A critical role for estrogen signaling in penis development

Abstract

Hypospadias, a developmental defect of the penis, is one of the most common congenital malformations in humans. Its incidence has rapidly increased over recent decades, and this has been largely attributed to our increased exposure to endocrine-disrupting chemicals. Penis development is primarily an androgen-driven process; however, estrogen and xenoestrogens are known to affect penis development in both humans and mice. Here, we investigated the role of estrogen in the developing penis. Using a novel penis culture system, we showed that exogenous estrogen directly targets the developing penis in utero to cause hypospadias. In addition, we also uncovered an unexpected endogenous role for estrogen in normal postnatal penis development and showed that a loss of estrogen signaling results in a mild hypospadias phenotype, the most common manifestation of this disease in humans. Our findings demonstrated that both androgen and estrogen signaling are intrinsically required for normal urethral closure. These findings confirmed that penis development is not an entirely androgen-driven process but one in which endogenous estrogen signaling also plays a critical role.—Govers, L. C., Phillips, T. R., Mattiske, D. M., Rashoo, N., Black, J. R., Sinclair, A., Baskin, L. S., Risbridger, G. P., Pask, A. J. A critical role for estrogen signaling in penis development.

In both mice and humans, development of the penis is primarily driven by androgens, which cause male-specific elongation and masculinization of the sexually indifferent genital tubercle (GT) (1–3). As the GT elongates in males, a urethral groove appears on the ventral surface (4). The urethra becomes internalized through the tightly androgen-regulated process of urethral closure, resulting in the urethra being central to the penis and glans and terminating at its distal tip (5–7). Any disruption to androgen signaling during development can prevent penis outgrowth and urethral closure, resulting in hypospadias, the abnormal placement of the urethral opening (8–10).

Defects in penis development are becoming increasingly common, with hypospadias now affecting up to 1 in every 125 live male births (11). Concerningly, the incidence of hypospadias has doubled over recent decades in most industrialized countries. This is not caused by increased reporting and is too rapid to be accounted for by genetic mutations (12, 13), instead implicating an environmental origin of this disease. It is well established that chemicals that block androgens can cause hypospadias (8–10). Interestingly, estrogen and estrogen-mimicking chemicals also impact penis growth and urethral closure, causing hypospadias in mice and humans alike (14–16). We are exposed to many chemicals in everyday life that impact our endocrine-signaling pathways. Some of the most pervasive of these endocrine-disrupting chemicals (EDCs) affect estrogen signaling (xenoestrogens) and include bisphenol A, phthalates, and genistein (17–22). EDCs are candidate contributors to rising hypospadias levels, and recent studies have established a direct link between the levels of estrogenic EDCs in fetal blood and the incidence of hypospadias in humans (22).

How EDCs cause hypospadias is still not clear. Our current understanding suggests that the effects of exogenous estrogen are mediated through the testis, where they can directly suppress androgen output from the Leydig cells (23, 24). This reduces androgen and then causes hypospadias. However, both estrogen receptors (ERs), ERα (encoded by the ESR1 gene) and ERβ (encoded by the ESR2 gene), are broadly expressed in the developing penis of rats (25, 26), mice (8, 27), and humans (28, 29). Furthermore, aromatase, the enzyme that irreversibly converts testosterone to estrogen, is also expressed in the developing rat (30) and human (31) penis, indicating the potential for local estrogen production and a role for estrogen signaling in normal penis development. Given the presence of ERs in the penis, we hypothesized that endogenous estrogen could play a role in normal penis development and that exogenous estrogen or xenoestrogens cause hypospadias by directly affecting developmental processes in the penis itself and not via suppressed testicular androgen output.

It is critical to determine the role of endogenous and exogenous estrogen signaling in penis development. Here, we report a previously undetected hypospadias phenotype in the ERα knockout (αERKO) mouse, demonstrating that endogenous estrogen signaling is critical for patterning and closure of the distal urethra. We next show that exogenous estrogen directly targets the developing penis to cause hypospadias, independent of its effects on testicular androgen output. Our findings highlight that penis development is not an entirely androgen-driven process and provide critical new insights into the endogenous and exogenous role of estrogen in penis development and the etiology of hypospadias. These findings provide a new framework for understanding the developmental and molecular causes of hypospadias.

MATERIALS AND METHODS

Animals

Mice were bred in the animal facility in The School of BioSciences at the The University of Melbourne. Mice were handled and humanely killed according to the guidelines established by the National Health and Medical Research Council of Australia (2004), and protocols were approved by The University of Melbourne Animal Experimentation and Ethics Committees (1413388). αERKO mice were obtained from Professor Gail Risbridger (Medicine, Nursing, and Health Sciences, Monash University) (32). These mice have been previously characterized extensively. Female homozygous mutants are infertile because of anovulation, and homozygous mutant males are infertile because of dysmorphogenesis of the seminiferous tubules and disruptions to spermatogenesis (33). Although no wild-type (WT) protein is detected by Western blot, knockout mice produce low levels of a truncated ERα isoform that is largely nonfunctional (34). Binding assays indicated ∼3% estradiol binding to the truncated isoform in uterine tissue in knockouts with no estradiol binding detected elsewhere. Despite residual amounts of a truncated ER isoform, estrogen treatment fails to rescue the knockout phenotype or induce known estrogen-responsive uterine markers, such as increased DNA synthesis, and transcription of the progesterone receptor, lactoferrin, and glucose-6-phosphate dehydrogenase genes. Furthermore, females display estrogen insensitivity, as indicated by significantly elevated estradiol (34).

αERKO mice and WT littermates were generated by mating C57BL/6J heterozygous mice expressing the disrupted ERα (32). For endogenous estrogen experiments, αERKO mice were kept on either standard mouse feed (Mouse Diet GR2, nonirradiated; Ridley, Melbourne, VIC, Australia) or a soy-free feed (SF06-053) supplied by Specialty Feeds (Glen Forrest, WA, Australia) to reduce the phytoestrogen content of the diet. Mouse embryos were collected from timed matings, with noon of the day on which the mating plug was observed designated as embryonic day (E)0.5 post coitum. For diethylstilbestrol (DES)-treated animals, pregnant dams were weighed and injected subcutaneously on d 12, 14, 16, and 18 of gestation with DES (Merck, Sydney, NSW, Australia) [100 ng/g body weight in ∼5 μl in sesame oil vehicle (35)]. For the control groups, CD-1 and C57BL/6 dams were injected with sesame oil (∼5 μl/g) following the methods of Mahawong et al. (20).

Tissue collection

To define the role of endogenous estrogen in urethral closure, penises for section histology were collected from age-matched αERKO and WT mice. Penises collected from time points covering prepuberty [postnatal day (P)0-P19] and puberty and early adulthood (P20-P36) were all kept internalized during dissection and a value of n = 1 was used for each genotype. Adult penises (P41 and older) were collected either with the external prepuce intact or with the penis externalized and the external prepuce removed for a total of 18 αERKO homozygotes and 19 WT. Penises for micro–computed tomography (micro-CT) analysis were collected when fully matured at P51 [n = 8 (WT) and n = 7 (αERKO)] and externalized during dissection to show the urethral meatus.

To define the role of exogenous estrogen, we used FVB/NJ male GTs collected at E14.5 for all culture experiments. Pregnant females were maintained on a standard diet before collection. Tissue collection included the dissection of the phallic cloaca and underlying mesenchyme to ensure the distal aspect of the urorectal septum was included.

Histologic and immunohistochemical analysis

αERKO and WT tissues for section histology were fixed overnight in 4% paraformaldehyde in PBS, dehydrated with ethanol, and processed in paraffin wax. Adult specimens were stored in 0.5 M EDTA for 4 d to decalcify the os penis before processing. Specimens were serially sectioned at 7 μm and stained with hematoxylin and eosin according to standard methods.

Micro-CT analysis

After dissection, penises were fixed overnight in 4% paraformaldehyde in PBS. Samples were washed 3 times for 30 min each in PBS at 4°C, and the samples were stored in 70% ethanol until scanning.

Tissue samples were mounted in small (0.2 ml) vials of ethanol and placed within narrow plastic tubing to prevent any sample movement during scanning. Micro-CT scanning was performed with a Phoenix Nanotom M (GE Healthcare, Waukesha, WI, USA), operated using XS Control and Phoenix Datos|x acquisition and reconstruction software (GE Healthcare). Samples were scanned for 30 min at a resolution ranging from 4.636 to 8.889 m with X-ray energy ranging from 40 to 50 kV and 290–880 A depending upon the number of samples mounted for scanning. A 0.2-mm aluminum filter was used in the higher-resolution scans of single specimens to reduce beam hardening.

Reconstructed volume data were processed using Avizo (Thermo Fisher Scientific, Waltham, MA, USA), with os penis bone, male urogenital mating protuberance (MUMP) tip, and ventral cleft (the urethral meatus) being segmented separately. The os penis bone length, MUMP length, and the distance from the tip of the MUMP to the proximal base of the ventral cleft (the proximal urethral meatus) were recorded in Avizo using a tool measuring the distance between 2 markers on the surface of segmented data. The os penis bone volumes were also determined from the segmented bone.

GT culture

Male GTs were cultured for 96 h in a hanging drop culture (36). Individual GTs were suspended in 50 µl of BGJb medium (Thermo Fisher Scientific) supplemented with 0.5 ml (10 mg/ml) ascorbic acid (MilliporeSigma, Burlington, MA, USA), 0.5 mg bovine insulin (MilliporeSigma), and 0.5 ml (10 mg/ml) penicillin-streptomycin on the lids of 16-well culture plates (MilliporeSigma). The drops were inverted to allow a free suspension of the tissue. GTs were cultured in carbogen (95% O₂) at 37°C and exposed to either 100 nM 4,5α-dihydrotestosterone (DHT; MilliporeSigma) alone (control treatment) to enable normal urethral closure or 100 nM DHT (MilliporeSigma) with the addition of 100 nM 17α-ethynylestradiol (EE2; MilliporeSigma) for estrogen treatment. Concentrations of both DHT and EE2 were empirically determined. One hundred nanomolars of DHT caused consistent masculinization over the culture period, whereas 100 nM EE2 caused a consistent block in male urethra formation. Media were replaced every 24 h throughout the 96-h culture period. All data were analyzed from 3 independent replicates of cultures with GTs derived from different litters. DHT was used in preference to testosterone because it cannot be converted to estrogen, ensuring control of estrogen availability in the cultures.

Morphologic and histologic analysis of cultured GTs

After 96 h in culture, male GTs were photographed using a Nikon Digital Sight DS-U3 (Nikon, Tokyo, Japan) and all images were processed through NIS Elements Analysis D v.4.300.00 64-bit software (Nikon). GTs were fixed for 12 h in Bouin’s Solution (MilliporeSigma) at 4°C with rotation and washed for 48 h in 70% ethanol. Tissues were processed into wax and transversely serially sectioned at 7 µm and stained with hematoxylin (MilliporeSigma) and eosin Y (MilliporeSigma) according to standard methods. To assess the extent of urethral closure, the number of sections which had an internalized urethra was divided by the total length of the GT measured from the distal tip of the glans to the proximal base of the GT. A 2-tailed Student’s t test assuming equal variance was performed comparing the placement of the mature urethral opening in GTs cultured in DHT alone as compared with GTs cultured in the presence of both DHT and estrogen.

Caspase-3 immunohistochemistry

Transverse sections from both DHT-control and estrogen-treated samples were stained with rabbit polyclonal anti–caspase-3 (9661S; Cell Signaling Technology, Danvers, MA, USA) at 1:500 overnight at 4°C. Following washes, slides were incubated with Alexa Fluor 544 donkey anti-rabbit IgG (A31572; Thermo Fisher Scientific) at a 1:1000 dilution. Sections were mounted with Fluoroshield Mounting Medium with DAPI (ab014139; Abcam, Cambridge, MA, USA). Sections were visualized using the Olympus DP70 Fluorescence Microscope (Olympus, Tokyo, Japan) and all images were processed with Adobe Illustrator software (Adobe, San Jose, CA, USA) for analysis.

RESULTS

Adult αERKO penises display distal hypospadias

The mouse urethral meatus terminates in a Y-shaped opening at the distal region of the mouse glans with a patent ventral cleft (4, 37, 38) (Fig. 1A–D, black arrowhead). The tubularized urethra lies central to the glans and ventral to the os penis bone, terminating at the MUMP ridge distal to the tip of the os penis (Fig. 1A, B).

Figure 1.

Differences in the placement of the urethral meatus (UM; ventral cleft) in a WT, αERKO, DES-treated mouse penis, and juvenile human penis with mild hypospadias. A) Sagittal cross section of the adult mouse penis (P57) with the external prepuce removed. The UM is directly underneath the MUMP (M), and the os penis (O) is more proximal than the UM. B–D) Transverse sections of adult mouse penises. In the WT mouse penis (P90), the UM (arrowhead) is present in the distal section (top panel) as a ventral cleft extending from the urethra (U) beneath the MUMP. In the more proximal section (bottom panel) showing the most distal tip of the os penis, the urethra is tubularized and internal to the penis (B). In a DES-treated adult penis (P57), the UM (black arrowhead) is present in the proximal section and thus opens more proximally than in the WT penis (C). An identical phenotype can also be seen in the αERKO penis with an open ventral cleft in proximal penis sections (P46) (D). E) Juvenile human penis with mild hypospadias, with the white arrow indicating the location of the UM. Scale bar, 500 µm.

In DES-exposed mice, the urethral meatus was more proximal than in the WT penis, as evidenced by the presence of the os penis bone in the same transverse section plane as a ventrally open urethra [Fig. 1C; Described in the Materials and Methods section (35)]. This distal (mild) hypospadias phenotype (35) is identical to that seen in αERKO (Fig. 1D) mice, with an 89% incidence rate (n = 18, αERKO). In contrast, only a 5% incidence of this phenotype was observed in WT mice (n = 20).

Loss of ERα causes accelerated delamination in the postnatal penis

To establish how a loss of endogenous estrogen affected urethral closure, we determined when in development homozygous αERKO mice first showed a morphologic difference to WT. Penises were sectioned in vivo at time points ranging from P0 (which were normal; unpublished results) through to adulthood (P41 and later). The penis was still undergoing masculinization from P0 to 32 and the mature urethral meatus on the glans was not fully developed until after P32 in the WT (Fig. 2C). At P12, the external prepuce is contiguous and fused to the internal prepuce, developing glans in the proximal regions, and is fused to the MUMP in distal sections (Fig. 2A). Maturation of the penis and the formation of the mature urethral meatus (which terminates in the glans) involves a series of delamination events to separate the MUMP from the external prepuce, the glans from the internal prepuce, and, finally, the internal prepuce from the external prepuce. All these events must be coordinated and are critical for mature urethral meatus formation and erection and protraction of the penis. The epithelial boundaries that will eventually delaminate to separate the MUMP from the glans (Fig. 2A, distal) and the glans from the internal and external prepuce (Fig. 2A, mid and proximal) can be clearly seen as early as P12.

Figure 2.

Age-matched WT and αERKO postnatal mouse penises showing temporal differences in delamination and ventral cleft formation. A) At P12, the MUMP (M) of both the WT and αERKO is surrounded by connective tissue (C) at the distal tip of the penis. In more proximal sections showing glans (G), the epithelial border (EB) that will later delaminate to form the MUMP ridge groove can be seen on the dorsal side of the glans. On the ventral side of the MUMP is the tubular urethra (black arrowhead). B) By P20, clear delamination (D) is seen in the αERKO, separating the MUMP from the external prepuce in distal sections as well as in midsections separating the glans (G) from the external prepuce. C) By P32, both the αERKO and the WT MUMP have completely separated from the external prepuce, such that the MUMP tip is not surrounded by connective tissue but rather by the preputial space (PS) through which the glans penis normally extends in erection. In more proximal sections of the WT, this delamination is not yet complete, because connective tissue remains between the glans and the external prepuce; however, in the αERKO, the glans is fully separated, even in more proximal sections. Black arrowhead denotes the urethra. Scale bar, 500 µm.

At P12, urine exits the penis via a distal lumen (Fig. 2A, black arrowhead) formed from the external prepuce and contiguous MUMP (Fig. 2A). This is an entirely different tissue composition and location to the mature urethral meatus, which terminates in the glans.

The first phenotypic difference was an earlier onset of delamination by cornification in the homozygous αERKO compared with WT mice (Fig. 2B). By P20, extensive delamination could be seen in the distal sections, with the MUMP almost completely separated from the external prepuce in homozygous αERKO mice, in contrast with the structures remaining fused in WT siblings. Delamination was also evident in the homozygous αERKO midsections on the dorsal aspect of the glans, which had begun separating from the prepuce. In contrast, all structures remained continuous in the distal, mid, and proximal sections of the WT mice, with no signs of delamination.

By P32, the accelerated delamination in the αERKO continued with complete delamination of the MUMP (Fig. 2C, distal and mid) and the glans from the external prepuce (Fig. 2C, proximal). This is again in contrast with the WT penis, wherein only the MUMP has begun delamination at this point in the distal aspect of the penis (Fig. 2C, distal) and the glans is still contiguous with the internal and external prepuce (Fig. 2C, mid and proximal).

In the WT P32 penis, the tissues that form mature urethral meatus in the glans are visible but have not yet delaminated to form the ventral cleft and terminal point of urine exit from the penis. In contrast, the αERKO ventral cleft is evident by P25, albeit not yet completely delaminated (Fig. 3), and forms the urethral meatus with a completely delaminated ventral cleft by P32 (Fig. 3; P32-36). In WT mice, the fully developed ventral cleft was not seen until P36 (Fig. 3), indicating accelerated delamination in the mutant. Hypospadias, as defined by a urethral meatus (formed from the delaminated ventral cleft) at the same level as the distal aspect of the os penis bone, only became evident in αERKO once all structures had fully matured.

Figure 3.

WT and αERKO postnatal mouse penises from P20-72 showing the delamination of the distal urethra and of the glans and prepuce. The ventral cleft (VC) can first be seen directly ventral to the urethral flaps in the P20 penis. At P32 in the WT mouse, the VC is still completely fused. By P36, the WT VC is completely delaminated, and the epithelial borders are no longer fused. The glans has also separated from the prepuce at this stage to form the preputial space (PS). In the P72 WT, the urethra is fully internalized, proximal to the tip of the os penis bone at the distal aspect of the penis. Although the VC persists more distally (Fig. 1B), the ventral mesenchyme (VM) forms a confluence of tissue across the midline, ventral to the urethra (U) at the tip of the os penis bone. In the αERKO P20, accelerated delamination (D) of the glans from the prepuce was observed in proximal sections. By P25, the αERKO VC is present but fused. By P36, the αERKO VC has completely delaminated. P72 shows a typical αERKO adult in which the VC remains unfused, proximal to the tip of the os penis bone (B) indicative of a mild hypospadias phenotype. Scale bar, 500 µm.

Loss of ERα affects development of the os penis and penis size

Our initial characterization of hypospadias in the αERKO mouse was based on the location of the urethral meatus relative to the os penis bone (35). However, estrogen is known to affect bone remodeling, and estrogen deficiency results in a higher bone turnover (39). We therefore investigated if the os penis itself was affected by ERα deficiency. We used micro-CT to measure the os penis bone length and volume, MUMP length, and the distance from the tip of the MUMP to the most proximal opening of the ventral cleft (the urethral meatus) in both mutant and WT litter mates. A heat map of the initial surface rendering of the WT and αERKO penis (Fig. 4A) shows an overall reduction in volume of the αERKO penis. A video showing the isolation of the os penis bones from the rendered 3-dimensional images via increasing the density threshold can be seen in Supplemental Data S1. The αERKO os penis has a significantly lower total volume than the WT os penis (Fig. 4B), and the os penis bone is slightly but significantly shorter (Fig. 4C). The level of bone porosity was also greater in the αERKO os penis (which was evident after segmenting the bone based on density) and contributes to its overall reduced volume when compared with WT (Fig. 4D). In contrast, the cartilaginous MUMP, as measured from the micro-CT, was not different, either in volume or length, between the WT and αERKO mice (Fig. 4E).

Figure 4.

Micro-CT analysis of WT and αERKO adult penis surface rendering and bone. A) Heat map of the micro-CT surface rendering of a WT and αERKO (KO) adult (P51) penis shown in sagittal view. B) The αERKO penis is significantly smaller in all aspects to the WT and also has a significantly smaller bone volume. C) Isolated os penis bones from WT and αERKO micro-CT scans showing reduced length of the αERKO os penis bone. D) Cross section of the os penis bones from WT and αERKO micro-CT scans shows increased porosity (shown by black within the os penis bone), as compared with WT. Although αERKO mutants did not have a reduction in MUMP length (E), there was a significantly longer distance from the MUMP tip to the urethral opening (F), leading to a mild hypospadias phenotype. *P < 0.05, ***P < 0.0001.

The opening of the ventral cleft was significantly more proximal (about 20% of total penis length more proximal than in WT) in homozygous αERKO mice (Fig. 4F). Taken together, these data demonstrated that estrogen affects penis size and os penis development; however, the longer distance to the urethral opening in the αERKO is not caused by the varying bone length and is a true proximal misplacement of the normal urethral opening.

Estrogen directly affects early stages of penis development

Having established that estrogen plays an endogenous role in urethral closure, we next examined if exogenous estrogen could also affect urethral closure through aberrant ER activation in the developing penis itself. GTs exposed to DHT in culture masculinized after 96 h showed an internalized urethra reflecting development in vivo (Figs. 5A and 6C). GTs cultured with DHT elongated, and the preputial swellings proliferated to form a prepuce (Fig. 5A). However, GTs cultured with DHT and EE2 showed disrupted morphologic development with a significantly (P < 0.05) shorter length (933 ± 29.74 µm) in DHT + EE2–exposed (n = 10) compared with 1034 ± 23.00 µm in DHT (n = 11; Fig. 5B). Additionally, GTs cultured in DHT and EE2 had a complete lack of ventral preputial development in all samples (Fig. 5A).

Figure 5.

Morphologic comparisons of GTs cultured in DHT and DHT + EE2. A) Gross morphologic comparisons of GTs cultured in DHT and DHT + EE2 for a total period of 96 h. The urethral-plate epithelium is marked by the arrowhead, with the preputial swellings (P) flanking the glans (G) of the GT. The asterisk indicates an exposed ventral surface in the DHT + EE2–treated GT after a period of 96 h in culture. Scale bar, 600 µm. B) The total length of GTs cultured in the presence of DHT (n = 11) and DHT + EE2 (n = 10). Tissue was measured from the distal tip of the glans to the proximal base of the GT, where it attaches to the body wall. *P < 0.05.

Figure 6.

Comparison of urethral internalization and quality of GTs cultured in DHT and DHT + EE2. A) Diagrammatic masculinized male GT signifying how the level of urethral internalization was determined (percentage from the proximal base of the GT), where it attaches to the body wall to the distal tip of the glans. B) The mean percentage of internalized urethra relevant to the total length of the GT for GTs cultured in DHT and DHT + EE2 for 96 h. C) Hematoxylin and eosin Y transverse section histology at the distal glans, middle prepuce region, and proximal basal region of GTs cultured in DHT only and DHT + EE2. The internalized urethra (IU) in the DHT-treated sample is indicated in the proximal basal section. The externalized urethral-plate epithelium in the DHT + EE2–treated GTs, as well as the invaginated urethral-plate epithelium in the DHT-treated GTs, is marked by the arrowhead. D) Hematoxylin and eosin Y transverse section histology of WT male E17.5 distal GT, where the clearly invaginated urethral-plate epithelium appears in the ventral region of the GT. E) Representative image of cleaved caspace-3 immunofluorescence staining, marking low levels of apoptosis in the GT after a culture period of 96 h. Scale bar, 500 µm. ***P < 0.0001.

Estrogen exposure prevents urethral closure, resulting in hypospadias

Serial sectioning in the transverse plane was used to determine the proportion of internalized urethra over the total length of the GT after 96 h in culture (Fig. 6A–B). GTs cultured in DHT had a significantly (P < 0.0001) greater mean proportion of internalized urethra of 28% (n = 8) of total GT length than GTs cultured in DHT + EE2, which had a mean proportion of internalized urethra of 12.4% (n = 9) of total GT length (Fig. 6B).

The urethral plate can be seen as an invaginated structure in the distal region of GTs cultured in DHT only (distal 50% of total GT length) (Fig. 6C). This is comparable to the normal morphology observed in vivo (Fig. 6D), indicating masculinization of the urethral plate in our in vitro system. In contrast, GTs cultured in DHT and EE2 did not show invagination of the urethral plate in the distal GT, and instead the urethral plate was continuous with and almost indistinguishable from the surface epithelium (Fig. 6C).

Cleaved caspase-3 staining was minimal and not different between DHT and DHT + EE2–exposed GTs, indicating long-term survival of the cultured tissues (Fig. 6E).

DISCUSSION

Our increased exposure to EDCs is concordant with the dramatic increase in incidence of hypospadias in industrialized countries. Defining how hormones regulate penis patterning and development is a critical first step in understanding hypospadias. Here, we describe the direct effects of both endogenous and exogenous estrogen in penis development and urethral closure. We describe an endogenous role for estrogen in urethral closure and show that a loss of estrogen signaling causes a distal (mild) hypospadias phenotype during postnatal development. Hypospadias was caused by accelerated rates of delamination in the estrogen-deficient penis. This not only gives important insights into the development of hypospadias but also provides a detailed description of the delamination events that occur during penis development in mice. In addition, using a novel GT culture system, we show that exogenous estrogen signaling directly inhibits urethral closure during early stages of penis development. Together, these data provide significant insight into the role of estrogen in normal development and in the etiology of hypospadias. Our data show that estrogen is required for urethral closure and that either too much estrogen during early urethra formation or too little during delamination events results in hypospadias.

Coordinated delamination events are critical for penis development in mice and humans

Much of the masculinization of the mouse penis occurs after birth (40). Here we define the development of the mature glans and prepuce structures through postnatal delamination, presenting the developmental events leading to the formation of the mature urethral opening in mice.

Delamination of the penile structures (including MUMP, glans, and ventral cleft) from the prepuce occurs in a distal-to-proximal direction and is complete by P36 in WT mice. Although the epithelial surfaces of the ventral cleft can be seen as early as P12, it does not serve as the urethral meatus until delamination of the glans is complete at ∼P36. Prior to delamination of the glans from the prepuce, the opening of the urethra is formed by the preputial space. An identical process of delamination occurs in humans to separate the glans from the prepuce through delamination (41). As in mice, this process is not complete in humans until the postnatal period and the urethral opening is similarly formed from the fused glans and preputial space (41). Although a role for estrogen in coordinating delamination of the foreskin has not been described in humans, estrogen plays an integral role in the cornification and opening of the analogous vaginal epithelium (42).

Endogenous estrogen signaling is required for distal urethral closure and normal penis development

The process of cornification in the αERKO penis occurs at an earlier age than the WT and results in a fully developed ventral cleft by P32, as compared with P36 when the mature ventral cleft can be seen in the WT penis. However, the hypospadias phenotype (as described by the urethral opening proximal to the os penis bone) could not be measured until the os penis had fully matured by P41. This measure of hypospadias was confounded by the finding that a loss of ERα also significantly affected os penis bone growth and density and the overall penis size, showing that endogenous estrogen signaling through ERα is required for many aspects of penis development. Consequently, we assessed the hypospadias independent of the os penis by defining the location of the urethral opening directly based on other landmarks. Even when controlling for the differences in size and os penis length, the urethral meatus was clearly and significantly more proximal in the αERKO mice than in WT. These data confirm that estrogen signaling is critical for the development of the distal urethra and suggest that the accelerated rate of delamination disrupts tissue patterning in early postnatal penis development, resulting in the mild hypospadias phenotype. Therefore, precisely timed delamination events are critical for the formation of the distal urethra, and disruptions are a likely cause of mild hypospadias.

A conserved role for estrogen in human and mouse mild hypospadias phenotypes

Both ERs and aromatase are expressed in the developing human penis (39, 40, 42), as in mice. Multiple human genome–wide association and linkage studies have identified mutations in both ERs (ESR1 and ESR2; ERα and ERβ, respectively) associated with an increased risk of hypospadias in humans (43–49). Several of these mutations are associated with repressed binding or signaling of the ERs (47), whereas other studies have noted a reduction in ER expression in the foreskin of patients with hypospadias (50). Together, these data suggest a similar causative association between a loss of ER signaling and hypospadias in humans, as we report here in mice.

Exogenous estrogen directly affects early stages of penis development and causes hypospadias

Having established an endogenous functional role of estrogen in urethral development during delamination events, we next examined if aberrant estrogen signaling could also directly affect urethral closure. The current dogma suggests that excess estrogenic exposures to the early embryo affect androgen output from the testis (23, 24), leading to hypospadias. However, the presence of functional ERs in the penis suggests that estrogenic endocrine disruptors could be directly affecting penis development. To examine this hypothesis, we developed a GT culture system that enables the process of urethral closure to occur in vitro in a completely hormonally defined environment. Previous studies have attempted culturing GTs but did not achieve normal masculinization and urethral closure, limiting their usage. Previous GT cultures consistently displayed high levels of apoptosis, abnormal GT outgrowth, disrupted prepuce development, and ambiguous urethral-plate epithelium development (51, 52). We developed a culture system that promoted both masculinization and urethral closure, therefore making it possible to analyze the direct effects of estrogen on penis development in vitro. Using DHT as the sole androgen source in our cultures eliminated any local estrogen production in the cultured tissue, whereas testosterone can be converted to estrogen through the action of aromatase. GTs cultured in the presence of DHT developed normally and, importantly, displayed male urethral closure. This is consistent with the phenotype in the αERKO mouse, which shows normal male development of the GT up to P20 (Fig. 2). However, GTs cultured in the presence of DHT and estrogen were smaller, developed hypospadias, and lacked ventral fusion of the prepuce. These phenotypes are consistent with the known effects of in vivo estrogen exposure to male embryos in both mice and humans, in each case resulting in hypospadias and a “hood” foreskin structure (8, 53). Our data demonstrate that these effects are directly manifested by aberrant estrogen signaling in the penis itself and are not caused by altered androgen output from the developing testis. The data presented in this study unequivocally demonstrate a direct and negative effect of exogenous estrogen on the masculinization of the GT.

Both estrogen excess and deficiency affect penis development similarly

It is striking that both a loss of estrogen signaling (through ERα deficiency) and excess of estrogen signaling (through DES in vivo or EE2 in vitro) cause similar defects to penis development. Both a loss and an increase in estrogen signaling result in reduced penis size and mild, distal hypospadias phenotype. Importantly, this correlates with the most common disease phenotype seen in humans [∼70% of all human hypospadias cases are considered mild or distal (54)]. This phenotype is also without any genetic diagnoses in most instances. Our data strongly implicate that altered estrogen signaling could be a factor in mild hypospadias.

Defining the role of endocrine disruptors in hypospadias

Endocrine disruptor exposures have now been unequivocally linked to increased hypospadias rates across several studies (17, 55–59). Given our findings that endogenous estrogen signaling through ERα in the penis is critical for distal urethral closure in mice and likely in humans too, future studies on the environmental causes of hypospadias should include anti-estrogenic as well as estrogen-mimicking EDCs. Our GT culture system provides a high-throughput method of analyzing the direct effects of EDCs on the phallus and in the development of hypospadias.

CONCLUSIONS

Our data demonstrate that endogenous estrogen is required for normal penile development and patterning of the distal urethra. A loss of endogenous estrogen signaling during delamination results in a mild hypospadias phenotype, which is the most common manifestation of this disease in humans. Furthermore, we show that mild hypospadias induced by exogenous estrogen during embryonic development is caused by the direct impact of excess estrogen in the penis itself and not by its impacts on testis development affecting androgen output. Therefore, hypospadias can be caused by EDCs that block or overstimulate the ERs as well as those that affect androgen signaling. The development of our novel GT culture approach provides a new system to examine the direct effect of EDCs on urethral closure in a rapid, high-throughput system.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

αERKO

ERα knockout

DES

diethylstilbestrol

DHT

4,5α-dihydrotestosterone

EDC

endocrine-disrupting chemical

EE2

17α-ethynylestradiol

ER

estrogen receptor

GT

genital tubercle

micro-CT

micro–computed tomography

MUMP

male urogenital mating protuberance

WT

wild type

AUTHOR CONTRIBUTIONS

L. C. Govers, T. R. Phillips, D. M. Mattiske, and A. J. Pask performed and designed the experiments; N. Rashoo and J. R. Black performed experiments; A. Sinclair, L. S. Baskin, and G. P. Risbridger designed the experiments; and all authors contributed to the writing and editing of the manuscript.