br We investigated the impact of NF B activation on
We investigated the impact of NF-κB activation on epithelial–me-senchymal transition and angiogenesis in CRC because they are im-portant for tumor proliferation and progression. In the present study, we evaluated epithelial–mesenchymal transition through the im-munohistochemistry of SNAIL and E-cadherin. There was a significant correlation between nuclear pNF-κB expression and SNAIL expression, but not between nuclear pNF-κB expression and E-cadherin expression. Angiogenesis is necessary for metastasis and the progression of malig-nant tumors , and hypoxia occurs when cell proliferation outpaces the rate of angiogenesis [29,30]. Hypoxia induces the expression of HIF-1α, which regulates several gene transcription factors . HIF-1α is frequently overexpressed in malignant tumors, and is associated with radiotherapy and chemoresistance [31–33]. HIF-1α is significantly correlated with poor survival in various cancers . However, de-tailed evaluation in the present study revealed that VEGF expression had no significant eﬀect on the rate of survival in CRC. The authors of a previous study reported that NF-κB expression is significantly corre-lated with HIF-1α expression and VEGF expression in stage III CRCs . However, in Filipin III to the results of that study , there was no significant correlation between NF-κB activation and VEGF expression in the present study. In addition, microvessel density was significantly correlated with VEGF expression, but not pNF-κB expression (P = 0.005 and P = 0.632, respectively; data not shown). Furthermore, in subgroup analysis based on pTNM stage, there was no significant cor-relation between pNF-κB expression and VEGF expression. As described above, NF-κB activation has been evaluated using various criteria in previous studies . Although our results may diﬀer from those of previous studies, NF-κB activation is important in epithelial–mesench-ymal transition and angiogenesis in CRC. In addition, because im-munohistochemistry using tissue microarray methods was performed in the present study, intra-tumoral heterogeneity could be present.
The authors of several previous studies, including a meta-analysis, have reported the prognostic role of NF-κB activation in CRC [34–37]. Our results were consistent with those of the meta-analysis. However, Lewander et al. reported that nuclear NF-κB expression was not corre-lated with worse survival rates , and O’Neil et al. used im-munohistochemical analysis of the NF-κB p50 subunit to evaluate NF-κB activation . The discrepancies in the results of the assessment of the prognostic value of NF-κB may be caused by variations in the po-pulations and tumor stage studied, and in the evaluation criteria used [22,30,34,36]. Lewander et al. reported that cytoplasmic pNF-κB ex-pression was significantly correlated with worse survival in CRC, but nuclear pNF-κB expression was not . As above described, because NF-κB is a transcription factor, the nuclear translocation of NF-κB is required. Therefore, the nuclear expression of NF-κB, regardless of whether it is in the phosphorylated or non-phosphorylated form, is comparable to NF-κB activation. Lewander et al. compared the cyto-plasmic and nuclear expression levels of NF-κB, and found that cyto-plasmic NF-κB expression was significantly correlated with nuclear NF-κB expression . Because the discrepancy between the rates of nu-clear and cytoplasmic NF-κB expression was 42.3%, it is unclear whe-ther cytoplasmic NF-κB expression is proportional to the transcriptional
Fig. 2. Kaplan-Meier curves for overall (A) and recurrence-free survivals (B) according to phosphorylated NF-κB p65 expression. Patients with positive (solid line) and negative (dotted line) phosphorylated NF-κB p65 expression showed overall and recurrence-free survivals.