Dicer1 Is Essential for Female Fertility and Normal Development of the Female Reproductive System

Abstract

The ribonuclease III endonuclease, Dicer1 (also known as Dicer), is essential for the synthesis of the 19–25 nucleotide noncoding RNAs known as micro-RNAs (miRNAs). These miRNAs associate with the RNA-induced silencing complex to regulate gene expression posttranscriptionally by base pairing with 3′untranslated regions of complementary mRNA targets. Although it is established that miRNAs are expressed in the reproductive tract, their functional role and effect on reproductive disease remain unknown. The studies herein establish for the first time the reproductive phenotype of mice with loxP insertions in the Dicer1 gene (Dicer1^fl/fl) when crossed with mice expressing Cre-recombinase driven by the anti-müllerian hormone receptor 2 promoter (Amhr2^Cre/+). Adult female Dicer1^fl/fl;Amhr2^Cre/+ mice displayed normal mating behavior but failed to produce offspring when exposed to fertile males during a 5-month breeding trial. Morphological and histological assessments of the reproductive tracts of immature and adult mice indicated that the uterus and oviduct were hypotrophic, and the oviduct was highly disorganized. Natural mating of Dicer1^fl/fl;Amhr2^Cre/+ females resulted in successful fertilization as evidenced by the recovery of fertilized oocytes on d 1 pregnancy, which developed normally to blastocysts in culture. Developmentally delayed embryos were collected from Dicer1^fl/fl; Amhr2^Cre/+ mice on d 3 pregnancy when compared with controls. Oviductal transport was disrupted in the Dicer1^fl/fl;Amhr2^Cre/+ mouse as evidenced by the failure of embryos to enter the uterus on d 4 pregnancy. These studies implicate Dicer1/miRNA mediated posttranscriptional gene regulation in reproductive somatic tissues as critical for the normal development and function of these tissues and for female fertility.

SUSTAINED REPRODUCTIVE function requires that the female reproductive tract, including the uterus, oviduct, cervix, and ovary, correctly develops and functions together. To obtain the precise temporospatial control over cellular protein expression that is necessary for organogenesis, a myriad of gene transcription, translation, and posttranslational regulatory mechanisms are invoked. Despite the general importance of the female reproductive system, our understanding of organogenesis and the molecular mechanisms that regulate the development and differentiation of these tissues lags behind other organ systems. Moreover, recent observations suggest that in addition to transcription, translation, and posttranslational modifications, posttranscriptional gene regulation may play a more pronounced role in cell, tissue, and organ function. Recently, micro-RNAs (miRNAs) have been demonstrated to play a novel, yet not thoroughly defined role in posttranscriptional regulation of gene expression.

miRNAs are a recently described class of small noncoding regulatory RNAs that regulate gene expression posttranscriptionally (1). miRNAs are proposed to be involved in diverse developmental and pathological processes (2). Biogenesis of miRNAs is a multistep process that culminates with the ribonuclease (RNase) III endonuclease, Dicer1, cleaving the 70- to 110-bp hairpin precursor miRNAs and forming 19–25 nucleotides long double-stranded miRNAs (3). Subsequently, one strand of this pair associates with the Argonaute proteins to form the RNA-induced silencing complex, which can then affect posttranscriptional gene regulation. In mammals the miRNA-RNA-induced silencing complex primarily binds the 3′untranslated regions of target mRNAs, with partial complementarity to either repress or enhance translation (4,5,6).

Dicer1 (through the generation of miRNAs and subsequent posttranscriptional regulation of specific gene products) has been proposed to play a role in the normal development of the lung (7), limbs (8), and skeletal muscle (9), as well as the female germ line (10). Recently, posttranscriptional regulation and miRNAs have been proposed to play a role in embryo implantation (11), as well as in the human endometrium and the pathophysiology of endometriosis (12). Collectively, these studies suggest that Dicer1 and its miRNA products play a pivotal role in the molecular regulation of multiple organ systems, which may include reproductive functions such as oocyte maturation, embryo implantation, and uterine pathophysiological conditions. To date, there is no information on the role of Dicer1 and miRNAs in the development and subsequent function of the female reproductive organs. As such, the objective of the current study was to examine the phenotypical consequences of conditional deletion of the Dicer1 gene product from the developing female reproductive system and its impact on female fertility.

Materials and Methods

Generation of conditional Dicer1 knockdown

The University of Kansas Medical Center’s Institutional Animal Care and Use Committee approved all procedures involving mice before use. Mice homozygous for loxP insertions flanking the second RNase III domain in the Dicer1 gene (Dicer1^fl/fl) (generously provided by Dr. Clifford Tabin, Harvard Medical School) were crossed with mice heterozygous for Cre recombinase knocked into the anti-müllerian hormone receptor 2 locus (Amhr2^Cre/+) (generously donated by Dr. Richard Behringer, Baylor College of Medicine) to produce mice that exhibit knocked-down expression of Dicer1 in the ovarian granulosa cells and the derivatives of the müllerian duct (i.e. oviduct, uterus, and cervix). The resulting progeny were genotyped as previously described (8,13).

Breeding studies

To assess fertility of these mice, female Dicer1^fl/fl;Amhr2^Cre/+ mice (42 d of age, n = 8) were mated with adult wild-type males of known fertility. Dicer1^fl/fl;Amhr2^Cre/+ females were continually exposed to males for a minimum of 5 months. Female mice were checked daily for the presence of a seminal plug to confirm mating.

To characterize the general morphology of the Dicer1^fl/fl;Amhr2^Cre/+ female reproductive tract, mice at several developmental ages, after natural mating or after a follicular stimulation protocol, were killed, and ovarian, oviductal and uterine function, and morphology were evaluated.

To assess ovarian and uterine function and examine fertility, adult littermate females (>42 d of age) of Dicer1^fl/fl;Amhr2^Cre/+ and wild-type control (i.e. Dicer1^fl/+;Amhr2^+/+ or Dicer1^fl/fl;Amhr2^+/+) genotypes were naturally mated, and killed on d 1, 3, 4, 6–7 postcoitus (d 1 = day seminal plug observed; n = 9, 6, 8, and 7 for wild-types, and n = 8, 6, 8, and 6 for Dicer1^fl/fl;Amhr2^Cre/+ on each respective day of pregnancy). Ovulation and fertilization rates were determined by counting the oocytes recovered from cumulus-oocyte complexes expressed from the oviducts of mice killed on d 1 pregnancy. After collection the fertilized embryos were cultured for 5 d as previously described to assess embryonic development (14,15). Total body, ovarian, and uterine weights were recorded for d-1 pregnant mice.

In vivo embryonic development, oviductal transport, and implantation rates were assessed on d 3, 4, 6–7 pregnancy, respectively. Embryos collected on d 3 and 4 pregnancy were classified for developmental stage as previously described (14,15). Oviductal transport was assessed by determining the location of the embryos within the oviducts on d 4. The oviducts and uterine horns were carefully bisected immediately below the uterotubal junction, and the oviduct and uteri flushed independently into separate collection dishes. It was found that oviducts of Dicer1^fl/fl;Amhr2^Cre/+ females could not be flushed, therefore, the oviducts were dissected along their entire length to release embryos. The embryos were counted, and the stage of development was recorded. On d 6–7 pregnancy, the number of implantation sites was determined by injection of Chicago Sky Blue dye into the tail vein 1 min before euthanization (16). Ovarian, oviductal, and uterine tissues from all mice were fixed and embedded in paraffin or flash frozen for Western blot analysis.

To examine uterine and ovarian function under controlled conditions, immature littermate females (22 d of age) of Dicer1^fl/fl;Amhr2^Cre/+ (n = 6) and control wild-type (i.e. Dicer1^fl/+or Dicer1^fl/fl lacking the Amhr2^Cre/+; n = 12) genotypes were administered 2 IU equine chorionic gonadotropin (eCG) for 46 h, followed by 2 IU human chorionic gonadotropin (hCG) for 16–17 h. Ovulation rates were determined by counting the oocytes recovered. Uterine and oviductal tissues were collected from the eCG plus hCG-treated mice, as well as from untreated immature d 10 and 26 female Dicer1^fl/fl;Amhr2^Cre/+ and littermate control mice for Western blot analysis and histological analyses. Serum blood samples were obtained from all adult and treated immature (d 25) mice for subsequent determination of progesterone and estrogen concentrations by RIA (17).

Western blot analysis

Combined uterine and oviductal tissues (d 10) or pooled oviductal tissues alone (d 25 immature eCG plus hCG treated mice) were homogenized in lysis buffer (Cell Signaling Technology, Inc., Danvers, MA). The resulting protein lysates were centrifuged at 16,000 × g for 5 min to pellet the cellular membrane debris. Supernatants were transferred to new tubes and stored at −80 C until use. Protein samples (10 μg), as determined by Bio-Rad Protein Assay (Bio-Rad Laboratories, Inc., Hercules, CA), were loaded onto 12% SDS-PAGE gels and transferred to polyvinylidene fluoride membranes using standard methods. Immunoblots were blocked with 5% milk solution and incubated overnight at 4 C with antibodies to β-catenin (BD Biosciences, San Jose, CA) and β-actin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). After washing, protein-antibody complexes were visualized using West Pico Chemiluminescent Substrate (Pierce, Rockford, IL) following the manufacturer’s protocol. ODs for the immunoblots were determined on a Gel-Pro Analyzer (Media Cybernetics, Inc., Bethesda, MD).

Histological analysis

Ovaries, oviducts, and uteri collected for histology were immediately fixed in either 4% paraformaldehyde or Bouin’s solution, before embedding in paraffin. Sections from the midregion of the uterus, and serial tissue sections (8 μm) from ovarian and oviductal tissues were stained with hematoxylin and eosin.

Statistical analysis

Statistical analysis was performed with GraphPad Prism (version 4; GraphPad Software Inc., San Diego, CA). Uterine and ovarian weights, ovulation rates, percentage of embryonic development, and progesterone levels were analyzed by the Student’s t test. P values less than 0.05 were considered significant.

Results

Female Dicer1^fl/fl;Amhr2^Cre/+ mice (n = 8) are infertile, as evidenced by the failure to produce offspring over a 5-month breeding study. The mating behavior of female Dicer1^fl/fl;Amhr2^Cre/+ mice appeared normal, as shown by the presence of seminal plugs after exposure to male mice. Moreover, several mice (not exposed to males) exhibited normal 4- to 5-d estrous cycles as detected by vaginal smears. Female wild-type mice used as controls (e.g. Dicer1^fl/fl or ^fl/+ lacking Amhr2^Cre/+) showed evidence of a pending pregnancy within 14 d of male exposure, whereas Dicer1^fl/fl;Amhr2^Cre/+ females failed to exhibit visible signs of pregnancy. Male Dicer1^fl/fl;Amhr2^Cre/+ mice (n = 5) were fertile, as evidenced by their ability to sire multiple litters with wild-type females.

To establish whether the cause of female infertility is related to a gross developmental defect, the reproductive tracts of untreated immature (d 25) Dicer1^fl/fl;Amhr2^Cre/+ and littermate control mice were examined (Fig. 1A). The length and diameter of the Dicer1^fl/fl;Amhr2^Cre/+ uterine horn was remarkably shorter compared with control uteri. Similar to the uterus, the oviducts of Dicer1^fl/fl;Amhr2^Cre/+ mice were truncated in length (less than one half the length), and both tissues appeared more transparent when observed through a dissecting microscope than corresponding tissues from control mice (compare Fig. 1, B and C). In addition, sac-like structures were observed within the oviduct of Dicer1^fl/fl;Amhr2^Cre/+ mice (Fig. 1C). Similar oviductal and uterine morphological findings for the eCG plus hCG-treated immature (d 25) and for the adult d-1 pregnant Dicer1^fl/fl;Amhr2^Cre/+mice were observed (data not shown). The uterine weights of the d-1 pregnant Dicer1^fl/fl;Amhr2^Cre/+ mice (1.61 ± 0.20 mg/g; uterine weight/total body weight) were (P < 0.05) reduced compared with age-matched mice (5.00 ± 0.28 mg/g). Again, distended sac-like structures filled with clear fluid (Fig. 1C) were concentrated at the end of the oviduct nearest the uterotubal junction, in contrast to the typical mucosal folding characteristic of oviducts from control mice.

Figure 1.

Morphological changes in the female reproductive tract of mice with anti-müllerian hormone receptor-2-Cre recombinase targeted deletion of Dicer1. A, Immature female reproductive tracts (magnification, ×1.5) from wild-type (WT) and Dicer1^fl/fl;Amhr2^Cre/+ [knockout (KO)] mice. The uteri of the Dicer1^fl/fl;Amhr2^Cre/+ mice are roughly two thirds the length of the wild-type mice. B and C, Magnified view (×6) of the oviduct and ovary from control (B) and Dicer1^fl/fl; Amhr2^Cre/+ mice (C). Note the extensive coiling of the wild-type oviduct (white arrow) and the lack of such coiling in the Dicer1^fl/fl;Amhr2^Cre/+ knockout oviduct, and the presence of fluid-filled sacs near the uterotubal junction (white arrowhead).

The oviducts of the immature and pregnant Dicer1^fl/fl;Amhr2^Cre/+ mice were extremely fragile and did not appear patent, as evidenced by unsuccessful attempts to flush the oviduct. Oviducts ruptured easily; thus, to collect embryos, oviducts were manually dissected rather than flushed. However, naked oocytes/embryos could be visualized through the transparent wall of the oviduct, implicating that the oviduct was capable of early gamete collection and transport. In contrast, wild-type control mice were readily flushed and contained a large number (10,11,12,13,14,15) of oocytes or embryos in immature and adult mice, respectively.

The uteri of Dicer1^fl/fl;Amhr2^Cre/+ mice did not appear to lack primary cell types because uterine glands, and stromal and myometrial tissues all appeared to be present (Fig. 2, A and B). However, the uteri of immature eCG plus hCG-treated Dicer1^fl/fl;Amhr2^Cre/+ mice did appear to have a thinner myometrial layer and reduced numbers of uterine glands than that observed in control littermates. Similar findings were observed for pregnant mice. However, the differences in uterine morphology and weights between the Dicer1^fl/fl;Amhr2^Cre/+ and wild-type mice were not attributable to changes in progesterone (d 1 pregnant mice; See Fig. 5) or estrogen levels (eCG plus hCG treated mice; data not shown). Similarly, the isthmus of the oviduct, the portion with the greatest musculature, exhibited an almost complete loss of smooth muscle tissue, and the characteristic mucosal folds were replaced by distended sac-like structures (Fig. 2, C–E). Histological evaluation of the oviduct also showed a generally disorganized epithelial cell layer (Fig. 2F).

Figure 2.

Histological analysis of the uterus and oviduct of immature Dicer1^fl/fl;Amhr2^Cre/+ and littermate wild-type control mice. A and C, Uteri and oviduct of wild-type mouse. B and D–F, Dicer1^fl/fl;Amhr2^Cre/+ uteri and oviductal tissue. Immature 25-d-old mice were sequentially administered eCG (46 h), followed by hCG (16 h) before tissue harvesting. Tissues were fixed in Bouin’s solution, paraffin embedded, sectioned (8 μm), and hematoxylin and eosin stained. The uteri (B) of the Dicer1^fl/fl;Amhr2^Cre/+ were reduced in size and appeared to contain much less smooth muscle [lateral and transverse smooth muscle are marked with brackets when compared with the wild-type (A)]. Uterine glands were prevalent in the wild-type uteri (A) and sparse in the Dicer1^fl/fl;Amhr2^Cre/+ uteri (arrowhead). The oviducts of wild-type mice were highly coiled, and the smooth muscle around the isthmus region was well developed (C, arrow). Similar regions in the Dicer1^fl/fl;Amhr2^Cre/+ were not observed because this region of the oviduct was populated with sac-like structures. D shows a region of the oviduct that is transitioning from a tubule-like structure to the sac-like structures seen more anterior in E, which resided near the uterotubal junction; arrowheads denote several of the sac-like structures. F shows a magnified view of epithelium from Dicer1^fl/fl; Amhr2^Cre/+ mice. Bars, 500 μm for A and B. Bars, 200 μm for C–F.

Figure 5.

Oviductal and uterine β-catenin levels in Dicer1^fl/fl;Amhr2^Cre/+ (KO) and wild-type (WT) female mice. Oviductal and uterine tissue pooled together from d 10 or oviductal tissue alone from d 26 mice was analyzed by Western blot for β-catenin protein levels. β-Actin was used as a loading control.

In addition to the gross morphological changes in the uterus and oviduct, ovarian weight was also reduced (P < 0.05) in the d-1 pregnant Dicer1^fl/fl;Amhr2^Cre/+ (0.22 ± 0.009 mg/g; ovarian weight/total body weight) mouse compared with control mice (0.27 ± 0.014 mg/g). Moreover, numbers of naturally ovulated cumulus-oocyte complexes were less (P < 0.05) in the Dicer1^fl/fl;Amhr2^Cre/+ mice (7.0 ± 1.1; mean ± sem) than control mice (10.7 ± 0.9). Similar to the naturally mated animals, the ovulation rate for the immature eCG plus hCG-treated wild-type females (n = 12) was significantly greater (P < 0.05) than that of Dicer1^fl/fl;Amhr2^Cre/+ females [(n = 6) 16.2 ± 1.4 vs. 3.67 ± 1.50, respectively]. However, gross morphological and histological examination of the pregnant ovaries from both genotypes showed no marked differences (data not shown). However, examination of immature Dicer1^fl/fl;Amhr2^Cre/+ mice given eCG alone (46 h) indicated that these ovaries contained fewer large antral follicles (data not shown). Overall, the reduced ovulation rates and ovarian weights, as well as histological observations, suggest that fewer preovulatory follicles developed in the Dicer1^fl/fl;Amhr2^Cre/+ mouse. However, in depth analyses of follicular dynamics and characterization of Dicer1 will be necessary to further address ovarian Dicer1 function.

Culture of the d-1 (pronuclear stage) embryos indicated that embryos from Dicer1^fl/fl;Amhr2^Cre/+ mice were capable of normal in vitro development (Table 1). Embryonic development through blastocyst formation and the percentage of hatching was not different across genotypes. In contrast, in vivo embryonic development assessed on d 3 pregnancy indicated that embryos from Dicer1^fl/fl;Amhr2^Cre/+ females were markedly delayed when compared with wild-type mice (Fig. 3 and Table 2). Moreover, the incidence of fragmentation and degeneration of these embryos increased in Dicer1^fl/fl;Amhr2^Cre/+ derived embryos after 1 d culture compared with wild-type derived embryos (Fig. 3). Embryos collected from Dicer1^fl/fl;Amhr2^Cre/+ mice on d 4 pregnancy also displayed increased fragmentation and degeneration (data not shown). To further evaluate oviductal function, the location of embryos within the reproductive tract was examined on d 4 pregnancy. All embryos in wild-type females (n = 9) were located in the uterus, whereas no embryos were found in the uterus of Dicer1^fl/fl;Amhr2^Cre/+ mice (n = 8; Table 3). The majority of the embryos within the oviduct of Dicer1^fl/fl;Amhr2^Cre/+ mice were found in the upper one third of the oviduct. The few embryos that had progressed through the oviduct to the isthmus were mostly fragmented, and some zonae pellucidae had been lost (data not shown).

Table 1.

In vitro embryonic development of embryos collected from d-1 pregnant Dicer1^fl/fl;Amhr2^Cre/+ and wild-type females

	Day of embryo culture
	2	3	4	5
Stage of embryo development	≥2 cells	≥4 cells	≥ Compacting morula	≥ Blastocyst	≥ Hatching blastocyst
Genotype of donor female (n = embryos; dams)
Wild-type (n = 75; 9)	100%	96%	90.7%	86.7%	54.7%
Dicer1fl/fl;Amhr2Cre/+ (n = 37; 7)	97%	86%	91.7%	72.2%	38.9%

Percentage of two or more cell embryos was calculated from total numbers of fertilized oocytes (n = 75 and 37) for each genotype. Percentages of four or more cell embryos and all subsequent groups were calculated from the number of two or more cell embryos present on d 2 culture. 

Figure 3.

Embryos from d-3 pregnant Dicer1^fl/fl;Amhr2^Cre/+ and wild-type female mice. A and C, Embryos from wild-type mice immediately after collection (A) and after 24 h culture (C). B and D, Embryos from Dicer1^fl/fl;Amhr2^Cre/+ mice immediately after collection (B) and after 24 h culture (D). Increased fragmentation and degeneration were observed in embryos collected from Dicer1^fl/fl;Amhr2^Cre/+ females compared with wild-type controls after 24 h culture. Bars, 100 μm.

Table 2.

In vivo embryonic development in d-3 pregnant Dicer1^fl/fl;Amhr2^Cre/+ and wild-type females

Genotype	Embryonic stage
	1 cell	4–6 cells	8 cells	>8 cells
Wild-type (n = 7)	0	0	54	7
Dicer1fl/fl;Amhr2Cre/+ (n = 8)	14	8	25	0

Numbers of embryos found in each region of oviduct and uterus from a wild-type (n = 7) and Dicer1^fl/fl;Amhr2^Cre/+ (n = 8) mice. 

Table 3.

Location of embryos on d 4 pregnancy in Dicer1^fl/fl;Amhr2^Cre/+ and wild-type females

Genotype	Oviduct				Uterus
	Upper	Mid	Lower	Total
Wild-type (n = 9)	0	0	0	0	91
Dicer1fl/fl;Amhr2Cre/+ (n = 9)	43	13	7	63	0

Numbers of embryos found in each region of oviduct and uterus from a total of nine mice of each genotype. 

Implantation was assessed in another group of mice at d 6 and 7 pregnancy. All wild-type mice exhibited implantation sites (9.8 ± 0.4 implantation sites per dam), whereas no evidence of implantation was observed in Dicer1^fl/fl;Amhr2^Cre/+ mice.

Serum progesterone concentrations indicated that Dicer1^fl/fl; Amhr2^Cre/+ and control mice had similar levels through d 4 pregnancy (Fig. 4). On d 6 pregnancy, Dicer1^fl/fl;Amhr2^Cre/+ mice exhibited a (P < 0.05) minor (24%) decline in progesterone levels compared with the controls. In a single embryo transfer experiment, in vitro fertilized oocytes derived from Dicer1^fl/fl;Amhr2^Cre/+ mice were able to establish pregnancy in recipient females with normal fetal development to at least embryonic d 15 (E15) (data not shown).

Figure 4.

Serum progesterone concentrations in pregnant Dicer1^fl/fl;Amhr2^Cre/+ (KO) and wild-type (WT) mice. Serum collected on d 1, 3, 4, and 6 pregnancy was analyzed for the concentration of progesterone by RIA. Data are presented4 as mean concentration of progesterone ± sem.

To determine whether the loss of Dicer1 affects Wnt signaling, we examined β-catenin expression. We observed reduced levels of β-catenin in combined uterine/oviductal tissues of immature d-10 Dicer1^fl/fl;Amhr2^Cre/+ mice compared with wild-type mice (Fig. 5). In addition, oviductal tissues from d 26 eCG plus hCG-treated Dicer1^fl/fl;Amhr2^Cre/+ mice also showed a loss in β-catenin protein expression. β-Catenin levels were 53 and 31% lower in the Dicer1 mice on d 10 and 26, respectively, after normalization with the loading control, β-actin.

Discussion

Conditional knockdown of the miRNA-processing enzyme, Dicer1, using the Amhr2-driven Cre-recombinase yielded female mice sterile, whereas having no effect on male fertility. A previous study using a similar approach to delete Dicer1 specifically in the oocyte via ZP3-driven Cre-recombinase, also rendered female mice infertile (10). In these later mice, oocyte development and folliculogenesis appeared normal until the resumption of meiosis, at which point defects in spindle formation and chromatin separation in the oocytes were observed. Female mice were also found to be infertile in a recent study in which Dicer1 expression was globally knocked down using a gene-trap method (18). In this study a defect in vascularization of luteal tissue was observed, leading to insufficient progesterone secretion and a failure to maintain pregnancy. Global knockdown of Dicer1 had no effect on ovulation rate, suggesting that granulosa cell function was not compromised in their model (18). Conversely, granulosa cell gene expression is modulated in the Amhr2-Cre recombinase model (19,20,21,22,23,24), and we observed reduced ovulation rates and preovulatory follicular development in our mice. Nevertheless, the infertile phenotype we observed in Dicer1^fl/fl;Amhr2^Cre/+ mice could not be attributed to reduced luteal function because progesterone levels were similar through the early stages of embryonic development (d 4) and only showed a minor decline on d 6 pregnancy. Our data instead point to a loss in oviductal development and function as the primary cause of Dicer1 mediated infertility in our model.

The marked gross morphological changes in the uterus and oviduct observed in our model of Dicer1 deletion suggest that either disruption of tissue development and/or function leads to female infertility. The Amhr2 promoter driving Cre recombinase expression has been used extensively to knock down gene expression in tissues derived from the müllerian duct (i.e. oviductal, uterine, and cervical tissues) (13,21). The Amhr2-driven Cre-recombinase activity has been detected in the müllerian duct mesenchyme as early as E12.5 in ROSA reporter mice (13) and as late as E15.5 d in mice using a fluorescent reporter (21). Moreover, the mesenchymal expression of the fluorescent reporter was also not uniform within the embryonic müllerian duct (21,25). Furthermore, postnatal Amhr2-lacZ and Amhr2-Cre expression was also greater in the circular smooth muscle cells of the myometrium (25,26). However, histological observations of the uteri and oviduct from Dicer1^fl/fl;Amhr2^Cre/+ mice suggest that the cellular layers normally comprising these tissues are present at decreased levels compared with wild-type mice. This was particularly true for the smooth muscle that is present in the isthmus of the oviduct and myometrium. Indeed, the isthmus region of the oviduct was not readily identifiable in the Dicer1^fl/fl;Amhr2^Cre/+ female mice. Therefore, the loss of smooth musculature in these tissues is consistent with elevated Amhr2 expression. Consistent with the loss and disruption of oviductal cell layers, the oviduct of the Dicer1^fl/fl;Amhr2^Cre/+ mouse was not able to support normal embryonic development, nor was it able to facilitate transport of embryos to the uterus. The ability of in vitro cultured pronuclear (d 1) embryos collected from Dicer1^fl/fl;Amhr2^Cre/+ donors to develop at a similar rate as those derived from wild-type females offers further support that oviductal function is disrupted. Finally, the ability of in vitro fertilized oocytes derived from Dicer1^fl/fl;Amhr2^Cre/+ donor mice to establish a pregnancy when transferred to wild-type recipients provides conclusive proof that the ovary is not the primary cause of infertility in this mouse.

Collectively, these observations suggest that Dicer1 and its product (miRNAs) play key roles in uterine and oviductal development. Disruptions of oviductal and uterine morphology have previously been seen in mice with deletions of the homeobox genes, Hoxa9, 10, 11, and 13, as well as genes in the Wnt pathway, including Wnt-7a and β-catenin (27,28,29). The loss of uterine musculature, lack of uterine glands, and failure of the oviduct to undergo coiling phenocopies some of the observations seen when Wnt-7a was knocked out (28,30,31) or when β-catenin was knocked out using Amhr2-Cre (21,25). However, both Wnt-7a and β-catenin had additional reproductive tract phenotypes that were not mimicked by the deletion of Dicer1. In these models, combined uterine and oviductal disruptions have been shown to occur mostly in early development. To determine whether the loss of Dicer1 might lead to disrupted Wnt signaling, immunoblot analysis of oviductal tissues from age-matched d-25 immature mice, and combined uterine/oviductal tissues from 10-d-old mice were analyzed. β-Catenin levels were lower in the Dicer1^fl/fl;Amhr2^Cre/+ uteri and oviducts. Ongoing studies are examining whether the loss of miRNAs is affecting the posttranscriptional gene regulation of this specific transcription factor within developing reproductive tracts. In conclusion, these studies indicate that posttranscriptional gene regulation in somatic tissues of the female reproductive system as regulated by Dicer1, and its product, miRNAs, plays an essential role in female fertility. The loss of Dicer1 resulted in developmental and functional consequences at both the reproductive tract and gonadal level. Ongoing studies are underway to identify the specific miRNAs and their target genes that affect both the development and function of these tissues.