Uterine Activin-Like Kinase 4 Regulates Trophoblast Development During Mouse Placentation

Abstract

The placenta is the first organ to develop after fertilization. It forms an interface between the maternal uterus and growing fetus to allow nutrient uptake, waste elimination, and gas exchange for a successful pregnancy in both mice and humans. In the past 2 decades, in vivo and in vitro approaches have been used to show that several members of the TGF-β superfamily regulate embryo implantation and placental development. Nodal, a TGF-β superfamily ligand, is essential for mesendoderm formation and left-right axis patterning during embryogenesis, and Nodal null mutants exhibit abnormal placental organization with expansion of trophoblast giant cells and a decrease of spongiotrophoblast and labyrinth. To better understand the importance of Nodal signaling in the uterus, we established a mouse model to conditionally ablate activin-like kinase 4 (ALK4; the Nodal type 1 receptor) using Cre recombinase driven by the progesterone receptor promoter sequences (Pgr-Cre). Alk4 conditional knockout females are subfertile due to placental abnormalities and fetal loss in pregnancy, with a placental disorganization phenotype similar to what is observed in Nodal null mice. Thus, Nodal likely functions as an indirect regulator of placental development by binding to type 1 and type 2 receptors on maternal decidual cells to stimulate expression of unknown regulators of placental development. Our findings not only describe the generation of a mouse model that enables study of Nodal signaling in placentation but also provides insights into the pathogenesis of pregnancy complications in humans, including spontaneous abortion, preeclampsia, intrauterine growth restriction, and preterm birth.

After the egg is fertilized, the placenta is the first organ to develop during mammalian embryogenesis. The placenta acts as the interface between the developing fetus and the maternal uterus allowing the exchange of nutrients and gases. In mice, placentation initiates from the formation of trophectoderm at embryonic day (E) 3.5. During implantation, mural trophectoderm gives rise to trophoblast giant cells that facilitate embryo attachment and invasion. These cells are characterized by polyploid nuclei due to endoreduplication. Polar trophectoderm develops into the extraembryonic ectoderm and ectoplacental cone (EPC) (1). The extraembryonic ectoderm develops into the labyrinth, whereas the EPC gives rise to a variety of trophoblasts including additional trophoblast giant cells, spongiotrophoblast, and trophoblast cells in the labyrinth (1).

The mature placenta, which appears at E10.5, is composed of three major regions: the outer maternal decidua, the middle junctional zone, and the innermost labyrinth. The maternal decidua and spiral arteries undergo invasion by trophoblast cells that promote vascularization and will replace maternal endothelial cells as the lining for the local maternal blood vessels (2). The junctional zone, which structurally supports the developing labyrinth, consists of spongiotrophoblasts separated from the decidua by a thin layer of trophoblast giants cells (2). The labyrinth layer, while still nascent at E10.5, will ultimately surround the sinusoidal blood space, line maternal spiral arteries with trophoblast cells, and serve as the location of nutrient exchange between the mother and fetus (2).

In the past 2 decades, a number of TGF-β superfamily members have been studied in vivo and in vitro to understand their potential roles in regulating placental and embryonic development during pregnancy (3). Activin-βA and activin-βB subunit single and double mutants have been generated and show no placental defects (4–6), indicating that these ligands play no intrinsic role during placental development. Paradoxically, several in vitro studies have reported that activins act directly on the cultured trophoblast stem cells to influence their proliferation and alter cell fate (7, 8). Nodal null mutants exhibit abnormal placental organization with loss labyrinth and expansion of giant cells (9, 10). Unlike activins, cultured trophoblast stem cells do not respond directly to Nodal, likely due to the absence of Cripto (a Nodal coreceptor) (8).

More recently, mouse models with conditional ablation of type 1 and type 2 receptors of bone morphogenetic protein signaling in the uterus were established in our group to investigate their physiological functions before or after implantation (11, 12). However, the potential roles of Nodal signaling in the uterus still remain largely unknown. Activin-like kinase 4 (ALK4) has been identified as the major type 1 receptor of Nodal signaling via the SMAD 2/3 pathways (13). To better understand the uterine roles of ALK4 in implantation, decidualization, and placentation, we established a novel mouse model. Due to the early embryonic lethality of Alk4 null mutants (14), Alk4 conditional knockout (cKO) mice were generated through expression of Cre recombinase under the control of the progesterone receptor locus (Pgr-Cre) (15). We found that Alk4 cKO mice exhibited excessive trophoblast giant cells during placental development, similar to previously reported embryonic and maternal Nodal mutants. Together with these studies, our data support a role for Nodal signaling via ALK4 in the uterus in regulating trophoblast cells and the development of the fetal-maternal interface.

Materials and Methods

Animals and ethics statement

All mouse lines were maintained on a hybrid C57BL/6J and 129S5/SvEvBrd genetic background. Animal handling and surgeries were performed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of Baylor College of Medicine.

Generation of Alk4 cKO female mice

Mice carrying the Alk4 floxed alleles (Alk4^flox/flox) have been described previously (16). Alk4^+/− mice were generated by crossing Alk4^flox/flox mice to EIIa-cre transgenic mice (17). The Alk4^+/− mice were bred to mice carrying Pgr-Cre knock-in alleles (Pgr^cre/+) (15) to generate Alk4^+/−Pgr^cre/+ mice. These mice were then bred to Alk4^flox/flox mice to generate Alk4^flox/− females as controls and Alk4^flox/−Pgr^cre/+ females as cKO (Figure 1A). Mice were genotyped by PCR analyses of genomic tail DNA using specific primers (Figure 1, B and C). Primer information is listed in Supplemental Table 1.

Figure 1.

Generation of Alk4 cKO mice. A, Schematic representation of the breeding strategy for generating control and Alk4 cKO mice. B, Illustration of the Alk4 conditional allele with exons 2 and 3 flanked by two loxP sites. Primer pairs P1/P2 and P3/P4 were used to detect the flanking regions including 5′ or 3′ LoxP sites, respectively. Primer pair P5/P4 was used to detect the recombined allele of Alk4. C, Genotyping results used tail genomic DNA samples to identify control (lane 1) and cKO (lane 6) mice. D–F, Relative Alk4 mRNA levels in the ovary (D), oviduct (E), and uterus (F) of control and cKO females. Data are presented as mean ± SEM (n = 3). *, P < .05, ***, P < .001 compared with controls. G–L, Immunofluorescence analysis was performed to examine ALK4 expression in the ovary (G–J), oviduct (G and H), and uterus (K and L) of control and cKO females. Panels I and J are higher magnifications of panels A and B. Scale bar, 500 μm (G and H); 100 μm (I–L).

Fertility analysis

To fertility of control and cKO mice, 6-week-old females were mated independently with sexually mature males for a 6-month period (n = 11). Cages were monitored daily, and the numbers of litters and pups were recorded.

Timed mating

Adult female mice were mated to fertile wild-type males, and the presence of vaginal plugs was monitored every morning. The morning of the plug was designated as 0.5 day post coitus (dpc). To visualize early implantation, 5.5 dpc pregnant mice were injected iv in the tail vein with 1% Chicago Blue B and killed 5 minutes after the injection.

Ovulation analysis

Three-week-old female mice were injected with 5 IU pregnant mare serum gonadotropin for 46 hours, followed by 5 IU human chorionic gonadotropin for an additional 18 hours. Cumulus-oocyte complexes were isolated from oviducts and collected into M2 medium (Sigma-Aldrich). The number of Cumulus-oocyte complexes per female was recorded, and data were presented as relative oocyte number.

Artificial induction of decidulization

The artificial decidualization of the uterus was performed as previously described (18). Briefly, 10 days after ovariectomy, female mice were injected with 100 ng estradiol (E₂) for 3 days. After 2 days of rest, mice were then treated with daily injections of 1 mg progesterone (P₄) and 6.7 ng E₂ for 3 days. After the last injection, one uterine horn was scratched by a needle on the antimesometrial lumen. The other horn was not scratched and served as a control. Mice were then treated with daily injections of 1 mg P₄ and 6.7 ng E₂. After 5 days, the mice were killed, and the wet weight of the scratched and control horns was recorded.

Histological analysis

Ovaries, oviducts, and uteri were dissected and fixed in 4% paraformaldehyde. Fixed tissues were processed through dehydration in graded ethanol, clearing in xylene, and embedding in paraffin blocks. Paraffin sections were stained with hematoxylin-periodic acid Schiff (PAS) or hematoxylin and eosin (H&E).

Immunofluorescence

Paraffin sections were deparaffinized, hydrated, and boiled for antigen retrieval. After blocking with 3% BSA for 1 hour at room temperature, sections were incubated overnight at 4°C with the primary antibodies: anti-ALK4 (Thermo Scientific), antiplacental lactogen-1 (PL-1; Santa Cruz Biotechnology), and antitrophoblast-specific protein alpha (TPBPA; Abcam). After washing with PBS, sections were incubated with Alexa Fluor 546- and 488-conjugated secondary antibodies (Life Technologies) for 1 hour at room temperature and mounted in mounting medium with 4′,6′-diamino-2-phenylindole (DAPI) (Vector Laboratories) to visualize chromatin.

Hormone analysis

Blood was collected from 2-month-old adult female mice by cardiac puncture. The serum was separated from the blood and stored at −80°C prior to hormone analysis. Serum P₄ and E₂ levels were measured by the Ligand Assay and Analysis Core at University of Virginia (Charlottesville, Virginia).

RNA isolation and real-time quantitative PCR

Tissues were collected and stored immediately at −80°C before RNA extraction. Total RNA was isolated using the RNeasy minikit (QIAGEN). Gene expression was analyzed by real-time quantitative PCR (RT-qPCR) with a SYBR Green detection system (Life Technologies). Primer information is listed in Supplemental Table 1. The relative fold change of transcript was calculated by the 2^−ΔΔCT method as described previously (19) and was normalized to Gapdh or 36B4 as an endogenous reference.

Statistical analysis

Differences among groups were analyzed for statistical significance by using one-way ANOVA followed by Tukey's multiple comparison test or two-way ANOVA followed by Bonferroni's posttest to compare replicate means. The data represent the mean ± SEM, and P < .05 was considered to be statistically significant.

Results

Generation of Alk4 cKO mice

Alk4 null mutants die by E9.5 due to the disorganization of extraembryonic ectoderm (14). To investigate the physiological roles of ALK4 in female reproduction, we generated an Alk4 cKO mouse model (Figure 1A) using Pgr-Cre, in which expression of Cre recombinase is driven by Pgr promoter sequences after knock-in insertion of Cre into the Pgr locus, to delete exons 2 and 3 of the gene (Figure 1B), causing a frameshift mutation that eliminates ALK4 protein expression (16). Previous studies indicated that Pgr-Cre is expressed postnatally in the anterior lobe of pituitary glands, granulosa cells of the preovulatory follicles, oviducts, epithelium, stroma, and myometrium of the uterus (15). We performed genotyping PCR to identify control and cKO female mice (Figure 1C). Primer pairs P1/P2 and P3/P4 were used to detect the flanking regions including 5′ or 3′ LoxP sites, respectively. Primer pair P5/P4 was used to detect the recombined allele of Alk4 (Figure 1, B and C).

Efficiency of Alk4 deletion was examined by RT-qPCR to compare the mRNA expression among the female reproductive organs. Significant decreases in Alk4 mRNA expression were detected in all the three organs: 50% reduction in ovary (Figure 1D), 70% reduction in oviduct (Figure 1E), and 85% reduction in uterus (Figure 1F). We performed immunofluorescence analysis to compare ALK4 protein expression between control and cKO mice in the same organs of the female reproductive system (Figure 1, G–L). ALK4 was expressed in oocytes at different stages of folliculogenesis, in the corpus luteum, and in the epithelium of oviducts. Whereas Alk4 mRNA expression was reduced in the ovaries and oviducts of Alk4 cKO mice, ALK4 protein levels in Alk4 cKO mice remained equivalent to control mice (Figure 1, G–J). In the uterus, ALK4 expression was detected in luminal epithelium, glandular epithelium, stroma, and myometrium (Figure 1K). Ablation of Alk4 by Pgr-Cre led to loss of ALK4 expression in those cellular compartments of uterus (Figure 1L).

Alk4 cKO female mice are subfertile but display normal ovulation and decidulization

To evaluate the fertility of Alk4 cKO female mice, we performed a continuous breeding study for control and Alk4 cKO female mice (Figure 2). Six-week-old female mice (n = 11 for each genotype) were mated with known fertile wild-type male mice for 6 months. The control females showed normal breeding activity during the test period. In contrast, Alk4 cKO females exhibited a significant reduction in fertility with decreases in total pups (Figure 2A), pups per month (Figure 2B), litters per female (Figure 2C), and pups per litter (Figure 2D) when compared with control females.

Figure 2.

Fertility analysis in control and cKO females. A, Total accumulated pups at each month during a 6-month breeding period. B, A time course of litter size per female during a 6-month testing period. C, Total litter number per female was compared between control and cKO. D, Litter size per female was compared between control and cKO. Error bars (B–D) represent ± SEM (n = 11). *, P < .05; **, P < .01; ***, P < .001.

To determine the causes of this subfertility phenotype, we first evaluated ovarian function in Alk4 cKO female mice. Alk4 cKO ovaries were histologically similar to controls and contained follicles of all developmental stages as well as corpora lutea (Figure 3, A and B). After pharmacological superovulation, control and Alk4 cKO mice also showed similar induction of ovulation (Figure 3, C and D) and normal cumulus expansion in preovulatory follicles (Figure 3, E and F). Furthermore, ovulation analysis performed in 3-week-old control and Alk4 cKO mice after superovulation found no significant differences in the number of ovulated oocyte (Figure 3G). Finally, our evaluation of hormone profiles in females of each genotype by the measurement of serum P₄ and E₂ levels found comparable levels of the 2 serum steroids in both control and cKO females (Figure 3H). Our analysis shows that Pgr-Cre driven knockout of Alk4 does not impair ovarian function, and loss of ALK4 in the ovaries cannot account for the observed subfertility phenotype.

Figure 3.

Characterization of ovarian and uterine phenotypes in control and cKO females. A and B, Control and Alk4 cKO ovaries and oviducts were stained with hematoxylin-PAS. C and D, Control and Alk4 cKO ovaries and oviducts were stained with hematoxylin-PAS after superovulation. Panels E and F are higher magnifications of panels C and D. G, Ovulation rate was tested in control and Alk4 cKO after superovulation. H, Serum E₂ and P₄ were measured in control and cKO females. I and J, Control and Alk4 cKO uteri were stained with H&E to examine the morphological change. K and L, Gross uterine morphology of control and Alk4 cKO mice was compared after artificial induction of decidualization. M, Ratio between wet weights of scratched and control horns were calculated for the control and Alk4 cKO mice. Scale bar, 2 mm (A–D, I, and J); 200 μm (E and F); 10 mm (K and L). Panels G, H, and M represent the mean ± SEM, and no significant difference was found between control and cKO mice.

We then performed histological H&E staining to check for uterine defects in Alk4 cKO mice. No morphological changes were found in the uteri of nonpregnant adult females (Figure 3, I and J). When the blastocyst attaches to the uterine epithelium, the uterine stroma at embryonic attachment sites undergoes proliferation and differentiation into decidual cells (20). To examine this uterine response in Alk4 cKO mice, we applied an artificial induction of decidulization by scratching uterine epithelium to mimic the stimulation of blastocyst attachment. Decidual responses were grossly normal in both genotypes (Figure 3, K and L), and all scratched uterine horns exhibited similar elevations in size and weight when compared with unscratched uterine horns (Figure 3M).

Expression of ALK4 in placental development

To further investigate the spatiotemporal localization of ALK4 in the pregnant uterus, we performed immunofluorescence analysis at sequential time points after implantation (Figure 4). We used double-immunofluorescence staining with antibodies to PL-1 (a molecular marker for parietal trophoblast giant cells) and ALK4 to localize the maternal-fetal interface after implantation. H&E staining is provided for reference (Figure 4, A–D). Whereas ALK4 was highly expressed in luminal and glandular epithelium of the nonpregnant uterus (Figure 1K), ALK4 protein expression was almost undetectable in uterine epithelium and stroma at 5.5 dpc (Figure 4, E and I). ALK4 expression was first observed in the antimesometrial stroma near the embryo at around 7.5 dpc (Figure 4, F and J) and reached its peak at 8.5 dpc, with the highest expression observed at the antimesometrial pole of the implantation site (Figure 4, G and K). Protein expression was dramatically lower after 8.5 dpc and was undetectable at 10.5 dpc (Figure 4, H and L). Thus, ALK4 is present in the stroma in proximity to the embryo after implantation, suggesting that signaling through ALK4 in the stroma might play a role in regulating placentation at this key time point.

Figure 4.

Localization of ALK4 expression in implantation sites during placentation. H&E staining and immunofluorescence analysis of ALK4 and PL-1 was performed in wild-type uteri at 5.5 (A and E), 7.5 (B and F), 8.5 (C and G), and 10.5 dpc (D and H). Panels I, J, K, and L are higher magnifications of panels E, F, G, and H. Scale bar, 200 μm (A–D); 300 μm (E–H); 50 μm (I–L). AM, antimesometrial; M, mesometrial. *, Embryo/fetus.

Increased trophoblast giant cells and resorbed embryos were observed in Alk4 cKO mice

To study the roles of ALK4 in regulating uterine function during pregnancy, female reproductive tracts were examined at sequential time points after implantation (Figure 5, A–F). At 5.5 dpc, tail vein injection of Chicago blue B dye was applied to visualize implantation sites (21). Alk4 cKO showed no differences in the number or apparent sizes of the implantation sites at 5.5 dpc (Figure 5, A and B), indicating that ALK4 is dispensable for embryo attachment and implantation. At 8.5 dpc, although most Alk4 cKO mice exhibited similar implantation sites as the control mice, smaller implantation sites were occasionally observed in Alk4 cKO mice (Figure 5, C and D). At 10.5 dpc, impaired implantation sites were observed in about half of Alk4 cKO mice (Figure 5, E and F). Statistically, there was no difference in the number of nonresorbed implantation sites between the two genotypes at either 8.5 or 10.5 dpc (Figure 5G). There was not a statistically significant difference in implantation site size at 8.5 dpc; however, cKO mice exhibited significantly smaller implantation sites at 10.5 dpc (Figure 5H). Compared with controls (Figure 5I), we observed embryonic defects in Alk4 cKO mice, typically a few smaller embryos in one litter (Figure 5J), but entire litters with resorbed embryos were occasionally seen (Figure 5K).

Figure 5.

Implantation site analysis in control and cKO females at sequential time points. A–F, Gross morphology of control and Alk4 cKO female reproductive tract was compared at 5.5 (A and B), 8.5 (C and D), and 10.5 (E and F) dpc. To visualize implantation sites at 5.5 dpc, animals were injected Chicago Blue B via tail vein (A and B). Resorbed implantation sites are highlighted by white arrows (D and F). G and H, Nonresorbed implantation site number and weight at 8.5 and 10.5 dpc were compared between control and cKO females. I–K, Litters of fetuses (or resorbed embryo) were dissected from implantation sites at 10.5 dpc. Scale bar, 10 mm (A–F); 1 mm (I–K). IS, implantation site. G and H, Error bars represent ± SEM. *, P < .05.

Consistent with the gross morphology, uterine histology appeared comparable between control and Alk4 cKO female mice at 8.5 dpc (Figure 6, A and B). Immunofluorescence analysis of ALK4 confirmed the ablation of ALK4 in cKO antimesometrial decidua (Figure 6, C–F).

Figure 6.

Histology and ALK4 immunofluorescence analysis of implantation sites at 8.5 dpc. A and B, Control and cKO implantation sites were stained with H&E. C and D, Immunofluorescence analysis of ALK4 was performed to compare the protein expression in the two genotypes. Panels E and F are higher magnifications of panels C and D. Scale bar, 500 μm (A and B); 200 μm (C and D); 100 μm (E and F). AM, antimesometrial; DAPI, 4′,6′-diamino-2-phenylindole; M, mesometrial. *, Embryo/fetus.

To examine the role of ALK4 during placentation, we performed immunofluorescence analysis of TPBPA, a molecular marker for spongiotrophoblast, and PL-1, a molecular marker for parietal trophoblast giant cells (Figure 7, A–H). In control mice, PL-1-positive trophoblast giant cells lined the implantation site as a single layer adjacent to the antimesometrial decidua (Figure 7, A and C); however, in Alk4 cKO mice, we observed an expansion of the parietal trophoblast giant cell layer, most notably in the antimesometrial region (Figure 7, B and D). TPBPA expression was observed in the chorion and EPC at 8.5 dpc in both control and cKO mice (Figure 7, E and F).

Figure 7.

Immunofluorescence analysis of implantation sites at 8.5 dpc. Immunofluorescence analysis of TPBPA and PL-1 was performed to examine trophoblast development in control and cKO genotypes at 8.5 dpc (A and B). Panels C and D are higher magnifications (white boxes) of panels A and B, showing the antimesometrial junction between decidua and parietal trophoblast giant cells. Panels E and F are higher magnifications (yellow boxes) of panels A and B, showing the EPC. White arrows in panels C and D indicate comparable points of the trophoblast giant cell layer. Yellow arrows in panels E and F indicate the EPC. Scale bar, 200 μm (A and B); 50 μm (C–F). AM, antimesometrial; M, mesometrial. *, Embryo/fetus.

To further characterize the effects of ALK4 ablation on placental development, we examined implantation sites at later time points in pregnancy. At 10.5 dpc, histology frequently revealed impaired implantation sites with hemorrhage in Alk4 cKO mice that were not observed in control mice (Figure 8, A and B). In control mice, histology showed that the placenta is, as expected, composed of three zones: maternal decidua, junctional zone (spongiotrophoblast and trophoblast giant cells), and labyrinth (Figure 8C). However, in cKO mice this normal organization was disrupted with the expansion of trophoblast giant cells and the reduction of spongiotrophoblast and labyrinth (Figure 8D). By immunofluorescence analysis, the junctional zone in control mice consisted of TPBPA-positive spongiotrophoblast cells forming a thin bilayer with PL-1-positive parietal trophoblast giant cells (Figure 8, E and G). In Alk4 cKO mice, we observed excessive numbers of PL-1-positive parietal trophoblast giant cells in a disorganized junctional zone with TPBPA positive spongiotrophoblasts (Figure 8, F and H). We assessed expression levels of Tpbpa, Pcdh12, and Nodal in 10.5 dpc implantation sites by RT-qPCR (Figure 9). In line with the histological observation of a reduced spongiotrophoblast layer, Tpbpa was down-regulated in cKO mice (0.09x, P < .05) (Figure 9A). Pcdh12, a specific marker for glycogen trophoblast cells, was also down-regulated (0.68x, P < .05) (Figure 9B); however, the decrease in Pcdh12 expression was far less dramatic than in Tpbpa, suggesting that the decrease in Tpbpa, which is expressed by glycogen trophoblast cells and spongiotrophoblast, is principally a consequence of reduced spongiotrophoblast. Nodal expression was down-regulated (0.11x, P < .05) (Figure 9C) at 10.5 dpc in cKO mice, suggesting signaling through maternal ALK4 promotes expression of fetal and/or maternal Nodal.

Figure 8.

Histology and immunofluorescence analysis of implantation sites at 10.5 dpc. A and B, Control and Alk4 cKO implantation sites were stained with H&E. Panels C and D are higher magnifications of panels A and B. E and F, Immunofluorescence analysis of TPBPA and PL-1 was performed to examine trophoblast development. Panels G and H are higher magnifications of panels E and F. Hemorrhage in implantation sites was highlighted with black arrows. AM, antimesometrial; d, maternal decidua; g, trophoblast giant cells; l, labyrinth; M, mesometrial; s, spongiotrophoblast. *, Embryo/fetus; Scale bar, 1 mm (A and B); 200 μm (C–F); 50 μm (G and H).

Figure 9.

Changes in gene expression at 10.5 dpc in cKO pregnancies. Relative mRNA levels of Tpbpa (A), Pcdh12 (B), and Nodal (C) were measured by RT-qPCR, normalized to levels of 36B4 mRNA, and compared across genotype. Error bars represent ± SEM (control n = 3, cKO n = 4). *, P < .05.

Together these observations show that the loss of ALK4 in the maternal decidua can cause placental disorganization, intrauterine growth restriction, and embryo resorption after implantation.

Discussion

In this study, we generated a mouse model with conditional ablation of ALK4 in the uterus. These mice are subfertile, likely due to the placental abnormalities we observed including expansion of trophoblast giant cells and a reduction of spongiotrophoblast and labyrinth at 10.5 dpc.

Although both ALK4 and ALK7 serve as the type I receptors for Nodal signaling (13), only homozygous mutant Alk4 mice phenocopy Nodal mutants, which die by E9.5 due to the disorganization of extraembryonic ectoderm (14). Alk7 null mice are also not subfertile (21), suggesting ALK7 is dispensable for both embryonic development and fertility in mice.

The phenotype we observed in our Alk4 cKO mice is similar to the placental abnormalities seen in fetal and maternal Nodal mutants. Deletion of Nodal in the fetus impairs embryonic development, but placental development is also disrupted with expanded trophoblast giant cells and the loss of spongiotrophoblast and labyrinth (14). Mice with hypomorphic Nodal alleles undergo embryonic development normally but suffer from aberrant placentation with expanded trophoblast giant cell and spongiotrophoblast layers with reduced labyrinth (9), showing that fetal Nodal plays a role in placental development independent from its role in embryonic development. Interestingly, maternal Nodal also plays an important role in placental development. When Nodal is specifically deleted in the uterus, placentation is disrupted with trophoblast giant cell expansion on the antimesometrial and lateral surfaces (22). Together with these studies, our data suggest that signaling through ALK4 via Nodal in both fetal trophoblast and maternal decidua regulates trophoblast cells and the formation of the fetal-maternal interface.

The mechanism by which signaling through ALK4 in the maternal decidua regulates trophoblast cells is unknown and presents an interesting question for future study. Because the ALK4 cKO phenotype we observed is milder than previously described fetal Nodal knockout and maternal Nodal cKO phenotypes, there may be limited redundancy with Nodal/ALK7/ACVR2B in the maternal decidua. Alternatively, it is possible that maternal Nodal plays a direct regulatory role independent of maternal ALK4 signaling. Paradoxically, whereas in vitro studies have supported an important role for activin signaling through ALK7 in regulating labyrinth formation (8), neither Alk7 null nor activin-β/α double-null mutants have placental defects (4–6, 21).

Our observation of reduced Nodal expression in implantation sites from 10.5 dpc Alk4 cKO mice raises several possible modes of the interaction between maternal ALK4 signaling and Nodal expression. First, the population of trophoblast cells that produce fetal Nodal could be diminished when signaling through maternal ALK4 is absent. Second, fetal Nodal production could be positively regulated by signaling through maternal ALK4 via unknown regulatory factors. Third, maternal Nodal production could be positively regulated by signaling through ALK4. Although our data cannot directly answer this question, we hypothesize that because Nodal in the implantation site is expressed primarily by spongiotrophoblast cells at 10.5 dpc (9), the decrease in Nodal expression is likely the result of the reduction of spongiotrophoblast that we observed. This leaves open the question of how maternal ALK4 signaling regulates the relative abundance of trophoblast subtypes.

Defects in the fetal-maternal interface at the placenta can contribute to preeclampsia and intrauterine growth restriction (23), both of which increase the risk of miscarriage and neonatal death. Our study highlights the importance of signaling through ALK4 in the maternal decidua for placentation. A clearer understanding of the mechanisms underlying placentation, particularly the maternal role, will hopefully enable screening to identify women at risk for placental defects and the development of treatments that can promote normal placentation.

Additional material

Supplementary data supplied by authors.