Detection of long non-coding RNA in archival tissue: correlation with polycomb protein expression in primary and metastatic breast carcinoma

Abstract

A major function of long non-coding RNAs (lncRNAs) is regulating gene expression through changes in chromatin state. Experimental evidence suggests that in cancer, they can influence Polycomb Repressive Complexes (PRC) to retarget to an occupancy pattern resembling that of the embryonic state. We have previously demonstrated that the expression level of lncRNA in the HOX locus, including HOTAIR, is a predictor of breast cancer metastasis. In this current project, RNA in situ hybridization of probes to three different lncRNAs (HOTAIR, ncHoxA1, and ncHoxD4), as well a immunohistochemical staining of EZH2, is undertaken in formalin-fixed paraffin-embedded breast cancer tissues in a high throughput tissue microarray format to correlate expression with clinicopathologic features. Though overall EZH2 and HOTAIR expression levels were highly correlated, the subset of cases with strong HOTAIR expression correlated with ER and PR positivity, while the subset of cases with strong EZH2 expression correlated with an increased proliferation rate, ER and PR negativity, HER2 underexpression, and triple negativity. Co-expression of HOTAIR and EZH2 trended with a worse outcome. In matched primary and metastatic cancers, both HOTAIR and EZH2 had increased expression in the metastatic carcinomas. This is the first study to show that RNA in situ hybridization of formalin fixed paraffin-embedded clinical material can be used to measure levels of long non-coding RNAs. This approach offers a method to make observations on lncRNAs that may influence the cancer epigenome in a tissue-based technique.

Figure 1. Expression levels in four different breast lesions.

Four different breast lesions, in rows, are stained each for hematoxylin and eosin (H&E), and RNA in situ probes for HOTAIR, ncHoxA1, and ncHoxD4, in columns. Expression levels of ncHoxA1, and ncHoxD4 are depicted as negative (−) and positive (+). For HOTAIR, expression levels are depicted as negative (−), weakly positive (+), and strongly positive (++).

Figure 2. Expression of HOTAIR and EZH2 in primary versus metastatic carcinoma.

Primary breast carcinomas with their matched metastases are stained with an RNA in situ hybridization probe for HOTAIR (rows 1 and 2) or immunohistochemical antibody for EZH2 (rows 3 and 4). Expression levels are indicated as negative (−), weakly positive (+), and strongly positive (++). As depicted, some metastatic carcinomas have increased expression compared to their matched primary carcinomas, while some metastatic carcinomas have decreased expression compared to their matched primary carcinomas.

Figure 3. Strong EZH2 and HOTAIR

coexpression trend with worse survival. Kaplan-Meier curve of overall survival based upon the primary carcinoma expressions of both EZH2 and HOTAIR stratified into five different groups: both EZH2 and HOTAIR negative scores (orange line), one negative and one positive score (black line), both EZH2 and HOTAIR weakly positive scores (green line), one weak positive and one strong positive score (blue line), and both EZH2 and HOTAIR strongly positive scores (pink line). The number of cases in each group is listed. As the expression of both EZH2 and HOTAIR become positive and stronger, the percent survival tends to decrease. Strong expression of both EZH2 and HOTAIR together correlate with a trend toward worse survival (logrank test for trend p-value 0.0739).