Main

Anorectal malformations include a broad spectrum of congenital defects that frequently require acute surgical treatment in the neonatal period, because of intestinal obstruction. One of the most common anomalies is imperforate anus, a condition associated with significant chronic morbidity, especially fecal incontinence. The most severe forms of anorectal malformation are the high type, which usually involve an abnormal connection between the intestine and the genitourinary tract, and a persistent cloaca, in which distal intestinal and genitourinary tracts remain in a common channel (1). Patients with the high type or cloaca have more significant, serious long-term medical problems, for example, dysfunctional defecation and gender assignment issues. Despite its clinical relevance, the causes of anorectal malformations remain poorly understood (2,3). Accordingly, it is important to reveal the causes of anorectal malformations to potentially improve clinical management.

The process of anal development has been shown based on anatomical and clinical observations (4–6). The molecular mechanisms underlying anal development and the pathogenesis of anorectal atresia have been studied in several model animals. Studies using knockout mice for Shh and Hox cluster genes have suggested that these gene products are essential for anorectal development. Shh is a secreted signaling molecule that mediates proliferation, survival, and differentiation during the development of various organs. Homozygous mutants of Shh exhibit massive defects in multiple organs including a persistent cloaca, complete agenesis of the vertebral column, heart-looping defects, tracheoesophageal fistula, and loss of distal limb structures (7–9). Concordantly, mutant mice lacking two zinc finger transcription factors participating in Shh signaling, Gli2 or Gli3, exhibit imperforate anus with rectourethral fistula and anal stenosis. Furthermore, persistent cloaca is also observed in Gli2+/−Gli3−/−, Gli2−/−Gli3+/−, and Gli2−/−Gli3−/− mice (10). Therefore, the Shh signaling cascade is essential for cloacal development. HNF-3β has been shown to regulate Shh expression, and its expression of HNF-3β in turn is regulated by Shh (11). Ptc1 encodes a transmembrane protein that functions as a receptor for Shh, and its expression is transcriptionally activated by Shh signaling (12,13). Hox cluster genes are also expressed during cloacal and anal development (14,15). Mutations in Hoxa13 cause hand-foot–genital syndrome, an autosomal dominant disorder that profoundly affects the development of limb and genitourinary structures including the external genitalia, uterus, bladder, ureter, cervix, and rectum (16). In fetuses lacking both Hoxa13 and Hoxd13, the cloacal cavity does not separate into a urogenital sinus and presumptive rectum, and subsequent development of the genital bud does not occur at all (17).

These observations suggest that the molecular circuitry operating in organogenetic processes in the kidney, lung, and intestine may also be active during anorectal development. Indeed, Bmp4 and Wnt5a are expressed in the mesenchymal cells surrounding the cloacal plate, whereas Fgf8 and Bmp7 are expressed in the cloacal plate itself (18–22). However, the functions of these signaling molecules during cloacal development have not yet been elucidated.

To address the role of these morphogenetic proteins during anorectal development, it is informative to investigate the expression of their genes during the pathogenesis of anorectal anomalies. We therefore attempted to address these issues by using two model animals of anorectal anomalies with distinct etiologies and different genetic processes, namely Danforth's short tail (Sd) and all-trans retinoic acid (ATRA)-treated mice. Sd is a spontaneous semidominant mutation on mouse chromosome 2 affecting the axial skeleton and the urogenital system. Sd heterozygotes and homozygotes display a broad range of abnormalities in the vertebral column and urogenital system. The anal opening is missing, and the embryonic cloaca persists in Sd homozygotes (23,24). ATRA administered to pregnant females at 9.5 days postcoitus (dpc) has been shown to induce anorectal malformations by toxic activity (25,26). Retinoic acid (RA) is essential for both normal embryonic development and correct tissue differentiation (27). The RA signal is transduced by RA receptors and retinoid-X receptors, which heterodimerize and bind to the upstream sequences of RA-responsive genes (28). However, the RA-responsive genes involved in anorectal development are mostly unknown although Hox cluster genes are possible candidates.

In this study, we compared the pathogenetic processes of anorectal malformation in two mouse models and further addressed the expression of these morphogenetic genes. The causes of anorectal malformation in the two mouse models are different. Therefore, any common signaling processes between these mice could help in the identification of genes responsible for anorectal malformation phenotypes.

METHODS

Mice.

Wild-type ICR mice were mated overnight, and females were examined the following morning for the presence of a vaginal plug. If the plug was found in the noon, the dpc was considered 0.5. Pregnant females at 9.5 dpc were dosed by oral gavage with ATRA (100 mg/kg) suspended in sesame oil. Male and female Sd heterozygotes were mated and embryos in Sd homozygotes were obtained. Embryos at 9.5, 10.5, 12.5, and 17.5 dpc were used. All experiments were performed at least three times. All procedures were conducted according to protocols and guidelines approved by the RIKEN Yokohama Institute.

Histologic analysis.

Tissues were dissected out and fixed overnight at 4°C in 4% paraformaldehyde. The fixed tissues were then dehydrated, embedded in paraffin wax, and cut into 7-μm sections. Next, the sections were dewaxed, rehydrated, and stained with hematoxylin and eosin.

Analysis of cell proliferation and cell death.

Pregnant females were injected i.p. at 10.5 dpc with 100 mg/kg body weight of 5-bromodeoxyuridine (BrdU) (Sigma Chemical Co.) in PBS. Two hours after injection the tissue was dissected and fixed. Fixed samples were sectioned and stained as described previously (29). TUNEL analysis for detection of apoptotic cells in the cloacal plate was performed using an in situ apoptosis detection kit (Roche).

In situ hybridization for gene expression.

Each embryo was pooled and rehydrated through a graded series of methanol. Digoxigenin-11-UTP-labeled riboprobes were prepared by in vitro transcription. Whole-mount in situ hybridization was performed as described previously with slight modifications (30). After hybridization, embryos were cleared and photographed.

RESULTS

Cloacal and anorectal development in wild-type, Sd homozygotes, and ATRA-treated mice.

We first histologically examined cloacal and anorectal development. Sagittal sections through the cloacal or anorectal region were prepared at 9.5, 10.5, 12.5, and 17.5 dpc and stained with hematoxylin and eosin. In the wild type, anal development is anatomically subdivided into at least three distinct phases. We defined the cloacal plate as consisting of the proximal and distal parts, and the distal part was subdivided into ventral and dorsal parts (Fig. 1L, M, N, and O). In the first, the cloacal plate is formed from the surface ectoderm and the hindgut endoderm at the caudal region of the allantois base around 9.5 dpc. By 12.5 dpc, the cloacal cavity is separated into the anorectum and the urogenital sinus by the urorectal septum, and further subdivided into functionally proximal and distal regions of the cloacal plate. Finally, the anus opens at the proximal part of the cloacal plate (Fig. 1A, C, F, and I) (5).

Figure 1
figure 1

Mid-sagittal sections of the cloaca or the genital tubercle were obtained and stained with hematoxylin and eosin at 9.5 (A, B), 10.5 (C–E), 12.5 (F–H), and 17.5 dpc (I–K). The images are pattern diagrams from a histologic study of the midsagittal and posterior view in the wild type (L–O). Normal development of the cloacal plate starts and the ectoderm and endoderm are defined in the cloacal plate at 9.5 dpc (A, L). From 10.5 dpc to 12.5 dpc, the cloacal plate grows to the proximal part of the genital tubercle. The cloacal plate consists of proximal and distal parts, with the distal part divided into ventral and dorsal parts. Around the same time, the urorectal septum grows along the anteroposterior axis to the proximal part of the cloacal plate (C, F, M, N). At 12.5 dpc, the distal part of the cloacal plate and the tip of the urorectal septum fuse, and the anus opens. The hindgut differentiates into the rectum, and the urogenital sinus differentiates into the bladder (I, O). At 9.5 dpc, the cloacal plate in Sd homozygotes grew normally compared with the wild type (B), but at 10.5 dpc, it did not grow (D). At 12.5 dpc, the distal part was present at the tip of the genital tubercle, but the proximal part did not grow. Therefore, the urorectal septum grew normally but the anus did not open (G). At 17.5 dpc, in Sd homozygotes, the anorectal malformation was classified as a rectovesicle or rectocloacal fistula (J). At 10.5 dpc in ATRA-treated mice, the cloacal plate grew normally compared with the wild type (E), but at 12.5 dpc, it did not grow (H). As in Sd homozygotes, at 12.5 dpc, the distal part present at the tip of the genital tubercle grew but the proximal part did not. Therefore, the urorectal septum grew normally but the anus did not open. At 17.5 dpc in ATRA-treated mice, the anorectal malformation was classified as a rectovesical or rectocloacal fistula (K). *, cloacal cavity; gt, genital tubercle; urs, urorectal septum; hg, hindgut; r, rectum; black bracket, cloacal plate; black arrow head, fistula; scale bars, 100 μm (A–H) and 500 μm (I–K); magnification, ×200 (A, B), ×100 (C–E), ×50 (F–H), and ×20 (I–K).

Full size image

We next examined anorectal development in the two model mice. In Sd homozygote compared with wild type, at 9.5 dpc, although the cloacal plate was formed, layers of ectodermal and endodermal cells were not regularly aligned (Fig. 1B). At 10.5 dpc, the morphologies of the cloaca were quite different between the two model mice. Extension of the cloacal plate along with the anteroposterior axis was significantly affected in Sd homozygotes, whereas it was morphologically intact in ATRA-treated mice (Fig. 1D and E). At 12.5 dpc, although there was outgrowth of the genital tubercles, the cloacal plate failed to extend proximodistally and, consequently, the proximal part of the cloacal plate was not formed in either case. Although the urorectal septum grew normally in these mice, their distal tips did not contact the proximal part of the cloacal plate; instead, an ectopic canal between the hindgut and urogenital sinus remained, which may contribute to the formation of a communication (fistula) between the rectum and bladder (Fig. 1G and H). At 17.5 dpc, there were very similar anorectal anomalies in the two model mice. These anomalies appeared very similar to the human high-type anorectal malformations. In the two model mice, the rectums were interrupted, the anal canals were completely missing, and an ectopic fistula between the rectum and bladder was present. The genital tubercles were also hypoplastic (Fig. 1J and K).

ATRA-induced cell death at the dorsal part of the cloacal plate.

In a previous study, caudal agenesis in ATRA-treated mice was found to correspond to extensive apoptotic outbursts in the tail bud mesenchyme, with a peak incidence at 24 h after ATRA treatment (31). Therefore, we examined the involvement of alterations in apoptotic outbursts and proliferation during cloacal plate extension in the two mouse models at 10.5 dpc. In the wild type, apoptotic cells were scattered throughout the ventral and dorsal parts of the cloacal plate (Fig. 2A). In Sd homozygotes, the cloacal plate was already shrunken, there were scattered TUNEL-positive cells, and the ventral and dorsal parts of the cloacal plate were not divided (Fig. 2B). In ATRA-treated mice, we consistently observed a noticeable increase in the number of TUNEL-positive cells in the dorsal part of the cloacal plate, while consistently observing no change in the number of TUNEL-positive cells in the ventral part of the cloacal plate (Fig. 2C). However, additional markers will have to be determined, which allow distinction of the ventral and dorsal part of the plate for quantitative analysis. BrdU labeling revealed that both the ectodermal and endodermal layers of the cloacal plate were mitotically active in the wild type (Fig. 2D). In both mouse models, we also found a substantial number of BrdU-labeled cells in the cloacal plate (Fig. 2E and F). It is therefore likely that defects in proximodistal extension of the cloacal plate in ATRA-treated mice are due to apoptotic depletion of cloacal cells in the dorsal part. The essentially normal frequency of apoptotic cells and mitotic cells in Sd homozygotes suggests that the Sd mutation may affect cloacal development at an earlier stage.

Figure 2
figure 2

Analysis of apoptosis and cell proliferation in the cloacal plate at 10.5 dpc. In the wild type, apoptotic cells were scattered throughout the ventral and dorsal parts of the cloacal plate (A). In Sd homozygotes, the cloacal plate was already shrunken and there were scattered TUNEL-positive cells; the ventral and dorsal parts of the cloacal plate were not divided (B). In ATRA-treated mice, a substantial increase in TUNEL-positive cells was observed in the dorsal part of the cloacal plate, whereas there was little change in the number of TUNEL-positive cells in the ventral part (C). BrdU labeling revealed that both the ectodermal and endodermal layers of the cloacal plate were mitotically active in the wild type (D). In Sd homozygotes and ATRA-treated mice, there were a substantial number of BrdU-labeled cells in the cloacal plate (E, F). *, cloacal cavity; black full bracket, dorsal part of the cloacal plate; black dotted bracket, ventral part of the cloacal plate; red bracket, entire region of the cloacal plate in Sd homozygotes; blue bracket, positive region of the cloacal plate; and scale bars, 100 μm; magnification, ×100.

Full size image

Expression of genes involved in genital tubercle development in two mouse models.

Our observations prompted us to explore the signaling molecules required for the formation of the proximal part of the cloacal plate. From 10.5 to 12.5 dpc, because the morphology of the cloacal plate changes markedly, the pattern of gene expression at 10.5 dpc is particularly important. We focused on the expression of Shh-, Bmp-, Wnt-, and Fgf-related, and Hox cluster genes at 10.5 dpc. Moreover, we sought to identify genes whose expression was commonly affected in the two model mice. We first examined the expression of genes involved in Shh signal cascade signaling. The expression of HNF-3β, which was observed in the cloacal plate, notochord, and floor plate of the neural tube in the wild type, was reduced in Sd homozygotes, whereas it was unaffected in ATRA-treated mice (Fig. 3A–C). Furthermore, Shh was expressed in the cloacal plate, hindgut endoderm, notochord, and floor plate in the wild type (Fig. 3D) (8,9). In Sd homozygotes, the expression of Shh was reduced in the cloacal plate, the endoderm of the cloacal cavity, and the caudal region of the ventral neural tube but not in the hindgut endoderm (Fig. 3E). The expression of Shh, however, was unaffected in ATRA-treated mice (Fig. 3F). Ptc1 was expressed at high levels in a broad domain of the mesenchyme surrounding the cloacal plate but it was not detected within the cloacal plate. The expression of Ptc1 was unchanged in the two model mice (Fig. 3G–I). Hoxa13 is expressed in the mesoderm and the endoderm of the cloaca including the cloacal plate in wild type. We observed a significant reduction in Hoxa13 expression in Sd homozygotes compared with the wild type, whereas its expression was unaffected in ATRA-treated mice (Fig. 3J–L). Finally, we addressed the expression of Bmp4, Bmp7, Fgf8, and Wnt5a because these gene products have been shown to play essential roles during outgrowth of the genital tubercle, which is associated with the extension of the cloacal plate (21). In the wild type, Bmp4 was expressed in the mesenchymal cells on either side of the cloacal plate (Fig. 3M), and Bmp7 was expressed in the caudal part of the cloacal plate (Fig. 3P). In addition, Fgf8 was strongly expressed in the cloacal plate (Fig. 3S), and Wnt5a was expressed in the mesenchymal cells surrounding the cloacal plate (Fig. 3V). The expression of Bmp4, Bmp7, and Fgf8 was substantially reduced in Sd homozygotes (Fig. 3N, Q, and T), whereas there was little change in ATRA-treated mice (Fig. 3O, R, and U). Strikingly, the expression of Wnt5a was substantially reduced in both mouse models (Fig. 3W and X).

Figure 3
figure 3

Molecular analysis of the cloaca in the mouse. Whole-mount images of lateral views to the right at 10.5 dpc (A–X). Black and blue brackets indicate the entire cloacal plate and the cloaca or endoderm around the cloacal plate, respectively. Black brackets indicate normal expression of genes, and blue brackets indicate weak or no expression. HNF-3β and Shh were expressed at the cloacal plate in the wild type (A, D) but not in Sd homozygotes (B, E). However, they were normally expressed in ATRA-treated mice (C, F). Ptc1 was expressed adjacent to the cloacal plate in the wild type and two model mice (G–I). Hoxa13 was expressed in the mesoderm and the endoderm of the cloaca including the cloacal plate in the wild type (J), weakly expressed in Sd homozygotes (K), and normally expressed in ATRA-treated mice (L). Bmp4 was expressed in the mesoderm on either side of the cloacal plate in the wild type (M), very weakly expressed in Sd homozygotes (N), and normally expressed in ATRA-treated mice (O). Bmp7 was expressed in the distal part of the cloacal plate in the wild type (P), weakly expressed in Sd homozygotes (Q), and normally expressed in ATRA-treated mice (R). Fgf8 was strongly expressed in the cloacal plate in the wild type (S), not expressed in Sd homozygotes (T), and normally expressed in ATRA-treated mice (U). Wnt5a was expressed in wild type (V) but not expressed in the two mouse models (W, X). Black bracket, entire cloacal plate; blue bracket, negative region of the cloacal plate; scale bars, 100 μm; and magnification, ×20.

Full size image

The expression of Axin2 was reduced in the dorsal part of the cloacal plate in both Sd homozygotes and ATRA-treated mice.

To gain an insight into the effect of reduced Wnt5a expression in both models, we examined the expression of Axin2, which is known to be regulated by Wnt and β-catenin signals, indicating cell proliferation (32). In the wild type at 12.5 dpc, Axin2 was expressed at the dorsal part of the cloacal plate and bilaterally in the ectoderm of the genital tubercle (Fig. 4A). The expression of Axin2 in the dorsal part of the cloacal plate was substantially reduced in both models at 12.5 dpc (Fig. 4B and C).

Figure 4
figure 4

Whole-mount images of posterior views of the cloacal plate at 12.5 dpc (A–C). Axin2 was expressed at the dorsal part of the cloacal plate in the wild type (A) but not expressed in the two mouse models (B, C). v, ventral part of the cloacal plate; d, dorsal part of the cloacal plate; p, proximal part of the cloacal plate; scale bars, 100 μm; magnification, ×40.

Full size image

DISCUSSION

Comparative analyses of Sd and ATRA-treated mice reveal that proximodistal elongation of the cloacal plate is a critical step for initiation of organogenesis in the anorectal region. Histologic examination indicates that defects in the cloacal plate in ATRA-treated mice appeared to occur between 10.5 and 12.5 dpc, and morphologic changes in Sd homozygotes were observed as soon as the cloacal plate formed at 9.5 dpc. Moreover, lack of the proximal part of the cloacal plate impaired its contact with the protruding urorectal septum and subsequent anal opening in the two mouse models of anorectal anomalies. Because anorectal anomalies in both models are phenocopies of high-type anorectal malformations in human patients, defects in this process may be responsible for some human forms of anal atresia, although the etiology may vary (24,25).

Our findings further suggest that induction of Wnt5a-expressing mesenchymal cells around the cloacal plate at 10.5 dpc is important for promoting its elongation. Because apoptotic cells predominantly accumulate in the dorsal part of the cloacal plate in ATRA-treated mice, it is likely that these cells promote elongation of the cloacal plate on the receipt of inductive signals from Wnt5a-expressing mesenchymal cells. The lack of Axin2 expression in the dorsal part of the cloacal plate at 12.5 dpc in both models suggests that Wnt5a is functional at least in the genital tubercle. Expression of genes is important in the dorsal part of the cloacal plate at 10.5 dpc. But in the wild type at 10.5 dpc, Axin2 in the dorsal part of the cloacal plate was expressed only weakly. The expression of Axin2 was not defined in either Sd homozygotes or ATRA-treated mice at 10.5 dpc, because it was assumed that Wnt5a expression began at 10.0 dpc. However, the lack of Axin2 expression in the dorsal part of the cloacal plate at 12.5 dpc suggests that the induction of Wnt5a-expressing mesenchymal cells adjacent to the cloacal plate may be a bottleneck in the process of proximodistal elongation of the cloacal plate. This may be the mechanism common to both the Sd mutation and exogenous ATRA phenotypes.

A loss-of-function mutation of Wnt5a affects the outgrowth of not only the axis and limb buds but also the genital tubercle (19). Conditional depletion of β-catenin in the genital tubercle leads to hypoplasia of the genital tubercle and its derivatives (33). Therefore, the dorsal part of the cloacal plate may be defined as the proliferating zone for elongation of the proximal part of the cloacal plate. In Hox13GFP mutants with hypospadius, the cloacal plate could be functionally divided into two regions during the process of urethra formation (34). Although the impact of Wnt5a on cloacal development has not been determined, it is possible that similar molecular circuitry operates during development of the cloacal plate and genital tubercle.

Different molecular pathways in the two models may be responsible for impaired induction of Wnt5a-expressing mesenchymal cells. In Sd homozygotes, Shh or Hoxa13, or both, may be involved. Because a lack of Shh affects cloacal development, a significant reduction of Shh expression in the cloacal plate in Sd homozygotes may participate. Notably, the absence of Shh signaling was previously reported to cause a substantial reduction of Fgf8, Bmp4, Bmp7, and Wnt5a expression in the genital tubercle (18–22). It is possible that reduced expression of Shh triggers the down-regulation of Fgf8, Bmp4, Bmp7, and Wnt5a expression in the cloacal plate and surrounding mesenchyme or causes the induction of cells expressing these genes. Indeed, in Shh null mutants, Wnt5a expression is substantially reduced in the genital tubercle (18). Because the expression of Hoxa13 is also under the regulation of Shh in the genital tubercle and limb buds, reduction of Hoxa13 expression in Sd homozygotes may also be due to reduced expression of Shh (35). In contrast, down-regulation of Wnt5a in ATRA-treated mice may not involve the Shh pathway because the expression of genes other than Wnt5a and Axin2 are not altered. It is possible that RA directly regulates the expression of Wnt5a, as has been demonstrated in various tumor cell lines (36,37). This may correspond to some extent with the complex clinical spectrum of human anorectal malformations.

In summary, our results indicate that the Sd mutation affects the Shh signaling cascade, by down-regulating the expression of Wnt5a, which may lead to disrupted development of the proximal part of the cloacal plate. Similarly, development of the proximal part of the cloacal plate is disrupted in ATRA-treated mice, which may be due to down-regulation of Wnt5a expression.