REPORT FOR PATHOLOGICAL SOCIETY SMALL GRANT SCHEME AWARD: GRANT REFERENCE NO. SGS 2010 10 08

Title:Characterisation of expression of microRNAs mir-125b, mir-146, mir-483 and mir-720 in HER2 expressing breast cancer

Name & Address:Helen Ingoldsby, Discipline of Pathology, National University of Ireland, Galway, Clinical Science Institute, Costello Rd., Galway, Ireland.

Background:As a pilot study, we had evaluated the expression of 837 microRNAs (miRNAs) in MCF7-HER2 and MCF7-Neo cell lines that differed only in the expression of HER2 using miRNA array analysis. Based on the results of this microarray analysis, we had proposed to evaluate the expression of miR-125b, miR-146, miR-483 and miR-720by in situ hybridisation (ISH) in HER2+invasive breast cancers (IBCs). However, qRT-PCR for miR-125b, miR-146, miR-483 and miR-720 using RNA from MCF7-neo and MCF7-HER2 cells did not match with the results of the microarray. In addition, we did not observe any significant difference in the expression level of these miRNAs in HER2+ (BT474 & SKBR3) vs. HER2- (MCF7 & T47D) human breast cancer cell lines. Therefore, we decided not to perform miRNA ISH for miR-125b, miR-146, miR-483 and miR-720.

In January 2011, a paper was published by Persson et al. in Cancer Research, reporting the presence of 5new miRNAs (miR4726, miR-4727, Candidate_296, miR-4728, and miR-4734) that map to 17q12[1]. Amplification of 17q12 is a clinically important genomic aberration that causes overexpression of the HER2 receptor. Two of these miRNAs (miR-4726 & miR-4728) were enriched in immunoprecipitated Ago2 fraction from MCF7 cells. Interestingly, miR-4728 is encoded within an intron of HER2, a gene that is frequently amplified and associated with more aggressive breast cancers and shorter patient survival. We chose to evaluate the expression of miR-4726 & miR-4728 by miRNA ISH.

Original Aims (copied from original application):

1. We will select formalin-fixed paraffin-embedded (FFPE) tissue blocks from ten cases each of luminal A (ER-positive, HER2-negative),luminal B (ER-positive, HER2-positive), HER2 expressing (ER-negative, HER2-positive) breast cancers from the pathology archive.

2. Tissue sections will be cut and tumour confirmed on H&E staining. A modified ISHmethod for detection of mammalian miRNAs in tissue sections will be used to detect expression of miRNAs, that combines the use of RNA oligonucleotide probes with highly specific wash conditions.

3. miRNA expression will be validated in a subset of cases, approximately 10, by using a Taqman qRT-PCR assay.

4. Statistical analysis will be used to detect associations between the expression of mir-125b, mir-146, mir-483 and mir-720 and HER2 status.

Results: In order to optimize the sensitivity and the performance of the miRNA ISH protocol, first we used the DIG-labelled probe against snRNA U6 (hereafter U6). The presence of U6 in virtually all nuclei validates whether the sequence of steps has allowed the development of ISH signal. Furthermore, it allows the sensitivity of the assay to be roughly estimated. We observed a strong nuclear signal in normal breast epithelial cells and in colonic crypt epithelial cells (Figure 1). A good performance of the U6 probe and a good signal obtained with the one-day protocol confirmed that all of the detecting reagents were performing well.

Next we performed ISH for miR-4726 and miR-4728 on FFPE tissue sections ofIBCs. Expression of miR-4726 and miR-4728 was evaluated in each of 4 cellular compartments in the tissue: benign epithelial cells, myoepithelial cells, tumour cells and stromal cells. The signal intensity was higher for miR-4726 compared to miR-4728. miR-4726 and miR-4728 were homogeneously expressed within cellular compartments. Therefore, scoring was weighted on staining intensity. We observed that mean miR-4726 and miR-4728 expression was highest in benign epithelial cells (Figure 2). Overall, mean expression was lower in tumour cells than in benign epithelial cells (Figure 2). For both miR-4726 and miR-4728, the expression appears to be downregulated in DCIS and IBCscompared to benign epithelial cells. We did not see any correlation between expression of miR-4726 and miR-4728 and HER2 in FFPE tissue sections of IBC (Figure 3). The expression of miR-4726 and miR-4728 has to be evaluated in a greater number of HER2+vs. HER2- breast cancer cases to evaluate its association with HER2.

Conclusions: The ISH technique for miRNA analysis has been optimised for FFPE tissue and an insight has been provided into the cellular localisation of miR-4726 and miR-4728 within tumour tissue samples.While results obtained were not of statistical significance, nonetheless, variations in miRNA expression between DCIS and IBC compared to benign epithelial cells were observed. Our results suggest a tumour suppressor role for miR-4726 and miR-4728 in breast cancer. More meaningful results could be obtained with evaluation of greater numbers of breast cancer cases.

In conclusion, the findings presented in this work need validation in large patient cohorts.Ideally, resources would include large tumour banks of frozen material, FFPE material andclinical outcome data. In this manner, research findings could be translated into clinicalpractice.

How Closely Have the Original Aims been Met:We have met all aims described in the original proposal with reasonable success. The only deviation from the original plan was the set of miRNAs whose expression was evaluated by miRNA ISH in FFPE sections of breast tumours. This change was made due to some unexpected results and a relevant paper published in Cancer Research [1]. Since this change was made quite early in the project, we were able to make significant progress in the project.

1.Persson, H., et al., Identification of new microRNAs in paired normal and tumor breast tissue suggests a dual role for the ERBB2/Her2 gene. Cancer Res, 2011. 71(1): p. 78-86.