MicroRNA miR-196a-2 and breast cancer: a genetic and epigenetic association study and functional analysis

doi:10.1158/0008-5472.CAN-09-0236

Cancer Res. Author manuscript; available in PMC 2010 July 15.

Published in final edited form as:

Cancer Res. 2009 July 15; 69(14): 5970–5977.

Published online 2009 June 30. doi: 10.1158/0008-5472.CAN-09-0236

PMCID: PMC2716085

NIHMSID: NIHMS119610

MicroRNA miR-196a-2 and breast cancer

a genetic and epigenetic association study and functional analysis

Aaron E. Hoffman,¹ Tongzhang Zheng,¹ Chunhui Yi,¹ Derek Leaderer,¹ Joanne Weidhaas,² Frank Slack,³ Yawei Zhang,¹ Trupti Paranjape,² and Yong Zhu¹

¹Department of Epidemiology and Public Health, Yale School of Medicine, Yale University, New Haven, Connecticut 06520

²Department of Therapeutic Radiology, Yale School of Medicine, Yale University, New Haven, Connecticut 06520

³Department of Molecular, Cellular and Developmental Biology; Yale University, New Haven, Connecticut 06520

Requests for reprints: Yong Zhu, Ph.D., Department of Epidemiology and Public Health, Yale University School of Medicine, New Haven, CT 06520. Phone: (203) 785-4844; Fax: (203) 737-6023; Email: yong.zhu/at/yale.edu

The publisher's final edited version of this article is available free at Cancer Res

See other articles in PMC that cite the published article.

Abstract

Increasing evidence has suggested that microRNAs (miRNAs) play an important role in tumorigenesis. As transcriptional regulators, altered miRNA expression may affect many cancer-related biological pathways, indicating that miRNAs can function as tumor suppressors and/or oncogenes. We first performed a genetic association analysis by screening genetic variants in 15 microRNA genes and detected that a common sequence variant in hsa-miR-196a-2 (rs11614913, C→T) was significantly associated with decreased breast cancer risk (for homozygous variant: OR=0.44, 95% CI, 0.28-0.70). Hypermethylation of a CpG island upstream (-700 bp) of the miR-196a-2 precursor was also associated with reduced breast cancer risk (OR=0.35, 95% CI, 0.15-0.81). By delivering expression vectors containing either wild-type or mutant precursors of miR-196a-2 into breast cancer cells, we demonstrated that this variant led to less efficient processing of the miRNA precursor to its mature form, as well as diminished capacity to regulate target genes. A whole genome expression microarray was preformed and a pathway-based analysis identified a cancer-relevant network formed by genes significantly altered following enforced expression of miR-196a-2. Mutagenesis analysis further showed that cell cycle response to mutagen challenge was significantly enhanced in cells treated with variant miR-196a-2 compared to cells treated with the wild-type. Taken together, our findings suggest that miR-196a-2 might have a potentially oncogenic role in breast tumorigenesis, and the function genetic variant in its mature region could serve as a novel biomarker for breast cancer susceptibility.

Keywords: miR-196a-2, Methylation, Network Analysis, Breast Cancer

Introduction

It was proposed over three decades ago that some non-protein coding transcripts may regulate gene activity, a missing piece to the central dogma of molecular biology (1). This novel group of small non-coding RNAs with posttranscriptional regulatory function, later termed microRNAs (miRNAs), has now been shown to negatively alter protein expression in most organisms, from C. elegans to humans (2-5). Interestingly, over 52.5% of the miRNAs elucidated thus far are located within fragile sites on chromosomes, suggesting that these transcripts may be relevant for carcinogenesis (6, 7).

Research linking miRNAs to cancer is still in its infancy, but the information obtained to date has suggested that miRNA may play an important role in a variety of carcinogenic processes (6, 8-11). miRNAs have been shown to act as tumor suppressors by repressing oncogene expression, or may themselves act as oncogenes by repressing other tumor suppressors (11-13). Recent evidence has also demonstrated that a global reduction in miRNA processing promotes carcinogenesis, and miRNA profiling data has been successfully applied to classify tumors and predict prognosis in a number of cancer types (14-16).

In addition, previous studies have suggested that CpG island methylation in miRNA regions may influence miRNA function; thereby altering carcinogenic processes. Weber et al. (17) found that nearly half of all identified miRNAs were associated with CpG sites, and an experimental assay revealed detectable methylation levels at several miRNA loci across normal and malignant cell lines. Lu et al. (18) also found that hypermethylation in the let-7a-3 miRNA gene was associated with survival in patients with ovarian cancer, and Lehmann et al. (19) suggested that aberrant hypermethylation of miRNAs may be an important early event in breast cancer development.

Given an important role for miRNAs in cancer development, we hypothesize that genetic and epigenetic variations in miRNAs may alter miRNA function and consequently influence susceptibility to cancer. To date, however, few published molecular epidemiological studies have investigated such associations at the population level in order to determine whether miRNAs may serve as valuable cancer biomarkers. In the current study, we tested this hypothesis in breast cancer by performing a genetic and epigenetic association study and functional analysis.

Patients, materials, and methods

Study population

The study population consisted of subjects enrolled in a previous breast cancer case-control study conducted in Connecticut. Details regarding subject recruitment and participant characteristics have been described in previous publications (20, 21). Briefly, cases were incident, histologically confirmed breast cancer patients (International Classification of Diseases for Oncology, 174.0 -174.9) between the ages of 30 and 80 with no previous diagnosis of cancer other than non-melanoma skin cancer, who were alive at the time of the interview. Cases were either identified from computerized patient information at Yale-New Haven Hospital (YNHH) in New Haven County, Connecticut, or from nearby Tolland County, Connecticut via hospital records by the Rapid Case Ascertainment Shared Resource at the Yale Cancer Center. YNHH controls were patients at YNHH who underwent breast-related surgery for histologically confirmed benign breast diseases. Tolland county controls were identified either through random digit dialing (for subjects younger than 65), or through the Health Care Finance Administration files (for subjects age 65 and older). Among YNHH subjects, participation rates were 71% for controls and 77% for cases, and among Tolland County subjects, participation rates were 61% for controls and 74% for cases. The study was approved by Institutional Review Boards at Yale University, the Connecticut Department of Public Health, and the National Cancer Institute. Participation was voluntary, and written informed consent was obtained. 441 cases and 479 controls had DNA samples available for genetic association analysis, and of these, 77 cases were untreated. Supplemental Table 4 presents the distribution of selected baseline characteristics for cases and controls.

SNP identification and genotyping

We used a recent release of the SNP database dbSNP b125 on NCBI Assembly B35 from HapMap Data Release #20/phaseII to search for genetic variation in miRNA genes. The miRNA base target positions on human chromosomes were first determined using the Search MicroCosm database produced by the Sanger Institute (http://microrna.sanger.ac.uk/sequences/). These positions were then explored using HapMap in order to locate SNPs within the region of a given miRNA gene. Genotyping for all SNPs was performed at Yale University’s W.M. Keck Foundation Biotechnology Research Laboratory using the Sequenom MassARRAY multiplex genotyping platform (Sequenom, Inc., San Diego, CA). Duplicate samples from 100 study subjects and 40 replicate samples from each of two blood donors were interspersed throughout each batch for all genotyping assays. The concordance rates for QC samples were over 95% for all assays. All genotyping scores, including quality control data, were re-checked by different laboratory personnel and the accuracy of each assay was confirmed.

CpG island identification and methylation analysis

Using the CpG Island Searcher web tool (http://www.cpgislands.com/), a CpG island was identified upstream (-700bp) of the region encoding pri-miRNA-196a-2. The MethPrimer program (http://www.urogene.org//methprimer) was then used to design methylation specific PCR primers within the identified CpG island region. The two methylated primers were: aatttcgatagtagttaatagaacg (left) and ctaaattctcgacaaacacga (right), and the two unmethylated primers were: ttttaattttgatagtagttaatagaaatg (left) and cctaaattctcaacaaacacaa (right). Genomic DNA samples were bisulfite treated using the EZ DNA Methylation Kit (Zymo Research, Orange, CA) according to the manufacturer’s protocol. The presence of methylation was determined by quantitative PCR using the Power SYBR Green Kit (Applied Biosystems), and the percentage of methylation was estimated for each sample using the formula: MI = [1 / (1+2^-(CTu-CTme)] x 100%, as previously described (18), where CTu = the average cycle threshold obtained from duplicate qPCRs using the unmethylated primers, and CTme = the average cycle threshold obtained using the methylated primers.

Cell culture and construction of miR-196a-2 expression vector

Human breast adenocarcinoma cells (MCF-7) were obtained from American Type Culture Collection (Manassas, VA), and were maintained in Dulbecco’s modified Eagle medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Invitrogen), 0.01 mg/ml bovine insulin, and 1% penicillin/streptomycin (Sigma-Aldrich, St. Louis, MO). An expression plasmid vector (pRNAT-U6.1/Neo) containing a Neomycin resistance gene and a coral green fluorescent protein (cGFP) marker under CMV promoter control was purchased from GenScript (Piscataway, NJ). Individuals known to be homozygous for either allele of pre-miR-196a-2 (C or T) were identified by genotyping (as detailed above), and genomic DNA from these subjects was amplified to generate pRNA-miR-196a-2-C and pRNA-miR-196a-2-T, containing the mature miRNAs, along with a flanking region on each side (130bp on left and 120bp on right). PCR products (360 bp) were purified and inserted into the pRNA vector according to the manufacture’s protocol (GeneScript). The ligation mixture was then transformed into competent DH5a cells (Invitrogen, Cat. 18258-012) and plated on LB-G418 plates. Positive clones were chosen and grown in liquid medium, and the constructed plasmid vectors were extracted from culture using a MiniPrep kit (Qiagen, Valencia, CA). Vectors were then sequenced to verify the accuracy of the insert, and each vector was transfected into cells using Lipofectamine 2000 transfection reagent (Invitrogen). Transfection efficiency was verified by fluorescence microscopy, and stable cell lines were established by adding G418 (Sigma-Aldrich) to the culture medium.

RNA isolation and miR-196 detection

Total RNA samples were isolated using the RNA Mini Kit (Qiagen, Valencia, CA), with on-column DNA digestion. The primers used for detecting the miR-196a-2 precursor were: (L: acccagcaacccaaagtcta, R: atctggaggagaagggaagg) and expression was assessed using the 2^-ΔΔCt method with RNA content normalized to the housekeeping gene hypoxanthine phosphoribosyltransferase 1 (HPRT1). To determine the levels of mature miRNA for both products of the miR-196a-2 stem-loop (i.e. miR-196a and miR-196a*), miRNA-specific qRT-PCR was done using the NCode kit (Invitrogen) which polyadenylates the miRNAs prior to cDNA synthesis. The cDNA was then amplified using one universal PCR primer targeting on the polyadenylated region of the miRNA (supplied with kit), and a miRNA-specific forward primer (miR-196a sequence: taggtagtttcatgttgttggg, miR-196a*-C: cggcaacaagaaactgCctgag, miR-196a*-T: cggcaacaagaaactgTctgag) which is complimentary to the miRNA cDNA. Amplification was measured by SYBR green and fold change in miRNA abundance relative to empty vector was calculated using the 2^-ΔΔCt method normalized to miR-16. All real-time quantitative PCR reactions were performed in triplicate.

Genome-wide expression microarray and pathway analysis

Gene expression differences in pre-miR-196a-C and pre-miRNA-196a-T transfected cells were interrogated by whole genome microarray (Agilent, Inc 41k chip, performed by MoGene, LC, St Louis, MO). RNA was isolated for array analysis from biological replicates of cells transfected with an empty vector (negative control), pre-miR-196a-C or pre-miR-196a-T. Fold changes in gene expression relative to the negative control, and false discovery rate adjusted p-values (FDR) were determined for each gene in cells from each treatment. Transcripts were identified as significantly influenced by each miRNA treatment if they fit the criteria of FDR <0.01 and fold change >|2.0|. Significantly altered transcripts were investigated for network and functional inter-relatedness using the Ingenuity Pathway Analysis software tool (Ingenuity Systems, www.ingenuity.com). This software scans the set of input genes to identify networks using information in the Ingenuity Pathways Knowledge Base, an extensive, manually curated database of functional interactions extracted from peer-reviewed publications (22). A Fisher’s exact test, based on the hypergeometric distribution, is then performed to determine the likelihood of obtaining at least the same number of molecules by chance (i.e. from a random input set), as actually overlap between the input gene set and the genes present in each identified network. A subset of breast cancer relevant genes from the top identified network were validated by quantitative PCR. All microarray data were uploaded to the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/projects/geo/); accession #GSE15576.

Cell cycle assay

Each cell treatment group (empty vector, pre-miR-196a-C, and pre-miR-196a-T) was treated with either 0.015% (v/v) methyl methanesulfonate (MMS, DNA damage-inducing agent) or mock treatment (0.015% PBS, pH 7.4) for one hour. MMS is an alkylating agent that interacts directly with DNA and creates lesions which are normally repaired via the base excision or nucleotide excision repair pathways (23). As such, after MMS treatment we expected to see a delay in the G2/M phase of the cell cycle as these repair mechanisms operate prior to cell cycle progression. Cells were then washed two times in PBS and incubated under normal conditions for 24 hours. Each cell population was then harvested and stained with propidium iodide (PI) in order to quantify DNA content for determination of cell cycle distributions. Cell fluorescence was determined by flow cytometry (FACSCalibur, Becton Dickinson, San Jose, CA), and gates were set to include cells with GFP activity only (i.e. to include only successfully transfected cells). Cell phases were determined with the CellQuest Software (Becton Dickinson) using the Watson-Pragmatic algorithm (24). All data are the result of three biological replicates of each treatment condition.

Statistical analysis

All statistical analyses were performed using the SAS statistical software (SAS Institute, Cary, NC) unless otherwise noted. A chi-square test was used to test for departures from Hardy-Weinberg equilibrium (HWE) for each SNP in the control population. Tests for trend were conducted by assigning the ordinal values 0, 1, 2 to the genotypes in rank order of wild type, heterozygous, and homozygous variant genotypes, respectively. Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were calculated in the overall population, and in pre-menopausal and post-menopausal women separately, by logistic regression to estimate the relative risk associated with each genotype. Adjusted odds ratios were calculated by unconditional logistic regression to control for age, ethnicity, family history of breast cancer in first-degree relatives, menopausal status, BMI, smoking status, family income, and study site. For analysis of methylation in the region upstream of miR-196a-2, the median MI in the controls was used to divide the population into two categories of “high” or “low” methylation, and an odds ratio for high methylation was calculated by logistic regression, controlling for the same covariates as the genotyping analysis. MI in cases and controls was also compared using the Wilcoxon rank sum test (normal approximation). Cell cycle response to mutagen challenge was measured by comparing the proportion of cells in each phase in the mutagen treated group to the same proportion in the mock treated group, based on triplicate experiments for each experimental condition. The normality of each of these sets of proportions was assessed using the Shapiro-Wilk test, and no p-value suggested a significant departure from the normal distribution at alpha = 0.05. Similarly, equality of variances was assessed using the Folded F method, and all tests suggested no departure from equal variances. As such, the two-sample pooled (homoscedastic) Student’s t-Test was used. All p-values are two-sided.

Results

A common SNP (rs11614913) in hsa-miR-196a-2 is associated with breast cancer risk

There were a total of 462 miRNA genes in the MicroCosm database produced by the Sanger Institute at the time of our search. By mapping the chromosomal positions of each miRNA on HapMap, we were able to initially identify 15 genetic variations in 15 different miRNA genes (Supplemental Table 1). These sequence variants were located in various regions of the miRNA gene, including the mature sequence, the stem-loop structure, and the primary miRNA region. Of the 15 original SNPs, five were rarely polymorphic (MAF<0.1), one failed to amplify in the genotyping assay and two had low quality amplification and allelic distributions that significantly departed from that which would be expected under HWE (p<0.05). These SNPs were thus excluded from further analysis. Each remaining SNP was analyzed in the full population, as well as in a secondary analysis with the population stratified by menopausal status (Supplemental Table 2). The homozygous variant genotype of a common SNP (rs11614913, minor allele frequency = 40%) located in miR-196a-2, was significantly associated with reduced breast cancer risk in the full population (OR=0.44, 95% CI 0.28-0.70), as were the combined variant genotypes (heterozygous + homozygous variant, OR=0.75, 95% CI 0.57-0.98, P for trend=0.002; Table 1). Similar results were obtained after stratifying by menopausal status, and when restricting the sample to Caucasians only (data not shown). Note that this SNP would remain significant at alpha=0.05 even after employing the conservative Bonferroni adjustment for multiple comparisons (p 0.002*15 = 0.030).

Table 1

Epidemiological analyses of a genetic variant in pre-miR-196a-2 and upstream CpG island methylation

Methylation of a CpG rich region upstream of hsa-miR-196a-2 is associated with breast cancer risk

In addition to the genetic variant identified in hsa-miR-196a-2, a CpG island was also identified in the region upstream of the miRNA precursor (Figure 1). The level of methylation in this region was determined by methylation-specific PCR for all women who had not undergone radio- or chemotherapy, due to the potential for these treatments to influence DNA methylation (N=77 controls and 73 cases). Cases were significantly less likely to have high methylation at this CpG site (OR=0.36 95% CI 0.16-0.82, P=0.02). Methylation indices were also compared using the Wilcoxon rank sum test, and again, methylation was significantly lower among cases (P=0.03; Table 1). A comparison of untreated cases to treated cases was performed, which revealed no significant differences in any patient characteristics (data not shown).

Figure 1

CpG island location and stem-loop structure of hsa-miR-196a-2, including mature sequences of miR-196a and miR-196a*, and the location of the genetic variant.

Impact of the variant on miR-196a levels in vitro

hsa-miR-196a-2 is comprised of two different mature miRNAs (miR-196a and miR196a*) which are processed from the same stem-loop (Figure 1). The SNP (rs11614913) lies in the mature sequence of miR-196a*, but may influence either miRNA by affecting processing of the pre-miRNA to its mature form. To investigate the influence of the SNP on miRNA processing and target binding efficiency, 3 separate expression vectors were constructed which contained either pre-miR-196a-C, pre-miR-196a-T, or no miRNA insert (empty vector/negative control). MCF-7 cells were vector transfected and selected for 2 weeks with G418, and RNA was isolated for subsequent analysis by qRT-PCR to determine the levels of precursor and mature miRNA. Relative to the empty vector, both the pre-miR-196a-C and pre-miR-196a-T treated cells had approximately equal levels of the precursor (9.8 and 11.1 fold increase, respectively). Mature miR-196a levels were increased 9.3 fold in cells transfected with pre-miR-196a-C compared to empty vector, but only 4.4 fold in cell transfected with pre-miR-196a-T. In cells transfected with pre-miR-196a-C and those transfected with pre-miR-196a-T, mature miR-196a* was increased 3.1 and 3.4 fold, respectively (Figure 2). These results indicate that miR-196a is the primary mature product of the pre-miR-196a-2 hairpin, and processing of this mature sequence is affected by the SNP. Note that the results shown for miR-196a* were obtained using a primer sequence complimentary to the C allele. Similar results were obtained when using a primer for the T allele (data not shown), which indicates insufficient specificity for the assay to discriminate between this single base substitution.

Figure 2

Successful vector transfection of MCF-7 cells, and increased expression of the miRNA precursor, mature miR-196a and mature miR-196a*. A) Cluster of cells visualized using both light (left panel) and fluorescence (right panel) microscopy. Expression of (more ...)

Impact of the variant on expression of miR-196a-2 mediated genes

A whole genome expression microarray was also done using RNA isolated from cells transfected with each of the vectors (empty vector, pre-miR-196a-C, and pre-miR-196a-T). RNA was isolated separately from two independently transfected populations (biological replicates), and genome-wide expression was determined in miR-196a-C cells compared to empty vector, and in miR-196a-T cells compared to empty vector. A total of 684 transcripts fit the criteria for altered expression (FDR<0.01 and fold change > |2|) following introduction of pre-miR-196a-C, while fewer than half that number (263) were significantly altered following introduction of pre-miR-196a-T. Since the epidemiologic evidence suggests that the variant allele is associated with protection from breast cancer risk, and the variant leads to diminished regulatory capacity and decreased mature miR-196a levels, this data indicate that miR-196a may have oncogenic properties in breast cancer

Cancer-related network formed by miR-196 mediated genes

In order to determine which biological pathways may be influenced by the products of pre-miR-196a-2, the set of genes with altered expression following introduction of the pre-miR-196a-C vector was examined using the Ingenuity Pathway Analysis software package. A list of all networks significantly associated with the input genes is available in the supplementary material (Supplemental Table 3). The most significant disease associated with the altered genes was cancer, with 137 cancer-related molecules identified. In addition, a significant pathway (P=1.0E^-48) was identified with known relevance for “Cancer, Cell Death, and Reproductive System Disease” (Figure 3). This network consists of 35 molecules, two of which (hCG and MAP2K1/2) refer to gene families with multiple subunits, leaving 33 molecules which can be identified uniquely in the microarray dataset (Table 2). In order to validate the microarray results, a subset of 9 breast cancer relevant genes identified in this network as differentially regulated by pre-miR-196a-C were validated by quantitative PCR. In all cases, the direction of the fold change was the same when analyzed by array or PCR, and there was excellent correlation between the two platforms (r=0.979, Supplemental Table 5).

Figure 3

Network of altered genes following miR-196a-C vector transfection. This network was identified by the IPA software as significantly associated with the set of genes with altered expression following introduction of pre-miR-196a-C (P=1.0E-48). According (more ...)

All genes in the network were then compared in the miR-196a-C and miR-196a-T populations to determine the effect of the SNP on target gene regulation. 29 of the 33 genes were significantly differentially expressed in the miR-196a-C population, while only 3 were significantly altered in the miR-196a-T cells, again suggesting diminished regulatory capacity of the variant form. Cancer-relevant transcripts in this network included Tumor Protein p63 (TP63), and the tumor suppressors Growth Arrest and DNA-Damage-Inducible, Gamma (GADD45G) and Inhibin, Beta B (INHBB). Also present in this network were two members of the S100 family of genes (S100A8 and S100A9) which are prognostic indicators for poor clinical outcome in some breast tumors. All of these genes were significantly altered following introduction of miR-196a-C, but were unchanged in the miR-196a-T cells, suggesting that this SNP has the potential to influence genes with relevance for breast tumor promotion and progression.

Cell cycle response to mutagen challenge affected by miR-196a-2

In addition to cancer generally, the network analysis identified 86 differentially expressed molecules with relevance for cellular growth and proliferation. To test whether miR-196a-2 may influence cell cycle phenotypically, we performed an analysis of cell cycle regulation using cell populations from each vector treatment group. After MMS treatment, we expected to see a delay in the G2/M phase of the cell cycle as these repair mechanisms operate prior to cell cycle progression. As expected, all cells exhibited a substantial G2 delay following mutagen challenge (Empty Vector: 4.1 times as many cells in G2 compared to mock treated, Student’s t-test, P=0.026; pre-miR-196-C: 2.6 times as many G2 cells in mutagen treated vs. mock treated, P=0.040; and pre-miR-196-T: 6.5 times as many G2 cells, P=0.002) (Figure 4). However, when comparing the response of pre-miR-196a-C treated cells versus those treated with pre-miR-196a-T, the G2 delay was significantly larger in the variant (T) population (P=0.020), again suggesting that the variant has a protective effect on tumorigenesis. Of note: while we may have expected intermediate G2 response in the variant population, given the intermediate level of mature miR-196a in this population, this group appears to in fact display an enhanced response. However, this difference in G2 delay in the variant population was not significantly different from control (P=0.218), and it remains unclear whether the sequence variant in mature miR-196a* influences targets relevant for cell cycle progression.

Figure 4

Cell cycle response to mutagen challenge affected by miR-196a-2. Cells transfected with empty vector, pre-miR-196a-C and pre-miR-196a-T were either mock treated or treated with a DNA damage-inducing agent. The expected G2 delay was significantly diminished (more ...)

Discussion

After analyzing fifteen SNPs identified from the public SNP database in DNA samples extracted from patients with incident breast cancer and controls, we report here a significant association between a common sequence variant in miR-196a-2 and predisposition to breast cancer. This association remains significant after correcting for multiple comparisons, and is corroborated by in vitro evidence, including: 1) The variant negatively influences endogenous processing of the miRNA precursor to its mature form; 2) The variant has a phenotypic impact on target gene expression, as fewer genes were altered following introduction of the pre-miR-196a-T vector relative to the pre-miR-196a-C vector; 3) A pathway analysis revealed 137 cancer-related transcripts with significantly altered expression following introduction of pre-miR-196a; and 4) Cell cycle response to mutagen challenge in the form of G2 delay was directly affected in cells transfected with pre-miR-196a-C, and this effect was abrogated in cells transfected with pre-miR-196a-T.

In addition to the in vitro findings, a methylation analysis of a CpG island located upstream of the miRNA indicated that increased methylation was associated with decreased breast cancer risk, as cases had significantly higher methylation levels than controls. However, since chemical and radiation therapy may influence methylation patterns, only samples from untreated cases could be used for this analysis, thus significantly reducing the available sample size. In addition, RNA samples were not available for these subjects, and thus expression levels cannot be directly associated with level of methylation. However, a previous study reported a strong correlation (r=0.80) between methylation levels near a miRNA gene and miRNA expression in a subgroup of breast cancer samples (19). If higher degree of methylation correlates with decreased miRNA expression, these data would be consistent with all other findings, which suggest that miR-196a has oncogenic properties. Further study will be necessary in order to confirm whether methylation at this site does in fact correlate with pre-miR-196a expression, and to determine whether this area may indeed serve as a novel epigenetic biomarker for breast cancer susceptibility.

A previous study of 5 common miRNA SNPs conducted in China also identified the same polymorphism (rs11614913) as significantly associated with breast cancer risk (25). Interestingly, although the same allele was associated with risk (C allele), this was the rare allele in the Chinese population, while it was the common allele in our, predominantly Caucasian, population. In another study investigating the same five miRNA SNPs and non-small cell lung cancer survival in China, the same group identified the C allele of rs11614913 as significantly associated with poor survival, indicating the possibility that this miRNA may have implications for prognosis, in addition to cancer risk (26). This study also suggests that the pathways influenced by miR-196a are not relevant only for breast cancer, but are also important for at least one other unrelated cancer type.

Several members of the HOX gene family are known to be targeted by miR-196a (27). While many of these genes showed up at poor intensity in our array, four members of the HOX family, HOXB2, HOXB3, HOXC13, and HOXB5 were significantly downregulated in cells treated with pre-miR-196a-C, with average fold changes of -10.21 (FDR<0.001), -3.23 (FDR<0.001), and -1.88 (FDR=0.001), and -2.18 (FDR=0.038), respectively. Of these, only HOXB2 was significantly downregulated in pre-miR-196a-T treated cells, although the magnitude of the effect was much smaller than that observed in the pre-miR-196a-C population (fold change=-3.69), again suggesting that the SNP results in decreased target gene repression.

The network analysis demonstrated the potential for miR-196a-2 to be operative in a number of biological pathways, including a high confidence network (P=1.0E^-48) with relevance for cancer. Genes in this network included at least two tumor suppressors which were significantly downregulated following pre-miR-196a-C introduction but were not significantly altered in the pre-miR-196-T cells. GADD45G, which was downregulated more than two-fold, is epigenetically silenced in a number of tumors, and acts as a tumor suppressor by negatively regulating cell proliferation (28). INHBB, which was also downregulated more than two-fold, is a negative regulator of growth which has been proposed as a tumor suppressor for some cancer types, and is expressed in normal breast tissue (29, 30). TP63, which was more than six-fold upregulated following pre-miR-196a-C introduction may act as an oncogene, despite having significant sequence similarity to p53 (31). In addition, two genes encoding calcium-binding proteins (S100A8 and S100A9) were also significantly upregulated, by fifteen- and 47-fold, respectively. These proteins form a heterodimer which has been implicated in tumor promotion, possibly mediated by its role as a proinflammatory cytokine (32). In addition, these genes are upregulated in a number of tumors, and their overexpression has been associated with poor prognosis in invasive ductal cell carcinoma of the breast (33). Downregulation of tumor suppressors and upregulation of oncogenes by introduction of pre-miR-196a-C would be consistent with overall oncogenic activity, and the diminished regulatory activity of pre-miR-196a-T is consistent with a protective role for the T allele of the SNP.

Although our findings demonstrate that the SNP does have a phenotypic impact, both in terms of pre-miRNA processing and target gene regulation, these data are not useful in determining which genes are direct targets of mature miR-196a and miR-196a*, and which are indirectly influenced. As such, an exploration into the nature of the physical interactions between these miRNAs and their potential targets should be the subject of future investigations. Furthermore, given the earlier finding demonstrating a role for this SNP in lung cancer prognosis, similar studies are warranted to determine whether there are prognostic implications for this variant among patients with breast cancer.

Table 3

Network of differentially expressed genes following enforced expression of pre-miR-196a-C and pre-miR-196a-T

Supplementary Material

Click here to view.^{(50K, doc)}

Acknowledgments

We are grateful to Ewa Menet at the Cell Sorter Facility and Irina Tikhonova at the Keck Center for assistance with the cell cycle assay, and Sequenom genotyping analysis, respectively. This work was supported by the National Institutes of Health (grants CA122676, CA110937, and CA108369).

References

1. Britten RJ, Davidson EH. Gene regulation for higher cells: a theory. Science. 1969;165:349–357. [PubMed]

2. Lee RC, Feinbaum RL, Ambros V. The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell. 1993;75:843–854. [PubMed]

3. Reinhart BJ, Slack FJ, Basson M, et al. The 21-nucleotide let-7 RNA regulates developmental timing in Caenorhabditis elegans. Nature. 2000;403:901–906. [PubMed]

4. Song L, Tuan RS. MicroRNAs and cell differentiation in mammalian development. Birth Defects Res C Embryo Today. 2006;78:140–149. [PubMed]

5. Pasquinelli AE. Demystifying small RNA pathways. Dev Cell. 2006;10:419–424. [PubMed]

6. Calin GA, Sevignani C, Dumitru CD, et al. Human microRNA genes are frequently located at fragile sites and genomic regions involved in cancers. Proc Natl Acad Sci U S A. 2004;101:2999–3004. Epub 2004 Feb 2918. [PubMed]

7. Calin GA, Croce CM. MicroRNA-Cancer Connection: The Beginning of a New Tale. Cancer Res. 2006;66:7390–7394. [PubMed]

8. Takamizawa J, Konishi H, Yanagisawa K, et al. Reduced expression of the let-7 microRNAs in human lung cancers in association with shortened postoperative survival. Cancer Res. 2004;64:3753–3756. [PubMed]

9. Calin GA, Dumitru CD, Shimizu M, et al. Frequent deletions and down-regulation of micro-RNA genes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia. Proc Natl Acad Sci U S A. 2002;99:15524–15529. Epub 12002 Nov 15514. [PubMed]

10. Gregory RI, Shiekhattar R. MicroRNA biogenesis and cancer. Cancer Res. 2005;65:3509–3512. [PubMed]

11. Esquela-Kerscher A, Slack FJ. Oncomirs - microRNAs with a role in cancer. Nat Rev Cancer. 2006;6:259–269. [PubMed]

12. Wijnhoven BP, Michael MZ, Watson DI. MicroRNAs and cancer. Br J Surg. 2007;94:23–30. [PubMed]

13. Lagos-Quintana M, Rauhut R, Meyer J, Borkhardt A, Tuschl T. New microRNAs from mouse and human. Rna. 2003;9:175–179. [PubMed]

14. Kumar MS, Lu J, Mercer KL, Golub TR, Jacks T. Impaired microRNA processing enhances cellular transformation and tumorigenesis. Nat Genet. 2007;39:673–677. [PubMed]

15. Volinia S, Calin GA, Liu CG, et al. A microRNA expression signature of human solid tumors defines cancer gene targets. Proc Natl Acad Sci U S A. 2006;103:2257–2261. Epub 2006 Feb 2253. [PubMed]

16. Iorio MV, Ferracin M, Liu C-G, et al. MicroRNA Gene Expression Deregulation in Human Breast Cancer. Cancer Res. 2005;65:7065–7070. [PubMed]

17. Weber B, Stresemann C, Brueckner B, Lyko F. Methylation of human microRNA genes in normal and neoplastic cells. Cell Cycle. 2007;6:1001–1005. [PubMed]

18. Lu L, Katsaros D, de la Longrais IA, Sochirca O, Yu H. Hypermethylation of let-7a-3 in epithelial ovarian cancer is associated with low insulin-like growth factor-II expression and favorable prognosis. Cancer Res. 2007;67:10117–10122. [PubMed]

19. Lehmann U, Hasemeier B, Christgen M, et al. Epigenetic inactivation of microRNA gene hsa-mir-9-1 in human breast cancer. J Pathol. 2008;214:17–24. [PubMed]

20. Zheng T, Holford TR, Mayne ST, et al. Risk of female breast cancer associated with serum polychlorinated biphenyls and 1,1-dichloro-2,2′-bis(p-chlorophenyl)ethylene. Cancer Epidemiol Biomarkers Prev. 2000;9:167–174. [PubMed]

21. Zhu Y, Brown HN, Zhang Y, Stevens RG, Zheng T. Period3 structural variation: a circadian biomarker associated with breast cancer in young women. Cancer Epidemiol Biomarkers Prev. 2005;14:268–270. [PubMed]

22. Calvano SE, Xiao W, Richards DR, et al. A network-based analysis of systemic inflammation in humans. Nature. 2005;437:1032–1037. [PubMed]

23. Viau CM, Guecheva TN, Sousa FG, et al. SnCl(2)-induced DNA damage and repair inhibition of MMS-caused lesions in V79 Chinese hamster fibroblasts. Arch Toxicol. 2009 [PubMed]

24. Watson JV, Chambers SH, Smith PJ. A pragmatic approach to the analysis of DNA histograms with a definable G1 peak. Cytometry. 1987;8:1–8. [PubMed]

25. Hu Z, Liang J, Wang Z, et al. Common genetic variants in pre-microRNAs were associated with increased risk of breast cancer in Chinese women. Hum Mutat. 2008 [PubMed]

26. Hu Z, Chen J, Tian T, et al. Genetic variants of miRNA sequences and non-small cell lung cancer survival. J Clin Invest. 2008;118:2600–2608. [PMC free article] [PubMed]

27. Yekta S, Shih IH, Bartel DP. MicroRNA-directed cleavage of HOXB8 mRNA. Science. 2004;304:594–596. [PubMed]

28. Ying J, Srivastava G, Hsieh WS, et al. The stress-responsive gene GADD45G is a functional tumor suppressor, with its response to environmental stresses frequently disrupted epigenetically in multiple tumors. Clin Cancer Res. 2005;11:6442–6449. [PubMed]

29. Reis FM, Luisi S, Carneiro MM, et al. Activin, inhibin and the human breast. Mol Cell Endocrinol. 2004;225:77–82. [PubMed]

30. Sharifi N, Lechleider RJ, Farrar WL. Transforming growth factor-beta receptor III downregulation in prostate cancer: is inhibin B a tumor suppressor in prostate? J Mol Endocrinol. 2007;39:329–332. [PubMed]

31. Westfall MD, Pietenpol JA. p63: Molecular complexity in development and cancer. Carcinogenesis. 2004;25:857–864. [PubMed]

32. Gebhardt C, Nemeth J, Angel P, Hess J. S100A8 and S100A9 in inflammation and cancer. Biochem Pharmacol. 2006;72:1622–1631. [PubMed]

33. Arai K, Takano S, Teratani T, et al. S100A8 and S100A9 overexpression is associated with poor pathological parameters in invasive ductal carcinoma of the breast. Curr Cancer Drug Targets. 2008;8:243–252. [PubMed]