Dicer is required for female reproductive tract development and fertility in the mouse

Morphological and histological abnormalities in Amhr2-Cre; Dicer conditional mutant female reproductive tract

The reproductive organs of adult control and mutant females and males were dissected. No gross abnormalities were observed in the mutant male reproductive organs. Mutant males were fertile and bred with Dicer fx/fx females to increase the probability of generating Amhr2-cre; Dicer fx/fx mice. In mutant males, no Müllerian remnants (uterus and oviducts) were observed, indicating that the Müllerian ducts regressed normally. However, abnormal phenotypes were observed in the mutant female reproductive tract. In adult mutant females, all the components of the female reproductive tract were present, including ovaries, oviducts, uterine horns, a cervix and the vagina. No gross morphological abnormalities were observed in the ovary. However, the uterine horns appeared shorter and thinner when compared to control littermates (Fig. 2A, B). The oviducts also had several abnormalities; they were shorter in length and less coiled (Fig. 2C, D). In the majority of the mutant females, the oviducts had spheroidal expansions near the uterotubal junction, resembling cysts (Fig. 3A). These cysts were found unilaterally (n=9, 11%) and bilaterally (n=9, 89%). The cysts were found in the isthmus, the region of the oviduct that connects the ampulla to the uterus. These were observed to develop from a transparent fluid-filled cyst into a large cyst filled with a yellowish soft mass that filled the entire oviduct (Fig. 3A, B). Histologically, the mass inside the oviduct contained immune/inflammatory cells (Fig. 3E-H). The mutant oviducts also showed a deficiency of smooth muscle (Fig. 3C, D, and F) perhaps causing a weakening in the oviductal wall, leading to cyst formation. In many unmated mutant females, oocytes were found inside the cysts (Fig. 3I). The oocytes were isolated and observed under the microscope. Most of the oocytes appear to retain the zona pellucida, the acellular surrounding membrane of the oocyte, however the cytoplasm appeared fragmented and crystalline structures where found inside the zona (Fig. 3J). These findings demonstrate that the mutant ovaries can ovulate.

Open in a separate window

Fig. 2

Gross morphology of the female reproductive tract

A-D, Reproductive tracts from 5 week-old control (A, C) and mutant (B, D) females. E-H, Reproductive tracts from postnatal day 4 control (E, G) and mutant (F, H) females. Ov, ovary; ovd, oviduct. Scale bar, 1 cm.

Open in a separate window

Fig. 3

Oviduct cysts and inflammation in Dicer mutant females

A, Inflammatory response inside the oviduct of a mutant female (white arrowhead) at postnatal day 28. B, At 8 weeks of age the inflammation response (white arrowhead) encompassed the entire oviduct. C, D, H&E stained histological section of postnatal day 28 oviducts from control (C) and mutant (D) females. The mutant showed a localized deficiency of smooth muscle (arrow). E-H, Chronic inflammation is observed in H&E stained cross sections of a 6 week-old mutant oviduct. F, There was a deficiency of smooth muscle (red arrow) in some regions of the oviduct when compared to the smooth muscle (black arrow) in other regions of the same oviduct. G, H, Higher magnification of panel E. Lymphocytes (black arrowhead), plasma cells (red arrowheads) and macrophages (black arrow) are indicative of chronic inflammation in the oviduct of mutant females. I, Oocytes are observed inside the cysts of unmated mutant females. J, Degenerate embryos (arrowheads) are found surrounded by the zona pellucida (arrow). Ovd, oviduct; zp, zona pellucida. Scale bar, 100 μm.

The female reproductive tract of control and mutant females were dissected at various times after birth to determine when the phenotypic differences were first apparent in the mutant females. On postnatal day 4 there were no significant differences observed in the length of the uterus of mutant females when compared to control females (Fig. 2E, F). However, the mutant uterus appeared thinner and the oviducts were less coiled than the control littermates (Fig. 2G, H). Similar phenotypes were observed on postnatal day 8 (data not shown). On postnatal day 21, there was a marked difference in the size of mutant reproductive tract when compared to control littermates. Expansion of the isthmus, forming oviductal cysts was first observed around postnatal day 28 (Fig. 3A). At this stage, small yellow masses, apparently the onset of an inflammatory response were observed (Fig. 3A). To determine the progression of the cysts and the extent of their growth, 8 week (8 control, 8 mutant), 4 month (3 control, 4 mutant), and 8 month-old (2 control, 2 mutant) females were sacrificed. The cysts were present at all three time points in the mutants. However, at 8 weeks the inflammation appeared to be more prominent than at 4 and 8 months because the cysts were generally smaller. Interestingly, in two 8 week-old mutant females with cysts, the uterus was apparently fused to the abdominal wall and intestines.

Histological analysis of Amhr2-Cre; Dicer conditional mutant female reproductive tract

To assess the cytological differences between control and mutant female reproductive tracts, a histological analysis initially using hematoxylin and eosin (H&E) staining was performed. Tissue sections of the control and mutant female reproductive tracts collected from postnatal day 4 and day 8 mice, showed no differences in uterine cytoarchitecture or in the amount of endometrial gland tissue. However, on postnatal day 21, less endometrial gland tissue was observed in the mutant female uterus when compared to controls (Fig. 4). The reduced amount of uterine gland tissue in the mutants was found at all later time points examined (Fig. 5).

Open in a separate window

Fig. 4

Histological and immunofluorescent analysis of control and mutant females

A-F, H&E staining (A, B) and immunofluorescent staining using an E-cadherin antibody against epithelial cells and (C-F) at P21 shows less endometrial gland tissue in mutant (B, D, F) females when compared to control (A, C, E) females. G, H, H&E sections of four month-old control (G) and mutant (H) females. In the mutants, endometrial glands were observed inside the myometrium (H) and not in controls (G). Arrowheads indicate endometrial glands. Ic, inner circular; ol, outer longitudinal. Scale bar, 100 μm.

Open in a separate window

Fig. 5

Quantification of uterine glandular tissue in Dicer mutant females

No difference in uterine glandular tissue is observed at postnatal day 8 (P8) (n=3, p= 0.23). However, less endometrial glandular tissue is observed in the mutant uterus at P35 (n=3, p< 0.001), 6 weeks (n=4, p= 0.04), and 4 months (n=4, p= 0.04).

On postnatal day 28, in the oviduct of mutant females, epithelial cells in the ampulla region appeared morphologically normal compared to controls. However, the epithelial cells lining the lumen of the cysts that developed in the isthmus region of mutant females appeared morphologically abnormal (Fig. 3D). Epithelial cells of the expanded isthmus contained a large number of vacuoles inside the cytoplasm, known as intracellular edema, there was a loss of typical columnar epithelium morphology, and epithelial folds were absent (Fig. 3D). Furthermore, inside the lumen of the isthmus a large number of white blood cells clumps were observed.

Tissue sections collected after 6 weeks of age showed the same major differences between control and mutant female reproductive tracts. There was an expanded isthmus with intracellular edema, loss of mucosal folding of the epithelial cells, an immune/inflammatory response inside the oviduct, and lower amount of glandular tissue in the uterus. Interestingly, in older females (> 4 month of age), many times the glands that were present in the mutant uterus were found within the myometrium (Fig. 4G, H).

E-cadherin is a convenient marker expressed in both the LE and GE of the uterus and epithelial component of the oviduct, therefore control and mutant female reproductive tracts were immunostained for E-cadherin and analyzed by fluorescent microscopy. This analysis confirmed the H&E results, showing that mutant females had significantly lower amount of endometrial gland tissue in the uterus compared to controls (Fig. 4C-F). No differences in LE morphology were detected between controls and mutants.

The Dicer mutant females should also express Cre in granulosa cells of the postnatal ovary (Jorgez et al. 2004), thus hormone produced by granulosa cells could potentially be altered, affecting the female reproductive tract. Serum estradiol and progesterone levels were measured. The data showed a slight difference between control and mutant females but was not significantly different between the two groups (Fig. 6). Although there was a high degree of variability in these hormone measurements, we can conclude that estradiol and progesterone levels in the mutants were not dramatically different from controls.

Open in a separate window

Fig. 6

Serum hormone measurements

Estradiol and progesterone levels between 6-8 week-old control (n=4) and mutant (n=7) females. The levels were not statistically different (p > 0.05).

Abnormal function of the conditional mutant oviduct, uterotubal junction, and uterus

In most of the mutant females, oocytes were found inside the cystoid expansions of the oviduct. Normally, oocytes are fertilized inside the ampulla, then peristaltic contractions and the cilia of the epithelial folds move the oocytes into the uterus, where they implant in the endometrium. If this transport mechanism is disrupted or the path is obstructed the embryos will be unable to reach the uterus. In most of the mutant oviducts collected at 6 week, 8 week and 4 months, there was obvious inflammation blocking the path of the oocytes (and embryos if sperm could reach the ampulla). However, mutant females sometimes also had an oviduct that was free of inflammation. Therefore, to test if the oviduct transport mechanism was disrupted in the mutants, known numbers of Affi-Gel Blue affinity chromatography beads that are similar in size to mouse preimplantation embryos (75-150 mm), were transferred into oviducts. For these experiments, 5 week-old females (2 control, 3 mutant) were selected because at this age the inflammatory response is at an early stage that might not interfere with embryo movements. 1-4 days after bead transfer the reproductive tracts were examined. Most of the beads transferred inside control oviducts had moved into the uterus or were located at the uterotubal junction, whereas beads transferred into mutant females remained in a clump at about the same position of the initial transfer inside the oviduct near the ampulla (Fig. 7A-D).

Open in a separate window

Fig. 7

Functional analysis of Dicer mutant female infertility

A-D, Bead transfers into the ampulla of 5 week-old control (A, C) and mutant (B, D) females. 4 days after transfer, by comparing the number of beads transferred and the number of beads observed, we estimate that most of the beads transferred into the control ampulla moved into the uterus (C), however all blue beads transferred into mutant ampulla remained as a cluster in the oviduct (D). E, F, Dye injections into the posterior lumen of the uterine horns of postnatal day 28 control (E) and mutant (F) females. The dye entered the oviduct of the mutant (F) but not the control (E). A, ampulla; is, isthmus; ov, ovary; utj, ovd, oviduct; u, uterus; utj, uterotubal junction.

Impaired ciliogenesis or unidirectional flow created by cilia pulsation was also examined. The oviducts of 5 week-old females, both control and mutant, were dissected and placed in tissue culture medium. A longitudinal incision was made along the wall of the oviduct, and the oviducts were placed in a glass bottomed Petri dish. Both control and mutant oviducts contained cilia, thus ciliogenesis was not impaired (data not shown). To test for unidirectional flow by cilia, a small drop of concentrated dye solution was added and the direction of the flow of dye was recorded using time-lapse imaging. Control and mutant cilia showed unidirectional flow as indicated by the dye, thus cilia of the oviductal epithelium appear to be functional in both control and mutant females (data not shown).

The uterotubal junction regulates embryo movement into the uterus and retrograde flow from the uterus into the oviduct (Newbold et al. 1983).To examine the uterotubal junction, the rostral reproductive tract separated from the cervix of 2 week and 4 month-old females was isolated and a blue dye was injected into the caudal uterine lumen. In control reproductive tracts, the injected dye moved up to the most rostral end of the uterine horn but did not enter the oviduct (Fig. 7E). In mutant reproductive tracts, the dye also flowed up to the most posterior part of the uterine horn, however, it was able to move past the uterotubal junction and enter the oviduct (Fig. 7F). Thus there is a defect in the uterotubal junction of the mutant females such that retrograde flow into the oviduct is not regulated.

Although the infertility of the mutant females is likely due to abnormal embryo transit in the oviduct, we also tested the function of the uterus to support embryo development by embryo transfer. Wild-type blastocyst stage (E3.5) embryos were collected from Swiss matings. Approximately 10-12 embryos were transferred into the most rostral part of the uterus of pseudopregnant (E2.5) surrogate mutant females (n=3). The reproductive tracts of the surrogate mice were dissected 11 days after embryo transfer but none were observed to be pregnant nor were there embryo reabsorption sites.

Female infertility in Dicer mutants

Fertilization, embryo transit, implantation and development require an elaborate synchrony of physiological processes (Guzeloglu-Kayisli et al. 2007). Synchronized epithelial-mesenchymal, epithelial-epithelial interactions, in conjunction with endocrine factors are all essential for a successful pregnancy. Reduction of Dicer expression using a hypomorphic gene trap allele resulted in female mouse infertility, showing impaired corpus luteum angiogenesis that led to low progesterone production (Otsuka et al. 2008). Alterations in systemic hormone levels could lead to the abnormal phenotypes observed in the Amhr2-Cre, Dicer mutant female reproductive tract. However, serum estradiol and progesterone levels in the Dicer mutants although quite variable were not significantly different from controls, suggesting that systemic hormones may not be the cause of the infertility.

The infertility of the Dicer conditional mutant females appears to be due to several oviduct and uterine abnormalities. Cilia in the oviduct are important for the movement of preimplantation embryos to the uterus (Halbert et al. 1976). Ciliogenesis and cilia function in the oviducts of the mutant females did not seem to be impaired, however blue beads transferred inside the ampulla failed to move into the uterus, suggesting a defect in the transport mechanism for oviductal movement of preimplantation embryos. Peristaltic contractions caused by oviductal smooth muscle also contribute to the movements of preimplantation embryos (Maistrello 1971). Interestingly, we observed regions of the mutant oviducts that were deficient in smooth muscle. The Müllerian duct mesenchyme is likely the progenitor tissue of the oviduct smooth muscle and uterine myometrium (Arango et al. 2008). Therefore, Dicer may be important for the generation of miRNAs that regulate smooth muscle differentiation or maintenance. Interestingly, Müllerian duct mesenchyme-specific knockout of beta-catenin results in smaller uteri and a deficiency of myometrium that is replaced by adipose tissue (Arango et al. 2005). These findings are consistent with the idea that miRNAs may be important for maintaining correct Wnt signaling for myometrial differentiation. The abnormalities in oviduct smooth muscle could also explain the development of the cysts because a weakness in the oviduct wall caused by smooth muscle abnormalities could lead to a weakness in the oviduct wall. Although the Dicer mutant females were successfully mated as judged by the presence of a vaginal plug, no pregnancies occurred. This could be due in part to the oviductal inflammation perhaps blocking access of the sperm to the oocytes. Finally, the mutant females had defects in the uterotubal junction. This structure functions to regulate embryo transit into the uterus and retrograde flow from the uterus into the oviduct (Newbold et al. 1983). Abnormalities in these uterotubal junction functions may also contribute to the infertility.

Abnormalities in the uterus of the Dicer mutant females also lead to female infertility. We showed that there are fewer endometrial glands in the mutant uterus. Endometrial glands are required for secretion and transport of nutrients, including leukemia inhibitory factor, necessary for embryo implantation and survival (Gray et al. 2001a). In a sheep uterine gland knockout model, embryos do not elongate and fail to implant (Gray et al. 2001b). Therefore, the reduced amount of endometrial gland tissue observed in our Dicer mice probably hinders embryo implantation and survival. Although the primary cause of the infertility in the mutant females remains to be determined, it is clear that Dicer regulates several aspects of oviduct and uterine development that are essential for fertility.

Histology and immunofluorescence

Reproductive tracts were dissected and fixed in 4% paraformaldehyde overnight at 4°C. Paraffin-embedded tissues were sectioned at 6 μm, and stained with Harris hematoxylin (Statlab) and 0.5% eosin Y (Sigma) (H&E). For tracts at P35 and earlier, the entire tract was serially-sectioned through the oviduct to the cervix. For later ages, the entire oviduct was serially-section and ∼300 μm into the anterior uterus. Histological sampling of the more posterior uterus showed phenotypes similar to the anterior uterus. Paraffin sections were immunostained for E-cadherin, using a mouse anti-E-cadherin monoclonal antibody (BD Biosciences) at a 1:200 dilution. The secondary antibody was an Alexa Fluor goat anti-mouse IgG (Molecular Probes) used at a dilution of 1:800. Slides were mounted with Vectashield mounting media containing DAPI (Vector Laboratories). The amount of endometrial gland tissue in the uterus was estimated by counting the number of glandular duct cross sections in each uterine section. Statistical analysis was performed using a two-sample t test.

Oviduct transfer, dye injections, and uterine transfer

10-15 Affi-Gel Blue affinity chromatography beads (75-150 mm) were transferred into the ampulla of the oviducts of control (n=2) and mutant (n=3) females (Nagy et al., 2003). The females were sacrificed 1 to 4 days after transfer and their reproductive tracts isolated to visualize the movement of the beads through the oviduct. To examine the function of the uterotubal junction, adult reproductive tracts were dissected and 0.25% bromophenol blue was injected using a 23-gauge needle into the posterior uterine horn. Blastocysts generated from Swiss × Swiss mouse crosses were transferred into the uterine horns of E2.5 pseudopregnant control (n=3) and mutant (n=3) females (Nagy et al. 2003). 11 days after embryo transfer the females were examined for pregnancy.

Non-radioactive in situ hybridization

8 mm thick paraffin sections were hybridized with digoxygenin-labeled RNA probes as described (Xu and Wilkinson 1999). A Wnt7a sense probe was used as a control.

We thank Hao Chang and Chuan-Wei Jang for assistance with oviduct transfers, C. Allison Stewart for assistance with in situ hybridization, Grant Orvis for initial guidance, and C. Allison Stewart for helpful comments on the manuscript. We thank Clifford Tabin for Dicer fx mice. Probes for in situ hybridization were obtained from Andrew McMahon, Kanako Hayashi, and the American Type Culture Collection. Supported by National Institutes of Health grant HD30284 and the Ben F. Love Endowment to R.R.B. Veterinary resources were supported by the National Institutes of Health (NIH) Cancer Center Support Grant CA16672.