High expression of RelA/p65 is associated with activation of nuclear factor-kappaB-dependent signaling in pancreatic cancer and marks a patient population with poor prognosis

Abstract

Activation of nuclear factor-kappaB (NF-kappaB) signaling was observed in pancreatic adenocarcinoma cell lines and tumours. However, information on the expression of RelA/p65, the major transcription activating NF-kappaB subunit, in these carcinomas and possible correlations thereof with NF-kappaB activation and patient survival is not available. To provide this missing translational link, we analysed expression of RelA/p65 in 82 pancreatic adenocarcinomas by immunohistochemistry. Moreover, we measured activation of the NF-kappaB pathway in 11 tumours by quantitative PCR for NF-kappaB target genes. We observed strong cytoplasmic or nuclear expression of RelA/p65 in 42 and 37 carcinomas, respectively. High cytoplasmic and nuclear expression of RelA/p65 had negative prognostic impact with 2-year survival rates for patients without cytoplasmic or nuclear RelA/p65 positivity of 41 and 40% and rates for patients with strong cytoplasmic or nuclear RelA/p65 expression of 22 and 20%, respectively. High RelA/p65 expression was correlated to increased expression of NF-kappaB target genes. The observation that high expression of RelA/p65 is correlated to an activation of the NF-kappaB pathway and indicates poor patient survival identifies a patient subgroup that might particularly benefit from NF-kappaB-inhibiting agents in the treatment of pancreatic cancer. Based on our findings, this subgroup could be identified by applying simple immunohistochemical techniques.

Figure 1

RelA/p65 expression in pancreatic tissue. (A) Weak cytoplasmic RelA expression in pancreatic acinar parenchyma and in pancreatic ducts (arrows). (B) Ductular proliferates in chronic pancreatitis with weak cytoplasmic but focal nuclear (arrows) RelA expression. Inset: Higher magnification of one duct with nuclear RelA positivity. (C) Ductal adenocarcinoma (arrows) with weak cytoplasmic RelA expression. Note smooth muscle cells in the tumour vicinity exhibiting nuclear RelA positivity (arrowheads). (D) Ductal adenocarcinoma (arrows) with moderate cytoplasmic but no nuclear RelA expression. Note strong cytoplasmic and nuclear RelA positivity in adjacent inflammatory cells (arrowheads). (E) Ductal adenocarcinoma with strong cytoplasmic and nuclear (arrows) expression of RelA. (F) PanIN III with strong cytoplasmic positivity for RelA (arrowheads). Note an invasive gland in the vicinity (arrow).

Figure 2

Kaplan–Meier survival curves in dependence of clinicopathological factors and RelA/p65 expression patterns. Overall survival dependent on nodal status (A) and grade (B). Overall survival dependent on cytoplasmic (C and E) and nuclear (D and F) RelA overexpression for the whole study cohort (C and D) as well as for the subgroup of node-negative patients (E and F).

Figure 3

Correlation of RELA expression with NF-κB target gene expression. Higher levels of RELA expression, as indicated by decreasing ΔC_t values, were significantly associated with higher expression of (A) NFKBIA (r²=0.4816, P<0.05) and (B) CCND2 (r²=0.5225, P<0.05). (C) A trend that higher expression levels of CCND1 (circles) and BCL2L1 (squares) correlated with higher RELA expression levels was observed as well. (D) NFKBIA expression correlated positively with CCND2 (squares, r²=0.7096, P<0.01), BCL2L1 (closed circles, r²=0.4635, P<0.05) and CCND1 (open circles, r²=0.4196, P=0.059) expression (see text).