Cdna Microarray Analysis and Immunohistochemistry Reveal a Distinct Molecular Phenotype in Serous Endometrial Cancer Compared to Endometrioid Endometrial Cancer

Yonghua Chen,Yuanyang Yao,Lili Zhang,Xiaoping Li,Yue Wang,Lijun Zhao,Jianliu Wang,Gongwei Wang,Danhua Shen,Lihui Wei,Jianqing Zhao
DOI: https://doi.org/10.1016/j.yexmp.2011.04.005
IF: 4.401
2011-01-01
Experimental and Molecular Pathology
Abstract:Purpose: The main objective of this study was to refine more precisely the gene expression patterns used to distinguish serous from endometrioid endometrial carcinoma.Methods: A low-density cDNA microarray containing 492 genes was designed and constructed. The gene expression profiles of 32 endometrioid and 5 serous endometrial cancer tissue samples were compared. The expression of 5 differentially expressed genes: NDC80, BUB1, FUT8, ANXA4 and BBC3 in endometrioid and serous adenocarcinoma samples was further evaluated by quantitative real-time PCR and immunohistochemistry.Results: Unsupervised cluster analysis revealed that the 5 serous adenocarcinomas clustered together. These were separated from the endometrioid adenocarcinomas which were further sorted into 3 additional clusters. A comparative analysis indicated that there was a significant difference in FIGO stage with no significant difference in depth of myometrial invasion among the 4 clusters. The FIGO ternary grading system could not distinctly separate the 3 clusters of endometrioid adenocarcinomas, but a binary grading system was able to do so. Using a supervised analysis, we have identified 46 genes exhibiting >2-fold differences that can be used to statistically differentiate serous adenocarcinomas from endometrioid adenocarcinomas. The directions of gene and protein expression change of five differentially expressed genes estimated by real-time PCR and immunohistochemistry are consistent with those estimated from microarray.Conclusions: Serous adenocarcinoma exhibits distinct gene expression profiles, compared with those of endometrioid adenocarcinoma. These differences make it feasible to validate microarray data by immunohistochemistry, and they will ultimately allow us to identify tumors according to their immunohistochemical phenotype. The accuracy of classifying endometrial tumors using a system based on their gene expression patterns is much higher than the accuracy of the FIGO grading system. Thus, this gene expression pattern-based system may prove to be crucial in developing novel treatment strategies for endometrial cancers at the molecular level in future. (C) 2011 Elsevier Inc. All rights reserved.
What problem does this paper attempt to address?