cohorts were significant higher than that in normal esophageal tissues. Further analysis demonstrated that high expression of H3K27me3 was associated closely with poor tumor differentiation and advanced local invasion in ESCC, which was in general agreement with findings revealed by Ching et al . In addition, we did further found that H3K27me3 expression correlated significantly with tumor locoregional progression rather than distant metastasis. These findings provide evidence that the up-regulation of H3K27me3 may provide a selective advantage for carcinogenesis and tumor progression of ESCC.
When focus on locoregional progression-free survival (LPFS) of ESCC, we found that high expression of H3K27me3 correlated with poor prognosis. Furthermore, the difference in H3K27me3 expression levels was evaluated to be a better predictor of LPFS when patients are stratified on the basis of T, N, M status and CRT response. It has been revealed that H3K27 methylation was mediated by the primary H3K27 methyltransferase, EZH2, a component of the Polycomb (PcG) complex that is involved in early carcinogenesis . As aforementioned, our results showed that H3K27me3 predicted LPFS predominately in T2-3, N0 and M0 stages of ESCC may partly support the link between PcG complex and H3K27me3 in early carcinogenesis. Similar results were also reported in surgical treated ESCC patients by Ching et al . In our CRT treated ESCC cohort, we further demonstrated that H3K27me3 could better stratify LPFS for CRT effective patients. In addition, ESCC patients with high expression of H3K27me3 were more likely to present with large tumor size and advanced T status; such may be at higher risk to suffer from tumor progression after CRT. Thus, our findings might be helpful to better understand the heterogeneity in the prognosis of ESCC patients within the same stages and even with a good response to CRT. Considering the potentially important role of H3K27me3 as a biological mechanism in locoregional progression of ESCC, the examination of H3K27me3 expression by IHC, therefore, might be used as an additional tool in risk assessment and therapy optimizing.
With regards to the biologic function of H3K27me3, it has been observed that, H3K27me3 was linked to PcG-mediated suppression of homeotic box (Hox) genes and maintenance of embryonic stem cell identity , whereas during embryogenesis and stem cell differentiation, consistent with the strong decrease in H3K27me3 levels associated with HOX genes expression, the level of EZH2 also declined [28, 29]. Our results supported the role of H3K27me3 in cell differentiation of ESCC. Since methylation of H3K27 mediated by EZH2 has been implicated in the aggressive phenotype of cancer cells through repression of a panel of tumor suppression genes [30, 31], the loss of function of these genes, in turn, locks stem/precursor cells into abnormal clonal expansion which begins a process of neoplastic initiation [32, 33]. This might provide a possible explanation for the association of H3K27me3 expression and local invasion in ESCC found by our study. Moreover, an imbalance of H3K27 methylation owing to overexpression of EZH2 has been implicated in metastatic prostate and aggressive breast cancers [8, 9], in which a highly significant overlap between PRC2- and H3K27me3-occupied genes was observed . To determine whether there was a potential correlation between the expression of EZH2 and H3K27me3 in ESCC, we evaluated the expression status of the two proteins by IHC in the same cohort of cases. Our results demonstrated that the expression level of H3K27me3 in high EZH2 expression group was significantly higher than that in low EZH2 expression group, which supported the view that up-regulated expression of H3K27me3 in ESCCs might be caused, at least partly, by increased expression of EZH2. In other types of human cancer, in contrast, Wei et al reported that loss of H3K27me3 correlated with poor prognosis of patients with breast, ovarian, and pancreatic cancers, in which no association was observed between expression of H3K27me3 and EZH2 . Considering that the mechanism by which EZH2-mediated H3K27 methylation leads to gene silencing may vary among gene targets and among organisms , discrepancies in different studies may arise. In addition, recent studies demonstrate that the H3K27me3 mark is more dynamic than previously anticipated and suggest that the levels of H3K27me3 is fine-tuned by opposing activities: demethylases and the PRC2 complex [27, 35]. These results, collectively, suggested that the regulation of H3K27me3 expression is quite complicated and its biologic/clinical significance in different human cancers may be tissue-specific. Clearly, further investigations are substantially needed.