Pluripotent stem cell-derived myogenic progenitors remodel their molecular signature upon in vivo engraftment

Significance

These studies provide mechanistic insights into the nature of in vitro-generated iPSC-derived myogenic progenitors and, importantly, what these cells become upon engraftment. The fact that iPSC-derived myogenic progenitors remodel their embryonic/fetal molecular signature upon in vivo exposure to the adult muscle environment has important implications for therapeutic applications, as attested by their robust and persistent regenerative potential.

Abstract

Optimal cell-based therapies for the treatment of muscle degenerative disorders should not only regenerate fibers but provide a quiescent satellite cell pool ensuring long-term maintenance and regeneration. Conditional expression of Pax3/Pax7 in differentiating pluripotent stem cells (PSCs) allows the generation of myogenic progenitors endowed with enhanced regenerative capacity. To identify the molecular determinants underlying their regenerative potential, we performed transcriptome analyses of these cells along with primary myogenic cells from several developmental stages. Here we show that in vitro-generated PSC-derived myogenic progenitors possess a molecular signature similar to embryonic/fetal myoblasts. However, compared with fetal myoblasts, following transplantation they show superior myofiber engraftment and ability to seed the satellite cell niche, respond to multiple reinjuries, and contribute to long-term regeneration. Upon engraftment, the transcriptome of reisolated Pax3/Pax7–induced PSC-derived myogenic progenitors changes toward a postnatal molecular signature, particularly in genes involved in extracellular matrix remodeling. These findings demonstrate that Pax3/Pax7–induced myogenic progenitors remodel their molecular signature and functionally mature upon in vivo exposure to the adult muscle environment.

Cell-replacement therapies hold great promise for the treatment of muscular dystrophies, since diseased muscle can be replaced with healthy and functional tissue regardless of the underlying genetic defects (1). The ideal therapeutic cell type should also seed the satellite stem cell compartment, thus providing long-term regeneration for the engrafted muscle (2, 3). Pluripotent stem cells (PSCs) represent an alluring option, as these cells can be expanded indefinitely while having the ability to differentiate toward any cell type in vitro (4, 5). We have shown that upon transplantation into dystrophic mice, both Pax3- and Pax7-induced mouse and human PSC-derived myogenic progenitors are able not only to generate new functional myofibers but also to seed the satellite cell compartment, thus providing long-term regeneration (6–8). Although these findings indicate the ability of in vitro-generated PSC-derived myogenic progenitors to behave functionally like satellite cells, we do not know whether they share the same molecular signature.

To accurately determine the molecular signature of PSC-derived myogenic progenitors in regard to the different stages of myogenic development, we compared their transcriptome profiles with those of primary skeletal myogenic progenitors isolated at different developmental stages. To test the hypothesis that exposure to the adult host skeletal muscle environment may induce molecular changes in transplanted cells, we performed transcriptome analysis on PSC-derived mononuclear cells (MNCs) reisolated after engraftment. Our findings reveal that in vitro-generated PSC-derived myogenic progenitors display a distinct signature, more closely related to the embryonic/fetal molecular signature, but that this shifts toward the early neonatal stage upon engraftment. This study provides a direct comparison between in vitro-generated PSC-derived myogenic progenitors and their in vivo-engrafted counterparts and, importantly, suggests that PSC-derived myogenic progenitors remodel their transcriptional profile upon interaction within the recipient environment.

Results

Molecular Signature of in Vitro-Generated Myogenic Progenitors.

Conditional expression of Pax3 or Pax7 in the appropriate developmental window allows for the generation of large numbers of skeletal myogenic progenitors endowed with significant in vivo regenerative potential, as their transplantation into dystrophic muscles results in new myofibers that incorporate normally into the host muscle and in the rescue of muscle contractility (6–10). To determine the developmental molecular signature of PSC-derived myogenic progenitors, we performed RNA sequencing on both inducible (i)Pax3 and iPax7 cell preparations, and compared them with four developmental stages: embryonic, fetal, neonatal, and adult. For the prenatal stages, we used the Myf5Cre Rosa26-YFP mouse strain to freshly isolate embryonic and fetal myogenic progenitors from E10.5 and E14.5 mouse embryos, respectively (11). For the postnatal groups, we utilized the Pax7-ZsGreen reporter mouse model (12), which allowed us to isolate satellite cells from neonatal (Neo) and adult mice at day 3 (P3) and at 3 mo old, respectively (Fig. 1A). We chose the Myf5Cre Rosa26-YFP model for the prenatal reference samples due to feasibility, since it allows for the reliable isolation of myogenic progenitors, as opposed to the Pax7-ZsGreen reporter, which, at these early time points, does not clearly separate the Pax7⁺ population from the negative fraction, in addition to only allowing for the recovery of very small numbers of myogenic precursors (SI Appendix, Fig. S1A). To validate our approach, we confirmed that Myf5Cre Rosa26-YFP sorted myoblasts (YFP⁺) express Pax7 abundantly, and also that the YFP⁻ population does not contain Pax7⁺ cells (SI Appendix, Fig. S1 B–E). Moreover, we observed that more than 95% of Pax7-ZsGreen⁺ cells express Myf5 at both embryonic and fetal stages (SI Appendix, Fig. S1 B–D). RNA sequencing on biological replicates for each group allowed us to identify 3,111 differentially expressed genes (DEGs; Fig. 1A), whose expression levels displayed a fold change higher than 4 and P value lower than 0.05 between iPax3/iPax7 samples and at least one of the reference groups. We conducted an unsupervised analysis of these DEGs and found that (i) iPax3 and iPax7 PSC-derived myogenic progenitors share very similar transcriptional profiles (Fig. 1 B and C), having only 12 DEGs (SI Appendix, Fig. S2A), and thus were subsequently grouped together (referred to as iPax hereafter), and (ii) iPax PSC-derived myogenic progenitor transcriptional profiles are closer to prenatal groups, in particular to embryonic myogenic progenitors, as shown by principal component analysis (PCA) of differentially expressed genes (Fig. 1B).

Fig. 1.

Molecular signature of Pax3/Pax7–induced PSC-derived myogenic progenitors. (A) Schematic representation of experimental groups: iPax3 and iPax7 in vitro-generated PSC-derived myogenic progenitors (iPax), Myf5⁺ embryonic (E10.5) and fetal (E14.5) myoblasts, and Pax7⁺ neonatal (P3) and adult (3 mo) satellite cells. (B) Principal component analysis shows the distribution of samples according to DEGs. PC1 and PC2: 65.7 and 18.5% variance, respectively. (C) Heatmap representing DEGs between iPax3/7 cells and reference samples. Analyzed clusters, identified based on hierarchical gene clustering, are highlighted in colored boxes: red box for genes overexpressed in all iPax samples (group a); orange for the subgroups of genes overexpressed in iPax3 and iPax7 cells (groups a.1 and a.2, respectively); yellow for genes overexpressed in iPax cells and embryonic and fetal myoblasts (group b); pink for genes overexpressed in iPax cells, fetal myoblasts, and neonatal satellite cells (group c); and green for genes overexpressed in iPax and satellite cells (group d).

Among the genes found overexpressed in the iPax samples (Fig. 1C), we identified four major subgroups based on hierarchical gene clustering (Dataset S1): (i) genes overexpressed in iPax samples only (group a; red square); (ii) genes overexpressed in iPax and both prenatal developmental stages (group b; yellow square); (iii) genes overexpressed in iPax, fetal myoblasts, and neonatal satellite cells (group c; pink squares); and (iv) genes overexpressed in iPax and adult satellite cells (group d; green squares). Using Panther overrepresentation enrichment analysis, we identified integrin signaling to be the most represented pathway in the iPax cell population (SI Appendix, Fig. S2B). To further investigate the identified subgroups, we conducted gene ontology (GO) functional annotation and evaluated the top enriched biological processes. Genes exclusively overexpressed in iPax samples (group a) belong to categories such as negative regulation of apoptosis and the tricarboxylic acid cycle, indicative of the mitochondrial respiratory chain (SI Appendix, Fig. S2C), which agree with the sustained in vitro expansion of PSC-derived myogenic progenitors (7, 8). Moreover, we found genes involved in cell migration, including Integrin (Itg) a2, a3, and b4 (SI Appendix, Fig. S2C). We then functionally annotated the genes with respect to developmental references. Genes selectively expressed in iPax and prenatal myogenic progenitors are involved in cell-cycle regulation and cell division (SI Appendix, Fig. S2D). This observation suggests that the proliferative capacity of in vitro PSC-generated myogenic progenitors is similar to earlier developmental stages and superior to neonatal/adult satellite cells. iPax progenitors were also found to express genes involved in adhesion and ECM organization, at levels equivalent to fetal myoblasts and neonatal satellite cells (SI Appendix, Fig. S2E). Analysis of the annotated genes overexpressed in iPax and adult satellite cells revealed only three major enriched categories (SI Appendix, Fig. S2F), confirming that in vitro, PSC-derived myogenic progenitors mainly resemble earlier developmental stages of skeletal muscle cells.

We were not surprised with the finding that iPax3 and iPax7 clustered together (Fig. 1B), since these progenitors behave similarly in vitro and in vivo (8). Accordingly, transcriptional differences between iPax3 and iPax7 are minimal (SI Appendix, Fig. S2A). Nevertheless, two subgroups appeared to be more expressed in iPax3 or iPax7 progenitors (Fig. 1C, subgroups a.1 and a.2, respectively, and SI Appendix, Fig. S2 G and H). Functional annotation revealed that most of the genes found overexpressed in iPax3 myogenic progenitors are involved in cell organization and migration (SI Appendix, Fig. S2G), while genes highly expressed in iPax7 samples are mainly involved in skeletal muscle development (SI Appendix, Fig. S2H), in accordance with the respective developmental roles of these two transcription factors (13). Altogether, these data indicate that iPax3/7 PSC-derived myogenic progenitors more closely recapitulate earlier developmental stages of skeletal myogenesis, including unique attractive features, such as migratory and proliferative potential.

To evaluate whether the expression of genes involved in migration in the iPax groups is functionally relevant, first we performed an in vitro cell-adhesion assay (14) in which we assessed the ability of iPax samples to adhere to a layer of endothelial cells in comparison with fetal progenitors and adult satellite cells. Using a fluorescent tracker, we observed that the ability of iPax3 and iPax7 PSC-derived myogenic progenitors to adhere to the endothelial layer was comparable to fetal progenitors but significantly superior to adult satellite cells (SI Appendix, Fig. S3 A and B). Next, we assessed the ability of iPax cells to migrate through a layer of ECM proteins in comparison with the same reference groups using a transwell migration assay. The migratory ability of iPax3 and iPax7 cells was found comparable to fetal progenitors, which is significantly superior to adult satellite cells (SI Appendix, Fig. S3 C and D). Of note, iPax3 cells showed greater migratory ability than iPax7 counterparts, thus supporting the functional annotation data (SI Appendix, Fig. S2G).

Unique Molecular Features of PSC-Derived Myogenic Progenitors.

Next, we singularly compared the transcriptional profile of iPax samples with each of the developmental reference groups. For each comparison, the identified DEGs were further distinguished between over- and underexpressed genes (SI Appendix, Fig. S4 A and D, respectively), and Venn diagrams were generated to identify overlapping genes. GO assessment revealed that PSC-derived myogenic progenitors uniquely express high levels of genes involved in cell adhesion and muscle development (SI Appendix, Fig. S4B). Among the cell-motility/adhesion genes, we found class III secreted Semaphorin 3E (Sema3e) (15) and Itga3 (SI Appendix, Fig. S4C) (16), which were expressed most highly in iPax3. We also identified Eyes absent homolog 2 (Eya2) and phosphorylatable myosin light chain (Mylpf), genes associated with skeletal muscle differentiation and development (17), and at higher levels in the iPax7 group (SI Appendix, Fig. S4A). Moreover, T-box transcription factor 1 (Tbx1), involved in limb and cranial mesoderm formation (17), and N-cadherin (Cdh2), associated with muscle differentiation (18), were highly expressed in iPax samples with respect to all reference groups. These data further highlight that the iPax samples express high levels of several genes involved in early myogenesis.

We then conducted the same analysis focusing on the underexpressed genes, and found 179 transcripts that were expressed at lower levels in iPax samples in comparison with all reference groups (SI Appendix, Fig. S4D). GO functional annotation showed that iPax samples express very low levels of genes involved in collagen organization and elastic fiber assembly (SI Appendix, Fig. S4E), such as Collagen type I alpha 1 chain (Col1a1) and Tenascin XB (Tnxb) (SI Appendix, Fig. S4F). Among others, genes involved in the ERK1/2 signaling cascade, including Insulin-like Growth factor 1 (Igf1) and Delta-Like 4 (Dll4), were also much lower expressed in iPax samples with respect to their in vivo reference samples, except for the adult stage. As Dll4 is also a ligand for Notch signaling, an important pathway in muscle stem cell activation (19), we looked for other interactors in the same pathway and found that the ligand Jag1 and the downstream effector Hey1 are also expressed at very low levels in PSC-derived myogenic progenitors (SI Appendix, Fig. S4F). Jag1 and Hey1 were found highly expressed in the adult, unlike Dll4, in accordance with previous literature (20).

PSC-Derived Myogenic Progenitors Have Superior in Vivo Regenerative Potential than Primary Prenatal Counterparts.

As previously reported, PSC-derived myogenic progenitors possess in vivo regenerative potential and the ability to repopulate the satellite cell compartment (7, 8). At the same time, our transcriptome analysis revealed that in vitro-generated myogenic progenitors resemble prenatal skeletal muscle progenitors. To elucidate the in vivo significance of these findings, we compared the regenerative potential of iPax cells with prenatal myogenic progenitors and adult satellite cells. We transplanted 300,000 myogenic progenitors from each group into irradiated and cardiotoxin (CTX)-injured tibialis anterior (TA) muscles of NSGmdx^4cv mice, with the exception of adult satellite cells, for which we transplanted 5,000 cells, a cell number previously shown to be optimal under this preconditioning setting (21). When comparing all nonadult cell preparations injected at the same cell number, robust engraftment, as demonstrated by dystrophin expression, was observed only upon transplantation of iPax progenitors, whereas embryonic and fetal myoblasts gave rise to many fewer myofibers (Fig. 2 A and B and SI Appendix, Fig. S5 A and B). Total engraftment of iPax cells was higher than that of satellite cells (SI Appendix, Fig. S5 A and B), although not on a per-cell basis, since the number of iPax-injected cells was much higher than the number of satellite cells. Next, we investigated satellite cell contribution. Immunostaining revealed the presence of comparable numbers of donor-derived Pax7⁺ cells under the basal lamina for the postnatal and iPax cohorts (Fig. 2 C and D and SI Appendix, Fig. S5 C and D), whereas muscles injected with fetal cells showed minimal satellite cell contribution, in accordance with lower levels of muscle engraftment (Fig. 2 A and B). These results were corroborated by flow cytometry. For prenatal myogenic progenitors, only 5.7 ± 2.3% of the Vcam1⁺Itga7⁺ satellite cell population (22–24) was YFP-positive (Fig. 2 E–G), while in iPax-transplanted muscle, more than 80% of this population was donor-derived (85.6 ± 2.4% and 84.2 ± 4.2% GFP⁺ for iPax3 and iPax7, respectively). This high level of engraftment was similar to our adult satellite cell reference group, in which recipients were transplanted with 3,000 Pax7-ZsGreen⁺ cells (86.5 ± 3.5%; Fig. 2 E–G). Of note, contrary to adult satellite cells, both freshly isolated E14.5 fetal progenitors and in vitro-generated iPax cells do not have a distinct Itga7⁺Vcam1⁺ subpopulation of cells before transplantation (SI Appendix, Fig. S5E). When reisolated iPax-derived Vcam1⁺Itga7⁺ cells are subjected to in vitro differentiation, resulting myotubes showed lower fusion ability than those from freshly isolated Vcam1⁺Itga7⁺ satellite cells, cultured under the same conditions (SI Appendix, Fig. S5F), suggesting that the in vitro assay does not fully capture the in vivo regenerative capacity of myogenic progenitors in response to injury.

Fig. 2.

Engraftment potential of iPax cells is superior to prenatal myoblasts. (A) Representative images show donor-derived dystrophin⁺ myofibers (Dys; green) in TA muscles of NSGmdx^4cv mice transplanted with the indicated groups. PBS-injected TA muscles represent control. Nuclei were counterstained with DAPI (blue). (Scale bar, 100 µm.) (B) Graph shows quantification of Dys⁺ myofibers shown in A (n = 4). (C) Representative images show Pax7 (red), Dys (green), and Laminin (Lam; white) staining of injected samples in NSGmdx^4cv mice. White arrowheads indicate Pax7⁺ cells. (Scale bar, 100 µm.) (D) Graph shows quantification of Pax7⁺ nuclei shown in C per field at 20× magnification (n = 4). (E) Representative FACS plots show gate settings for isolation of lineage negative (Lin⁻), Itga7⁺Vcam1⁺, YFP⁺ (fetal), GFP⁺ (iPax3 and iPax7), and ZsGreen⁺ (adult) MNCs reisolated from injected muscles. PI, propidium iodide. (F and G) Percentage of Lin⁻Itga7⁺Vcam1⁺ (F) and fluorescent reporter⁺ (G) of analyzed groups from E (n = 3). Error bars represent SEs; ***P < 0.001 and ****P < 0.0001.

To further confirm the extent of the regenerative potential of iPax progenitors, we labeled these cells with a luciferase-encoding lentiviral vector to assess the ability of engrafted cells to respond to multiple rounds of injury. In vivo imaging confirmed the presence of donor-derived myofibers up to 20 wk from a single cell injection, with or without three rounds of CTX reinjuries (SI Appendix, Fig. S6A). Importantly, engrafted iPax myogenic progenitors are able to respond to these multiple reinjuries by proliferating and eventually generating new luciferase-positive myofibers, as shown by bioluminescence measurements and dystrophin staining (SI Appendix, Fig. S6 B and C). These data indicate that iPax PSC-derived myogenic progenitors, despite having a molecular signature closer to earlier developmental stages, are phenotypically distinct from these cells, being able to achieve significant and persistent in vivo engraftment.

iPax Progenitors Change Their Molecular Signature Upon in Vivo Engraftment.

Next, we sought to uncover the molecular determinants behind the superior engraftment potential of PSC-derived progenitors with respect to prenatal myoblasts, and the regenerative behavior comparable to adult satellite cells. For this, we transplanted GFP-labeled iPax cells, and reisolated for transcriptional analysis GFP⁺ donor-derived MNCs also expressing CD34. CD34 was barely detectable in the in vitro-generated iPax population but expressed by the majority of the GFP⁺ MNC fraction (SI Appendix, Fig. S7A). Three replicates for both iPax3 and iPax7 GFP⁺CD34⁺ MNCs were transcriptionally profiled along with freshly isolated adult satellite cells from Pax7-ZsGreen mice (Fig. 3A). Unsupervised analysis showed that iPax3 and iPax7 share a very similar transcriptional profile, and therefore were grouped together for further analyses. Over the 21,214 expressed transcripts, only 397 genes were found to be differentially expressed between iPax MNCs and satellite cells (Dataset S2), a much lower proportion with respect to the genes that were differentially expressed between in vitro-generated iPax myogenic progenitors and satellite cells (2,160 DEGs). We compared the list of DEGs between adult satellite cells and iPax myogenic progenitors before and after engraftment, and found that 1,580 genes were no longer differentially expressed (Fig. 3B and Dataset S3). This result indicates that molecular changes occur upon in vivo engraftment. To evaluate the extent of such changes, we projected the reisolated iPax MNCs and adult satellite cells onto the PCA plot shown in Fig. 1B, and observed that in vivo-reisolated iPax samples acquire a molecular signature that clusters more closely to neonatal and adult satellite cells than prenatal counterparts (Fig. 3C, red circle).

Fig. 3.

iPax progenitors undergo molecular signature changes upon engraftment. (A) Scheme shows the experimental approach for the reisolation of donor-derived CD34⁺GFP⁺ MNCs. (B) Heatmaps show the comparative analysis of DEGs between iPax progenitors and satellite cells before (in vitro) and after (in vivo) engraftment. (C) PCA shows the merged distribution of samples from in vitro and in vivo transcriptome analyses. In vivo-reisolated iPax MNCs are circled in red. PC1 and PC2: 59.5 and 36% variance, respectively. (D) Scatter plot shows t-distributed stochastic neighbor embedding (t-SNE) of adult and iPax3 PSC-derived myogenic progenitors upon single-cell RNA sequencing. (E) Scatter plot shows t-SNE of cells shown in D overlapped with the expression of selected genes. (F) Representative images show donor-derived Dys⁺ myofibers (green) in TA muscles of secondary recipient NSGmdx^4cv mice transplanted with iPax3 and iPax7 MNCs. Nuclei were counterstained with DAPI (blue). (Scale bar, 100 µm.) (G) Graph shows quantification of Dys⁺ myofibers shown in E between iPax3 (n = 3) and iPax7 (n = 5) groups. PBS-injected muscle is shown as negative control (n = 3). Error bars represent SEs; ns, nonsignificant.

To understand whether these molecular changes are due to the in vivo niche selecting for a rare subset of cells with a more adult-like phenotype, we performed single-cell RNA sequencing (scRNA-seq) on iPax3 PSC-derived myogenic progenitors, and compared these with scRNA-seq performed on freshly isolated Pax7-ZsGreen⁺ adult satellite cells. As shown in Fig. 3D, the two samples do not overlap. To validate this finding, we first investigated for the expression of two control transcripts: the housekeeping Actb (positive control), which shows homogeneous expression in both populations, and the neural marker Sox2 (negative control), which is not expressed in either fraction (Fig. 3E). Pax3 and Pax7 clearly distinguish the iPax3 cells from the adult satellite cells, respectively (Fig. 3E). Myf5 and Vcam1 are more evenly distributed among the two populations, while Itga7 and Cd34 are highly expressed in adult satellite cells; note that Cd34 is absent in in vitro iPax3 cells. These data suggest that exposure to the in vivo satellite cell niche is shaping the observed molecular changes rather than selecting an existing, adult-like, in vitro subpopulation.

To investigate the extent of these changes in more detail, we analyzed the 840 genes that were found to be up-regulated following transplantation. We identified signaling pathways that are fundamental for satellite cell function, such as JAK/STAT (25), Integrin (26), and TGFβ (27), among others (SI Appendix, Fig. S7B). Of note, we found that Notch1 and 3 and Jag1 were no longer among the genes differentially expressed with respect to adult satellite cells and, similar to representative genes from the above-mentioned pathways, they all display fold changes lower than 2 between iPax MNCs and adult satellite cells (SI Appendix, Fig. S7C). We confirmed this for several genes by RT-qPCR (SI Appendix, Fig. S7D). These data further indicate that iPax cells change their molecular signature upon engraftment.

We also looked into those genes that remained differentially expressed, and found 73 genes whose expression was differentially regulated between iPax and satellite cells in both analyses. While only 36 genes maintained their relative differential expression compared with satellite cells (DN-DN and UP-UP; SI Appendix, Fig. S7E) or were down-regulated in iPax MNCs (UP-DN), we found an interesting group of genes whose expression is higher in iPax cells after exposure to the endogenous skeletal muscle environment, compared with satellite cells (DN-UP). Among the 37 genes in this group, we found Dll4, Vegfa, Col1a1, Pdgfra, and Mmp19. Some of these genes were found at low levels in iPax myogenic progenitors and at very high levels in neonatal satellite cells (SI Appendix, Fig. S4C), further confirming that iPax cells acquire more mature characteristics with regard to their in vitro counterparts.

Finally, we transplanted 15,000 freshly isolated donor-derived MNCs into irradiated and CTX-injured TAs of NSGmdx^4cv mice (20-fold fewer cells than primary transplantations; Fig. 2 A and B) and observed an impressive engraftment with both iPax3- and iPax7-derived MNCs (Fig. 3 F and G). Of note, when we reduced the number of injected primary iPax cells to 30,000 (10-fold less), this produced fewer than 20 myofibers (SI Appendix, Fig. S7 F and G). These data further confirm that iPax myogenic progenitors undergo significant molecular changes upon exposure to the adult muscle environment.

Taken together, these data indicate that iPax PSC-derived myogenic progenitors are able to remodel their molecular signature upon in vivo engraftment by shifting from an early developmental to a postnatal signature. These molecular changes ensure proper homing to the satellite cell niche, which is well-demonstrated by the ability of engrafted iPax MNCs to promote regeneration upon multiple injuries and upon secondary transplantation.

Discussion

In the present study, we dissected the transcriptional profile of Pax3- and Pax7-induced PSC-derived myogenic progenitors both in vitro and in vivo to determine the molecular determinants underlying the myogenic regenerative potential as well as the muscle stem cell behavior shown by these progenitors. We focused on mouse PSC-derived samples because it allowed us to directly compare them with species-specific references isolated from mice at different developmental stages. We chose our reference samples to capture the major stages of the skeletal myogenesis program during development: E10.5 embryonic myoblasts and E14.5 fetal myogenic progenitors, representing the first and second waves of myogenesis, respectively (28); and early neonatal and adult satellite cells, based on the evidence that a transitional molecular signature exists from early to adult postnatal skeletal myogenic progenitors (29). These transcription analyses allow us to conclude that iPax progenitors display a molecular signature more similar to prenatal myogenic progenitors, including expression of genes involved in skeletal muscle development and structure, proliferative capacity, and adhesion/motility. The later results are supported by functional in vitro assays demonstrating the superior adhesion and migratory abilities of iPax cells compared with adult satellite cells. These data are not surprising, since PSC derivatives are by nature more embryonic, as previously described (30–33). However, our data show that the in vivo regenerative potential of iPax myogenic progenitors is much superior to embryonic and fetal myogenic progenitors. Importantly, iPax progenitors are able to robustly repopulate the satellite cell compartment, whereas prenatal myogenic progenitors do so minimally under exactly the same experimental conditions. Fetal myogenic progenitors have been shown to possess a decreased ability to reenter the satellite cell niche when transplanted into irradiated NOD/SCID mice (34). Using serial transplantations, Tierney et al. demonstrated that ECM remodeling might be the key determinant for efficient homing and reentry into the satellite cell niche. In light of this, we analyzed the transcriptional profile of engrafted reisolated iPax cells 1 mo after transplantation. Unlike in vitro-generated iPax myogenic progenitors, donor-derived CD34⁺ MNCs display molecular characteristics more similar to postnatal developmental stages. The number of genes differentially expressed in regard to adult satellite cells was lower, suggesting that the molecular signature of PSC-derived myogenic progenitors changes upon in vivo engraftment. This is supported by the finding that most of the genes differentially expressed between iPax cells and adult satellite cells were no longer differentially expressed upon exposure to the in vivo environment. Of interest, some of the genes found at higher levels in reisolated iPax MNCs are involved in ECM structure and remodeling, such as collagens, metalloproteinases, and fibrillary proteins. The remodeling of the ECM is an attractive mechanism in the context of satellite cell niche homing, since recent reports have shown that molecular changes altering stiffness, topology, and signaling within the ECM contribute to satellite cell behavior (35–38).

Altogether, these findings demonstrate that PSC-derived myogenic progenitors have the ability to adapt their transcriptional program and functionally mature upon in vivo exposure to the adult muscle environment. Interestingly, in vitro-generated PSC-derived myogenic progenitors do not seem to contain a more mature subpopulation of cells, as shown by scRNA-seq, and therefore the interaction with niche components alone may be triggering the observed molecular changes. Future studies aimed at dissecting the molecular mechanisms underlying the environment-induced maturation of PSC-derived myogenic progenitors will be instrumental to understand this switch, which could involve genetic manipulation of the genes whose expression changes upon in vivo exposure.

Materials and Methods

Cell Culture and Differentiation.

Inducible Pax3 and Pax7 mouse ES cell lines were generated as previously described (39). Full protocols are available in detail in SI Appendix, Materials and Methods.

Primary Cell Harvesting and Isolation.

Animal experiments were carried out in strict accordance with protocols approved by the University of Minnesota Institutional Animal Care and Use Committee. Full procedures are described in detail in SI Appendix, Materials and Methods.

RNA Isolation and Sequencing.

Samples for RNA-seq were resuspended in RA1+TCEP lysis buffer (Macherey-Nagel), and RNA was isolated using the NucleoSpin RNA XS Kit (Macherey-Nagel) following the manufacturer’s instructions. Full procedures are described in detail in SI Appendix, Materials and Methods.

Transplantation Studies.

All experiments were approved by the University of Minnesota Institutional Animal Care and Use Committee. Full procedures are described in detail in SI Appendix, Materials and Methods.

Immunofluorescence.

Immunofluorescence staining was performed on fixed cultured cells and on unfixed TA cryosections as described (9). Full procedures are described in detail in SI Appendix, Materials and Methods.

Data Availability.

Raw and processed RNA-seq, microarray, and single-cell RNA-seq data have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO) database and are accessible under GEO accession nos. GSE121639 and GSE123595.