Dysregulation of Uterine Signaling Pathways in Progesterone Receptor-Cre Knockout of Dicer

Abstract

Epithelial-stromal interactions in the uterus are required for normal uterine functions such as pregnancy, and multiple signaling pathways are essential for this process. Although Dicer and microRNA (miRNA) have been implicated in several reproductive processes, the specific roles of Dicer and miRNA in uterine development are not known. To address the roles of miRNA in the regulation of key uterine pathways, we generated a conditional knockout of Dicer in the postnatal uterine epithelium and stroma using progesterone receptor-Cre. These Dicer conditional knockout females are sterile with small uteri, which demonstrate significant defects, including absence of glandular epithelium and enhanced stromal apoptosis, beginning at approximately postnatal d 15, with coincident expression of Cre and deletion of Dicer. Specific miRNA (miR-181c, −200b, −101, let-7d) were down-regulated and corresponding predicted proapoptotic target genes (Bcl2l11, Aldh1a3) were up-regulated, reflecting the apoptotic phenomenon. Although these mice had normal serum hormone levels, critical uterine signaling pathways, including progesterone-responsive genes, Indian hedgehog signaling, and the Wnt/β-catenin canonical pathway, were dysregulated at the mRNA level. Importantly, uterine stromal cell proliferation in response to progesterone was absent, whereas uterine epithelial cell proliferation in response to estradiol was maintained in adult uteri. These data implicate Dicer and appropriate miRNA expression as essential players in the regulation of multiple uterine signaling pathways required for uterine development and appropriate function.

Dicer, a ribonuclease III essential for microRNA (miRNA) synthesis, is responsible for the cytoplasmic cleavage of the miRNA stem-loop precursor to form the double-stranded mature approximately 22-nucleotide miRNA and its corresponding star form (1, 2). miRNA have been shown to function as repressors of gene expression either at the level of mRNA translation or mRNA transcript degradation (2, 3). Dicer is essential for mammalian development because deletion of Dicer through traditional knockout mouse technology causes embryonic lethality in early development secondary to a defect in formation of embryonic stem cells (4). To overcome this early defect and study the global roles of miRNA in mammals, multiple groups have created Dicer floxed alleles (5, 6) and generated Dicer conditional knockout (cKO) mice. Using these mice, conditional ablation of Dicer has shed light on the role of Dicer in reproduction. For example, in male reproduction, Dicer has been shown to play roles in spermatogenesis, retrotransposon activity, and primordial germ cell proliferation (7). Our group and others have also demonstrated that Dicer plays an important role in the reproductive tract of female mammals, including women (8–17). Additionally, miRNA profiles have identified abnormally regulated miRNA in ovarian cancers and cell lines (18–23) and benign and malignant uterine diseases (2, 24–31). Although no particular miRNA has been deleted in the female reproductive tract, other studies have used zona pellucida 3 (Zp3)-Cre to delete Dicer in the oocyte (13–15) and anti-Müllerian hormone receptor 2 (Amhr2)-Cre to ablate Dicer in ovarian granulosa cells, muscle layers of the oviduct and uterus, and partially from the stromal cells of the uterus (8, 10, 12, 16). These Dicer-deleted mice are sterile, although by different mechanisms. Conditional deletion of Dicer with Zp3-Cre in the oocyte demonstrated roles of small interfering RNA (siRNA) and not miRNA, resulting in meiotic defects due to defective spindle formation (13, 14). Conditional deletion of Dicer with Amhr2-Cre by three independent groups led to decreased ovulation rates and oviductal diverticuli, the latter preventing embryos from reaching the uterus (8, 10, 12, 16). Although deletion of Dicer from the mesenchymal cells of the uterus (stroma and smooth muscle) resulted in shorter uteri, these mice have a functional uterine decidual response (8) but lack functional implantation (12). The lack of a role of Dicer in the uterus of these models may have been the result of inefficient ablation of the conditional Dicer allele using this specific Cre model. A more efficient uterine expressed Cre model may shed light on the role of Dicer in uterine biology. To accomplish this, we used the progesterone receptor (PR)-Cre mouse model (32).

The PR-Cre mouse model has Cre recombinase inserted into the PR gene. Cre is expressed in cells which express PR. This includes all compartments of the uterus, in the ovarian granulosa cells just before ovulation, in the anterior pituitary and mammary gland (32, 33). In the uterus, Cre recombinase is expressed in cells beginning at 2 wk postnatally in the luminal and glandular epithelial cells of the uterus and the surrounding stroma with little expression in myometrium. As the mice age, PR expression increases in subepithelial stroma and myometrium and tall columnar cells of the oviduct. The PR-Cre mice have proven useful for studying female reproduction. The PR-Cre line has been used to study the Indian hedgehog signaling pathway (34, 35), aberrant phosphatase and tensin homolog/K-ras signaling in endometrial cancer (36), β-catenin signaling (37), TGFβ signaling (38), steroid receptor coactivators (39) in the uterus, and peroxisomal proliferator-activated receptor-γ signaling in the ovary (40). Using the PR-Cre mice, we show that deletion of Dicer in the uterus leads to severe uterine defects and sterility. Using next-generation sequencing and other molecular approaches, we demonstrated that postnatal deletion of Dicer in the uterus leads to reciprocal decreases in the miRNA expression and increased expression of the genes important in this phenotype and for uterine function, ultimately leading to dysfunctional signaling pathways.

Results

Uterine Dicer cKO mice are sterile

To examine the global roles of miRNA in the stroma and epithelium of the postnatal uterus, a uterine Dicer cKO mouse was created by crossing Dicer^flox/flox mice (5) with PR-Cre mice (32). Dicer^flox/flox mice (5) have exon 23, containing one of the two ribonuclease III domains of Dicer, flanked by LoxP sites. Previous studies have shown that this strategy deletes Dicer function (5, 8).

Dicer cKO mice (PR^cre/+;Dicer^flox/flox) and control mice (Dicer^flox/flox) were born healthy. To confirm tissue-specific recombination and ablation of Dicer expression, we performed immunofluorescent staining for Dicer on control and Dicer cKO uteri. Dicer is significantly expressed in the cytoplasm of postnatal d 15 (P) 15 uteri in control mice (Fig. 1, A and B), with highest expression in the epithelium and stroma. Dicer cKO uteri have no detectable expression of Dicer in the luminal epithelium and significantly less expression in the uterine stroma (Fig. 1, C and D). Previously published work showed PR expression, and thus PR-Cre function, present mainly in the luminal and glandular epithelium at P14 (32, 33). Because we had a significant loss of Dicer expression in the subepithelial stroma at P15, we examined the expression of PR specifically at P15. PR expression at P15 in the control uteri shows high expression in the nucleus of the luminal and glandular epithelium and lower but still consistent expression in the nucleus of the surrounding stroma (Fig. 1, E and F), in agreement with published results (32, 33). At P15, analysis of the Dicer cKO uteri demonstrates that PR is not detectable in the luminal epithelium and has decreased expression in the uterine stroma (Fig. 1, E and F).

Fig. 1.

Dicer is deleted in the uteri of Dicer PR-Cre cKO mice. Comparison of immunofluorescence staining for Dicer in P15 control at ×25 (A) and control at ×40 (B) and Dicer cKO at ×25 (C) and Dicer cKO at ×40 (D). Merged images of Dicer (green) and PI (red) are shown. Dicer expression is present in luminal epithelium in the cytoplasm of control (A and B). Dicer protein is absent from the luminal epithelium and stroma of the Dicer cKO (C and D) uterus. Progesterone receptor staining (PR) in strong in the luminal and glandular epithelium in P15 control uteri (E). However, PR staining is negative in the luminal epithelium of the Dicer cKO uteri (F). Merged images of PR (green) and PI (red) are shown. L, Lumen; PI, propidium iodide staining.

Dicer cKO and control adult female mice (6 wk of age) were individually placed with wild-type males of proven fertility and allowed to mate naturally. Control mice had normal fertility, giving 453 pups in 53 litters over 6 months (mean 8.5 pups/litter ± 0.6; mean 0.98 litters/month ± 0.06). In contrast, Dicer cKO mice did not deliver any pups over this 6-month period and are considered sterile (P = 1.1 × 10⁻⁹⁰ by χ² analysis). Thus, Dicer expression in the uterine epithelium and/or stroma is essential for uterine function and fertility in mice.

Dicer cKO mice have small uteri with severe histological defects

The PR-Cre mice target Dicer deletion in the uterus and because the reproductive tract is important for embryo quality and survival (8), the infertility is most likely due to the uterine defects. However, in addition to the reproductive tract, the PR-Cre also ablates Dicer in the pituitary and the corpus luteum of the ovary (32). Because problems with each of these components of the reproductive axis could contribute to the infertility of the Dicer cKO mice, we undertook a systematic approach to examine each.

To examine gonadotropin hormone function of the pituitary in these mice, we examined serum hormone levels of adult control and Dicer cKO mice. FSH and LH serum levels from unstimulated adult female mice (6–8 wk) were examined in randomly cycling mice. No significant differences were found in FSH levels between control (9.74 ± 2.3 ng/ml, n = 11) and Dicer cKO (5.04 ± 0.62 ng/ml, n = 10, P = 0.08) mice or LH levels in control (0.28 ± 0.10 ng/ml, n = 11) or Dicer cKO (0.32 ± 0.14 ng/ml, n = 8, P = 0.82) mice. Thus, in early adult mice, gonadotropin hormone levels are not affected by Dicer deletion. Histological examination of pituitary glands revealed normal histology in both control and Dicer cKO females (data not shown).

Next, to evaluate steroid hormone production, serum estradiol levels from unstimulated adult mice were examined in randomly cycling mice. No statistically significant differences were found between control (28.0 ± 7.3 pg/ml, n = 7) and Dicer cKO (16.4 ± 1.5 pg/ml, n = 9, P = 0.15) mice. Thus, estradiol levels are not affected by Dicer deletion. When we examined serum progesterone levels of stimulated pregnant mice 4.5 d after conception, we found that progesterone levels produced by Dicer cKO females were similar to those of control females, indicating normal luteal function (43.90 ± 7.12 ng/ml, n = 4) and Dicer cKO (47.61 ± 13.44 ng/ml, n = 4, P = 0.40) mice. Additionally, both control and Dicer cKO mice had histologically normal corpora lutea (data not shown). Therefore, corpus luteum function should support pregnancy. Because ovaries of adult and prepubertal Dicer cKO mice appeared grossly and histologically normal (data not shown), we next examined their ability to respond to gonadotropins and produce fertilizable oocytes. Four-week-old control (n = 6) and Dicer cKO (n = 3) female mice were pharmacologically superovulated with pregnant mare serum gonadotropin and human chorionic gonadotropin, and mated with wild-type males of proven fertility. The total number of embryos obtained from the oviducts of control (41 ± 5.6) and Dicer cKO (56.7 ± 15.3) mice was similar (P = 0.27). Although the percentage of embryos that progressed to the two-cell stage in overnight cultures was lower for embryos derived from Dicer cKO females (control: 75.2%; Dicer cKO: 54.2%), the differences were not statistically significant (P = 0.10). Additionally, ovaries and oviducts from unstimulated Dicer cKO mice showed no obvious histological defects (data not shown).

Because Dicer cKO mice had normal serum gonadotropin levels, steroid hormone levels, oocyte ovulation and fertilization rates, and embryo development, a uterine phenotype was suspected to be the primary cause of the infertility. To explore this possibility, uteri from unstimulated mice at various postnatal ages were examined grossly and histologically. Mice and uteri were weighed. Grossly, uteri from adult Dicer cKO mice were significantly smaller than control uteri (Fig. 2A). Comparison of uterine weight to body weight at prepubertal and adult ages showed a progressive decrease in the size of the Dicer cKO uteri compared with control uteri (Fig. 2B). Body weights were not significantly different between control and Dicer cKO mice (data not shown). Histological changes in Dicer cKO uteri were also progressive. Grossly, uteri from P15 Dicer cKO mice looked similar to control uteri. However, histological examination of uteri from P15 mice showed a decrease in glandular epithelium (Fig. 3, C and D) compared with control (Fig. 3, A and B) but normal myometrium, columnar luminal epithelium, and stroma at this age. Gene expression analysis by real-time quantitative PCR (qPCR) of Forkhead box A2 (Foxa2), a protein expressed specifically in the glandular epithelium (41) (Fig. 3I), showed a 12.5-fold decrease in Foxa2 expression in the Dicer cKO uteri, confirming the reduction in glandular epithelium. Histological changes were significantly more striking in uteri from adult Dicer cKO mice. By 8 wk postnatally, the Dicer cKO uteri lacked glandular epithelium and showed a drastic decrease in uterine stroma, with significant preservation of luminal epithelial architecture (Fig. 3, G and H) compared with control uteri (Fig. 3, E and F). Additionally, the uterine myometrium appeared intact in adult Dicer cKO uteri, consistent with lack of expression of PR-cre at this site. Overall, the results indicate that the infertility phenotype was most likely due to a severe, progressive postnatal uterine defect, which was most apparent in adult mice. However, the histological phenotype begins about P15 with a decrease in glandular epithelium. Moreover, the drastic decrease in uterine size in adult mice was most likely due to the significant decrease in uterine stroma, although the Dicer cKO uteri also have a significant loss of glandular epithelium but not luminal epithelium.

Fig. 2.

Dicer cKO uteri are significantly smaller as adults. Postnatal mice were weighed and dissected, and uteri were examined grossly (A). Uteri were weighed (B) and normalized to body weight. *, P < 0.001, Student's t test, sem.

Fig. 3.

Uterine histology of Dicer cKO mice is severely altered and age dependent. Uteri from P15 (A and B) and 8-wk control mice (E and F) display normal histology. Uteri from P15 Dicer cKO uteri (C and D) have significant loss of glandular epithelium. Uteri from 8-wk-old mice (G and H) have disorganized and decreased uterine stroma. Uteri from P15 mice have severely decreased levels of Foxa2 (E) gene expression by qPCR, consistent with lack of glandular epithelium. *, P < 0.01, Student's t test, sem. M, Myometrium; S, stroma; LE, luminal epithelium; G, glandular epithelium; RQ, relative quantity.

Apoptosis is observed in the uterine stroma of the Dicer cKO mice

To begin to decipher the mechanisms behind the significant decrease in uterine stroma, uteri were examined for apoptosis using the terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay. Because the Dicer cKO adult mice had uteri with significant stromal decrease, we chose an earlier time point before obvious cell loss. Uteri from P15 mice were chosen because the uterine size and histology were similar between the control and Dicer cKO uteri. As described above, uterine stroma and luminal epithelium in the Dicer cKO appeared normal at P15 with only the loss of glandular epithelium. Uteri from P15 mice displayed significant apoptosis in the Dicer cKO stroma (Fig. 4, C and D) but not in the control (Fig. 4, A and B). Additionally, stromal apoptosis appeared to be in the stroma closest to the epithelium in which Dicer was initially deleted.

Fig. 4.

Uterine stroma of Dicer cKO uteri undergoes apoptosis early during uterine postnatal development. Apoptosis was detected by TUNEL assay in the subepithelial stroma of the Dicer cKO uteri at P15 (C and D) compared with little apoptosis in controls (A and B). Using a delayed implantation model (44), Dicer cKO uteri respond with appropriate luminal epithelium response to estrogen (F) compared with control (E). However, stromal proliferation as a progesterone response is absent in the Dicer cKO uteri (H) compared with control (G). Arrows, BrdU-stained cells.

Surprisingly, Dicer deletion leads to a significant decrease in PR in the luminal epithelium. PR staining in uteri from P15 control mice was strong in the luminal and glandular epithelium, consistent with published results (32) (Fig. 1, E and F). PR staining was nearly absent in the luminal epithelium in the Dicer cKO uteri (Fig. 1, G and H), suggesting that PR-positive epithelial cells have already been depleted from the Dicer deletion or that the Dicer deletion directly or indirectly causes PR expression to be decreased. Thus, Dicer deletion in luminal epithelium leads to stromal apoptosis, possibly due to a defective cellular signaling pathway.

Multiple human studies have suggested that loss of Dicer and/or Drosha is associated with endometrial cancer (42, 43). To examine the proliferative effects of Dicer deletion in the uterus, we performed uterine proliferation experiments as previously described (44). Generally, treatment with estradiol alone should cause proliferation as indicated by bromodeoxyuridine (BrdU) incorporation in the luminal epithelium, whereas treatment with progesterone should cause stromal proliferation (44). Dicer cKO mice treated with estradiol (Fig. 4G) responded similarly to controls (Fig. 4E), with BrdU staining representing highly proliferative cells detected in the nuclei of the luminal epithelium. However, Dicer cKO mice treated with both estradiol and progesterone showed no detectable stromal cell proliferation, consistent with a loss of progesterone signaling (Fig. 4H), compared with stromal BrdU staining in the control (Fig. 4F). Thus, Dicer cKO mice have a significant stromal cell defect responsible for their infertility.

Apoptosis and defective uterine signaling pathways arise from dysregulated miRNA and mRNA

Because Dicer is a key enzyme in miRNA biosynthesis (3), we hypothesized that deletion of Dicer would globally decrease miRNA levels in the uterus. Additionally, we hypothesized that changes in miRNA expression may play a role in defective cellular signaling pathways by affecting gene expression in the uterus. Therefore, we initiated miRNA and mRNA profiling to elucidate the molecular mechanisms behind this severe uterine phenotype. We chose uteri from P15 mice because most of the uterine cell types were still present in the Dicer cKO compared with the control uteri at that age, but Dicer deletion had occurred in the epithelium and stroma and apoptosis was occurring only in the uterine stroma. Pools of five uteri (n = 2) of the same genotype were used for RNA extraction to obtain sufficient RNA for miRNA sequencing using Illumina GA-IIx Next-Generation sequencing (Illumina, San Diego, CA). Due to the large amounts of high-quality input RNA required for these studies, we used a small number of genotyped, pooled samples for our discovery cohort and larger numbers of independent samples of individual mice for validation studies. Using our discovery cohort, we generated 7,367,734 and 5,255,918 small RNA sequences for the Dicer cKO uteri and 7,096,446 and 7,312,778 small RNA sequences for the control uteri. The proportion of small RNA sequences that mapped to known miRNA sequences (miRBase 16.0, see Ref. 84) was similar for Dicer cKO (81.5%) and control (83.7%). Of the 1040 known mouse miRNA from miRBase 16.0, examination of the entire control mouse uterus revealed 135 miRNA expressed in all components of the mouse uterus inclusive of luminal and glandular epithelium, stroma, and myometrium (Supplemental Table 1, published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org). This is significantly fewer than the 240 miRNA detected in human endometrium (21). Thus, there is an even more limited subset of miRNA expressed in the mouse uterus, even when myometrium is included, compared with human endometrium. Similar to the human endometrium (21), the let-7 family is the most abundant miRNA family, comprising approximately 51% of the reads. When we analyzed the differentially expressed miRNA in the Dicer cKO uteri, we found 37 miRNA that were down-regulated and four miRNA that were up-regulated using a fold-change cutoff of 1.2 (Table 1). Therefore, by deleting Dicer in a portion of the cells of the mouse uterus, we decreased specific miRNA levels.

Table 1.

Differentially expressed miRNA in P15 Dicer cKO uteri

Precursor	Fold change up
mmu-mir-1	1.40
mmu-mir-10b	1.26
mmu-mir-379	1.24
mmu-let-7g	1.21
Precursor	Fold change down
mmu-mir-29a	1.21
mmu-mir-30e	1.21
mmu-mir-192	1.24
mmu-mir-33	1.24
mmu-mir-199a-3p	1.26
mmu-mir-199b	1.26
mmu-mir-214	1.26
mmu-mir-106b	1.27
mmu-mir-196b	1.27
mmu-mir-669c	1.28
mmu-mir-152	1.29
mmu-mir-434–3p	1.30
mmu-mir-196a	1.31
mmu-mir-191	1.32
mmu-mir-185	1.32
mmu-mir-130a	1.32
mmu-mir-433	1.33
mmu-mir-10a	1.35
mmu-mir-181a	1.36
mmu-mir-181c	1.40
mmu-mir-495	1.41
mmu-let-7c	1.46
mmu-mir-181b	1.46
mmu-mir-181d	1.53
mmu-mir-200c	1.54
mmu-mir-203	1.55
mmu-let-7d	1.56
mmu-mir-31	1.57
mmu-mir-101b	1.58
mmu-mir-376a	1.62
mmu-mir-503	1.67
mmu-mir-872	1.94
mmu-mir-298	2.08
mmu-mir-676	2.71
mmu-mir-200a	2.82
mmu-mir-200b	3.88
mmu-mir-429	4.95

miRNA target specific mRNA sequences for repression through binding of the 5′ seed sequence (nucleotides 2–8) of the miRNA to the 3′ untranslated region (UTR) of target genes (21). miRNA are grouped into families based on seed sequence (45), and miRNA in the same family should target the same genes. Thus, our working hypothesis is that deletion of Dicer decreased the miRNA in the uterus, and consequently, miRNA-targeted genes were up-regulated and could be responsible for the infertility phenotype. To test this hypothesis, the same RNA discovery cohort samples underwent genome-wide gene expression analysis using microarray. Using our discovery cohort samples, analysis of differential gene expression between control and Dicer cKO uterine pools from P15 females showed 153 genes up-regulated and 282 genes down-regulated more than 1.5-fold (P < 0.05) (Supplemental Table 2). Thus, Dicer deletion in the uterus causes a misregulation of a fraction of both miRNA and mRNA in the uterus.

Examination of up-regulated genes using DAVID Bioinformatical Analysis (85, 86) showed several genes in the positive regulation of apoptosis category such as Bcl2l11, Skp2, Aldh1a3, Cdkn1a, and Mal. Increased expression of Bcl2l11 and Aldh1a3 were validated using independent samples and qPCR (Fig. 5). Aldehyde dehydrogenase family 1, subfamily A3 (Aldh1a3), was increased 18-fold in Dicer cKO uteri compared with control (Fig. 5A). BCL2-like 11 (apoptosis facilitator), Bcl2l11, was significantly increased in Dicer cKO uteri at P15 compared with control (Fig. 5B). To identify potential functional miRNA-mRNA pairs, we used our previously published SigTerms software (46) with three online publically available target prediction algorithms, TargetScan 5.1 (www.targetscan.org), miRanda (www.miRNA.org), and PicTar (www.pictar.mdc-berlin.de/). Examination of up-regulated genes that were predicted to be targeted by down-regulated miRNA in all three algorithms revealed that 23 up-regulated genes were predicted targets of 16 down-regulated miRNA (Supplemental Table 3). Interestingly, Bcl2l11, a proapoptotic gene (also known as BIM), was the predicted target of the largest number of down-regulated miRNA, although other genes in the positive regulation of apoptosis category were predicted miRNA targets (Table 2). Specifically, members of the miR-101 and miR-181 family were predicted to target Bcl2l11. Knockout mouse models of Bcl2l11 showed that T lymphocytes do not undergo apoptosis in response to appropriate stimuli, and BCL2L11 appears to function by inactivating survival proteins (47, 48). Thus, BCL2L11 is important in facilitating apoptosis, and we believe that increased expression of Bcl2l11 and increased apoptosis is in part due to the decrease in specific miRNA that target Bcl2l11, miR-101b, and miR-181c. Importantly, members of the miR-181 family directly target members of the BCL2 gene family, including BCL2L11 in human astrocytes (49). Thus, the down-regulation of these specific miRNA may lead to the subsequent up-regulation of Bcl2l11 and apoptosis, resulting in the subsequent severe uterine phenotype in the Dicer cKO mice. Decreased expression of miRNA expression was validated in independent samples (Fig. 5). The Dicer cKO uteri at P15 confirmed a 1.6-fold decrease in miR-101b expression, a 9-fold decrease in miR-200b, and a 1.2-fold decrease in let-7d expression by qPCR. Although not statistically significant, miR-181c and miR-503 showed a similar trend toward a the decreased expression by qPCR compared with next-generation sequencing (Fig. 5C).

Fig. 5.

Altered mRNA and miRNA expression in Dicer cKO uteri. Dicer cKO uteri at P15 showed significant up-regulation of Aldh1a3 (A) and Bcl2l11 (B), both positive regulators of apoptosis. miRNA that target either or both Aldh1a2 or Bcl2l11 or other dysregulated genes involved in positive regulation of apoptosis (Table 2) were confirmed to be differentially expressed by TaqMan miRNA assays (C). Dicer cKO uteri at P15 also show significant changes in gene expression of signaling and cell cycle pathway genes (Acvr2b, Ccnd1, Wnt4, and Wnt7a) (D–G) and tight junction signaling pathway genes (Cldn3, Cldn23) (H and I). Acute progesterone response was molecularly defective in Dicer cKO uteri as evidenced by lack of Cyp26a1 and Ihh induction with progesterone treatment (J and K). *, P < 0.01, Student's t test, sem.

Table 2.

Wnt/β-catenin signaling genes dysregulated in Dicer cKO

Symbol	Entrez gene name	Fold change
Acvr2b	Activin A receptor, type IIB	1.640
Ccnd1	Cyclin D1	1.576
Fzd10	Frizzled homolog 10 (Drosophila)	1.784
Ppp2rb	Protein phosphatase 2 (formerly 2A), regulatory subunit B, β-isoform	−3.717
Rac3	Ras-related C3 botulinum toxin substrate 3 (ρ-family, small GTP binding protein Rac3)	−1.905
Sox17	SRY (sex determining region Y)-box 17	−1.876
Sox5	SRY (sex determining region Y)-box 5	−1.609
Wnt4	Wingless-type MMTV integration site family, member 4	2.393
Wnt10a	Wingless-type MMTV integration site family, member 10A	1.662
Wnt7a	Wingless-type MMTV integration site family, member 7A	1.796
Wnt7b	Wingless-type MMTV integration site family, member 7B	1.943

Little is known about the specific localization of the above miRNA in the mouse uterus. The miR-200 family is highly expressed in the mouse uterine myometrium and is very important for parturition (50, 51). Additionally, the miR-200 family is thought to be highly expressed in human endometrial epithelium with loss of miR-200 leading to epithelial-to-mesenchymal transition and endometrial cancer (52–56). Additionally, miR-200 expression changes with the phase of the human menstrual cycle (57). The let-7 family is expressed in both epithelium and stroma, and its expression changes with mouse embryo implantation (58). To begin to examine potential localization of these miRNA, we started with P21 uteri from wild-type mice. Using P21 wild-type mouse uteri, we were able to separate the luminal epithelium from the stromal epithelium and examined the expression of these particular miRNA in the various fractions. As expected, miR-200b was most highly expressed in the epithelial cell compartment. Let-7d was equally expressed in both epithelial and stromal cell compartments. Although miR-101b and miR-181c were more highly expressed in the epithelial cell compartment, they were also expressed in the stroma (Supplemental Fig. 1). Because P15 Dicer cKO uteri had normal amounts of uterine stroma and luminal epithelium and only loss of glandular epithelium, we do not believe that the changes in our miRNA expression by next-generation sequencing or by qPCR are due to loss of a particular cell type.

Because Dicer cKO uteri had decreased miRNA and both miRNA and mRNA were misregulated, we hypothesized that Dicer cKO uteri would have dysregulated signaling pathways. Therefore, we performed pathway analysis using gene expression profiling data from our pools of P15 Dicer cKO and control uteri. Using Ingenuity Pathway Analysis (Ingenuity Systems, Inc., Redwood, CA), we discovered several important biological functions were dysregulated in the Dicer cKO uteri. Not surprisingly, reproductive system disease, cell death, and cell-to-cell signaling and interaction were in the top five biological functions dysregulated in the Dicer cKO uteri. Additionally, we found dysregulation in the Wnt/β-catenin (genes are listed in Table 2) and tight junction canonical signaling pathways. We used qPCR to validate specific genes involved in these canonical signaling pathways (Fig. 5, D–G). Expression of the activin type 2B receptor (Acvr2b) showed a 2-fold increase in the Dicer cKO uteri at P15 (Fig. 5D). Cyclin D1 (Ccnd1) expression was significantly increased by 1.5-fold (Fig. 5E). Two members of the wingless-related mouse mammary tumor virus (MMTV) integrated site family, Wnt4 and Wnt7a, were both significantly increased in the Dicer cKO uteri by 5- and 2-fold, respectively (Fig. 5, F and G). In the tight junction canonical signaling pathway genes, Cldn3 and Cldn23 were significantly down-regulated by 2- and 4-fold, respectively (Fig. 5, H and I). The increase in Wnt/ß-catenin signaling is likely to be a compensatory mechanism for loss of glandular epithelium and not a direct effect of miRNA dysregulation through loss of Dicer. Ablation of Dicer using Amhr2-Cre also showed increased expression of members of the WNT family (8, 12).

To examine the effect of miR-101 and miR-181 on endogenous gene expression, we used a human in vitro system as previously described (21) because no benign mouse endometrial cell line is available. Previously published work has shown that the miR-181 family directly targets BCL2, including BCL2L11, in primary astrocytes (49). Using primary cultures of human endometrial stromal fibroblasts (HESF), increased expression of miR-101 using miRNA mimics resulted in a significant decrease in gene expression of histone methyltransferase enhancer of zeste homolog 2 (EZH2), a validated direct target of miR-101 in HeLa cells (45), gastric cancer cells (59), transitional cell carcinoma (60), and prostate cancer (61) (Fig. 6). Importantly, increased expression of miR-101 decreased endogenous expression of BCL2L11 in HESF (Fig. 6). Similar repression of gene expression of BCL2L11 was observed with increased expression of miR-181 through miRNA mimics in HESF (Fig. 6). Together these data support the hypothesis that Bcl2l11 is a direct target of miR-101 and miR-181. Thus, loss of miR-101, leading to increased Bcl2l11 gene expression, fits with our proapoptotic phenotype in P15 uteri in Dicer cKO mice.

Fig. 6.

miR-101 and miR-181 repress endogenous gene expression of BCL2L11 in human endometrial stromal fibroblasts. Independent cultures of human endometrial stromal fibroblasts were transfected with miR-101, miR-181, or negative control no. 1 (NT). Repression of endogenous gene expression of BCL2L11 occurred with either miR-101 or miR-181 (A). Repression of EZH2 occurred with miR-101 (B). *, P < 0.05, n = 3, Student's t test, sem.

To further expand on important signaling pathways in the mouse uterus, we looked at the ability of the Dicer cKO uteri to respond to progesterone. Previous work has described the acute and chronic progesterone-responsive genes in wild-type mice at 8–10 wk of age. In previously published work, mice were starved of steroid hormones by ovariectomy at 8 wk followed by addition of steroid hormones and then treated either acutely or chronically with progesterone (P₄) (62). Due to the severe histological abnormalities in the 8-wk-old Dicer cKO uteri, we shortened the time course for our progesterone-responsive gene studies. Mice underwent ovariectomy at P21, were allowed to rest for 1 wk, and then were treated with 1 mg of P₄ or sesame oil for 6 h, uteri dissected, and RNA isolated from individual uteri. The progesterone response in the control mice showed a significant increase in expression of both cytochrome P450, family 26, subfamily A, polypeptide 1 (Cyp26a1) and Indian hedgehog (Ihh) with progesterone treatment similar to mice treated at adult ages (62). Importantly, the progesterone response on a molecular level in the Dicer cKO uteri was abnormal, as evidenced by the lack of an increase in Cyp26a1 and Ihh levels in response to acute progesterone treatment (Fig. 5, J and K). Because IHH is an important mediator of progesterone signaling in the mouse uterus (63), these results suggest that Dicer cKO uteri may have an abnormal response to progesterone in the acute phase and abnormal IHH signaling.

Discussion

Consistent with multiple papers highlighting the significant biological function in mammalian development of Dicer, a key enzyme in small RNA synthesis (21), we discovered that Dicer is essential for postnatal development of the mouse uterus and fertility. In our study herein, we found that Dicer cKO mice are infertile due to severe uterine abnormalities that stem from abnormal cellular and molecular signaling. Although multiple studies have shown the importance of Dicer in other organ systems, this is the first example showing the importance of Dicer specifically in the postnatal uterus.

Although our mouse model has a very simple infertility phenotype, deciphering the mechanism behind the phenotype is challenging due to the severity of the uterine defect. Postnatally the uterine stroma of this PR-Cre Dicer cKO mouse progressively undergoes apoptosis, disappears, and is unavailable for further study. Thus, Dicer is crucial for proper uterine postnatal development. Recent mammalian studies have suggested that Dicer knockdown leads to apoptosis in neural crest cells (64) and neuronal stem cells through an increase in proapoptosis proteins (65). Additionally, loss of Dicer in the hippocampus leads to increased apoptosis and significant embryonic and postnatal morphological defects (66). Furthermore, other studies suggest that Dicer is important for DNA fragmentation during apoptosis in Caenorhabditis elegans (67). In the ovary, deletion of Dicer leads to increased apoptosis and follicle atresia (8). Thus, deletion of Dicer in the uterus leading to apoptosis is a very consistent biological phenotype. This apoptosis is likely facilitated through abnormal up-regulation of expression of BCL2L11, a proapoptosis protein. Recent work in hematopoietic malignancies has shown that BCL2L11 is regulated by miR-32 (68). Studies in astrocytes have shown that the miR-181 family targets multiple genes in the BCL2 family, and there is direct functional evidence that miR-181c targets BCL2L11 in astrocytes (49). Although we did not find miR-32 dysregulated in our Dicer cKO uteri, we did discover dysregulation of miR-181 and other miRNA such as miR-101 that are predicted to target multiple positive regulatory apoptotic genes, including Bcl2l11. Our in vitro studies in human cell cultures suggest that both miR-101 and miR-181 target BCL2L11 (Fig. 6).

Consistent with other studies showing the importance of Dicer in the female reproductive tract (8–18), the PR-Cre Dicer cKO mice are infertile. A comparison of the PR-Cre Dicer cKO phenotype to the Amhr2-Cre Dicer cKO uterine phenotype (8, 10, 12) reveals significant differences. The key phenotypic difference is that the Amhr2-Cre Dicer cKO has significant oviductal diverticuli (8, 10, 12), leading to trapped embryos, whereas the PR-Cre Dicer cKO has a severe uterine developmental defect. Although Amhr2-Cre targets mesenchymal cells of the female reproductive tract, conflict exists regarding the exact cells targeted because two different reporter mice showed slightly different expression patterns (69, 70). In general, Amhr2-Cre should target Dicer deletion to the stroma and smooth muscle layers of the uterus (70). Although all three published reports on the Amhr2-Cre Dicer cKO phenotype showed infertility, oviductal cysts or diverticuli, and short uteri (8, 10, 12), only one showed a decrease in uterine glandular epithelium at P28 (12). The molecular mechanism in the uterus was not studied in any of these models (8, 10, 12). In addition, deletion using the Amhr2-Cre model occurs during fetal development, and many of the phenotypes may be a result of disrupted development. PR-Cre deletion is observed beginning in the luminal and glandular epithelium and subepithelial stroma at P15 (32), although it must result in earlier deletion in the glandular epithelium, which is absent by P15 in our Dicer cKO.

Abnormal Wnt/β-catenin signaling, especially up-regulation of Wnt4 and Wnt7a, was an interesting finding. Wnt4 cKO females are subfertile due to implantation defects. Additionally, uteri from Wnt4 cKO also have a loss of glandular epithelium, suggesting that WNT4 plays a critical role in glandular epithelium development (71). Alternately, expression of a stable form of β-catenin in a similar PR-Cre mouse model showed significant glandular hyperplasia and enlarged glands at P14. Although the conditional ablation of β-catenin showed significantly smaller uterus than control mice, the uterine epithelium displayed significant squamous metaplasia (37). Both of these mouse models of β-catenin dysregulation displayed increased epithelial proliferation (37), which we did not observe in Dicer cKO uteri. It is likely that the increased expression of Wnt4 and other members of the Wnt/β-catenin signaling pathway are a compensatory mechanism to overcome the lack of glandular epithelium in the Dicer cKO uteri. Recent miRNA screens have suggested that multiple miRNA play a role in mammalian Wnt signaling (72). Additionally, loss of Dicer in the mesenchymal cells of the uterus also resulted in ectopic expression of Wnt4, Wnt5a, and Wnt11 at 8 wk and increased expression of Wnt7a at P14 in the uterus (12) and Wnt5a and Wnt7a in the oviduct. Thus, miRNA regulation likely plays a critical role in Wnt/β-catenin signaling pathways, potentially through cross talk between the uterine stroma and epithelium.

The loss of endometrial glands may be in part due to altered hedgehog signaling. Progesterone given to neonatal mice cause loss of uterine glands and this process is mediated through Ihh (73). Also constitutive activation of Smoothened (Smo), a mediator of hedgehog action results in loss of uterine glands (35). Therefore, the increased Ihh expression in this Dicer cKO model may explain the loss of endometrial glands.

The infertile phenotype observed by the conditional ablation of Dicer using the PR-Cre was due to a direct effect on the uterus. Ovarian histology and steroid hormone production as well as fertilization rates were normal in prepubertal Dicer cKO mice. However, a small, not statistically significant decrease in embryonic development was observed. This may be due to loss of histotrophic factors produced by the uterine glands.

One interesting observation in this study was that compared with women, a smaller number of miRNA were detected in the mouse uterus. The number and magnitude of changes in miRNA expression were also small in our comparison study. This phenomenon could be attributed to differential miRNA stability in the absence of Dicer or the fact that Dicer is being deleted in a subset of distinct cell types, the epithelial and subepithelial stromal cells at P15. Surprisingly, with ablation of Dicer, we found increased expression of some miRNA by next-generation sequencing. Mmu-miR-1 is one of the most abundant miRNA in the mouse uterus, comprising 8% of Dicer cKO miRNA and 5% of control. miR-1 is considered a skeletal muscle miRNA (74). However, little is known about its expression in the uterine myometrium. From our data set, miR-10b is also significantly abundant in the mouse uterus. Interestingly, miR-10b is increased and miR-10a is decreased. Because both miR-10b and miR-10a belong to the same miRNA family and should target the same genes, the impact of this change in expression on actual function is likely to be minimal. Let-7 g is also expressed very highly in both control and Dicer cKO uteri. Because the let-7 family is fairly ubiquitously expressed, this small change is not likely a significant finding. miR-379 has very low read counts in both samples; its expression goes from barely detectable (500 copies) to slightly increased (700 copies) but still very low levels of expression. Thus, the miR-379 findings are likely biologically insignificant. Alternately, it is possible that mature miRNA may have been processed before Dicer deletion; loss of Dicer can increase the expression of certain transcription factors, which may lead to an increase in miRNA precursors (primary miRNA). It may also be possible that a number of mature miRNA are produced in the uterus by Dicer-independent mechanisms, for instance through Argonaute 2 (Ago2) processing (75, 76).

Recent evidence has shown that Dicer is also important for endogenous siRNA biosynthesis. Although endogenous siRNA are important in the oocyte (13, 14), the role of these small RNA sequences (if any) in the uterus has not been characterized. Our next-generation sequencing studies were focused on miRNA, excluding other small RNA from analysis.

Loss of miRNA should result in a loss of repression of transcription, or de-repression. Thus, we anticipated that many more genes would be up-regulated than down-regulated in the Dicer cKO uteri. However, the microarray data presented here are only a single snapshot in time, and they reflect both direct and indirect consequences of miRNA loss. It is likely that many of the genes that are up-regulated, especially those in the Wnt/B-catenin family, form part of a compensatory mechanism to the lack of glandular epithelium and not from direct miRNA targeting. Similarly, the tight junction signaling pathway genes that are dysregulated may reflect a significantly defective epithelial-stromal cross talk. Thus, we believe that the dysregulated genes and signaling pathways are secondary effects and not directly due to specific miRNA loss. However, we cannot discount the idea that some miRNA function to increase their target gene expression, and thus, loss of important activator miRNA would lead to a down-regulation of those key target genes. Additionally, the Dicer cKO uteri show a significant loss of glandular epithelium. Genes that are expressed in glandular epithelium such as Foxa2 should be significantly down-regulated as a consequence of structural changes in the uterus. Consistent with the absence of glandular epithelium in the Dicer cKO uteri, we found a significant decrease in Foxa2 expression using both microarray and independent samples by validated qPCR.

To examine the direct function of miR-101 and miR-181 in the uterus, we performed in vitro studies using human endometrial stromal fibroblasts since there are no useful benign mouse endometrial cell lines. Our data show that miR-101 and miR-181 overexpression leads to decreased endogenous BCL2L11 gene expression. Although this study was not done in the mouse, it was done in a biologically plausible endometrial culture system. Thus, we believe that loss of miR-101 and/or miR-181 through Dicer ablation leads to increased apoptosis through loss of de-repression of Bcl2l11.

Other interesting miRNA were dysregulated in our Dicer cKO model. A member of the miR-29 family, miR-29a, is down-regulated in the Dicer cKO uteri. Previous work from our group has shown that miR-29c is highly expressed in endometriomas. Overexpression of miR-29c in human endometrial stromal fibroblasts affects endogenous target gene expression as well as biological function, such as in an in vitro model of decidualization (21). Additionally, miR-29b and miR-29c are thought to be progesterone responsive miRNA because their expression is highest during the midsecretory phase of the human menstrual cycle (57). Because our mouse model shows significant progesterone resistance, the decreased expression of miR-29a in Dicer cKO uteri fits with human miRNA data. Previously published work has shown that another member of the miR-29 family, miR-29b, directly targets the 3′ UTR of Bcl2l11 in mouse brain (77). Several genes of the Wnt/β-catenin signaling pathway are validated targets of specific dysregulated miRNA, although none of the experiments were done in uterus. For instance, using mouse granulosa cells, expression of miR-503 resulted in decreased expression of endogenous expression of Acvr2b (16). Additionally, miR-101 is a validated direct target for the 3′ UTR of ACVR2B in humans (78). Our findings showing Acvr2b up-regulation and down-regulation of miR-503 and miR-101b, a member of the miR-101 family in Dicer cKO uteri are in agreement with those studies. Multiple members of the let-7 family were down-regulated, and one member, let-7 g, was up-regulated in the Dicer cKO uteri. By extrapolating to other members of the let-7 family that were not changed in our mouse model, we can begin to hypothesize what these molecules are doing in our model system. Previously published work has shown that a member of the let-7 family, let-7b, directly targets the 3′ UTR of Ccnd1 (77). This direct effect on the 3′ UTR of CCND1 was also shown in human melanoma cell lines (79). Using two different human cell lines, the direct effect of miR-503 on endogenous gene expression and 3′ UTR repression was confirmed (80). Lastly, miR-1 was validated as a miRNA that targeted CCND1 (81). Using human cells, miR-1 directly targets Sox5 (81). In our Dicer cKO uteri, we found up-regulation of miR-1 and down-regulation of Sox5.

In summary, deletion of Dicer in the mouse uterus leads to globally decreased miRNA levels in the uterus, apoptosis and disappearance of uterine stroma, and abnormal cellular and molecular signaling, all leading to a sterility phenotype.

Materials and Methods

Generation and genotyping of Dicer cKO mice

The Dicer conditional allele (Dicer^flox/flox) mice have been described previously (5) and were maintained on a C57BL/6J;129S5/Brd mixed hybrid background. Dicer^flox/flox mice were bred to Pgr^cre/+ (herein called PR-Cre) mice (32) to generate Dicer^flox/+; PR^cre/+ mice. These mice were then crossed to generate Dicer^flox/flox; PR^cre/+ (designated as Dicer cKO throughout) and Dicer^flox/flox; PR^+/+ mice (designated as control). All breeding were performed by the Baylor College of Medicine Genetically Engineered Mouse Core (Houston, TX). Mice were genotyped from tail genomic DNA using PCR primers as described (32). All mice were maintained in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals under an approved protocol.

Fertility analysis

For fertility studies, 10 6-wk-old Dicer cKO and 10 control females were bred to wild-type C57BL/6J;129S5/Brd hybrid males with proven fertility. One control mouse died. The numbers of litters and pups were recorded over a 6-month period. A χ² analysis was performed to determine statistical significance.

Steroid hormone treatment, tissue collection, and histological analysis

Mice were genotyped at 12–14 d of life. Control or Dicer cKO mice underwent ovariectomy followed by rest for 1 wk and were treated with either sesame oil or 1 mg of progesterone for 6 h or as described for the delayed implantation response (44) (all reagents from Sigma, St. Louis, MO). Reproductive organs were dissected and fixed for histology or snap frozen for RNA isolation. Uteri were fixed in 4% paraformaldehyde or Bouin's solution; ovaries were fixed in 10% neutral buffered formalin. Tissue processing and paraffin embedding were performed by the Department of Pathology Core Services laboratory. Sections were cut at 5 μm and stained with hematoxylin-eosin using standard techniques. For pooled samples, at P15, mice were dissected and five uteri of each genotype were placed into tubes and snap frozen together.

TUNEL and immunohistochemical staining

Three sections from each of three independent control and Dicer cKO uteri were analyzed in parallel. Analysis of apoptosis in uteri from P15 mice was carried out by TUNEL assay using the ApopTag Plus Fluorescein in situ apoptosis detection kit (Chemicon International, Temecula, CA). A TUNEL assay was performed according to the manufacturer's instructions, and slides were mounted in Vectashield (Vector Laboratories, Inc., Burlingame, CA) containing propidium iodine to visualize chromatin. Progesterone receptor (clone A0098; Dako, Carpentaria, CA) antibody was used as described previously (36). Dicer antibody (Santa Cruz Biotechnologies, Santa Cruz, CA) was used with the diaminobenzidine peroxidase substrate kit (Vector Laboratories). The GE Healthcare cell proliferation kit (Piscataway, NJ) was used per the manufacturer's instructions for BrdU staining. Slides were analyzed by confocal or light microscopy.

Superovulation and isolation of oocytes/embryos

Superovulation experiments were carried out as described (82). Twenty P21 Dicer cKO and control female mice were injected ip with 5 IU pregnant mare serum gonadotropin (Calbiochem, San Diego, CA) for 46 h, followed by ip injection with 5 IU Novarel (Ferring Pharmaceuticals, Parsippany, NJ). Mice were then bred to wild-type males with known fertility. Eggs and/or embryos were recovered from the ampulla of the oviduct 20 h after Novarel, collected in M2 medium (Sigma) containing 1 mg/ml hyaluronidase (Sigma) to dissociate cumulus cells, counted, and cultured overnight in M16 medium (Sigma). The numbers of one-cell, two-cell, and fragmented oocytes/embryos were recorded.

Serum analysis

Mice were anesthetized by isoflurane inhalation (Abbott Laboratories, North Chicago, IL), and blood was collected in microtainer tubes (Becton Dickinson, Franklin Lakes, NJ) by closed cardiac puncture. Serum was separated by centrifugation and stored at −20 C until further use. FSH, LH, and P₄ measurements were performed by the University of Virginia Ligand Assay and Analysis Core. Estradiol measurements were performed on a Beckman Access II automated platform (Beckman-Coulter, Inc., Brea, CA) at the Laboratory for Male Reproductive Research and Testing at Baylor College of Medicine. A Student's t test was performed on log-transformed data.

RNA extraction, next-generation miRNA sequencing, and whole-genome expression analysis

Uteri from P15 Dicer cKO and control female mice were dissected, pooled by genotype (n = 2 pools per genotype; five mice per pool), and snap frozen. Total RNA was extracted using the mirVana miRNA isolation kit (Applied Biosystems, Inc., Foster City, CA). RNA quality control was performed on an Agilent 2100 BioAnalyzer (Palo Alto, CA) through the microarray core facility at Baylor College of Medicine. Small RNA library construction was performed using the DGE-Small RNA sample prep kit (Illumina) according to the manufacturer's protocol. Purified cDNA was quantified using the Quant-iT PicoGreen dsDNA kit (Invitrogen, Carlsbad, CA) and diluted to 10 nm for sequencing on the Illumina GA-IIx genome analyzer at the University of Houston. All unique sequence reads with a minimum read count of 10 reads were aligned to mouse miRNA sequences in the miRNA database (miRBase version 15.0) as described previously (21). For each sample, the counts were tabulated and normalized to the total number of small RNA sequences. For each individual miRNA, the average number of counts was compared between the genotypes. miRNA with fold change greater than 1.2 and read counts greater than 300 in the control group were considered differentially expressed. Whole genome-wide expression analysis was performed using an Illumina-mouse whole genome array at the Texas Cancer Center Genomics Core. Gene expression arrays were analyzed using GeneSpring software (Agilent) to define differentially expressed genes. Array data have been deposited into the Gene Expression Omnibus (GSE39181).

miRNA target predictions

TargetScan (version 5.0) was used to identify potential mRNA targets for miRNA that were differentially expressed between P15 Dicer cKO and control uteri. Predictions were compiled from a Microsoft Excel annotation worksheet consisting of all putative miRNA-mRNA interactions for TargetScan as described previously (21).

qPCR for mRNA and miRNA

Uteri from Dicer cKO and control mice were dissected and snap frozen. Total RNA was extracted using the mirVana miRNA isolation kit (Applied Biosystems) as above. RNA was treated with Turbo DNase (Applied Biosystems, Inc.) according to manufacturer's protocol. Deoxyribonuclease (DNase)-treated RNA (1000 ng) was reverse transcribed in a 50-μl reaction using 250 U Superscript III reverse transcriptase (Invitrogen) or random primers (Invitrogen). Samples were diluted to 100 μl, and 2 μl was used for each qPCR reaction. For TaqMan miRNA assays, 20 ng of DNase-treated RNA underwent reverse transcription using the miRNA reverse transcriptase kit (Applied Biosystems). qPCR was performed on the ABI StepOnePlus using either predesigned TaqMan Gene or miRNA expression assays (Applied Biosystems) or custom primers designed using Primer Express software (Applied Biosystems) for SYBR Green. Expression levels of mouse Rpl13a or snoRNA202 were used as endogenous controls for mouse or RPL19 for human. TaqMan PCR was performed using TaqMan universal PCR master mix (Applied Biosystems), and PCR with custom primers was performed using SYBR Green PCR master mix (Applied Biosystems) in 10 μl. The reaction conditions were as follows: 2 min at 50 C, 10 min at 95 C, followed by 40 cycles of 15 sec at 95 C (denaturation) and 1 min at 60 C (annealing/extension). Each sample was analyzed in duplicate or triplicate, and a nontemplate control (nuclease-free water) sample was included on each plate for each primer-probe set. All custom primers had efficiency of 85–110%. All SYBR Green runs had dissociation curves to detect potential primer-dimers. The relative quantity of transcript was calculated using the 2^−ΔΔCt method (83) and plotted as mean ± sem. A Student's t test was used to generate P values for statistical significance.

Institutional review board approval, collection of tissues, creation of primary cultures, and transfection of cells

All human tissues collected were collected with Baylor College of Medicine Institutional Review Board approval with written informed consent, and cell cultures were created as described previously (21). Cells were transfected as described previously using 150 nm of miRNA mimic or negative control no. 1 (Sigma) (21) and harvested at 48 h. RNA was extracted using the mirVana kit as above.

†

Annotations provided by Nuclear Receptor Signaling Atlas (NURSA) Bioinformatics Resource. Molecule Pages can be accessed on the NURSA website at www.nursa.org.

Acknowledgments

We thank Julio Agno for technical assistance with the mouse colony, Lang Ma for technical assistance with immunohistochemistry, the Baylor College of Medicine Laboratory for Male Reproductive Research and Testing (Houston, TX) for estradiol assays, the University of Virginia Ligand Assay and Analysis Core (Specialized Cooperative Centers Program in Reproduction Research, Charlottesville, VA) for performing serum hormone assays, Dr. Alexander Tarakhovsky for the gift of the Dicer1^flox/flox mice, and Yiqun Zhang for technical assistance with bioinformatic analysis.

This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development/National Institutes of Health through Cooperative Agreements U54HD007495 (to S.M.H., J.-W.J., P.H.G., F.J.D., and M.M.M.) and U54HD028934 (to the University of Virginia Center for Research in Reproduction Ligand Assay and Analysis Core, Charlottesville, VA) as part of the Specialized Cooperative Centers Program in Reproduction and Infertility Research, the Women's Reproductive Health Research Program Grant 5K12HD050128 (to S.M.H.), and the Herman L. and LeNan Gardner Research Fund in Obstetrics and Gynecology (to S.M.H.).

Disclosure Summary: The authors have no conflicts of interest to disclose.

Abbreviations

 
• BrdU

  Bromodeoxyuridine

 
• cKO

  conditional knockout

 
• DNase

  deoxyribonuclease

 
• HESF

  human endometrial stromal fibroblast

 
• IHH

  Indian hedgehog

 
• miRNA

  microRNA

 
• MMTV

  mouse mammary tumor virus

 
• P

  postnatal day

 
• P₄

  progesterone

 
• PR

  P₄ receptor

 
• qPCR

  quantitative PCR

 
• siRNA

  small interfering RNA

 
• TUNEL

  terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling

 
• UTR

  untranslated region.