Necessity of Notch3 signaling in myofiber maturation in a pluripotent stem cell transplant model

Abstract

Background

Pluripotent stem cell-derived myogenic progenitors change from an embryonic to a postnatal molecular signature upon engrafting as satellite cells, which coincides with upregulation of Notch3. Since a role for Notch3 in skeletal muscle maturation is unknown, here we investigate whether Notch3 is required for this in vivo molecular maturation switch.

Results

Our results show that lack of Notch3 in transplanted progenitors (N3KO) does not impact degree of engraftment, but leads to increased numbers of embryonic myofibers. Conversely, transplantation of Notch3 overexpressing (N3OE) myogenic progenitors results in lower numbers of embryonic myofibers, but diminished muscle grafts when compared to empty vector (EV) controls. Secondary transplantation studies confirmed these effects, whereby Notch3 overexpression significantly reduced secondary engraftment. Further characterization of N3OE donor-derived satellite cells revealed reduced proliferation and downregulation of cell cycle genes. Importantly, secondary grafts from N3KO satellite cells had increased numbers of embryonic myofibers compared to N3OE and EV controls.

Conclusions

Taken together, these findings demonstrate that Notch3 signaling is required for myofiber maturation, and that constant activation of Notch3 impairs proliferation and muscle regeneration. Transcriptional profiles of N3OE donor-derived satellite cells suggest that dampened regeneration may be driven by inhibitory alterations in cell cycle regulation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13395-025-00403-4.

Background

Embryonic and neonatal myosin heavy chain (MHC) isoforms are transiently expressed during muscle development and regeneration, before being replaced by the adult MHC isoforms [1]. In postnatal skeletal muscle, the transient re-expression of these developmental MHC isoforms serves as an established indicator of muscle repair, identifying newly formed myofibers during muscle regeneration [1, 2]. However, the mechanisms driving the switch between developmental and adult myosin isoforms remain poorly understood.

Studies involving the in vitro differentiation of pluripotent stem cells (PSC) into the myogenic lineage have documented the immature nature of in vitro generated myogenic cells [3–6]. However, upon transplantation, these myogenic progenitors undergo maturation switch as they acquire a postnatal molecular signature [4, 6, 7], thus this system represents a valuable model to study the molecular mechanisms underlining skeletal muscle maturation. Our earlier RNA-sequencing studies of reisolated donor-derived satellite cells revealed significant upregulation of many signaling pathways, among which, several components of Notch signaling were identified, including Notch1, Notch3, and Jag1 [4]. In vivo upregulation of Notch3 was subsequently observed by another research group [7].

The Notch signaling pathway is well-established as a key regulator of satellite cell function [8–13] but a role in skeletal muscle maturation has yet to be reported. Notch signaling entails the interaction between one of the Notch receptors (Notch 1–4 in mammals) and one of the ligands (Dll1 and −4 and Jag1 and −2), which leads to the cleavage of the Notch intracellular domain (NICD), which then translocates to the nucleus [14]. There, NICD binds to RBP-J_Κ and activates the transcription of downstream target genes [15]. Notch1 and Notch2 have the highest homology with each other, whereas Notch3 is structurally distinct, lacking the transactivation domain [16]. Although Notch3 lacks a transactivation domain, NICD3 retains the RBP-J_Κ associated molecule (RAM) and ankyrin repeat domains required for RBP-J_Κ binding, and therefore, NICD3 can still regulate canonical target genes, such as Hey1 and Hey2 [17–19]. As a result, Notch3 exhibits functional differences compared to Notch1 and Notch2. Whereas loss-of-function mutations in Notch1 [20] or Notch2 [21] are embryonic lethal, Notch3-deficient mice show normal development but postnatal muscle hyperplasia when exposed to repetitive muscle injuries [22], suggesting a unique role for Notch3 compared to others Notch receptors. Notch3 is expressed in satellite cells [22] and myoblasts [23], but its function remains to be fully elucidated. It has been proposed that Notch3 regulates myogenic differentiation via Mef2c [24].

The present study focuses on the role of Notch3 in the engraftment and maturation of PSC-derived myogenic progenitors. Here, we find that Notch3 is required for the in vivo maturation of PSC-derived skeletal muscle as transplantation of myogenic progenitors lacking Notch3 resulted in elevated numbers of myofibers expressing embryonic MHC, whereas overexpression of NICD3 led to the opposite phenotype. These results provide evidence that Notch3 is required for fiber type maturation. Moreover, serial transplantation studies show that NICD3 overexpression inhibits skeletal muscle regeneration. Transcriptional profiling of NICD3 donor-derived satellite cells indicates that this regeneration defect may be mediated by inhibitory alterations in cell cycle regulation.

Methods

Cell culture and differentiation

iPax3 ES cells were maintained in a 1:1 mixture of ES medium and 2 inhibitors (2i) medium, as previously described [25]. To facilitate in vivo tracking, iPax3 ES cells were labeled with a lentiviral vector encoding the histone 2B-red fluorescent protein (H2B-RFP) fusion protein (LV-RFP plasmid; Addgene #26,001). For embryoid body (EB) differentiation, iPax3 ES cells were cultured in suspension in EB differentiation medium, consisting of IMDM (Invitrogen), 15% FBS, 1% penicillin–streptomycin, 2 mM Glutamax, 50 μg/ml ascorbic acid (Sigma-Aldrich), and 4.5 mM monothioglycerol, on an orbital shaker (80 RPM) at a density of 40,000 cells/mL. Pax3 induction started on day 3 of EB differentiation by adding doxycycline (dox; Sigma-Aldrich) to the medium at 1 μg/mL. The PDGFRα + Flk1- cell fraction that identifies myogenic progenitors [26, 27] was purified on day 5 of EB differentiation. Briefly, EBs were disaggregated, treated with Fc Block (1 μL/million cells; BD Biosciences) for 5 min, and incubated for 20 min with PE- and APC-conjugated antibodies to PDGFRα and Flk-1-PE antibodies, respectively from E-bioscience and BD Biosciences. After washing with PBS, cells were resuspended in MACS buffer (Miltenyi Biotec), and sorted for the PDGFRα + Flk-1- cell fraction using a FACSAria II (BD Biosciences). Purified PDGFRα + Flk-1- cells were re-plated on gelatin-coated dishes in EB differentiation medium with 1 μg/mL dox and 10 ng/mL mouse basic FGF (Peprotech) and cultured for 3 passages prior to subsequent studies. For differentiation into myotubes, we discontinued dox and replaced the EB medium differentiation medium. After 3 days in culture, cells were evaluated by immunofluorescence.

Notch3 engineering

For Notch3 overexpression, we amplified the murine NICD3 open reading frame (ORF) by PCR and cloned it into the pRRL-T2AGFP lentiviral vector (plasmid #20185, Addgene). pRRL-NICD3-T2AGFP (NICD3) and empty vector control pRRL-T2AGFP were co-transfected with pVSV-G and Δ8.9 packaging plasmids into HEK293T cells using Lipofectamine LTX-Plus reagent (Invitrogen). Lentiviral supernatants were collected 48 h post-transfection, filtered, and used for the spin infection (90 min at 2500 rpm at 30 °C) of PDGFRα + Flk1- myogenic progenitors at passage 0. GFP + cells were sorted by FACSAria II, re-plated, and expanded in myogenic medium until passage 3.

For Notch3 knockout, a CRISPR-Cas9 construct targeting exon 3 of Notch3 was generated by cloning the sgRNA (5’-CAGGTGCCTGCCAGGCTGGG-3’) into pSpCas9(BB)−2A-GFP (PX458) vector (Plasmid #48138, Addgene). The sgRNA/Cas9 plasmids were introduced into iPax3 ES cells via nucleofection (Lonza), and Notch3 knockout clones isolated by single-cell sorting based on GFP expression. Notch3 deletion was confirmed by DNA sequencing.

Mice and cell transplantation

All animal experiments were approved by the University of Minnesota Institutional Care and Use Committee. NOD-scid IL2Rgamma^null (NSG) mice (Jackson Laboratories, strain #005557), at 6–12 weeks of age, were used for transplantation studies. Conditioning consisted of irradiation of both limbs with a 12-Gy single dose, followed by muscle injury of both TA muscles with 10 μM cardiotoxin (CTX) at 24 h later. Next day, injured muscles were injected with 3 × 10⁵ myogenic progenitors (left TA) or PBS (right TA). Engraftment assessment was performed four weeks later.

Immunofluorescence staining

Immunofluorescence staining was performed on muscle cryosections and cultured cells. Muscle samples were embedded in Tissue-Tek O.C.T. compound (Sakura), snap-frozen in pre-cooled isopentane, and cut into 10 μm sections on glass slides. Samples were fixed with 4% paraformaldehyde (PFA) for 30 min, washed with PBS, and permeabilized with 0.3% Triton X/PBS for 15 min. After blocking with 3% BSA/PBS for 1 h, samples were incubated overnight at 4 °C with primary antibodies. Antibodies used included RFP (rabbit, 1:500, ab62341, Abcam), dystrophin (Dys, mouse, 1:20, DYS1-CE, Leica), laminin α−2 (Lam, rat, 1:200, Sc-59854, Santa Cruz), Pax7 (mouse, 1:10, DSHB), Ki67 (mouse, 1:500, 550,609, BD Biosciences), MHC (MF20, mouse, 1:100, DSHB), and the following MHC isoforms (all from DSHB): type I (BA-D5, 1:100), type IIA (sc-71, 1:30), type IIB (BF-F3, 1:20), and type I + IIA + IIB (BF-35, 1:20), DSHB). Controls antibodies consisted of mouse IgG isotype (1:20; Novus Biologicals) and mouse IgM isotype (1:30; Novus Biologicals). After washing with PBS, samples were incubated with Alexa Fluor-conjugated secondary antibodies (Thermo Fisher Scientific) and DAPI (Santa Cruz) for 1 h at room temperature. Cultured cells were then left in PBS, while muscle sections were mounted using Prolong Gold (Invitrogen). The fusion index was calculated as the percentage of nuclei within myotubes relative to the total number of nuclei per microscopic field. Myotube area was calculated by determining the proportion of the field occupied by positive pan-MHC staining, relative to the total field area. For quantification of engraftment, the percentage of RFP/Dystrophin-positive area relative to total engraftment area was calculated using stitched images of the whole TA muscle, processed with Zen software. Myofiber type and satellite cell quantifications were performed on five sections per muscle sample. Myofiber type was quantified as the percentage of RFP/MHC positivity relative to the engraftment area. Each MHC isoform was stained on separate tissue sections due to technical limitations. Our transplants cells express both GFP and RFP, which limited out ability to perform multiplex immunostaining with multiple MHC isoform and DAPI on single slide. Therefore, each quantification was performed independently on serial sections, which can lead to discrepancies in total fiber counts and proportions when comparing across markers. The satellite cells were quantified as the percentage of Pax7 + RFP + cells. Images were captured using an Axioimager M1 fluorescence microscope (Zeiss) and analyzed with Fiji software.

Donor-derived satellite cell isolation

Donor-derived satellite cells were purified by FACS at 4 weeks post-transplantation based on the expression of RFP (donor tracking) and the surface marker profile CD31-CD45-Itga7 + VCAM +, as previously described [4, 28].

Western blotting

Proteins were extracted using RIPA buffer (150 mM NaCl, 50 mM Tris–HCl, pH 7.5, 1 mM EDTA, 1% Triton X-100, 1% Sodium Deoxycholate, and 0.1% SDS) supplemented with 1% protease inhibitors (Complete, Millipore). Lysates were centrifuged at 15,000 rpm for 5 min at 4 °C. The total protein concentration was determined using the Bradford assay (Sigma-Aldrich). Protein samples were prepared in Laemmli buffer, loaded onto SDS-PAGE gels, and transferred to PVDF membranes (Millipore). Membranes were blocked for 1 h at room temperature using Odyssey Blocking Buffer (Li-Cor Biosciences), followed by overnight incubation at 4 °C with primary antibodies for Notch3, Notch1, Notch2 and α/β-Tubulin (1:1000; 2148; Cell Signaling). For Notch3, we used the D11B8 antibody (1:1000; Cell Signaling), which detects full-length Notch3 (270 kDa) and the membrane-tethered Notch3 (NTM; 90 kDa). For Notch1, we used the D6F11 antibody (1:1000; Cell signaling), which detects full-length Notch1 (300 kDa) and the membrane-tethered Notch1 (NTM;120 kDa). For Notch2, we used the D76A6 antibody (1:1000; Cell signaling), which detects full-length Notch2 (300 kDa) and the membrane-tethered Notch2 (NTM;110 kDa). After incubation with infrared fluorescent secondary antibodies (Li-Cor Biosciences) for 1 h at room temperature, membranes were visualized using the Li-Cor Odyssey Infrared Imaging System. Images were quantified using Fiji ImageJ software.

RNA isolation and analysis

For in vitro samples containing high RNA input, we extracted RNA using Trizol (Invitrogen) followed by the RNA extraction Kit (Invitrogen), according to the manufacturer’s protocol. For low RNA input samples, reisolated donor-derived satellite cells, we extracted RNA using Buffer RLT (QIAGEN) and nucleic acid beads. We carried out cDNA synthesis using the Superscript VILO cDNA Synthesis Kit (Invitrogen) for high input samples, and the NEBNext Single Cell/Low Input cDNA Synthesis & Amplification Kit (New England Biolabs) for low input samples.

Gene expression analysis was performed using 10 ng of cDNA per reaction. RT-qPCR was conducted with the Premix Ex Taq Probe qPCR Master Mix (Takara) and TaqMan probes from Applied Biosystems (Notch3, Mm01345646_m1; Hes1, Mm01342805_m1; Hey1, Mm00468865_m1; HeyL, Mm00516558_m1; Cdk1, Mm00772472_m1; Cdkn1a, Mm00432448_m1; Ccnd1, Mm00432359_m1 and GAPDH, Mm05724508_g1). Gene expression levels were normalized to GAPDH.

For RNA sequencing, libraries were generated using the Nextera XT DNA Library Preparation Kit (Illumina) and processed at the University of Minnesota Genomics Core (UMGC). Libraries were sequenced on a paired-end run using the NovaSeq (Illumina) platform.

RNA-sequencing analysis

Bulk RNA-sequencing datasets from paired-end libraries using the Illumina platform were downloaded from the gene expression omnibus (GEO) database using the Sequence Read Archive (SRA) toolkit for 9 samples of mouse iPax3 control donor-derived satellite cells [29]. We followed a well-established pipeline, as previously [30] described. Data alignment was performed using the CHURP pipeline at the University of Minnesota Genomics Center (UMGC). A total of 23 biological samples (2 × 150 bp FASTQ paired-end reads, averaging ~ 20 million protein-coding mapped reads per sample) were trimmed using Trimmomatic (v0.33), with the optional “-q” option enabled for 3 bp sliding-window trimming from the 3' end (minimum Q30), and an additional 3 bp headcrop for samples processed with Takara Bio during library preparation. Quality control on the raw sequence data was performed using FastQC (RIN ≥ 9.7). Read mapping was conducted using HISAT2 [31] (v2.1.0) with the mouse genome (GRCh39.106) as the reference. Gene quantification was done using FeatureCounts for raw read counts. Batch correction for library preparation differences (TruSeq Stranded vs. SMARTer Stranded) and batch effects was performed with the CombatSeq package (v3.46.0) [32] in R. Differentially expressed genes (DEGs) were identified based on the following criteria: p-value < 0.05, FDR < 0.05, and absolute log(fold change) |1.5|. Gene set enrichment analysis (GSEA) and gene ontology (GO) analysis were conducted using R packages clusterProfiler (v4.8.2) and ReactomePA (v1.44.0). Data visualization was performed with the before mentioned R packages, along with pheatmap (v1.0.12) and enrichplot (v1.20.0).

Statistical analysis

Statistical differences between two independent samples were performed using the unpaired two-tailed Student’s t-test. For comparisons between three groups, one-way or two-way ANOVA was used, followed by post hoc correction for multiple comparisons using Tukey’s or Dunnett’s test. A p-value < 0.05 was considered statistically significant. GraphPad Prism v9 software (GraphPad Software, LLC) was used for statistical analyses.

Data availability

Raw and processed RNA-seq data have been deposited as SuperSeries to the NCBI Gene Expression Omnibus (GEO) database and are accessible under the GEO Accession Number GSE286218.

Results

Generation of Notch3 knockout myogenic progenitors

To investigate the role of Notch3 in the regenerative potential of embryonic stem (ES) cell-derived iPax3 myogenic progenitors, we generated Notch3 knockout ES cells (N3KO) using CRISPR/Cas9. This approach employed a vector expressing a guide RNA targeting Notch3, which was tagged with GFP to facilitate selection, thereby introducing a loss-of-function mutation in the Notch3 gene. An empty vector (EV) tagged with GFP was used as control (Fig. 1A). FACS-purified GFP + transduced N3KO and control EV ES cells (Supplementary Fig. S1A) were subjected to myogenic differentiation, as previously described [26]. PDGFRα is a well-established marker of paraxial mesodermal, while Flk-1 marks lateral plate mesoderm, thus the PDGFRα + Flk-1- population is commonly used to purify myogenic progenitors in mouse ESC-derived cultures [26, 27, 33–35]. The frequency of PDGFRα + Flk-1- myogenic progenitors was comparable between N3KO and EV control cultures (Supplementary Fig. S1B).

Fig. 1.

Regenerative potential of N3KO myogenic progenitors. A Schematic outline of Notch3 loss-of-function studies in ES cell-derived iPax3 myogenic progenitors. B Western blot for the intracellular domain (ICD) of Notch3 (90 kD) in N3KO myogenic progenitors and their respective EV control, both in the absence (left panel) and presence (right panel) of Notch ligands Jag1 or Dll4. α/β-tubulin (55 kD) was used as a loading control. Representative gels from three independent experiments. C, D, E Myofiber engraftment. Representative images show myofiber engraftment in muscles that had been transplanted with N3KO and respective EV control myogenic progenitors, as indicated by staining for RFP (red) and dystrophin (DYS, white). DAPI in blue stained nuclei (C, E). Graph shows quantification of engraftment shown as the percentage of RFP + DYS + engrafted area relative to the total muscle area (D). Data are presented as mean ± SEM from two independent experiments (n = 6 per cohort). F, G Myofiber type composition. F Representative images show staining for different MHC isoforms (AF647 signal, rendered in green), RFP (red), and laminin (white). Scale bar = 100 µm. G Bar graphs show quantification (from F) of the ratio between RFP + engraftment area and MHC expression (n = 6 for each cohort). ****p < 0.0001 by unpaired Student t test

In accordance with our previous findings showing that in vitro-generated iPax3 myogenic progenitors express very low levels of Notch3 [4], we detected Notch3 only in Notch-ligand (Dll4 or Jag1) stimulated EV control myogenic progenitor cultures (Fig. 1B and Supplementary Fig. S1C-D). Notch3 was absent in N3KO cultures, regardless of signaling pathway activation with Notch ligands (Fig. 1B and Supplementary Fig. S1C-D), thus confirming the loss of Notch3. Accordingly, gene expression analysis by qPCR revealed downregulation of Notch3, as well as the downstream target genes Hey1 and HeyL, in unstimulated N3KO myogenic progenitors compared to control counterparts, whereas Hes1 was unaltered (Supplementary Fig. S1E). To assess a potential compensatory role by other Notch receptors, we examined protein levels of Notch1 and Notch2 in EV and N3KO myogenic progenitors under unstimulated conditions. Notch4 was not tested since it is not expressed in skeletal muscle [36]. Comparable levels of Notch1 and Notch2 were observed in EV and N3KO cultures (Supplementary Fig. S1F-G), indicating that the downregulation of Hey1 and HeyL in N3KO cells likely reflects the specific loss of Notch3 signaling. When subjected to terminal differentiation conditions, N3KO myogenic progenitors gave rise to MHC-expressing myotubes, similar to EV controls (Supplementary Fig. S1H-J), indicating that Notch3 is not required for in vitro myotube differentiation.

Comparable engraftment but distinct maturation by N3KO myogenic progenitors

To assess the in vivo regenerative potential of N3KO myogenic progenitors, we transplanted equal numbers of N3KO or EV control myogenic progenitors into irradiated and cardiotoxin (CTX)-injured tibialis anterior (TA) muscles of NSG mice. Four weeks after transplantation, we evaluated myofiber engraftment by immunostaining for RFP (donor-tracking) and dystrophin (DYS). Absence of Notch3 did not affect the ability of myogenic progenitors to produce myofibers, as comparable myofiber engraftment was observed between N3KO and EV controls (Fig. 1C-E and Supplementary Fig. S2A).

To investigate whether Notch3 has a role in myofiber maturation, we evaluated myofiber composition of muscle grafts by immunostaining for the different MHC isoforms: eMHC, oxidative slow-twitch type I (MHC-I), oxidative fast-twitch type IIA (MHC-IIA), as well as glycolytic fast-twitch type IIB (MHC-IIB), and glycolytic fast-twitch type IIX (MHC-IIX; by exclusion). While no differences were observed for MHC-I, IIA, IIB, and IIX, our data revealed that N3KO muscles grafts display higher expression of eMHC when compared to EV controls (Fig. 1F-G), suggesting that loss of Notch3 delays the progression of nascent fibers toward a more mature state. Because the contribution from endogenous muscle satellite cells is significantly blunted with irradiation and cardiotoxin (Supplementary Fig. S2C-D), there is minimal fusion or contribution from host cells in the engrafted area.

Notch3 gain-of-function in iPax3 myogenic progenitors

Based on the effect of Notch3 loss-of-function on myofiber maturation, we overexpressed NICD3 in iPax3 myogenic progenitors (N3OE) to determine whether activation of the Notch3 pathway would conversely enhance the maturation status of myofibers produced by PSC-derived iPax3-myogenic progenitors. As outlined in Fig. 2A, iPax3 ES cells differentiated into myogenic progenitors were transduced with lentiviral vectors constitutively expressing NICD3-tagged with GFP or an empty vector (EV) also tagged with GFP. Transduced cells were purified based on GFP positivity (Supplementary Fig. S3A), expanded and used for these studies. We validated NICD3 overexpression in N3OE myogenic progenitors by western blot (Fig. 2B and Supplementary Fig. S3B). Gene expression analysis further confirmed upregulation of Notch3, and its downstream target genes Hes1, Hey1 and HeyL, in N3OE myogenic progenitors when compared to EV control counterparts (Supplementary Fig. S3C). Upon in vitro terminal differentiation, these cells produced enlarged MHC + myotubes when compared to control counterparts (Supplementary Fig. S3D). Of note, while N3OE did not show a difference in fusion index (Supplementary Fig. S3E), displaying a similar number of nuclei accretion to controls, an increase in MHC + area was observed (Supplementary Fig. S3F).

Fig. 2.

Regenerative potential of N3OE myogenic progenitors. A Schematic representation of Notch3 gain-of-function studies in ES cell-derived iPax3 myogenic progenitors. B Western blot for Notch3 ICD (90 kD) in N3OE myogenic progenitors compared to the respective EV control. α/β-tubulin (55 kD) served as the loading control. Representative gels from three independent experiments. C, D, E Myofiber engraftment. Representative images show staining for RFP (red) and Dys (white) in muscles that had been transplanted with EV and N3OE myogenic progenitors. DAPI stained nuclei (C, E). Graph shows the percentage of RFP + DYS + engrafted area relative to the total muscle area (D). Data are presented as mean ± SEM from two independent experiments (n = 10 per cohort). ****p < 0.0001 by unpaired Student t test. F, G. Myofiber type composition. (F) Representative images showing staining for different MHC isoforms (AF647 signal, rendered in green), RFP (red), and laminin (white). Scale bar = 100 µm. (G) Bar graphs show quantification (from F) of the ratio between RFP + engraftment area and MHC expression (n = 6 for each cohort). *p < 0.05 by unpaired Student t test

Blunted muscle engraftment and lower expression of eMHC by N3OE myogenic progenitors

To assess the effect of Notch3 activation on the in vivo regenerative potential of iPax3 myogenic progenitors, we transplanted equal numbers of N3OE or EV control myogenic progenitors, as described above for N3KO. Assessment of engraftment revealed significant reduction by N3OE myogenic progenitors compared to controls (~ 44% vs. 85%, respectively), as indicated by quantification of donor-derived RFP + DYS + myofiber area (Fig. 2C-E and Supplementary Fig. S3G). This data indicate that Notch3 activation inhibits the engraftment ability of iPax3 myogenic progenitors.

When we looked at myofiber composition, we found that N3OE muscle grafts exhibited lower expression levels of eMHC compared to EV controls (Fig. 2F-G). No significant changes in adult myosin isoforms were observed (Fig. 2F-G). Together, these results indicate that Notch3 activation results in blunted myofiber engraftment, but that these donor-derived myofibers are less immature.

Changes in Notch3 signaling do not affect satellite cell engraftment by iPax3 myogenic progenitors

To investigate the effect of Notch3 gain- or loss-of-function on the ability of PSC-derived myogenic progenitors to seed the satellite cell compartment, we assessed donor-derived satellite cell engraftment by immunofluorescence staining and flow cytometry. Comparable numbers of RFP + PAX7 + cells (under the basal lamina and in the interstitial space) were observed in all cohorts (Fig. 3A-C). Flow cytometry analysis confirmed no differences in the frequency of donor-derived RFP + CD31-CD45-ITGA7 + VCAM-1 + satellite cells across all groups (Fig. 3D-E). These findings suggest that Notch3 is not required for iPax3 myogenic progenitors to repopulate the stem cell pool. Additionally, we confirmed that donor-derived satellite cells retain the characteristics of in vitro-engineered Notch3 gain- or loss-of-function upon transplantation, as evidenced by the respective upregulation and downregulation of Notch3 and downstream target genes (Fig. 3F-G).

Fig. 3.

Satellite cell engraftment of N3KO and N3OE myogenic progenitors. A-C Satellite cell engraftment. A Representative images show donor-derived satellite cells as evidenced by the presence of RFP + PAX7 + cells beneath the basal lamina. RFP staining is shown in red, PAX7 in green, and Laminin (LAM) in white. DAPI in blue stains nuclei. Yellow circles highlight RFP + PAX7 + cells. Scale bar = 25 µm. B, C Graphs show quantification of the percentages of RFP + PAX7 + cells under the basal lamina (B) and within the interstitial space (C). Data are presented as mean ± SEM from two independent experiments (n = 8 for EV controls, n = 5 for N3KO, and n = 4 for N3OE). D Representative FACS plots show the frequency of RFP + CD31-CD45 − Itga7 + Vcam-1 + donor-derived satellite cells in EV, N3KO, and N3OE muscle grafts. E Graph displays the percentage of donor-derived RFP + CD31-CD45-Itga7 + Vcam-1 + cells in the total mononuclear cell (MNC) fraction. Data are presented as mean ± SEM from two independent experiments (n = 11 for EV controls, n = 4 for N3OE, and n = 4 for N3KO). F, G Bar graphs show gene expression levels of Notch3 and its downstream target genes Hes1, Hey1, and HeyL in N3KO (F) and N3OE (G) reisolated satellite cells compared to respective EV controls. Results are normalized to GAPDH expression. Data are presented as mean ± SEM from two independent experiments (n = 5–10 per sample). **p < 0.01, ***p < 0.001 and ****p < 0.0001 by unpaired Student t test

Secondary transplantation reveals lower myofiber engraftment by N3OE satellite cells and persistence of eMHC in muscle grafts by N3KO satellite cells

To test the regenerative capacity of Notch3-engineered satellite cells, we performed secondary transplantation assays. We transplanted 5,000 freshly reisolated donor-derived RFP + CD31-CD45-Itga7 + VCAM-1 + satellite cells from each cohort into irradiated and CTX injured TA muscles of NSG mice (Fig. 4A). Our results showed comparable myofiber engraftment by N3KO and EV control satellite cells, while the N3OE cohort displayed significantly reduced myofiber engraftment (Fig. 4B-C and Supplementary Fig. S4A). This data suggests that constitutive NICD3 overexpression also inhibits the ability of satellite cells to produce myofibers in vivo, as observed with PSC-derived myogenic progenitors (Fig. 2C-D).

Fig. 4.

Myofiber engraftment of reisolated donor derived N3OE and N3KO satellite cells. A Overview of secondary transplantation studies. B, C Myofiber engraftment of secondary grafts. B Representative images show myofiber engraftment in muscles that had been transplanted with EV control, N3KO and N3OE reisolated donor-derived satellite cells. Muscle sections were stained for RFP (red). C Graph shows quantification of the percentage of RFP + DYS + myofiber engraftment relative to the total muscle area (from B). Data are presented as mean ± SEM from two independent experiments (n = 22 for EV controls, n = 10 for N3KO, and n = 6 for N3OE). ****p < 0.0001 by ANOVA with Dunnett’s correction. D, E Myofiber type composition of secondary grafts. D Representative images show staining for various MHC isoforms (AF647 signal, rendered in green), RFP (red), and laminin (white). Scale bar = 100 µm. E Bar graphs show quantification (from D) of the ratio between RFP + engraftment area and MHC expression. Error bars represent standard errors (n = 11 for EV, n = 4 for N3KO, and n = 6 for N3OE). **p < 0.01 and ****p < 0.0001 by ANOVA with Dunnett’s correction

Fiber type composition of secondary muscle grafts showed comparable and lower expression of eMHC in EV and N3OE groups relative to primary grafts, while the N3KO cohort still displayed higher expression of eMHC (Fig. 4D-E). While no differences in the expression of adult isoforms were found in N3KO grafts, we observed an increase in MHC-I and a decrease in MHC-IIX in N3OE grafts when compared to EV control (Fig. 4D-E).

Assessment of satellite cell engraftment in secondary transplants showed the presence of RFP + PAX7 + cells under the basal lamina in all groups (Supplementary Fig. S4B). However, quantification revealed lower frequency in the N3KO cohort (Supplementary Fig. S4C). This coincided with a higher number of RFP + PAX7 + cells in the interstitial space when compared to EV controls (Supplementary Fig. S4D), suggesting that Notch3 signaling may play a role in satellite cell homing.

Disruption of cell cycle progression in Notch3-overexpressing satellite cells

To gain insights on the potential mechanism(s) underlining the reduced myofiber engraftment observed in the context of NICD3 activation, we performed RNA-sequencing analysis of donor-derived satellite cells from N3OE and EV control cohorts (purified as outlined in Fig. 4A). Comparing the transcriptomes of reisolated N3OE satellite cells with their EV control counterparts, we found that the majority of differentially expressed genes were downregulated in the N3OE cohort (863 downregulated vs. 349 upregulated) (Fig. 5A). Gene set enrichment analysis further supported these findings, showing that 96% (71/74) of all significantly altered pathways (adjusted p-value < 0.05; Reactome) were also downregulated in N3OE satellite cells. We found that N3OE satellite cells exhibit lower levels of genes associated with eukaryotic translation initiation, striated muscle contraction, extracellular matrix organization, degradation of the extracellular matrix, and key cell cycle processes. A closer examination of the pathways altered by Notch3 overexpression revealed a recurring theme related to cell cycle regulation and skeletal muscle development (Fig. 5B). Among the most notable and potentially biologically relevant to the observed impaired engraftment phenotype were pathways involving cell cycle checkpoints at various stages. Specifically, pathways such as “Cell Cycle Checkpoints,” “M Phase,” “Regulation of APC/C activators between G1/S and early anaphase,” and “Cyclin A/B1/B2 associated events during G2/M transition” were prominently altered, suggesting a significant impact of Notch3 on cell cycle progression and muscle differentiation.

Fig. 5.

Transcriptional signature of donor-derived N3OE satellite cells. A Heatmap depicting the hierarchical clustering of the top 1,212 differentially expressed genes (DEGs) from reisolated N3OE satellite cells compared to EV controls. B Bar graph depicting the normalized enrichment score (NES) for key pathways of interest after gene set enrichment analysis (GSEA) for EV vs. N3OE. All pathways were downregulated. C Overview of cell cycle regulation and key cell cycle regulators affected by Notch3 overexpression. Values represent average log-fold change for gene when comparing N3OE against EV. D Bar plots showing gene expression levels for representative genes annotated for cell cycle regulation. Data is presented as mean ± SEM (n = 9 for EV in pink and n = 4 for N3OE in purple). **p < 0.01 and ****p < 0.0001 by ANOVA with Dunnett’s correction. E, F Proliferative status of RFP + cells. E Bar graphs show gene expression levels of Cdk1, Cdkn1a and Ccnd1 in N3OE reisolated satellite cells compared to EV control counterparts. Results are normalized to GAPDH expression. Data is presented as mean ± SEM (n = 9 for EV in pink and n = 4 for N3OE in purple). *p < 0.05 and ****p < 0.0001 by ANOVA with Dunnett’s correction. F Representative images show staining for RFP (red) and Ki67 (green) in reisolated donor-derived EV and N3OE satellite cells maintained in culture for 2 days. Scale bar = 100 µm. G Bar graph shows respective quantification (from F). Data is presented as mean ± SEM (n = 4 for each group). **p < 0.01 by unpaired Student t test

Based on these results, we next examined the cell cycle status of N3OE satellite cells on a gene-by-gene basis to gain a more direct understanding. Notably, many key regulators of cell cycle transitions were downregulated in the N3OE cells compared to their EV control counterparts (Fig. 5C). Specifically, cell cycle activators such as cyclins A2, B1, B2, D1, and E1 showed significant downregulation with Notch3 overexpression (logFC of −1.76, −2.30, −1.53, −2.96, and −1.27, respectively). This downregulation of cyclins aligns with the observed upregulation of Cdkn1c (p57; logFC of 2.47), a cyclin-dependent kinase (CDK) inhibitor that can modulate multiple CDKs throughout the cell cycle. Additionally, the expression of the proliferation marker Mki67 was also downregulated in the N3OE satellite cells (logFC −2.17). This dampened cell cycle program was consistent across all N3OE samples surveyed, as shown by the transcript abundance (counts per million, cpm) for the selected key genes (Fig. 5D). qPCR validation confirmed the downregulation of Cdk1, Cdkn1a, and Ccnd1 in donor-derived satellite cells expressing N3OE (Fig. 5E).

Next, we investigated whether the downregulation of cell cycle genes correlates with reduced proliferation in N3OE reisolated satellite cells, as assessed by Ki67 staining. Donor-derived RFP + N3OE satellite cells displayed a significant reduction in the percentage of Ki67 + compared to controls (Fig. 5F-G), confirming the decreased proliferative activity of N3OE reisolated satellite cells. These findings suggest that Notch3 overexpression inhibits cell cycle progression and proliferation of satellite cells, which may underline the main mechanism for their reduced regenerative capacity.

Discussion

In this study, we investigated the role of Notch3 in the regenerative capacity and maturation of muscle grafts from PSC-derived myogenic progenitors. Results from primary and secondary transplants show comparable myofiber engraftment between N3KO and control cohorts, indicating that Notch3 is not required for muscle regeneration. While no differences were observed in satellite cell engraftment of primary recipients (Fig. 3A), transplantation of Notch3-deficient satellite cells into secondary recipients resulted in a higher proportion of interstitial RFP + PAX7 + cells (Supplementary Fig. S4B-D). These data suggest a role for Notch3 in satellite cell homing, as observed for Notch1 [29], although to a lesser degree (~ 30% vs. 80%, respectively). Kitamoto and Hanaoka [22] have reported a normal skeletal muscle phenotype in Notch3 knockout mice, not inconsistent with our findings, however when Notch3-deficient muscles were exposed to multiple rounds of CTX-injury, they developed hyperplasia, which the authors attributed to increased satellite cell numbers [22]. These contrasting results in satellite cell numbers may be due to the differences in experimental systems, since our conditioning prior to transplants consisted of a single CTX injury along with irradiation, not multiple rounds of CTX injury.

While our data show that Notch3 is dispensable for muscle regeneration, Notch3 gain-of-function significantly impaired myofiber engraftment in primary and secondary recipients, with no evident effect on satellite cell repopulation. To dissect the molecular mechanisms underlining this defective muscle regeneration, we performed RNA sequencing of reisolated N3OE satellite cells, which revealed an inhibition of cell cycle progression, as evidenced by the downregulation of key cell cycle regulators (Cyclins A2, B1, B2, D1, and E1) and the upregulation of the cell cycle inhibitor Cdkn1c. These changes suggest that Notch3 overexpression suppresses cell proliferation, potentially limiting myogenic progenitor expansion and engraftment capacity. Reduced proliferation was also seen in in vitro studies of NICD3 overexpression in C2C12 and primary satellite cells [22, 23]. It has been proposed that Notch target genes Hes/Hey mediate growth arrest through an interaction between cyclin-dependent kinase inhibitors and myogenic regulatory factors to promote cell cycle exit [37]. Similarly, activation of Notch1 or Notch4 in endothelial cells has been reported to induce cell cycle arrest via P21 repression [38]. Consistent with these observations, our data demonstrate that Notch3 overexpression constrains proliferation through modulation of the cell cycle.

The major finding from this study is that Notch3 signaling is required for muscle maturation. We have previously shown that transplantation of PSC-derived iPax3 or iPax7 myogenic progenitors into irradiated CTX-injured muscles results in myofiber grafts expressing high levels of embryonic MHC, but this is no longer observed in secondary recipients [39]. Our results here show that engraftment by N3KO produced elevated numbers of embryonic myofibers in both primary and secondary transplants, which in both circumstances were significantly higher than EV control counterparts. These data demonstrate that Notch3 is specifically required for the transition from immature eMHC-expressing fibers to more mature fiber types. Remarkably, this critical role is unique to Notch3 as our recent studies show that Notch1 is dispensable for the myofiber maturation switch [29]. Consistently, activation of Notch3 signaling in primary grafts resulted in lower numbers of fibers expressing eMHC, while secondary recipients showed no differences to controls, with both displaying comparable reduced numbers of embryonic fibers, but an increase in type I and a decrease in type IIX fibers. Although our study did not assess differences in differentiation kinetics or fusion efficiency in vivo, these could also contribute to the observed phenotype. Overall, our findings suggest that Notch3 not only regulates myofiber engraftment but also influences myofiber composition.

In summary, this is the first study to report a necessity for Notch3 in skeletal muscle maturation. Our findings demonstrate that Notch3 signaling is required for the myofiber maturation switch from embryonic to adult, but not for engraftment, whereas its overexpression results in blunted regeneration due, at least in part, to defects in cell cycle progression. Further studies are warranted to further dissect the mechanisms by which Notch3 plays these different roles in muscle regeneration.

Authors’ contributions

A.M.S.Y. designed and performed experiments, analyzed the data, and wrote the manuscript. S.B.C. analyzed the data and wrote the manuscript. H.K. designed and performed experiments and contributed to data analysis. B.I.G., and J.E.A. analyzed the data. K.A., D.B., P.A., S.H.D.M.F. and A.A. performed experiments. R.C.R.P. contributed with experimental design, interpretation of the data and wrote the manuscript.

Funding

This project was supported by NIH, NIAMS grants R01 AR078571, R01 AR078624, R01 AR081882, and R56 AR055299 (R.C.R.P.), R01 AR081228 (D.B.), NHLBI T32 HL144472 (S.B.C.), F30 HL151138 and NIGMS T32 GM008244 (B.I.G.).

Data availability

The datasets generated from this study are deposited and publicly available. Materials used in this study are commercially available.

Declarations

Ethics approval and consent to participate

All procedures were approved by the University of Minnesota Institutional Animal Care and Use Committee (protocol #2212-40650A) and were performed in accordance with the National Research Council’s Guide for the Care and Use of Laboratory Animals (2011).

Consent for publication

Not applicable.

Competing interests

R.C.R.P. is co-founder and has equity in Myogenica, Inc. All other authors have no competing financial interests.