Expression and function of the miR-143/145 cluster in vitro and in vivo in human breast cancer

MicroRNAs (miRNAs) are small non-coding RNAs that function as post-transcriptional regulators of gene expression and are dysregulated in cancer. Studies of miRNAs to explore their potential as diagnostic and prognostic markers are of great scientific interest. Here, we investigate the functional properties and expression of the miR-143/145 cluster in breast cancer (BC) in vitro and in vivo. The ER positive MCF7, the HER2 positive SK-BR-3, and the triple negative cell line MDA-MB-231 were used to assess cell proliferation and cell invasion. Expression of miRNA in 108 breast cancers in the Norwegian Women and Cancer Study and 44 benign tissue controls were analyzed by microarray and validated by RT-PCR. Further, in situ hybridization (ISH) was used to study the cellular and subcellular distribution of the miRNAs. In vitro, miR-143 promoted proliferation of MCF7 and MDA-MB-231 cells, whereas miR-145 and the cotransfection of both miRNAs inhibited proliferation in all three cell lines. The cells’ invasive capacity was reduced after transfection and cotransfection of the miRNAs. In line with the tumor suppressive functions in vitro, the expression of miR-143 and miR-145 was lower in malignant compared to benign breast tissue, and lower in the more aggressive tumors with higher tumor grade, loss of ER and the basal-like phenotype. ISH revealed miR-143 to be cytoplasmatic and predominantly expressed in luminal cells in benign tissue, whilst miR-145 was nuclear and with strong staining in myoepithelial cells. Both miRNAs were present in malignant epithelial cells and stromal fibroblasts in BC. This study demonstrates that miR-143 and -145 have functional properties and expression patterns typical for tumor suppressors, but the function is influenced by cellular factors such as cell type and miRNA cotransfection. Further, the nuclear functions of miR-145 should be explored for a more complete understanding of the complexity of miRNA regulation and function in BC.

Introduction (REKnord 2010(REKnord /1931(REKnord , 2013(REKnord /2271. The Data Inspectorate has also approved the storing of relevant, not identifiable data and the linkage to national registries. In addition, ethical aspects have been considered within the project to ensure the most efficient and accurate use of the material collected and data generated, in accordance with national and international guidelines and laws.

Functional studies
The potential function of miR-143 and miR-145 in tumorigenesis was investigated by a series of in vitro experiments. The experiments were performed by introducing miR-143 mimic or miR-145 mimic, alone or in combination, alongside a miRNA negative control into various BC cell lines. In this study, cell proliferation and cell invasion were assessed.
Total RNA isolation. Total RNA was isolated from cell lines using the miRNeasy Mini Kit (cat.# 217004, Qiagen, Hilden, Germany) according to the manufacturer's protocol. In short, cells were lysed in 700 μl QIAzol Lysis Reagent before homogenization and 5 minutes incubation at room temperature. 140 μl chloroform were added, and the samples were shaken before incubation at room temperature for 3 minutes. Samples were centrifuged for 15 minutes at 12000 g at 4˚C, and the upper aqueous phase was transferred and mixed thoroughly with 100% ethanol. The samples were transferred into the RNeasy1 Mini column and washed in several steps before elution with 50 μl ddH 2 O. Isolated total RNA samples were stored at -70˚C. cDNA synthesis. First strand cDNA synthesis was performed using the miScript II RT Kit (cat.# 218160, Qiagen, Hilden, Germany) according to the manufacturer's protocol. Briefly, 100 ng of total RNA was mixed with 4 μl 5x miScript HiSpec Buffer, 2 μl 10x Nucleics Mix, 2 μl miScript Reverse Transcriptase Mix, and RNase-free water to a total volume of 20 μl. Samples were incubated for 60 minutes at 37˚C, and subsequently incubated for 5 minutes at 95˚C to inactivate enzymes. Finally, samples were diluted up to a total volume of 200 μl in RNase-free water and stored at -20˚C.
Proliferation assay. The BC cell lines' ability to proliferate after transfection was evaluated using the real-time cell analyzer system xCelligence, RTCA DP (cat#05469759001, ACEA Biosciences, San Diego, USA) fitted with the E-plate 16 (cat#05469830001, ACEA Biosciences, San Diego, USA). Prior to analysis on the xCelligence platform, cell lines were trypsinized until detached, resuspended in complete growth media, and counted. Initial titration experiments estimated approximately 8000 cells per well to be optimal. In accordance with the manufacturer's protocol, cells were seeded in quadruplicates into an E-plate after baseline measurements. The E-plate containing cells was incubated for 30 minutes at room temperature before positioned in the RTCA DP instrument, which was located in an incubator preserving the same conditions as used for routine cultivation of cell lines. The instrument denotes the cellular growth rate as 'Cell Index', which is an arbitrary unit reflecting the cellsensor impedance. The cell index was recorded by the instrument every 30 minutes. Growth curves were calculated with the RTCA software version 1.2.1 (ACEA Biosciences, San Diego, USA). A minimum of three independent experiments were performed for each cell line.
Invasion assay. The cell lines invasiveness after transfection was tested using the CytoSelect TM 96-well Cell Invasion Assay, Basement membrane (cat.# CBA-112, Cell Biolabs, San Diego, USA) according to the manufacturer's protocol. Briefly, 50000 pretransfected and serum starved cells were seeded in the upper chamber of a modified Boyden chamber. The chamber was coated with a basement membrane consisting of a protein matrix isolated from Engelbreth-Holm-Swarm tumor cells, and the cells were allowed to invade for 24 h towards the bottom chamber containing media+10% FBS. Cancer cells able to invade the basement membrane and pass through the porous membrane to the bottom side of the membrane were lysed, stained, and fluorescence was measured at 480/520 nm using the CLARIOstar1 microplate reader (BMG LABTECH, Ortenberg, Germany). All experiments were performed in quadruplicates, and a minimum of three independent experiments were performed for each cell line.

Patient material and tumor classification
The patient samples were collected from the NOWAC postgenome cohort [25]. The NOWAC participants included in this study were diagnosed with breast cancer at the Department of Pathology at the University Hospital of North Norway in Tromsø, or the Nordland Hospital in Bodø in the years 2004-2010. Archived formalin-fixed paraffin-embedded (FFPE) tissue blocks, and hematoxylin and eosin stained slides were collected. Histological grading of tumors was based on the criteria modified by Elston and Ellis [26] and immunohistochemical (IHC) analyses of ER, progesterone receptor (PR) and HER2 were done on needle biopsies as part of routine diagnostics. The cut-off value for ER positivity was ! 1%, for PR ! 10% and a HER2 score of 3+ was considered positive, a score of 0-1+ negative whereas a score of 2+ lead to silver in situ hybridization (SISH) where HER2 was considered negative if HER2/chromosome 17-ratio was < 2. IHC staining for the proliferation marker Ki67 was done on slides from the primary surgery, and the expression evaluated in at least 500 tumor cells in the most proliferative areas of the tumor and the result reported as a percentage of positive tumor cells. Subtyping of the tumors according to molecular profile were based on the surrogate markers ER, PR, HER2 and Ki67 according to recommendations by the St Gallen International Expert Consensus and previous publications [27,28]. The subtyping was performed as follows: luminal A (ER+ and/or PR+, HER2-and Ki67 30%), luminal B (ER+ and/or PR+, HER2-and Ki67 > 30% or ER+ and/or PR+ and HER2+), HER2 positive (ER-and PR-and HER2+) and basal-like (ER-, PR-and HER2-). Histopathological data were collected from the original pathology reports, and reevaluated and completed according to updated criteria by a breast pathologist (L.M.). As benign tissue controls, FFPE tissue cores from 44 breast reduction surgery specimens were included in the study.

miRNA microarray
Total RNA was extracted from FFPE tissue cores from both malignant and benign breast tissue using the RecoverAll Total Nucleic Acid Isolation kit (Life Technologies, Grand Island, NY, USA) following the manufacturer's instructions. RNA quality and quantity was assessed using the NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE). Microarray hybridization and analyses were performed as a bought service by Exiqon (Vedbaek, Denmark). In short, the miRCURY LNATM microRNA Hi-Power Labeling Kit (Exiqon) was used to label 250 ng total RNA from samples and reference with Hy3TM and Hy5TM, respectively. The Hy5TM -labeled reference RNA contained an equal aliquot of all RNA species included in the study. Labeled samples and reference RNA were mixed before hybridization to the 7th generation miRCURY LNA microRNA array (Exiqon), using a Tecan HS4800 hybridization station (Tecan, Austria). The microarray contained capture probes for miRNAs in human, mouse and rat as annotated in miRBASE version 19.0. The slides were scanned on the Agilent G2565BA Microarray Scanner System (Agilent technologies Inc., USA) and the Ima-Gene 9.0 software (BioDiscovery Inc., USA) was used for image analysis. The quantified signals were background corrected and normalized using quantile normalization method and detection threshold set as 1.2 times the 25th percentile of the overall signal intensity of the individual slides.

Validation of microarray and quantification of miRNAs by RT-qPCR
Microarray miRNA analyses were validated using RT-qPCR. 40 tumor samples representing the four major molecular subtypes of cancer included in the study, and 20 of the benign breast tissue controls were included in the PCR validation done by Exiqon. In short, RNA was extracted from FFPE tissue cores using the Qiagen miRNeasy FFPE kit according to the manufacturer's instructions (Qiagen, Hilden, Germany). 10 ng RNA was reverse transcriped using the miRCURY LNA Universal RT microRNA PCR, Polyadenylation and cDNA synthesis kit (Exiqon) and PCR-reactions performed on 100 x diluted cDNA using ExiLENT SYBR Green master mix. The amplification was done in a Light Cycler 480 Real-Time PCR System (Roche) in 384 well plates. All reverse transcription reactions were done in duplicates. Based on stable expression across the data set, the most suitable reference miRNAs were evaluated by Exiqon using the Normfinder software. Of the suitable reference miRNAs, miR-664a-3p was detected in all samples and was used for normalization. Normalized expression values for each miRNA in each sample were calculated using the quantification cycle (Cq) from PCR analyses and the formula: average Cq (all samples)-assay Cq (sample).

In situ hybridization
In order to study the cellular and subcellular location of miR-143 and miR-145 in benign and malignant breast tissue, we analyzed the miRNA in situ hybridization (ISH) staining in full histological slides of 16 tumors with adjacent normal tissue. Buffers and detection reagents were purchased from Roche (Basel, Switzerland) and labelled locked nucleic acid (LNA) modified probes purchased from Exiqon (Vedbaek, Denmark). The chromogen ISH was performed in the Ventana Discovery Ultra instrument for IHC and ISH (Ventana Medical Systems Inc, Arizona, USA) with deparaffinization, pretreatment, hybridization, chromogen staining and counterstaining automatized in the instrument. In short, 4 μm tissue sections were incubated overnight at 60˚C to attach tissue to Super Frost Plus slides. To ensure good distribution of reagents and protect sections from drying, liquid coverslip oil (Roche) was added during incubation. Sections were deparaffinized in EZ Prep buffer (Roche) at 68˚C (3 x 12 min), followed by heat-mediated retrieval pretreatment at 95˚C with CC1 buffer (Roche) for 40 minutes and rinsing with Reaction Buffer (Roche) followed by RiboWash SSPE buffer (Roche). In this study, we used 5 nM miR-145-5p target probe, 10 nM miR-143-3p target probe, 10 nM scramble miR negative control probe and 0.5 nM U6 positive control probe. Positive and negative tissue controls for both miRNAs were included by using a TMA multi-organ slide.
All slides were denaturated for 8 min at 90˚C, hybridization with probes took place for 60 min at 50˚C for miR-145, 55˚C for miR-143, 57˚C for scramble miR and 55˚C for U6. Stringent washes were done 2 x 8 minutes with 2.0X RiboWash SSPE, followed by rinsing with Reaction Buffer and blocking against unspecific binding with blocking solution (Roche) for 16 minutes at 37˚C. Immunological detection was done with prediluted alkaline phosphatase (AP)-conjugated anti-DIG (Roche) at 37˚C for 20 minutes. The sections were rinsed with Reaction Buffer and EZ Prep before the substrate enzymatic reactions were carried out with NBT/BCIP (CromoMap Blue kit, Roche) for 60 minutes at 37˚C. Sections were rinsed with Reaction Buffer and counterstained for 4 minutes with Red Stain II (Roche). Dehydration of the sections was performed by increasing gradients of ethanol and finally the tissue sections were mounted with glass cover slips.
Scoring of ISH staining intensity. ISH stained full slides of selected tumors were used to collect information on staining intensity and density in tumor cells, stromal fibroblast and adjacent normal breast tissue. The tumors were randomly selected from histological slides of good quality with well preserved invasive carcinoma present. For each slide, three areas of tumor tissue and tumor-associated stromal tissue were evaluated using a microscope at 200x magnification. By morphologic criteria, the tumor cells and stromal fibroblasts were scored for staining intensity with the dominant staining intensity scored as: 0 = negative, 1 = weak, 2 = moderate, 3 = strong. From the observed ISH staining pattern, both tumor cells and stromal fibroblasts stained diffusely and homogenously and hence staining density was not scored since it did not give any additional information. All samples were independently scored by two experienced pathologists (L.M. and S.M.D.).

Statistics
The miRNA microarray and PCR expression data from the breast cancer samples in the NOWAC study were analyzed using the Limma package in R (Linear Models for Microarray and RNA-Seq Data). Moderated F-statistics were applied, with p-values corrected for multiple testing by controlling the false discovery rate using the method of Benjamini & Hochberg. Descriptive statistics, non-parametric tests and correlation analysis were performed using Stata, version 14.
For the qPCR results from the in vitro experiments, the standard error was calculated using all four technical replicates from a representative biological experiment. For the proliferation study, each proliferation curve was tested against the miRNA control using one-way ANOVA with p-values corrected for multiple testing by controlling the false discovery rate using the method of Benjamini & Hochberg. For the invasion study, the standard error was calculated using all four technical replicates from a representative biological experiment.

Relative expression of miR-143 and miR-145 in breast cancer
The endogenous expression of miR-143 and miR-145 in the studied BC cell lines was quantified relative to the non-cancerous cell line MCF-10A (Fig 1). Relative to MCF-10A, endogenous expression levels of miR-143 and miR-145 were downregulated in all BC cell lines. This pattern was also evident in the NOWAC patient material, as described later.

Functional studies on miR-143 and miR-145 in vitro
The potential functional role of miR-143 and miR-145 in breast cancer tumorigenesis was explored by a series of in vitro experiments. Proliferation and invasion were assessed after miR-143 promotes proliferation in vitro. The proliferation rate of the BC cell lines was assessed using the real-time monitoring system xCelligence, fitted with a proliferation plate. Interestingly, transfection of miR-143 led to increased proliferation in the ER+ cell line MCF7 and the TN cell line MDA-MB-231 (Fig 2A and 2C). The HER2+ cell line, however, did not demonstrate any significant change in proliferation when transfected with miR-143 (Fig 2B).
miR-145 inhibits proliferation in vitro. All three studied BC cell lines demonstrated a dramatic drop in proliferation when transfected with miR-145 (Fig 2A-2C). The shift in proliferation capacity occurred between 24-48 hours after transfection, depending on the cell line.
miR-143 and miR-145 inhibits invasion in vitro. The cells' invasive abilities were studied using a Boyden chamber assay. Pretransfected cells were seeded in serum-free media in the upper chambers, and allowed to invade through a basement membrane extract towards the bottom chamber containing media with 10% serum. Invading cells were quantified 20-24 hours after being seeded to the upper chamber. Both miR-143 and miR-145 had a profound effect on cell invasion (Fig 2D). All cell lines demonstrated reduced invasive capacity 24 hours after transfection.
Cotransfection of miR-143 and miR-145 results in a tumor suppressor phenotype. In addition to investigating the functional effects of each individual miRNA, we also wanted to study the effects of cotransfecting miR-143 and miR-145. The BC cell lines were simultaneously transfected with 50 nM of both miR-143 and miR-145. The proliferation of cells cotransfected with miR-143 and miR-145 was dramatically reduced in all three cell lines and was similar to the proliferation pattern of cells transfected with miR-145 alone (Fig 2A-2C). The proliferation promoting effects observed for miR-143 in the ER+ cell line and the TN cell line were cancelled by the simultaneous transfection of miR-145 (Fig 2A and 2C). Cotransfection of miR-143 and miR-145 had an inhibitory effect on invasion, in line with the observations made for each individual miRNA (Fig 2D). The cumulative effect of cotransfection using both miRNAs was not significantly different from experiments where only one miRNA was used.

Patient material
A total of 108 NOWAC postgenome cohort participants were diagnosed with BC at the pathology departments in Northern Norway in the years 2004-2010 and included in the study. Of these, one case had no FFPE tissue block with enough tumor tissue for further analyses. Five cases and six of the benign tissue controls had poor RNA quality, leaving 102 BC surgery specimens and 38 of the 44 benign breast specimens to be included in the miRNA microarray. After microarray miRNA analyses, an additional case and two of the controls were identified as outliers and excluded from further statistical analyses. The histopathological variables for the study cohort are presented in Table 1.

Expression of miR-143 and miR-145 in benign and malignant breast tissue and according to histopathological parameters: microarray and PCR-results
Microarray miRNA analyses demonstrated that miR-143 and miR-145 were significantly downregulated in BC tissue compared to benign breast tissue (p<0.001 for both comparisons) and the downregulation was validated and confirmed by PCR (p<0.001) (Fig 1). Of note, microarray-and PCR-based expression levels were significantly correlated for both miR-143 (r = 0.60, p<0.001) and miR-145 (r = 0.72, p<0.001). Additionally, the expression levels of miR-143 and miR-145 were highly correlated (R = 0.88, p<0.001, Fig 3).
Based on PCR measurements, both miR-143 and miR-145 demonstrated significantly higher expression in tumors that typically have a better prognosis compared to the rest. There was higher expression of both miR-143 and miR-145 in low and intermediate grade tumors compared to the high grade tumors, as shown in Tables 2 and 3 and Fig 4, and in ER-positive compared to ER-negative tumors (Table 4). MiR-145 displayed higher expression in luminal A tumors compared to the other molecular subtypes, and the expression of both miRNAs was lower in basal-like tumors compared to the other major subtypes (Table 5 and Fig 5). The means plot for miR-143 and miR-145 in Fig 6A and 6B, respectively, illustrates the distribution of expression across molecular subtypes. The same trends were observed in the microarray results, but the differences were not statistically significant. There were no significant differences in miR-143 and miR-145 expression between tumor groups stratified according to tumor size or lymph node metastases.

ISH expression of miR-143 and miR-145 in benign and malignant breast tissue
The cellular and subcellular expression of miR-143 and miR-145 in benign and malignant breast tissue was evaluated in full histological slides of 16 tumors with adjacent benign breast tissue. MiR-143 ISH staining was mainly cytoplasmatic and found predominantly in luminal cells in benign breast tissue. Noteworthy, the staining intensity in benign breast ducts and lobuli was strong and homogenous, and appeared stronger in benign tissue compared to adjacent tumor tissue, also in the tumors with moderate to strong staining intensity (Fig 7). As in benign tissue, miR-143 was expressed in the cytoplasm of tumor cells and stromal fibroblasts. However, in the tumors with high staining intensity in stromal cells, the staining was both cytoplasmatic and nuclear.
In contrast to the cytoplasmatic staining pattern for miR-143, miR-145 was expressed in the nuclei and with the strongest staining intensity in the myoepithelial cells in benign breast tissue (Fig 7). MiR-145 staining intensity was strong in benign breast tissue and tended to be stronger in benign compared to malignant breast tissue, but the difference was not as clear as for miR-143 (Fig 7). Also in tumor cells and stromal fibroblasts, miR-145 staining was predominantly nuclear.
Tumor cells and stromal fibroblasts were scored for staining intensity as illustrated in  Study of the miR-143/145 cluster in breast cancer stromal fibroblasts was found to be positively correlated for both miR-143 (p = 0.006) and miR-145 (p = 0.006). Stromal ISH expression of miR-143 and miR-145 was also correlated (p = 0.049) ( Table 6).

Discussion
This study characterizes the functional properties and expression pattern of the miRNA cluster miR-143 and miR-145 in BC. Using microarray and PCR, we found a significant downregulation of miR-143 and miR-145 in malignant tumors compared to normal tissue, which is in line with previously published studies in breast cancer [29][30][31] and other tissues, especially colorectal carcinomas [32]. However, in our study the results were also verified by ISH, which added interesting data on the cellular and subcellular location of these miRNAs in benign and malignant breast tissue. Using ISH, we found the highest expression of miR-145 in myoepithelial cells in benign breast tissue. This is in line with other smaller studies using FISH or laser microdissection and PCR-analyses of miRNA expression [33,34]. Further, we observed a distinct nuclear enrichment of mature miR-145, as has been previously reported in both breast and other tissues [34][35][36]. This is interesting, as the general dogma of miRNA biosynthesis and function involves post- Study of the miR-143/145 cluster in breast cancer transcriptional regulation of mRNA in the cytoplasm [37]. Although the nuclear functions of mature miRNAs remain elusive, there is growing evidence of specific miRNAs localized to the nucleus [38][39][40]. Park et al. suggest that nuclear miRNAs could uncover an entirely new role for this family of non-coding RNAs [40] where transportation across the nuclear membrane could regulate miRNA storage and function. Further, it is suggested that nuclear miRNA could be involved in post-transcriptional gene silencing via the nuclear RNA induced silencing complex (nRISC), transcriptional gene activation via recruitment of transcriptional activators, and influence splicing decisions at specific exons (alternative splicing) [41].
Using microarray and PCR, we found miR-145 to be significantly higher in ER+ tumors, which is in line with a previous study on lymph-node negative tumors where ER status was based on gene expression using microarray [42]. In our study, the expression of both miR-143 and miR-145 was elevated in the least aggressive tumor types, which is in line with the tumor suppressor functions described for these miRNAs in previous publications [15][16][17][18][19]43]. Further, the expression of the two miRNAs seems to correlate in benign and malignant breast tissue and between tumor types, an observation that could partly be contributed to their transcriptional regulation, where the two miRNAs are located within the same chromosomal region and are transcribed as a cluster. However, miRNA functions are complex. In our in vitro experiments, miR-143 was revealed to increase the proliferative capacity of the ER+ cell line MCF7 and the TN cell line MDA-MB-231, whilst having no significant effect on proliferation in the HER2+ cell line SK-BR-3. The effect was pronounced (Fig 2A and 2C), and reproduced in three biological replicates. This finding is in contrast to the majority of studies regarding the effect of miR-143 on proliferation [15][16][17][18][19]. Interestingly, we found that miR-143 had a significant inhibitory effect on cell invasion in all cell lines. The overexpression of miR-145 led to inhibition of both proliferation and invasion, which is in accordance with the general consensus [19,20,[22][23][24][44][45][46][47][48][49][50][51]. Noteworthy, the coexpression of miR-143 and miR-145 led to inhibition of proliferation and invasion, which is also described in previous publications [52][53][54][55]. At present, we do not have enough information to offer any explanatory model for the effects demonstrated by miR-   Study of the miR-143/145 cluster in breast cancer through exosome-mediated delivery of miRNA to breast cancer cells [57]. Estradiol has been published to downregulate expression of miR-143 in an ER-dependent manner in the ER+ cell line MCF7, whilst having no effect on miRNA levels in the TN cell line MDA-MB-231 [43]. Downregulation of miR-143 appears to contribute to the estradiol-mediated upregulation of bcl-2 and survivin. These observations underline that effects of signaling molecules, where miRNAs could be included, are highly dependent on the cellular context. Further, the stimulation of proliferation and, at the same time, inhibition of invasion, argue for several regulated target genes which could have opposing effects in tumorigenesis, as indeed demonstrated by the target genes for miR-143 listed in S1 Table. Spizzo et al. verified miR-145 to directly target ER, and at the same time activate TP53, and in turn apoptosis [51]. These results may in part explain some of the inhibitory effects we    Study of the miR-143/145 cluster in breast cancer progression is wide-ranging (S2 Table), and the dominant active pathway(s) at any given time would be challenging to predict. S1 and S2 Tables present the validated mRNA targets for miR-143 and miR-145, based on searches using the online database miRTarBase [58]. Synergistic effects of miR-143 and miR-145 on BC cell proliferation have been demonstrated in MCF7 and MBA-MD-231 cells and in a mouse model using MCF7 xenografts [53]. Interestingly, both miRNAs, although not sharing sequence homology, were demonstrated to bind to the 3'-UTR region of HER3, the most potent effects were demonstrated by miR-145. Further, high HER3 expression could rescue the cells from the inhibitory effects of miR-143 and miR-145, underlining the importance of the cellular context and active signaling pathways on the net effect of miRNA up-or downregulation. Different results in different cell lines may be explained by miRNAs effects on target genes that are operating or driving the oncogenic phenotype in specific cell types where ER would be the classical example in ER+ BC. Noteworthy, miRNA-profiling of BC cells has demonstrated that miRNA-based clusters follow ERbased dichotomization of cell lines, indicating that a significant number of miRNAs may directly, or indirectly, be regulated by ER-signaling [59].
The results in this study, including the differential effects in vitro of miR-143 in BC cell lines, the effects of cotransfecting miR-143 and miR-145, and the nuclear enrichment of miR-145 in breast tissue, underline the complexity of miRNA regulation and function. Studies of miRNA expression in tissues and cancer types should be supplied by functional research of individual miRNAs where non-canonical effects of miRNAs are being explored in greater detail.
Supporting information S1