MiR-21 Is under Control of STAT5 but Is Dispensable for Mammary Development and Lactation

Development of mammary alveolar epithelium during pregnancy is controlled by prolactin, through the transcription factors STAT5A/B that activate specific sets of target genes. Here we asked whether some of STAT5's functions are mediated by microRNAs. The miR-21 promoter sequence contains a bona-fide STAT5 binding site and miR-21 levels increased in HC11 mammary cells upon prolactin treatment. In vivo miR-21 was abundantly expressed in mammary epithelium at day 6 of pregnancy. Analysis of mice lacking miR-21 revealed that their mammary tissue developed normally during pregnancy and dams were able to nurse their pups. Our study demonstrated that although expression of miR-21 is under prolactin control through the transcription factors STAT5A/B its presence is dispensable for mammary development and lactation.


Introduction
Cytokines control a plethora of biological functions, many of them mediated by the JAK-STAT signal transduction pathway. The transcription factors Signal Transducer and Activator of Transcription 5A and 5B (collectively referred to as STAT5) are key executors of cytokine signaling, including prolactin, growth hormone, erythropoietin, and interleukins [1]. STAT5 controls different facets of mammary development, from the establishment and maintenance of alveolar progenitors during puberty to alveolar differentiation during pregnancy [2][3][4]. The complexity of defects observed in the absence of STAT5 suggests the presence of a network of downstream mediators that effects a variety of physiological functions [5]. MicroRNAs are small non coding RNA molecules, which modulate cell functions via post transcriptional regulation of mRNAs [6]. Like most genes, microRNA genes are regulated by transcription factors and are expressed in tissue and developmental stage specific patterns. For example, the expression of the miR-21 gene is suppressed by STAT3 in glioma cells [7] but upregulated by STAT3 in multiple myeloma cells [8]. These findings open the possibility that the same microRNA can be differentially controlled by the same transcription factor in different tissues or development stages. STAT5 is essential for prolactin induced mammary gland development during pregnancy [2,3]. While multiple STAT5 regulated genes have been discovered [1], only a few microRNAs, among them miR 17-92 and miR-15/16 [9,10] have been linked directly to STAT5. MicroRNA-21 (miR-21) is conserved in mammals and it is located within the gene encoding the transmembrane protein 49 (VPM1) in human [11] and mouse [8]. MicroRNA profiling of different cancer types revealed that miR-21 is up regulated in several cancers, including those of breast, prostate and lung [12,13]. MiR-21 has been shown to target the expression of phosphatase and tensin homologue (PTEN) in hepatocellular cancer [14], programmed cell death 4 (PDCD4) in colorectal cancer [15] and Sprouty2 (SPRY2) in cardiocytes [16]. Notably, transgenic mice expressing active PTEN specifically in mammary epithelium exhibited a decrease in the number of alveolar epithelial cells [17]. Moreover, it has been suggested that low levels of the functional PDCD4 protein may be essential for cell proliferation, and that accumulation of the protein in nuclei may negatively regulate cell proliferation [18,19]. SPRY2 has been described as a negative regulator of branching morphogenesis [16]. A recent study demonstrated that IL-6 through STAT3 activates the miR-17/92 cluster [20] and miR21 [21]. In contrast, expression of the miR-21 gene is suppressed by STAT3 in glioma cells [7]. Here we asked whether the STAT3 target gene miR-21 is also recognized and controlled by STAT5.

Mice
Animals were handled and housed in accordance with the guidelines of NIH and all experiments were approved by the Animal Care and Use Committee of The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (Permit: ASP K089-LGP-11). All efforts were made to minimize suffering. Generation of miR-21 2/2 mice was previously described [22]. Mice with a deletion of the Stat5 locus in mammary epithelial cells have previously been reported by Yamaji et al [2].

MicroRNA extraction and analysis
miRNA was extracted with miRNeasy system for purifying total RNA (Qiagen). Expression analysis was performed by using TaqMan probes (Applied Biosystems; Foster City, CA) for realtime PCR. cDNA was synthesized from total RNA using genespecific primers according to the TaqMan MicroRNA Assay protocol (Applied Biosystems; Foster City, CA). Reverse transcriptase reactions contained 10 ng of RNA samples, 50 nM stem-loop RT primer, 16 RT buffer, 0.25 mM each of dNTPs, 3.33 U/ml MultiScribe reverse transcriptase and 0.25 U/ml RNase inhibitor (cDNA Archive kit from Applied Biosystems). The 7.5 ml reactions were incubated for 30 min at 16uC, 30 min at 42uC, 5 min at 85uC and then held at 4uC. Real-time PCR was performed using an Applied Biosystems 7300 Sequence Detection system. The 20 ml PCR included 0.67 ml RT product, 16 TaqMan Universal PCR master mix and 1 ml of primers and probe mix of the TaqMan MicroRNA Assay protocol (Applied Biosystems; Foster City, CA). The reactions were incubated in a 96-well optical plate at 95uC for 10 min, followed by 40 cycles of 95uC for 15 sec and 60u for 10 min. The CT values were determined using default threshold settings. The threshold cycle (CT) is defined as the fractional cycle number at which the fluorescence passes the fixed threshold.

RNA extraction and analysis
Total RNA was extracted with the miRNeasy system (Qiagen). Expression analysis was performed using TaqMan probes (Applied Biosystems; Foster City, CA) for real-time PCR. Real-time PCR was carried out on an ABI Prism 7900HT (Applied Biosystems). Individual PCRs were performed in triplicate on samples using the housekeeping gene mouse beta-actin for normalization and experimental probes to obtain average threshold cycle (CT) values. The threshold cycle is defined as the fractional cycle number at which the fluorescence passes the fixed threshold. Average beta-actin CT values were subtracted from experimental CT values to obtain delta CT values. Delta CT values were then calculated by subtracting experimental sample delta CT values from the control sample delta CT. Relative gene expression was calculated using 2 delta CT.

HC-11 cell culture
Mouse mammary HC-11 epithelial cells were cultured in DMEM/F-12 medium supplemented with 5% FBS, 5 mg/ml insulin, 100 ng/ml hydrocortisone, 100 U/ml penicillin and streptomycin. Cells were plated on 100-mm dishes at a density of 1.0610 6 cells and grown at 37uC for approximately 24 h to reach 60-70% confluence. Cells were then cultured for 24 h in DMEM/F-12 medium supplemented with 5 mg/ml insulin and 100 U/ml penicillin. After 24 h prolactin (0.5 m mg/ml; source?) was added to the cultured medium. After 0.5, 1, 2 and 4 h the cells were harvested and total RNA was extracted.

RNA-Seq analysis
RNA was prepared for sequencing using the TruSeq RNA Sample Prep Kit, Set A (Illumina catalog # FC-122-1001) according to manufacturer's recommendations. Each sample was prepared from 1 mg total RNA. cDNA was synthesized using random hexamer primers and SuperScript II (Invitrogen). The resulting cDNA libraries were quantified using the Quanti-Fluor TM -ST system and checked for quality and size using the Agilent 2100 Bioanalyzer (High Sensitivity DNA kit, Agilent catalog # 5067-4626). A portion of each library was diluted to 10 nM and stored at 220uC. 2 ml of the 10 nM libraries were diluted and denatured, according to instructions from Illumina. The samples were loaded onto the cBot for clustering on a flow cell according to manufacturer's instructions (Illumina Inc., San Diego, CA) and sequenced using HiSeq 2000 (Illumina). The single-end sequenced tags from biological replicates of both wildtype and miR-21 2/2 samples were aligned to the mouse reference genome (mm9 assembly) using the TopHat program [23]. Transcript abundances were estimated by means of FPKM (fragment per kilobase of exon per million fragments mapped) and differentially expressed genes (DEGs) were initially identified by using the Cufflinks program [24]. To remove potential falsepositives, we only regarded the identified genes showing more than 5 FPKM in either miR-21 +/+ or miR-21 2/2 samples as reliable DEGs. All statistical analyses were performed using a two-tailed unpaired Student's t test. A p-value of 0.05 or less was considered significant. The RNA-seq data are deposited in GEO under accession number (GSE50068).

Transplantation of mammary epithelia
Transplantations were performed as previously described [4]. In brief, athymic nude mice (3-wk-old) were anesthetized and the proximal part of the inguinal gland containing the mammary epithelium was excised. Small pieces of mammary tissue from mature virgin female miR-21 +/+ or miR-21 2/2 mice were transplanted into the cleared fat pads. To assess the complete removal of host epithelia, the excised endogenous glands were processed for whole mount staining. Six weeks after transplantation, fat pads were harvested from virgin hosts and processed for whole mount staining in carmine alum.
Locked nucleic acid (LNA)-modified-in situ hybridization for miR-21 Locked nucleic acid (LNA)-modified, digoxigenin (DIG)-labeled probes complementary to mouse mature miRNAs were obtained from Exiqon, Inc (Woburn, MA). Sections of Stat5a/b fl/fl mammary tissue from 6 days pregnant mice were deparaffinized and treated with proteinase K for 20 min (10 mg/ml). After washing in PBS, sections were re-fixed in 4% paraformaldehyde for 10 min, washed twice in PBS, and prehybridized for 1 h in hybridization buffer (Roche, Mannheim, Germany). Tissues were hybridized overnight in the presence of 0.4 pmol/ml probe at 50uC. Slides were washed twice in 26 SSC at 37uC and finally washed in a high stringency 50% formamide-26 SSC at the hybridization temperature. Immunological detection was carried out using anti-DIG Fab conjugated to alkaline phosphatase (Roche, Mannheim, Germany) according to the manufacturer's protocol.

Statistics
All statistical analyses were performed using a two-tailed unpaired Student's t test. A p-value of 0.05 or less was considered significant.

Mir-21 expression in mammary epithelium
MiR-21 promotes cell transformation [26] and proliferation [27,28]. We chose to analyze miR-21 levels in mammary tissue from day 6 of pregnancy as this stage is characterized by extensive cell proliferation [29]. Compared to virgin tissue, miR-21 levels in control mammary tissue at pregnancy day 6 had increased ,1.7 fold (Figure 2b). In contrast, miR-21 levels did not change in tissue from Stat5 fl/fl;MMTV-Cre mice, in which STAT5 was deleted in the entire mammary epithelial compartment (Figure 2a). Increased expression of beta-casein (Csn2) served as a control for proper STAT5 activation. While Csn2 mRNA levels where elevated ,4.5 fold in tissue from pregnant Stat5 fl/fl mice (Figure 1d) they did not change in tissue from Stat5 fl/flMMTV-Cre mice (Figure 2c). In order to validate miR-21 is expression we used in situ hybridization with locked nucleic acid (LNA) nucleotides, which detects mature miRNAs in tissue sections. Mammary tissues from day 6 pregnant Stat5 fl/fl mice were used to perform the hybridization with miR-21, U6 and scrambled double digoxigenin labeled LAN oligonucleotides. The results demonstrated that mir-21 is preferentially expressed in the epithelial compartment (Figure 2e).

STAT5a binds to the miR-21 gene promoter
The miR-21 gene promoter is located in an intron of the TMEM49 gene and contains two bona fide STAT3 binding sites at approximately 800 bp upstream of the transcriptional start site [8,11]. Since STAT3 and STAT5 recognize the same GAS motif (TTCnnnGAA) this site likely is also recognized by STAT5. Furthermore, reanalysis of previous genome-wide STAT5A and H3K4me3 ChIP-seq datasets (GSE40930) performed in mammary gland at parturition, clearly showed that STAT5A binds to the miR-21 promoter and aligns with the presence of H3K4me3 [30] ( Figure 3a). In order to confirm the STAT5A binding to the miR-21 gene promoter in mammary tissue, primary mammary epithelial cells were stimulated with prolactin for 1 h followed by ChIP analysis. Real-time PCR analysis of DNA extracted after immunoprecipitation with anti-STAT5A antibodies, confirmed a 2.5 fold increased binding of STAT5A to the GAS motif in the miPPR21 compared to the untreated sample (Figure 3b). Binding to GAS sites in the promoter of the Igf1 and Socs2 genes, which both are STAT5 target genes, was enhanced by ,2.5 fold and ,5 fold, respectively (Figure 3c).

Assessing the role of miR-21 during mammary gland development
Next we investigated mammary development in miR-21 2/2 mice from puberty to lactation. MiR-21 2/2 females were able to nurse their pups until weaning. Mammary whole mount analyses from 8 week-old virgins and dams at their first day of lactation confirmed normal mammary alveolar structures (Figure 4 a-d). Histological analysis of miR-21 2/2 mammary epithelium at parturition displayed all features of secretory epithelium (Figure 4 e-f). Notably, fewer secondary ductal branches were seen in 8 week-old miR-21 2/2 mice (Figure 4 a-b). In order to examine epithelial cell autonomy, miR-21 2/2 mammary epithelium was transplanted into control mice. When harvested 6 weeks after transplantation, the miR-21 2/2 transplants did not exhibit any defects in the secondary branching pattern suggesting that systemic defects are responsible for this delay ( Figure 5).

Discussion
Development of mammary tissue during pregnancy is fully dependent on prolactin and its downstream transcriptional executor STAT5, which activates a differentiation program culminating in lactation [3,32]. Here we identify miR-21 as a STAT5 target gene both in a mammary epithelial cell line (HC-11) and in the developing mammary gland. We also show STAT5 binding to the miR-21 promoter (miPPR-21) in primary mammary cells. The ability of STAT5 to bind to the miPPR-21 gene and   . Mammary gland whole mounts and H&E analysis. Mammary glands from 8 weeks old virgin mice revealed fewer secondary ductal branching (marked with white arrows) in miR-21 2/2 (a) compared to miR-21 +/+ mice (b). No differences were detected in mammary glands harvested at the first day of lactation from miR-21 2/2 (c) and miR-21 +/+ dams (d). No differences were detected in the H&E staining in mammary gland harvested at the first day of lactation from miR-21 2/2 (e) and miR-21 +/+ animals (f). White arrows indicate lumina. Scale bar indicates 2 mm in a, b, c and d, 100 mm in e and f. doi:10.1371/journal.pone.0085123.g004 activate its transcription was also reported for Jurkat cells, an immortalized line of T lymphocytes [33]. Based on our study and other reports [8], it is likely that the miPPR-21 gene is recognized by either STAT3 or STAT5 and thus can be activated by several distinct cytokines, which even might initiate opposing biological programs. For example, prolactin activates STAT5 to control proliferation, differentiation and survival [2], while STAT3 is activated by IL-6 and LIF to control cell death during involution [34]. If STAT3 as well as STAT5 activate miR-21 gene expression during diametrically opposed developmental programs (survival versus death) it can be predicted that miR-21 serves as a common STAT cytokine-induced microRNA that targets different genes at different developmental stages during mammary gland development. Although the miR-21 gene is under STAT5 control, pregnancy mediated mammary development of miR-21 deficient mice (miR-21 2/2 ) was normal and these mice delivered and lactated normally and no abnormalities were detected in pup growth.
While attenuated secondary mammary ductal branching was observed in miR-21 2/2 mice during puberty, miR-21 2/2 epithelium transplanted into fat pads of wild type mice developed normally. It is therefore likely that either suboptimal ovarian function or impaired signaling from the stroma is responsible for impaired ductal branching. Analysis of RNA-seq data which was performed on 8-10 week-old miR-21 2/2 and miR-21 +/+ mice did not exhibit any differences in well-known miR-21 target gens such as Spry2, Pdcd4 and Pten (Table 1, GSE50068). Furthermore, the distribution of mammary epithelial cell lineages (luminal, myoepithelial and stem cells) did not show any difference between the miR-21 2/2 mammary and control tissue (data not shown).
In summary, STAT5 controls different aspects of mammary development, from establishment and maintenance of alveolar progenitors during puberty to alveolar differentiation during pregnancy. While some of the developmental steps are controlled by STAT5-dependent protein-encoding genes, others might be under the influence of microRNAs. However, miR-21, which is regulated by STAT5, is not involved in the regulation of the mammary gland development or lactation.