Deficiency and overexpression of Rtl1 in the mouse cause distinct muscle abnormalities related to Temple and Kagami-Ogata syndromes

Temple and Kagami-Ogata syndromes are caused by maternal and paternal disomy of chromosome 14 [Upd(14)mat and upd(14)pat], respectively. The individuals with Temple syndrome exhibit certain muscle-related symptoms, such as muscle hypotonia, feeding difficulty/poor sucking function in early childhood, in contrast to neonatal lethality due to respiratory problems associated with a bell-shaped thorax and abdominal wall hernia in the Kagami-Ogata syndrome (Kagami et al., 2005, 2015; Ogata and Kagami, 2016; Kotzot, 2004; Ioannides et al., 2014). Retrotransposon gag like 1 (RTL1) was first identified as PEG11 in a DLK1-DIO3 imprinted region in sheep (Charlier et al., 2001), and later was formally renamed RTL1 as one of the genes derived from a suchi-ichi-related LTR retrotransposon in humans and mice. Human RTL1 is a paternally expressed imprinted gene located in the same DLK1-DIO3 imprinted region on human chromosome 14; its mouse ortholog Rtl1 is on mouse distal chromosome 12 (Fig. S1). Interestingly, its mRNA is regulated by seven microRNAs (miRNAs) processed from maternally expressed RTL1as (RTL1 antisense transcript) through an RNAi mechanism (Davis et al., 2005) (Fig. S2A). We have previously demonstrated that mouse Rtl1 is one of the major genes responsible for the placental abnormalities characteristic of these two syndromes, and also that the severity of the phenotypes of Kagami-Ogata syndrome, such as a bell-shaped thorax and neonatal lethality associated with respiratory problems, is associated with the degree of overproduction of human RTL1 (Kagami et al., 2008). Beyond the essential involvement of Rtl1 in maintaining placental fetal capillaries, the role of Rtl1 in the respiratory failure and other muscular problems observed in neonates remains unknown, although recent studies report that the ectopic expression of ovine RTL1 leads to muscle hypertrophy in mice, which mimics the sheep callipyge phenotype (Charlier et al., 2001; Fleming-Waddell et al., 2009; Byrne et al., 2010; Xu et al., 2015; Mikovic et al., 2018).

In this study, we directly addressed this issue and extensively examined the role of Rtl1 in fetal muscle development at both cellular and tissue levels using two mouse models: Pat-Rtl1Δ (loss of Rtl1) and Mat-Rtl1Δ (overproduction of Rtl1) mice. As described in the original paper in detail (Sekita et al., 2008), the lethal phenotypes of the Pat- and Mat-Rtl1Δ mice are entirely dependent on the genetic background. The first generation [C57BL/6×129/Sv (B6/129) F1] did not exhibit any lethality, but a pre- and postnatal growth retardation phenotype was observed in Pat-Rtl1Δ mice. Subsequently, KO mice were continuously crossed to B6 mice and the lethality became apparent after the F3 and F6 generations in the Pat-Rtl1Δ and Mat-Rtl1Δ mice, respectively. The lethal phenotype became stronger in the later generations and finally all of the F10 individuals on the B6 genetic background died. For example, in the F5 generation, one half of the Pat-Rtl1Δ mice exhibited mid- to late fetal lethality because of the severe placental defect of destruction of the fetal capillary network (Sekita et al., 2008; Kitazawa et al., 2017) and another half exhibited late fetal growth retardation and were born small (∼80% of WT pups). Among them, one half (a quarter in total) died within 24 h of birth and another half were viable. In the case of the Mat-Rtl1Δ mice, neonatal lethality was predominant after the F6 generation, always in association with placentomegaly (150% by weight compared with the wild type). The placental fetal capillaries were also severely damaged, but in a different manner from those in Pat-Rtl1Δ. These results are consistent with those obtained using other Rtl1 and miR-127 deletion mutants, although both of those exhibited milder phenotypes than in our models (Ito et al., 2015). Although both the Pat- and Mat-Rtl1Δ mice had severe placental problems, we think it is necessary to address the abnormalities of the neonates themselves, especially the fetal/neonatal muscle phenotypes, and then the relationship to the Temple and Kagami-Ogata syndromes. In both cases, we detected severe but distinct abnormalities in several neonatal muscles that are essential for respiration, such as the intercostal, abdominal and diaphragm muscles. This is the first demonstration that an LTR retrotransposon-derived Rtl1 plays an important role in fetal and neonatal muscle development, strongly suggesting the crucial involvement of RTL1 in several muscle symptoms observed in the Temple and Kagami-Ogata syndromes. We further discuss an evolutionary role of the domestication of RTL1 in fetal/neonatal muscle development as a eutherian-specific gene from an LTR retrotransposon.

We used knockout (KO) mice to elucidate Rtl1 function in muscle development, as previously reported (Sekita et al., 2008) (Figs S1 and S2). In the KO allele, most of the Rtl1 region was removed, so six out of seven microRNAs in Rtl1as, including miR-434-3p, miR-434-5p, miR-127, miR-433-3p, miR-433-5p and miR-431, were also removed (Fig. S2B). The mice with the deletion upon paternal transmission exhibited a loss of Rtl1 expression phenotype (Pat-Rtl1Δ and Rtl1 KO), while the mice with maternal transmission exhibited overexpression of the paternal Rtl1 allele because of a loss of expression of most of the maternal microRNAs (Mat-Rtl1Δ and antiRtl1as KO). As the Pat- and Mat-Rtl1Δ mice were obtained from different mating combinations, we employed two control wild-type mice, WT(p) and WT(m), respectively (Fig. S2C). We used Pat-Rtl1Δ and Mat-Rtl1Δ mice in the F8 and F9 generations, respectively, because most of the pups exhibited neonatal lethality.

Rtl1 exhibits temporal muscle expression exclusively in the fetal to neonatal stages. In the embryo, its expression was detected by at least embryonic day 12.5 (E12.5), while in the skeletal muscles it was detected in E16.5 and E18.5 fetuses and neonates [postnatal day 0 (P0)]; however, its expression was gradually reduced after birth and almost entirely disappeared by P15 (Fig. 1A, Figs S3 and S4). In the heart, there was no Rtl1 expression throughout development (Fig. 1B and Fig. S3). Immunohistochemical experiments using an anti-Rtl1 antibody clearly demonstrated the presence of RTL1 protein in E14.5 fetal myofibers (Fig. S5) and almost all of the neonatal skeletal muscles in the wild-type and Mat-Rtl1Δ mice, but not in the Pat-Rtl1Δ mice (Fig. 1B). In the western blotting experiment using E16.5 fetal diaphragm and hind limb muscle, the RTL1 protein was detected mainly as a 250 kDa band and a concomitant minor band of around 200 kDa was also detected (Fig. S6). These values are roughly consistent with the estimated molecular weight of 199 kDa from the full-length 1744 amino acid sequence of the mouse RTL1 protein. RTL1 expression was apparently higher in Mat-Rtl1Δ compared with the wild type. In the Pat-Rtl1Δ fetal diaphragm, a very small amount of the lower minor band was detected, consistent with the mRNA expression data shown in Fig. S4B, suggesting leaky expression from the maternal alleles.

Hematoxylin and Eosin (HE) staining, as well as anti-laminin antibody staining revealed apparent structural abnormalities of the skeletal muscles in both the Pat-Rtl1Δ and Mat-Rtl1Δ mice. The Pat-Rtl1Δ mice exhibited significantly thinner muscle fibers compared with the wild-type controls (Fig. 2A,B, left), whereas the Mat-Rtl1Δ mice had a significantly larger muscle fiber size, as estimated by the anti-laminin antibody staining of non-fixed cryosectioned samples (Fig. 2A,B, right). Georgiades et al. also reported similar features in the diaphragm and forelimb muscles using mUPD12 and pUPD12 mice (Table 1: Georgiades et al., 2000). However, after fixation with Super Fix, the muscle fibers of the Mat-Rtl1Δ mice displayed severe shrinkage and became detached from the extracellular matrix (ECM) in the intercostal (Fig. 2C and Fig. S7A, Elastica van Gieson staining), diaphragm (Fig. S7B), abdominal (Fig. S7C) and limb (Fig. S7D) muscles. Therefore, a substantial muscle fiber cross-sectional area of the Mat-Rtl1Δ mice had shrunk as in the case of the Pat-Rtl1Δ mice (Fig. 2D and Fig. S7E), suggesting some unknown structural abnormality in the Mat-Rtl1Δ myofibers. Notably, the proportion of centronuclear muscle fibers was significantly high in the intercostal and limb muscles of the Mat-Rtl1Δ mice (P=0.000224 and P=0.000143, two-tailed Student's t-test, respectively), also suggesting defects in muscle fiber maturation (Georgiades et al., 2000; Fürst et al., 1989) (Fig. 2C,E, Fig. S7D,F). We did not observe such differences in the diaphragm or abdominal muscles (Fig. S7F).

Immunofluorescence experiments using both longitudinal and cross-sections of these skeletal muscles demonstrated the localization of the RTL1 protein in the muscle fibers and the sarcolemma as well as around the muscle fiber nuclei (Fig. 3A,B). We found that RTL1 protein is expressed in a typically striped pattern in the striated muscle, displaying a pattern strikingly similar to that of α-actinin, which is located at the sarcomeric Z-disc (Clark et al., 2002; Sjöblom et al., 2008), although the signals of RTL1 were not evidently merged with α-actinin under high magnification (Fig. S8A,B). However, RTL1 seemed to be partially merged with desmin, which is known as an intermediate filament as well as one of a sarcomeric cytoskeleton showing some of the links between membranes (sarcolemma and nuclear membrane) and sarcomeres at the Z-disc. This confirms the precise location of RTL in the adjoining area of the Z-disc (Goldfarb and Dalakas, 2009) (Fig. 3A). Importantly, in the cross-section, both RTL1 and desmin localized on the surface of myofibrils at the Z-disc in an alternated manner but were not merged, demonstrating that these two proteins are located next to each other (Fig. 3B and Fig. S8C). It has been reported that desmin (Des) is the earliest marker in muscle development, detected around E9.0 (Fürst et al., 1989; Milner et al., 1996). However, Des KO mice are normal at birth but exhibit severe skeletal and cardiac muscle defects around 2 weeks after birth, leading to partial lethality (Milner et al., 1996; Li et al., 1997). This corresponds with the time point when Rtl1 expression ceases in the skeletal muscles (Fig. 1). Thus, it is likely that RTL1 plays a specific role in the fetal/neonatal muscle fibers, such as stabilizing the muscle contractile apparatus and/or regulating muscle constriction, instead of (or together with) desmin. This seems consistent with the severe shrinkage of the muscle fibers observed in the Mat-Rtl1Δ mice after fixation (Fig. 2D and Fig. S7E).

Consistent with the in vivo data, Rtl1 expression was observed in undifferentiated proliferating muscle satellite cells (SCs), which are stem cells for skeletal muscle regeneration, obtained from mice at 3 weeks of age, as well as in SC-differentiated myoblasts (Fig. 4A) where levels of the muscle stem cell marker Pax7 and the early muscle differentiation markers Myf5 and Myod1 were typically decreased, whereas the levels of the late muscle differentiation markers MCK (Ckm) and Myh4 were increased (Relaix et al., 2005; Wang and Rudnicki, 2011; Buckingham and Rigby, 2014) (Fig. 4B). Pax7, Myf5, MCK and Myh4 levels significantly increased in Pat-Rtl1Δ compared with wild type.

Importantly, Rtl1 affects SC proliferation and the structural strength of the SC-differentiated myoblasts: SCs from the Pat-Rtl1Δ mice, which did not exhibit a substantial Rtl1 expression, proliferated at a significantly higher rate (1.5-fold) than the control cells (Fig. 4C,D), whereas the SCs from the Mat-Rtl1Δ mice proliferated at a significantly lower rate (×0.8) (Fig. 4C,D). This result demonstrates that Rtl1 has a repressive role in satellite cell proliferation, and its expression is required for normal growth. This finding was confirmed by EdU incorporation, which was significantly higher (×1.6) and lower (×0.7) than that in the controls (P=0.000857 and P=0.0369, two-tailed Student's t-test, respectively, Fig. 4E,F). We then counted the number of Pax7⁺ satellite cells in the muscles in the wild type, Pat-Rtl1Δ and Mat-Rtl1Δ mice, but there was no significant difference in their number (Fig. S9). Furthermore, the myoblasts that differentiated from the Pat-Rtl1Δ and Mat-Rtl1Δ SCs evidently displayed a weak or low structural strength of myoblast cells, because some of the myoblasts were detached from the culture dishes and exhibited a rounded shape (Fig. 4G, arrows, H). This finding may also explain why both the Pat-Rtl1Δ and Mat-Rtl1Δ mice had an abnormal neonatal muscle structure.

We confirmed that the expression of other imprinted genes in the same region was almost normal in these muscles (Figs S1 and S10). All of these results provide strong evidence that the loss and overproduction of Rtl1 are responsible for the muscle abnormalities observed in the Pat-Rtl1Δ and Mat-Rtl1Δ mice, respectively, as well as the muscle symptoms observed in Temple and Kagami-Ogata syndromes (Kagami et al., 2005, 2008, 2015; Ogata and Kagami, 2016; Kotzot et al., 2004; Ioannides et al., 2014; Georgiades et al., 2000; Sekita et al., 2008; Kitazawa et al., 2017). Overall, there is a robust correlation between the muscle phenotypes observed in individuals with Temple and Kagami-Ogata syndromes and those in the Pat-Rtl1Δ and Mat-Rtl1Δ model mice (Kagami et al., 2005, 2008, 2015; Ogata and Kagami, 2016; Kotzot et al., 2004; Ioannides et al., 2014; Georgiades et al., 2000; Sekita et al., 2008; Kitazawa et al., 2017) (Table 1 and Tables S1-S3). The Pat-Rtl1Δ mice and mUPD12 mice exhibit late fetal/neonatal lethality, whereas individuals with Temple syndrome show no lethality; however, this is an apparent genetic background effect in the mice (Georgiades et al., 2000; Sekita et al., 2008; Kitazawa et al., 2017). The individuals with Temple syndrome frequently exhibit muscular hypotonia and feeding difficulty/poor sucking function in early childhood, consistent with the muscle abnormalities in the Pat-Rtl1Δ mice. Although there were no data reported, we speculate that abnormalities similar to those in Pat-Rtl1Δ mice may be observed in the intercostal muscle, diaphragm and abdominal muscle of human UPD(14)mat and mUPD12 mice. The muscle abnormalities in the Mat-Rtl1Δ mice are nearly consistent with those in the individuals with Kagami-Ogata syndrome, such as frequent abdominal wall hernia and diastasis recti. Although a bell-shaped narrow thorax is a typical symptom of Kagami-Ogata syndrome, only a mild rib bone phenotype was observed, even in the pUPD12 mice (Georgiades et al., 2000). We speculate that severe damage to the intercostal muscle might lead to the deformation of the rib bone in humans, because muscle contractions could control skeletal morphogenesis during development (Shwartz et al., 2012, 2013). It seems reasonable that the abnormalities observed in the Rtl1 Pat- and Mat-Rtl1Δ mice would also be detected in the diaphragm, intercostal and abdominal muscles of human UPD(14)mat and UPD(14)pat patients.

The DLK1-DIO3 imprinted region is of considerable interest in terms of muscle development, but it is also very complicated. Many reports have indicated a relationship between imprinted genes in this region and muscle hypertrophy (the callipyge phenotype in sheep) and/or muscle development, especially paternally expressed DLK1, maternally expressed MIRG, a non-coding mRNA encoding the miR-379/miR-410 cluster containing 39 miRNAs, and paternally expressed RTL1 and RTL1as encoding the miR-127/miR-136 cluster containing the seven miRNAs in this study. The miR-379/miR-544 cluster (the anterior half of the miR-379/miR-410 cluster) is highly expressed in neonatal muscle. Mice with a deletion of the miR-379/miR-544 region upon maternal transmission exhibited significantly larger fast-switch muscles and MyHC type II-B fibers at P10 (Gao et al., 2015). In particular, miR-329 in the miR-379/miR-544 cluster has a regulatory role in Dlk1 expression (Gao et al., 2015). In this case, the expression of Rtl1 mRNA and RTL1 protein in the muscle tissues was comparable with wild type, implying no involvement of RTL1 in the hypertrophic phenotypes in these KO mice. Wüst et al. have also reported miR-1/miR133a-mediated inhibition of the Dlk1-Dio3 mega gene cluster. They demonstrated that the deletion of miR-1/miR133a clusters caused abnormally high expression of the miR-379/miR-544 cluster (encoded by Rian and Mirg in this report) as well as the Meg3 (Gtl2) miRNA cluster during muscle differentiation mediated via Mef2a upregulation. As several miRNAs in the miR-379/miR-544 cluster have the ability to target multiple mitochondria genes, the miR-1/miR133a deletion results in an abnormal metabolic maturation of the skeletal muscles (Wüst et al., 2018). However, by comprehensive miRNA analysis during skeletal muscle lineage progression, Castel et al. proposed that the miR-127/miR-136 cluster in Rtl1as may be more important for the differentiation, proliferation, commitment to myogenesis and self-renewal of SCs, because the SCs isolated from mice with a deletion of the entire miR-379/miR-410 cluster exhibit no differences in these abilities (Castel et al., 2018). We previously analyzed several patients with Kagami-Ogata syndrome-like phenotype without upd(14)pat and demonstrated that epimutations (hyper-DNA methylation) of the imprinting control region, the inter-genetic differentially methylated region (IG-DMR) of the DLK1-DIO3 imprinted region and the maternal microdeletion of the IG-DMR also cause Kagami-Ogata syndrome-like phenotypes (Kagami et al., 2008). The IG-DMR deletions sometimes extend into neighboring imprinted gene region. We also demonstrated that the severity of Kagami-Ogata syndrome exhibits good correlation with the degree of RTL1 overexpression but does not have any apparent correlation with the expression levels of DLK1, MIRG or other miRNA clusters (Kagami et al., 2008).

This study provides strong support that the loss and overproduction of human RTL1 are the major cause of the muscle symptoms observed in Temple syndrome and Kagami-Ogata syndrome, respectively. We cannot rule out the possibility that some of the miRNA(s) in the miR-127/miR-136 cluster in Rtl1as have different target gene(s) other than RTL1 and are involved in muscle development. For example, it has been reported that miR-431 in the miR-127/miR-136 cluster promotes myogenic differentiation by targeting Pax7 (Wu et al., 2014). The authors showed that miR-431 regulates the Pax7 levels during muscle development and regeneration using two miR-431 TG strains of mice (4- and 20-fold overexpression of miR-431, respectively), but there was no difference in Rtl1 expression in the skeletal muscle of miR-431 transgenic and wild-type mice. They concluded that miR-431 regulates myogenic differentiation independently of Rtl1. However, in our models, the Pax7 expression level was not affected in the Mat-Rtl1Δ SCs (without miR-431 and other five miRNAs), whereas it was upregulated in the Pat-Rtl1Δ SCs (Fig. 4B), indicating that the phenotypes we observed in this study are independent of the miR-431-Pax7 pathway.

In our mouse models, the damage to the muscles essential for respiratory function, such as the intercostal, diaphragm and abdominal muscles, was more severe than in the limb muscles in neonates. It is highly likely that these defects were the major cause of neonatal lethality in knockout mice models, thus indicating that the loss of RTL1 is a major cause of muscular hypotonia and feeding difficulty/poor sucking function seen in Temple syndrome, and that the overproduction of RTL1 induced by loss of RTL1as is a major cause of the respiratory problems and diastasis recti damage in Kagami-Ogata syndrome.

As mentioned above, individuals with Temple syndrome have muscle abnormalities consistent with those in the Pat-Rtl1Δ mice (Kotzot, 2004; Ioannides et al., 2014; Kagami et al., 2017). The phenotypes of the Pat-Rtl1Δ mice are more severe than those in individuals with Temple syndrome because the Pat-Rtl1Δ mice exhibit late fetal/neonatal lethality (Sekita et al., 2008; Kitazawa et al., 2017), whereas the Temple syndrome patients show no lethality (Kagami et al., 2008). However, this is an apparent genetic background effect in the mice because the severity gradually increases with the ratio of the B6 genotype, whereas F1 mice crossed with B6 and other strains do not exhibit any neonatal lethality, but instead exhibit both neo- and postnatal growth retardation patterns similar to those in individuals with Temple syndrome (Sekita et al., 2008; Kitazawa et al., 2017). Individuals with Temple syndrome lack any DLK1 and RTL1 expression. It is known that Dlk1 KO mice also exhibit growth retardation and neonatal lethal phenotypes to some extent (Moon et al., 2002), and overexpression of Dlk1 causes muscular hypertrophy (Davis et al., 2004). Loss of paternally inherited DLK1 is reported to frequently cause precocious puberty in females with Temple syndrome (Dauber et al., 2017). Therefore, it is clear that combined and differential involvement of DLK1 and RTL1 should be considered when investigating the symptoms of Temple syndrome.

As we mentioned in the Results, the muscle abnormalities in the Mat-Rtl1Δ mice are also consistent with the fact that abdominal wall hernia and diastasis recti frequently occur in individuals with Kagami-Ogata syndrome. Although a bell-shaped narrow thorax is a typical symptom of Kagami-Ogata syndrome (Kagami et al., 2005, 2008, 2015), only a mild rib bone phenotype was observed, even in the pUPD12 mice (Georgiades et al., 2000). We have previously demonstrated that the severity exhibits good correlation with the degree of RTL1 overexpression (Kagami et al., 2008). We speculate that severe damage to the intercostal muscle might lead to the deformation of the rib bone in humans, because muscle contractions control skeletal morphogenesis during development (Shwartz et al., 2012, 2013). Thus, the intercostal, diaphragm and abdominal muscle defects, along with the resulting bell-shaped narrow thorax, presumably reflect the evident neonatal respiratory problems. Recently, Loo et al. reported that RTL1 is involved in muscle regeneration under the control of the linker of nucleoskeleton and cytoskeleton (LINC) complex component SUN1 (Loo et al., 2019). We also observed the re-expression of RTL1 in regenerating muscles (Fig. S11), indicating an involvement of RTL1 in the basic eutherian muscle generation program. We also observed the re-expression of Rtl1 in regenerating muscles, suggesting a possible role for Rtl1 in muscle regeneration as well as in muscle generation in eutherians, although the mechanism of RTL1 in these processes is unknown and future studies should be directed to mechanistic biochemical and genetic investigations.

Our study clearly demonstrates that RTL1 is of crucial physiological significance because it plays a major role in the maturation and maintenance of fetal muscle fibers (Figs 2 and 4); therefore, its loss and overproduction affect the muscle phenotypes of Temple and Kagami-Ogata syndromes, respectively (Table 1 and Tables S1-S3). How, then, does RTL1 work in the fetal/neonatal muscle fibers? An immunostaining experiment revealed that RTL1 is closely located to desmin at the level of the Z-disc (Fig. 3). Desmin is part of the sarcomeric cytoskeleton and is involved in linking membranes and sarcomeres at the Z-disc, acting as a sub-sarcolemmal protein that is part of the costamere – a structural-functional component of the striated muscle cells at the periphery of the Z-disc (Goldfarb and Dalakas, 2009). This muscle-specific complex plays an important role in connecting the force-generating sarcomeres with the sarcolemma, which helps to couple the sarcomere to the extracellular matrix (ECM). Thus, it is likely that the RTL1 protein plays a specific role in the function of fetal/neonatal muscle fibers, such as stabilizing the muscle contractile apparatus and/or regulating muscle constriction with desmin. Therefore, its loss and overproduction affect the strength of muscle fibers, as shown in Figs 2G and 4G. We also speculate that, at normal expression levels, the RTL1 protein also plays a role as a suppressor of desmin to prevent fast and vigorous muscle movement in the fetal/neonatal stages by interacting or interfering with desmin. This speculation is supported by the fact that desmin KO mice exhibit no gross abnormality in the fetal and neonatal periods; rather, they exhibit severe muscle defects around 2 weeks after birth (Milner et al., 1996; Li et al., 1997). Thus, it is possible that desmin starts functioning at the Z-disc with the disappearance of the RTL1 protein at the Z-disc. This suggests that no Rtl1 expression is observed in the heart throughout development (Fig. 1A,B), because the cardiac muscles must continuously function after the formation of the heart at E8.5. This view appears to be consistent with the observation that human babies and mouse pups exhibit slow and weak muscle movement immediately before and after birth. This type of inhibitory function of RTL1 in the fetal and neonatal muscles seems advantageous for both mothers, fetuses and neonates to ensure a safe pregnancy and to aid child-rearing in the current eutherian reproductive system. Our study also demonstrates that the eutherian skeletal muscle is unique because of the recruitment of RTL1 in fetal and neonatal muscle development, implying that it was one of the adaptations to the mammalian viviparaous reproduction system.

Despite its evolutionary origin as an LTR retrotransposon, RTL1 also plays an essential role in the maintenance of placental fetal capillaries in adapting to the long gestation period in the current eutherian reproduction system as a eutherian-specific gene (Sekita et al., 2008; Kagami et al., 2008; Kitazawa et al., 2017; Charlier et al., 2001; Edwards et al., 2008). Thus, this work provides additional evidence that LTR retrotransposons have exerted a profound impact on a variety of eutherian-specific traits, including the skeletal muscles as well as the placenta and the brain (Sekita et al., 2008; Kitazawa et al., 2017; Edwards et al., 2008; Ono et al., 2006; Naruse et al., 2014; Irie et al., 2016) and that ‘gene acquisition from exogenous DNAs’ is a crucial driving force in therian/eutherian evolution (Irie et al., 2015; Gould and Vrba, 1982; Brosius and Gould, 1992; Kaneko-Ishino and Ishino, 2012, 2015).

All animals and experimental procedures were approved by the Animal Ethics Committees of Tokyo Medical and Dental University. The Rtl1 KO mice were generated by using ES cells (CCE) of 129/SvEv mouse origin, as previously described (Sekita et al., 2008). The Rtl1 KO lines (Mat-Rtl1Δ and Pat-Rtl1Δ) were maintained by continuous crossing with male and female C57BL/6J mice (WT), respectively, and mice in the F8 (Rtl1 KO: Pat-Rtl1Δ) and F9 (antiRtl1as KO: Mat-Rtl1Δ) generations were used in this study.

The genomic DNA and total RNA samples were prepared from fetuses, neonates and 1- to 2-week-old mice by using TRIzol Regent (Life Technologies). cDNA was synthesized from 1-5 μg of total RNA using SuperScript III Reverse Transcriptase (Invitrogen) with the following oligo-dT+ adaptor primer: 5′-CTGATCTAGAGGTACCGGATCCGACTCGAGTCGACATCGTTTTTTTTTTTTTTTTT-3′. For the RT-PCR analysis of Rtl1, 10 ng of cDNA in a 25 μl reaction mixture containing 1×KOD FX buffer (KFX-101, TOYOBO), 200 μM of each dNTP, primers and 0.5 units of KOD FX were subjected to 30 cycles at 98°C for 15 s, 69°C for 30 s and 74°C for 30 s using a C1000 Touch thermal cycler (BioRad). The following primer sequences were used: Rtl1, 5′-TCCAAGGAGCATTCGACGTACCAGTGTGACTTACC-3′; an adaptor primer for both genes, 5′-AGAGGTACCGGATCCGACTCGAGTCGACATCG-3′; and Gapdh, 5′-CACTCTTCCACCTTCGATGC-3′ and 5′-CTCTTGCTCAGTGTCCTTGC-3′.

The quantitative real-time PCR was performed using 5 ng of cDNA in a THUNDERBIRD SYBR qPCR Mix (QPS-201, TOYOBO). The cycle conditions were 95°C for 1 min, followed by 40 cycles of 95°C for 10 s, 60°C for 20 s and 72°C for 10 s using the LightCycler 480 apparatus (Roche). The gene expression levels were normalized to Gapdh. The following primer sequences were used for this study: Rtl1, 5′-GAGTACTGTGCCAAGGAGCC-3′ and Adaptor primer; Dlk1/Peg9, 5′-TTACCGGGGTTCCTTAGAGC-3′ and 5′-TGCATTAATAGGGAGGAAGGG-3′; Meg3/Gtl2, 5′-TTGCACATTTCCTGTGGGAC-3′ and 5′-AAGCACCATGAGCCACTAGG-3′; Meg8/Rian, 5′-TCGAGACACAAGAGGACTGC-3′ and 5′-ATTGGAAGTCTGAGCCATGG-3′; Meg9/Mirg, 5′-TTGACTCCAGAAGATGCTCC-3′ and 5′-CCTCAGGTTCCTAAGCAAGG-3′; and Dio3, 5′-TCGAGATAGGGAAAGGGTGG-3′ and 5′-GAACCTCGCAGATTGATTCC-3′.

The protein samples were extracted from E16.5 fetuses using 8 M urea, 2% CHAPS and cOmplete EDTA-free protease inhibitor cocktail (Roche). We loaded 3 μg of total protein into the well of the 7.5% SDS-PAGE gel and ran the gel for 35 min at 200 mV. We transferred the protein from the gel to the PVDF membrane and the PVDF membrane was blocked with 5% skim milk in TBS-T at 4°C overnight. We incubated the membrane with rabbit anti-Rtl1 antibody (1:200; Sekita et al., 2008) in TBS-T as the primary antibody at room temperature for 2 h and with anti-rabbit IgG, with an HRP-linked antibody as the secondary antibody (1:5000; Cell Signaling Technology, #7074) in TBS-T at room temperature for 1 h. Western blotting was performed using ECL Select Western Blotting Detection Reagent (GE Healthcare).

Mouse fetuses and neonates were fixed by using Super Fix (KURABO), soaked in 5% formic acid in 70% ethanol at 4°C overnight for 2 nights; dehydrated in 70% and 90% ethanol for 2 h each, in 100% ethanol for 2 h three times, and in xylene for 2 h four times; and finally embedded in paraffin wax. The paraffin blocks were sectioned at 5 μm with a microtome and mounted on Superfrost Micro Slides (Matsunami Glass). The sections were deparaffinized three times in xylene for 20 min, three times in 100% ethanol for 5 min, and in 90% and 70% ethanol for 5 min each; the sections were then stained with Hematoxylin for 10 min, and washed with tap water for 2 min. After being stained with Eosin for 1 min, the sections were immersed once in 70% and 80% ethanol for a few seconds each and three times in 100% ethanol for 3 min, then dehydrated three times in xylene for 3 min and mounted with Malinol mounting medium (MUTO).

Another section was stained with resorcin-fuchsin solution for 1 h, immersed in 100% ethanol for a few seconds and washed with tap water for 10 min. After being stained with Weigert iron Hematoxylin solution for 5 min, the sections were washed with tap water for 7 min. The sections were then stained with van Gieson solution for 10 min and immersed once each in 70% and 80% ethanol for a few seconds and three times in 100% ethanol for 3 min, then dehydrated three times in xylene for 2 min and mounted with Malinol mounting medium.

For antigen retrieval, the sections were boiled in Immunosaver (1:200; Nissin EM) at 98°C for 40 min and then immersed (dehydrated) in ice-cold methanol at −30°C overnight. After being air dried, the sections were blocked with 10% goat serum, 1% bovine serum albumin (BSA: Sigma Aldrich) and 0.1% Triton-X 100 (WAKO) in PBS at room temperature for 1 h.

For the immunohistochemistry analysis, anti-Rtl1 antibody (1:200) was used as the primary antibody and was prepared in 1% BSA and 0.1% Triton-X 100 in PBS at 4°C overnight (for more than 20 h). This primary reaction was developed with a biotinylated goat anti-rabbit IgG secondary antibody (1:200; Vector Laboratories) for 2.5 h, then incubated with alkaline phosphatase (AP) complex (1:200; Vector Laboratories) for 1 h. The histochemical detection of the alkaline phosphatase activity was performed with BCIP/NBT (Vector Laboratories) in 100 mM Tris-HCl at pH 9.8 and mounted with VectaMount AQ Mounting Medium (Vector Laboratories). The images were captured using BIOREVO (Keyence).

The mouse neonates were corrected and embedded in OCT compound (Sakura Finetek). The OCT blocks were sectioned using a cryostat (MICROME), at 14 μm and mounted on Superfrost Micro Slides. The cryosections were fixed in 4% paraformaldehyde (PFA: Nacalai tesque) for 10 min at room temperature and washed three times with PBS for 5 min, then stained with Hematoxylin for 2 min and washed with tap water for 2 min. After being stained with Eosin for 1 min, the slides were immersed once in 70% and 80% ethanol for a few seconds each and three times in 100% ethanol for 2 min, then dehydrated three times in xylene for 3 min and mounted with Malinol mounting medium.

The cryosections were fixed in 4% PFA for 10 min at room temperature and washed three times with PBS for 5 min. For the antigen retrieval, the sections were boiled in 0.01 M citric acid solution (pH 6.0) at 80°C for 10 min and then immersed (dehydrated) in ice-cold methanol at −30°C for 6 min. After being air dried, the sections were blocked with 10% goat serum, 5% BSA and 0.5% TritonX-100 in PBS at room temperature for 1 h, and then incubated with the primary antibody, which was prepared in 5% BSA and 0.5% TritonX-100 in PBS at 4°C overnight (∼16 h).

For the immunofluorescence staining, anti-Rtl1 antibody (1:200, Sekita et al., 2008), anti-desmin antibody (Roche, DE-R-11, non-diluted), anti-Laminin antibody (1:200; Sigma Aldrich, L9393), anti-α-actinin (sarcomeric) antibody (1:100; Sigma Aldrich, A7811) and Pax7 (1;100; Santa Cruz, sc-81648) were used as the primary antibodies. Alexa Fluor 488-conjugated anti-mouse IgG (1:1000; ThermoFisher Scientific, A11008) and 555-conjugated anti-rabbit IgG (1:1000; ThermoFisher Scientific, A28180) or 488-conjugated anti-rabbit IgG (1:1000; ThermoFisher Scientific, A11034) were used as the secondary antibodies and stained with DAPI (1:1000; Wako, 340-07971) for 1 h. The slides were mounted in VectaShield (Vector Laboratories). The images were captured with an LSM710 confocal imaging system (Zeiss).

Satellite cells were isolated from the anterior tibialis anterior (TA) and gastrocnemius muscles of 3-week-old male mice. The muscle fragments were incubated in medium A [GlutaMax-DMEM (Gibco), containing 1% penicillin-streptomycin (Sigma Aldrich)] at 37°C for 30 min, then minced well. The minced muscles were digested in medium B [medium A containing 0.14% protease (Sigma Aldrich)] at 37°C for 10 min. After centrifugation at 400 rpm (30 g) for 30 s, the supernatant was discarded. The pellet was resuspended in medium B and incubated at 37°C for 10 min. After centrifugation at 400 rpm (30 g) for 30 s, the supernatant was collected. This procedure was repeated two additional times. The collected supernatant was filtered through a 100 μm cell strainer (Falcon), mixed with medium C [medium A containing 10% fetal bovine serum (FBS; Gibco)] and centrifuged at 2000 rpm (400 g) for 5 min. The pellet was resuspended in 10 ml of medium C and centrifuged one more time. The pellet was then resuspended in growth medium [GM; DMEM-GlutaMax with 20% FBS, 1% penicillin-streptomycin, 1% chicken embryonic extract (CEE; US Biological) and 10 ng/ml basic fibroblast growth factor (bFGF; Falcon)] and seeded onto Matrigel Matrix (Falcon)-coated dishes. The cells were cultured at 37°C under 5% CO₂ in a humidified chamber, and the medium was changed every 2 days.

Total cellular RNA were isolated using a Qiagen RNeasy kit according to the manufacturer's protocol. The cDNA was synthesized from 250 ng of total RNA using SuperScript III Reverse Transcriptase with the oligo-dT+ Adaptor primer. For the RT-PCR, 10 ng of cDNA in a 25 μl reaction mixture containing 1× Ex Taq buffer, 200 μM of each dNTP, primers and 0.5 units of Ex Taq Hot Start were subjected to 30 cycles at 96°C for 15 s, 60°C for 20 s and 72°C for 30 s in a C1000 thermal cycler (BioRad). The following primer sequences were used: Pax7, 5′-AGGCCTTCGAGAGGACCCAC-3′ and 5′-CTGAACCAGACCTGGACGCG-3′; Myf5, 5′-TGAAGGATGGACATGACGGACG-3′ and 5′-TTGTGTGCTCCGAAGGCTGCTA-3′; Myod1, 5′-GGCTACGACACCGCCTACTA-3′ and 5′-GAGATGCGCTCCACTATGCT-3′; myogenin, 5′-AGTGAATGCAACTCCCACAG-3′ and 5′-ACGATGGACGTAAGGGAGTG-3′; MCK, 5′-CTGACCCCTGACCTCTACAAT-3′ and 5′-CATGGCGGTCCTGGATGAT-3′; Myh4, 5′-GCTTGAAAACGAGGTGGAAA-3′ and 5′-CCTCCTCAGCCTGTCTCTTG-3′; Gapdh, 5′-CACTCTTCCACCTTCGATGC-3′ and 5′-CTCTTGCTCAGTGTCCTTGC-3′.

To induce muscle regeneration, 100 μl of 10 mM cardiotoxin (CTX; C9759, Sigma-Aldrich) were injected intramuscularly into the tibialis anterior (TA) muscle of anesthetized 10-week-old male mice using a 29 G syringe. Regenerating muscles were isolated 2, 4 and 7 days after CTX injection.

We thank Drs Y. Sekita (Kitasato University, School of Science, Tokyo, Japan), D. Endo (Nagasaki University, Graduate School of Biochemical Science, Japan), H. Shiura (University of Yamanashi, Kofu, Japan) and J. Lee (Tokyo Medical and Dental University, Japan) for their helpful advice.

The peer review history is available online at https://dev.biologists.org/lookup/doi/10.1242/dev.185918.reviewer-comments.pdf