PEG10 viral aspartic protease domain is essential for the maintenance of fetal capillary structure in the mouse placenta

PEG10 is a therian-specific gene encoding GAG- and POL-like proteins that exhibit at most ∼20-30% homology with the long terminal repeat (LTR) sushi-ichi retrotransposon (Ono et al., 2001). Certain characteristic traits, such as the presence of a CCHC RNA-binding domain in the GAG-like region, a DSG viral aspartic protease motif in the POL-like region and a −1 ribosomal frameshift that results in a PEG10-open reading frame (ORF) 1 and an ORF2 fusion protein (PEG10-ORF1/2), are highly conserved among all of the eutherian and marsupial orthologs of PEG10 (Shigemoto et al., 2001; Clark et al., 2007; Suzuki et al., 2007) (Fig. S1), although PEG10 has lost the LTR sequences at both ends of its structure. These unique features, together with its emergence in therian mammals, indicate that PEG10 was domesticated in a common therian ancestor from an extinct retrovirus ∼166 million years ago (Suzuki et al., 2007; Warren et al., 2008). We previously demonstrated that Peg10 knockout (KO) mice exhibit early embryonic lethality before 10.5 days post-coitus (dpc) owing to severe placental dysplasia, including the loss of trophoblast cells in the labyrinth and spongiotrophoblast layers (Ono et al., 2006). The labyrinth layer is an essential part of the mouse placenta in which nutrient and gas exchange occurs between the fetal and maternal blood, and trophoblast cells are unique to the placenta (Rossant and Cross, 2001). These features imply that the acquisition of PEG10 was a crucial event in the emergence of the viviparous reproduction system in the common therian ancestor(s) (Kaneko-Ishino and Ishino, 2012, 2015).

In what manner is PEG10 involved in placenta development? It has been reported that PEG10 is involved in the development and progression of cancer (Okabe et al., 2003; Tsou et al., 2003; Li et al., 2006). Unfortunately, the biochemical functions of the PEG10 protein have remained obscure since its discovery in 2001, possibly because of the range of its structural and functional varieties resulting from the properties of the GAG and POL proteins. GAG and the GAG-POL fusion proteins are digested by the aspartic protease activity of POL to form several distinct parts, each with a specific role or function. These include three structural proteins (matrix, capsid and nucleocapsid proteins) from GAG along with four enzymatic proteins (aspartic protease, reverse transcriptase, RNase H and DNA integrase) from POL (Kirchner and Sandmeyer, 1993; Dunn et al., 2002; Pettit et al., 2004). Similarly, PEG10 is reported to be digested by the highly conserved aspartic protease in the POL-like ORF2 (Clark et al., 2007). Therefore, it is conceivable that PEG10 plays multiple roles during the course of development, not only in the placenta, but also in other organs and tissues. Which part and/or motif of PEG10 is essential for early placenta formation associated with trophoblast differentiation and growth, and what are the functions of other parts and/or motifs of PEG10? It is clear that a systematic approach is necessary to solve this puzzle in a step-by-step manner, so we generated a series of Peg10 mutant mice harboring a mutation for each of the highly conserved traits. In this study, we analyzed Peg10-protease-motif mutant mice and found that PEG10 is expressed in the trophoblast cell layers at the feto-maternal interface and the mice exhibited perinatal lethality owing to severe defects in feto-maternal circulation in the placenta. This result clearly shows that Peg10 plays different roles in placental development in a stage-dependent manner – placenta formation via trophoblast differentiation and growth in the early gestational stage and maintenance of the fetal capillary network from mid- to late-gestation – and that PEG10 protease activity is indispensable for the latter process.

Interestingly, retrotransposon Gag like 1 (RTL1, also called PEG11), another paternally expressed imprinted gene, exhibits homology to the same sushi-ichi retrotransposon as PEG10 (Charlier et al., 2001; Seitz et al., 2003; Youngson et al., 2005). It is a eutherian-specific gene and plays an essential role in the maintenance of the fetal capillary network in mid- to late-gestation on the endothelial cell side (Sekita et al., 2008; Kitazawa et al., 2017). Thus, it is likely that there exists some interaction between the two retrovirus-derived genes Peg10 and Rtl1, along with their effects at the feto-maternal interface in the mouse placenta for maintaining the fetal vasculature, via their functions in trophoblast and endothelial cells, respectively.

The DSG protease domain in the POL-like PEG10-ORF2 is highly conserved among therian mammals (Fig. S1), suggesting the importance of the function(s) of this domain. In an effort to determine what function the PEG10 DSG protease plays in placental development, we generated a mutant strain harboring a point mutation in the DSG domain by replacing aspartic acid (D) with an alanine (A) residue using the CRISPR-Cas9 system (Fig. 1A; Fig. S2). It has been reported that the PEG10 DSG protease exerts self-cleavage activity, producing several fragments, and that the same amino acid substitution (DSG to ASG) disrupts the protease activity of the human PEG10 protein in vitro (Clark et al., 2007). As Peg10 is a paternally expressed gene, mice carrying a paternally transmitted mutant allele (+/ASG) (hereafter called Peg10-ASG mice) were used for phenotypic analysis throughout this study. First, we examined whether Peg10-ASG mice exhibit early embryonic lethality like Peg10 null KO mice. Unexpectedly, the mutant embryos and placenta exhibited an apparently normal appearance at 12.5 dpc (Fig. S3). In western blotting experiments (Fig. 1B), a PEG10 self-cleavage product (∼75 kDa) was observed only in the wild-type (+/+) placenta but was undetectable in the Peg10-ASG placenta, thus confirming the fact that the mutation from DSG to ASG disrupts protease activity, as expected. These results demonstrate that loss of the DSG protease function is not responsible for the early embryonic lethality associated with the differentiation and growth defects of placental trophoblast cells observed in Peg10 null KO mice.

At the weaning stage, the total proportion of mutant pups was extremely low, only 14% (5 mutants/35 total pups) compared with the expected value of 50% (Table 1), indicating that most of the Peg10-ASG pups died during the course of perinatal development and growth. Approximately 50% of the recovered mutant fetuses were found dead at 18.5 dpc (22 dead fetuses/44 total fetuses) (Fig. 1C, Table 1). Moreover, the surviving mutant fetuses had small overall body size and had small placentas compared with normal mice (Fig. 1C). The mutant embryos were of normal weight before 13.5 dpc, but their growth retardation started around 15.5 dpc and reached over 20% reduction at 18.5 dpc (Fig. 1D), whereas the placental weight reduction was already evident at 12.5 dpc and persisted to term (Fig. 1E). Given that the Peg10 expression levels are unaffected by this mutation (Fig. S4), these results demonstrate that the DSG protease activity itself is essential for placental growth in mid- to late-gestation and that the inactive ASG mutation apparently caused placental hypoplasia, which consequently led to fetal growth retardation and perinatal lethality.

Evident morphological abnormalities were observed in the labyrinth layers of the Peg10-ASG placenta. The fine, mesh-like structure of the fetal vascular network was severely damaged. The fetal vasculature is composed of fetal capillary endothelial cells and three surrounding layers of trophoblast cells: two syncytiotrophoblast (SynT-I and II) layers and one mononuclear sinusoidal trophoblast giant cell (s-TGC) layer. On sections of 18.5 dpc placenta, alkaline phosphatase (AP) staining, which was employed to distinguish the maternal blood sinus (Adamson et al., 2002) (Fig. 2A; Fig. S5A), and immunohistochemistry for the endothelial marker CD31 (PECAM1) (Fig. 2B; Fig. S5B) revealed severe endothelial cell deformation and collapse of most of the fetal capillaries in the labyrinth layers, resulting in a loss of the fine, mesh-like structure of the fetal vasculature. We performed simultaneous immunofluorescence staining using antibodies against CD31 and the pan-trophoblast marker cytokeratin (CK) (Fig. 2C; Fig. S5C) to confirm a physical relationship between the fetal endothelial cells and trophoblast cells. In the normal placenta, the endothelial cells and trophoblast layers closely aligned along the fetal capillaries. In contrast, this alignment became irregular and their respective locations separated from each other in the Peg10-ASG placenta. Under magnified views, the fetal endothelial cell nuclei were observed to have accumulated in narrow spaces and as a result the fetal blood spaces had become clogged (Fig. 2D; Fig. S5D). Taken together, these results suggest that abnormal organization of the fetal vasculature causes severe impairment of feto-maternal exchange in the labyrinth layer, leading to the perinatal lethality of the Peg10-ASG fetuses and pups.

Next, placental morphology was assayed by Hematoxylin and Eosin (HE) staining (Fig. 3A; Fig. S6, Fig. S7A). We found that, in addition to a significant reduction in labyrinth size (Fig. S6), the areas deeply stained in shades of pink-red containing considerable endothelial cell nuclear debris were distributed throughout the labyrinth layer in the dead mutants, and were also observed to a certain extent in the living mutants (Fig. 3A; Fig. S7A). This staining pattern showing a deposition of deep-pink pigment around the damaged capillaries is characteristically seen in one of the necrotic forms of cell death, fibrinoid necrosis (Damjanov, 2009). Additionally, immunostaining for the pan-leukocyte marker CD45 (PTPRC) revealed that the number of leukocytes was markedly increased in Peg10-ASG placentas obtained from deceased embryos, and a large increase in leukocytes was also observed in certain areas of the living mutant labyrinth (Fig. 3B; Fig. S7B). These results suggest that severe inflammation was induced in the fetal vasculature of the Peg10-ASG placenta, which then led to the development of fibrinoid necrosis-like lesions in the labyrinth layers, ultimately resulting in severe fetal damage and/or death.

In this study, it was demonstrated that the disruption of the protease activity of the DSG domain in PEG10 results in perinatal lethality associated with fetal and placental growth retardation, probably because of the severe damage inflicted on the fetal vascular network. It is likely that such extensive fetal capillary damage would result in a disturbance of the feto-maternal circulation in the placental labyrinth. This indicates that the viral-type aspartic protease domain in PEG10 is essential during the mid to late stage of placental development, implying that the protease activity that is intrinsic to the retrovirus/LTR retrotransposon was successfully domesticated (or exapted) to the placental functions in eutherians.

Importantly, as shown in Fig. S8A, PEG10 protein exhibited high expression in all layers of the placenta except for the maternal decidua, including the labyrinth layer, and expression patterns were the same in the Peg10-ASG placentas. Intriguingly, we found that PEG10 did not exhibit any localization in the CD31-positive fetal endothelial cells that were damaged in the Peg10-ASG placental labyrinth (Fig. 3C; Fig. S8B). Detailed analysis demonstrated that, at 12.5 dpc, the PEG10 protein signals were uniformly distributed in the region between the fetal capillaries and the maternal blood sinus, indicating that PEG10 protein was expressed in all of the trophoblast cell layers, the two syncytiotrophoblast (SynT-I and II) layers and the single s-TGC layer. After 15.5 dpc, PEG10 protein signals became prominent in cells with large nuclei facing the maternal blood sinus, which are presumed to be s-TGCs (Fig. S8B, Fig. S9). It is thus strongly suggested that the trophoblast cells in the fetal capillary, especially s-TGCs, have some defense mechanism to protect the fetal endothelial cells against non-specific hazardous events and that direct or indirect effects of PEG10 DSG protease activity is largely involved in this process. We have demonstrated in previous studies that Rtl1 (or Peg11), which is derived from the presumable same retrovirus as Peg10, is essential for the maintenance of the fetal capillaries in the placenta (Sekita et al., 2008; Kitazawa et al., 2017). It should be noted that, in contrast to Peg10, Rtl1 is only expressed in endothelial cells (Sekita et al., 2008). This demonstrates that the two retrovirus-derived genes Peg10 and Rtl1 function on either side of the feto-maternal interface, i.e. in trophoblasts and endothelial cells in the fetal capillaries, respectively, and presumably act in a cooperative manner in order to maintain a normal fetal vasculature for the exchange of nutrients/waste and O₂/CO₂ gas between the fetal and maternal blood (Fig. S10). Recent phylogenic analysis has revealed that the mouse and human-type hemochorial placentas, which have a feto-maternal interface in which the trophoblast surface has direct contact with the maternal blood, is ancestral eutherian placenta (Wildman et al., 2006; Roberts et al., 2016). Therefore, it is highly likely that domestication of PEG10 and RTL1, before and after divergence of the eutherians and marsupials, respectively, must have been critical events and exerted a driving force in the evolution of the eutherian viviparous reproduction system via their effect on the eutherian placenta. How then do PEG10 and RTL1 act as guardians of the feto-maternal interface? One possibility is that PEG10 protease activity in trophoblast cells contributes to the immune tolerance of the mother to embryo antigens: for instance, the PEG10 protease may induce an aggressiveness of trophoblasts toward maternal immune cells in order to protect the fetal capillaries. Under such a scenario, RTL1 would play a defensive role in the fetal capillary endothelial cells against a collateral attack by trophoblast cells. The cooperation between PEG10 and RTL1 would thus provide the essential architecture for the feto-maternal interface in the placenta. The trophoblast cells invariably localize between the fetal capillary endothelial cells and the maternal tissues in all three types of the eutherian placenta, the endotheliochorial (e.g. cats and dogs), epitheliochorial (e.g. horses and pigs) and hemochorial (e.g. humans and mice) placentas. They are associated with invasiveness and access to maternal blood flow (Nakaya and Miyazawa, 2015; Roberts et al., 2016). The fact that the protease domain in PEG10 is highly conserved in all eutherian species (Fig. S1) suggests that the protease activity of PEG10 plays an important role. Accordingly, the relationship between PEG10 and RTL1 evolved to provide an adequate feto-maternal interface in a placental type-specific manner. Future studies are needed to determine the molecular mechanisms by which PEG10 and RTL1 exert a cooperatively beneficial effect on protecting the fetal capillaries via their crucial involvement in the immune tolerance system during pregnancy.

PEG10 and RTL1 are expressed in many other tissues and organs besides the placenta (Abed et al., 2019; Kitazawa et al., 2021). Recently, we demonstrated that mouse Rtl1 plays other important roles during the course of development, such as in fetal/neonatal skeletal muscle maturation as well as in functional activity of the central nervous system (Kitazawa et al., 2020, 2021). Therefore, it is possible that PEG10 has a multiplicity of functions across a range of organs and tissues. Future studies are needed to show how PEG10 contributed to the evolution and acquisition of various other therian-specific traits, and it may turn out that both PEG10 and RTL1 were domesticated in the same process and function cooperatively in current eutherian developmental systems, as in the case of the maintenance of the fetal vascular network in the placenta.

All animal and experimental procedures were approved by the Animal Ethics Committees of Tokyo Medical and Dental University and the University of Yamanashi. Animals were allowed access to a standard chow diet and water ad libitum and were housed in a pathogen-free barrier facility with a 12:12 h light-dark cycle.

The plasmids expressing hCas9 and sgRNA were prepared by ligating oligos into the BbsI site of pX330 (Addgene plasmid #42230). The 20 bp sgRNA recognition sequence was:

Peg10-ORF2-sgRNA (5′-GTCCGAGCTATGATTGATTC-3′).

The oligo DNA for co-injection with the plasmid into mouse zygotes was:

ACCTGCAAGTGATGCTCCAGATTCATATGCCGGGCAGACCCACCCTGTTTGTCCGAGCTATGATTGCTTCTGGTGCATCTGGCAACTTCATTGATCAAGACTTTGTCATACAAAATGCAATTCCTCTCAGAAT.

To produce the hCas9 mRNA, the T7 promoter was added to the hCas9 coding region of the pX330 plasmid by PCR amplification, as previously reported (Wang et al., 2013). The T7-Cas9 PCR product was gel-purified and used as the template for in vitro transcription (IVT) using the mMESSAGE mMACHINE T7 ULTRA kit (Life Technologies). The T7 promoter was added to the Peg10-ORF2-sgRNA region of the pX330 plasmid by PCR amplification using the following primers:

Peg10-ORF2-IVT-F, TTAATACGACTCACTATAGGTCCGAGCTATGATTGATTC; IVT-R, AAAAGCACCGACTCGGTGCC.

The T7-sgRNA PCR product was gel-purified and employed as the template for IVT using the MEGAshortscript T7 kit (Life Technologies). Both the Cas9 mRNA and Peg10-ORF2-sgRNA were DNase-treated to eliminate template DNA, purified using the MEGAclear kit (Life Technologies), and eluted into RNase-free water.

B6D2F1 female mice were superovulated and in vitro fertilization was carried out using B6D2F1 mouse sperm. The synthesized hCas9 mRNA (50 ng/μl) and Peg10- ORF2-sgRNA (25 ng/μl) with oligo DNA (10 ng/μl) were injected into the cytoplasm of fertilized eggs at the indicated concentration. The eggs were cultivated in KSOM overnight, then transferred into the oviducts of pseudopregnant ICR females. The genotype was determined by direct sequencing of the PCR product using the following primers: F primer, GGAAGGTCTCAACCCAGACA; R primer, GTATCTCACGGTGGTCTCCC. The F-primer was used for direct sequencing.

The assays were performed in triplicate and the copy number of the genes examined was calculated with a StepOnePlus Real Time PCR system (Thermo Fisher Scientific) using the THUNDERBIRD SYBR qPCR Mix (TOYOBO). PCR was carried out under the following conditions: 95°C for 15 s, 65°C for 30 s and 72°C for 30 s. The primers used are as follows:

Peg10-F, TTGGTCCCTTACCCCTACCAAC; Peg10-R, CCCTTGAGTTAATTCCCAGAGCC; Actb-F, AAGTGTGACGTTGACATCCG; Actb-R, GATCCACATCTGCTGGAAGG.

The recovered mouse placentas were directly embedded in OCT compound (Sakura Finetek) or first fixed in 4% paraformaldehyde, incubated sequentially in 10% sucrose in PBS for 2 h, 15% sucrose in PBS for 2 h then 25% sucrose in PBS overnight at 4˚C, and ultimately embedded in OCT compound (Sakura Finetek). The OCT blocks were sectioned at a thickness of 7 μm with a cryostat (Thermo Fisher Scientific) and mounted on Superfrost Micro Slides (Matsunami Glass).

Anti-PEG10-ORF1 antibody was generated according to the protocol of the iliac lymph node method (Kishiro et al., 1995). Briefly, after immunizing with the whole PEG10-ORF1 via rat hind footpads for 3 weeks, cells from iliac lymph node were fused with SP2/O-Ag14 mouse myeloma cells. Cell lines with high ability for anti-PEG10-ORF1 antibody production were then screened. The specificity of anti-PEG10-ORF1 antibody in immunofluorescence analysis was validated by using the tissues from Peg10 null KO mice as a true negative control for our experiments. Anti-PEG10-ORF2 antibody was produced from rabbit serum by immunizing with the synthetic PEG10 peptide (NH2-CPSGHLYSMSESEMN-COOH) and purified by antigen-specific affinity purification. The antibodies were validated in a western blot by checking whether the band size was generally consistent with the expected molecular weight and using the lysate from Peg10 null KO cells as a true negative control for our western blot analysis.

Frozen sections were fixed in 4% paraformaldehyde for 10 min at room temperature and washed three times with PBS for 5 min. The sections were then subjected to HE staining, immunodetection (immunofluorescence and immunohistochemistry) or endogenous alkaline phosphatase activity detection.

For HE staining, the sections were stained with HE, and mounted with Malinol mounting medium (Muto Pure Chemicals).

For immunodetection, after incubation in blocking buffer (1% bovine serum albumin in PBS with 0.1% Triton X-100) for 60 min, the sections were reacted with primary antibody in blocking buffer at 4°C overnight. Primary antibodies used were: rat monoclonal anti-CD31 (1:100; BD Biosciences, 550274), rabbit polyclonal anti-CD31 (1:50; Abcam, ab28364) or rabbit recombinant multiclonal anti-CD31 (1:100; Abcam, ab281583); rabbit polyclonal anti-cytokeratin (1:200; Dako, Z0622); rat monoclonal anti-CD45 (1:100; BD Biosciences, 561037); rat monoclonal anti-PEG10-ORF1 (1:200; produced in-house). After washing with PBS three times for 5 min each, the sections were incubated with a secondary antibody conjugated with a fluorescence label for immunofluorescence [goat anti-rat IgG conjugated with Cy3 (1:500; Thermo Fisher Scientific, A10522) or goat anti-rabbit IgG conjugated with Alexa 488 (1:500; Thermo Fisher Scientific, A11008)] or biotin for immunohistochemistry [donkey anti-rat IgG conjugated with biotin (1:500; Jackson ImmunoResearch, 712-065-153)] in blocking buffer at 4°C overnight. For immunofluorescence, after washing with PBS three times for 5 min each, the slides were mounted with VECTASHIELD-Hardset (Vector Laboratories) containing 1 μg/ml DAPI. The sections were imaged under confocal or widefield fluorescence microscopy (LSM510, Carl Zeiss; or EVOS M5000, Thermo Fisher Scientific). For immunohistochemistry, after washing with PBS three times for 5 min each, the slides were incubated with ABC-AP reagent (VECTASTAIN ABC-AP Kit, Vector Laboratories) with 200 nM levamisole for 30 min at room temperature. The sides were washed in PBS, followed by incubation with 100 mM Tris-HCl (pH 9.5) for 10 min at room temperature. Chromogenic detection was performed by NBT-BCIP staining (BCIP/NBT AP substrate, Vector Laboratories) according to the manufacturer's instructions. After rinsing with 100 mM Tris-HCl (pH 9.5), the sections were counterstained with Nuclear Fast Red (Vector Laboratories). For endogenous alkaline phosphatase activity detection, after being incubated with 100 mM Tris-HCl (pH 9.5) for 10 min at room temperature, chromogenic detection was performed as described above.

Placenta tissue lysates were resolved by SDS-PAGE and transferred to a PVDF membrane (Bio-Rad) before immunoblotting. The membrane was then probed with the rabbit polyclonal anti-Peg10-ORF2 antibody (1:1000; produced in-house), followed by incubation with horseradish peroxidase-conjugated donkey anti-rabbit IgG (1:5000; GE Healthcare, NA9340). The blots were developed using the ECL system according to the manufacturer's instructions (Thermo Fisher Scientific).

We thank T. Usami (MRI, TMDU) for generation of Peg10-ASG mice. Pacific Edit reviewed the manuscript prior to submission.

The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.199564