Adipogenic-Myogenic Signaling in Engineered Human Muscle Grafts used to Treat Volumetric Muscle Loss

Abstract

Tissue engineered muscle grafts (TEMGs) are a promising treatment for volumetric muscle loss (VML). In this study, we employed human myogenic progenitors (hMPs) cultured on electrospun fibrin microfiber bundles and evaluated the therapeutic potential of engineered hMP TEMGs in the treatment of murine tibialis anterior (TA) VML injuries. In vitro, the hMP TEMGs expressed mature muscle markers by 21 days. Upon implantation into VML injuries, the hMP TEMGs enabled remarkable regeneration. To further promote wound healing and myogenesis, we incorporated human adipose derived stem/stromal cells (hASCs) as fibroadipogenic progenitor (FAP)-like cells with the potential to secrete pro-regenerative cytokines. We investigated the impact of dose and timing of seeding the hASCs on in vitro myogenesis and VML recovery using hMP-hASC TEMGs. The hASCs increased myogenesis of hMPs when co-cultured at 5% hASCs: 95% hMPs and with delayed seeding. Upon implantation into immunocompromised mice, hMP-hASC TEMGs increase cell survival, collagen IV deposition, and pro-regenerative macrophage recruitment, but resulted in excessive adipose tissue growth after 28 days. These data demonstrate the interactions of hASCs and hMPs enhance myogenesis in vitro but there remains a need to optimize treatments to minimize adipogenesis and promote full therapeutic recovery following VML treatment.

GRAPHICAL ABSTRACT

Summary of Results: In vitro co-culture of hMPs with hASCs resulted in decreased expression of p53, CASP3, Ki67, and the pro-inflammatory gene (CCL2) in hASCs. Conversely, hASCs stimulated increased expression of Ki67, and maturation markers (increased total MHC and decreased MYOG) in hMPs. In vivo, implantation of hMP-hASC TEMGs in an immunocompromised VML environment resulted in excessive adipose tissue formation, increased ECM deposition (Col IV), increased cell survival, and increased pro-regenerative immune skewing (increased ratio of CD206+/F4/80+ to iNOS+/F4/80+ macrophages).

Introduction/Background

Volumetric Muscle Loss (VML) is the loss of a large portion of skeletal muscle that overwhelms the muscle’s innate mechanisms for repair and results in lifelong functional impairment.^1–3 VML can result from traumatic injury or as a byproduct of tumor resection, but in either case the current standard for treatment of VML is an autologous tissue graft and physical therapy. These treatment options are limited by donor site morbidity and minimal functional recovery due to a lack of complete regeneration.^2,4 Tissue engineered muscle grafts (TEMGs) provide an opportunity to mitigate these complications. The source of cells is an important consideration in TEMGs and often include a population of cells that has myogenic potential so that the formation of new muscle can be supplemented and/or induced via the secretion of pro-regenerative factors by the implanted TEMGs. C2C12 immortalized mouse myoblasts have yielded robust muscle regeneration in TEMG treatments of VML but are not translatable for clinical use and do not adequately reproduce the growth or behavior of primary myogenic cells. Mesenchymal stem/stromal cells^5,6 and adipose-derived stem/stromal cells^7–10 can be patient specific, are readily procurable, but have low myogenic differentiation efficiencies.^9,11 Human pluripotent stem cells can be patient specific and differentiate along the myogenic lineage,^12,13 but present challenges with reproducibility, and are limited by insufficient regenerative capacity.^14–16 Primary myogenic progenitors readily differentiate along the myogenic lineage and can be autologous, making them a promising myogenic cell source for TEMGs used in clinical skeletal muscle regeneration.

We have previously developed electrospun fibrin microfiber bundles which have topographical structure that induces alignment of cells, mechanical stiffness similar to native skeletal muscle, and are biocompatible (being made from fibrin) making it a strong candidate for creation of TEMGs.¹⁷ These fibrin microfiber bundles were used with C2C12 immortalized mouse myoblasts,^18–20 human adipose-derived stem/stromal cells (hASCs),⁹ and human pluripotent stem-cell derived myogenic progenitors,¹⁶ to create TEMGs that promote histological regeneration of skeletal muscle in NOD-SCID IL2rγ^null (NSG) mice. C2C12s were first used to establish the injury model and engineering strategies in an NSG mouse model with the expectation of working with human cells in the future which would require an NSG mouse model to avoid rejection.^18–20 TEMGs fabricated with hASCs⁹ and human pluripotent stem-cell derived myogenic progenitors¹⁶ were not rejected by the NSG host but displayed sub-optimal regenerative outcomes which motivated our testing of human myogenic progenitors (hMPs). In comparison to myogenic progenitors from other sources, hMPs have been insufficiently evaluated in TEMG treatment of VML. Previous homogenous hMP TEMGs were created by seeding hMPs onto decellularized muscle and implanting into immunocompromised animals, and post-implantation showed excess ECM around regenerating fibers resembling fibrosis.^10,21 Decellularized muscle is neither fully degraded nor fully incorporated into the native muscle leaving excess ECM at the site of injury, while our electrospun fibrin microfiber bundles have been shown to degrade leaving minimal fibrosis after regeneration.^18–20 We hypothesize that using hMPs with our electrospun fibrin microfiber bundles would promote regeneration without excess ECM/fibrosis.

Prior hMP-based TEMGs have been shown to achieve only partial regeneration of lost muscle volume after VML^10,21 potentially because skeletal muscle is a complex tissue and contains a diversity of non-myogenic cells. Nerves, vasculature, adipose tissue, and immune cells all play important roles in normal muscle homeostasis and in regeneration upon injury. Thus, to enhance native repair mechanisms and mimic native skeletal muscle, non-myogenic cells such as fibroblasts,^21–28 endothelial cells,^21–29 macrophages,³⁰ and fibroadipogenic progenitors (FAPs),^16,21,23 have been incorporated into TEMG constructs along with myogenic cells to provide a more pro-regenerative niche. FAPs have been identified as CD31^-/CD45^-/SM/C-2.6^-/PDGFRα⁺ or CD31^-/CD45^-/Integrin-α7^-/Sca-1⁺ skeletal muscle resident cells that are normally quiescent but become activated upon injury,^31,32 and are a promising population to include in a construct due to their significant involvement in skeletal muscle regeneration.^33–35 Although FAPs are known to be one of the main contributors to fibrosis and fatty infiltration of muscle,^31,32,36 they are also a source of pro-myogenic and wound healing signaling. Recently, we showed that unsorted human myogenic progenitor cells (differentiated from human embryonic stem cells) which contained a population identified as FAPs exhibited better survival when implanted in TEMGs in vivo than the Pax-7-sorted myogenic population,¹⁶ implicating FAPs potential benefits when included in a TEMG. It was recently reported that activated FAPs are transcriptionally more similar to ASCs than to inactivated/resident FAPs, and migrate from adipose tissue upon injury.³⁷ Further, when the nearby adipose tissue is removed, skeletal muscle regeneration is impaired, directly implicating ASCs involvement in repair.³⁷ Previous characterizations of ASCs also show similar surface markers to FAPs⁹, and having high expression of platelet derived growth factor receptor-α (PDGFRα).³⁸ Additionally, when looking at common secretions of ASCs, FAPs share at least 41 major secretory factors^39–42 (Table. S1). ASCs are more easily procured than FAPs and in significantly higher numbers giving them an advantage with regards to subsequent clinical translation. Consequently, in our study we assessed whether inclusion of hASCs as an FAP-like support cell in our hMP TEMGs could enhance myogenesis and regeneration.

Our objective was to assess the growth and myogenesis of hMPs on electrospun fibrin microfiber bundles over time in vitro and assess the effectiveness of hMP seeded fibrin microfiber bundles and cultured for up to 21 days to form hMP TEMGs for the treatment of VML. We hypothesized that hMP TEMGs would promote regeneration.^9,16,18–20 Since FAPs have pro-myogenic and pro-wound healing signaling during regeneration^33–35 but result in fibrosis and fat infiltration when dysregulated^31,32,36, we hypothesized that there would be an optimal concentration of hASCs for maximizing myogenesis and multinucleation of hMPs and that delaying their seeding into the TEMGs would be beneficial as it allows the population of myogenic cells to stabilize in vitro. We assessed the growth and myogenesis of hMP-seeded fibrin microfiber bundles when also seeded with hASCs added at different concentrations (0.5%, 5%, and 50%), at different times (0, 3, and 7 days post-hMP seeding), and assessed the effectiveness of hMP-hASC TEMGs for the treatment of VML. To assess the treatment potential of hMP-hASC TEMGs, grafts were implanted into the male NSG mice tibialis anterior (TA) VML model and compared to hMP TEMG treatment and no treatment histologically at 28 DPI.

Results

hMP TEMGs promote volume retention and muscle regeneration in a murine (NSG) model of TA VML

To determine the minimum appropriate culture period for creation of TEMGs, hMPs were seeded at 900,000 cells/fibrin microfiber bundle and cultured for 10, 14, and 21 days in vitro (Fig. 1A & B). These TEMGs were then stained with Desmin and MHC. Desmin, an intermediate filament protein expressed in cells of the skeletal muscle lineage, was used to identify all myogenic cells. Myosin heavy chain (MHC) is expressed in late myotubes/myofibers and identifies mature myogenic cells that are primed for contraction. The normalized area of MHC+ myofibers significantly increased over 21 days in vitro (Fig. 1C). The normalized area of Desmin+ hMPs showed a significant increase in fiber coverage from day 10 that plateaued on days 14 and 21 in vitro (Fig. 1D). These fibers were also stained for Sarcomeric α-Actinin and Laminin (Fig. S1A). Laminin and Collagen IV are major components of the skeletal muscle basement membrane. Sarcomeric α-Actinin is a skeletal muscle specific microfilament protein which is turned on at the initiation of myotubes/myofiber formation. The normalized area of α-Actinin+ myofibers increased over 21 days (Fig. S1B) while the normalized area of Laminin+ myofibers showed no differences over time (Fig. S1C). To evaluate the effectiveness of the hMP TEMGs for treatment of VML, after 21 days of in vitro culture, they were implanted into TA VML injuries (Fig. 1E). At 28 days post implantation (DPI) (Fig. 1F), TAs treated with the TEMGs exhibited a comparable normalized area of MHC+ myofibers compared to the uninjured contralateral TA and was significantly higher than untreated VML (Fig. 1G). The area of Desmin+ myofibers normalized by the area of the TA cross-section showed no significant difference between any of the groups (Fig. 1H). Quantification of the H&E stained muscle cross-sections (Fig. S2A) revealed that the increase in cross-sectional area between the no treatment VML and TEMG-treated VML was significant, and confirmed that the cross sectional area of the hMP TEMG-treated VML was statistically equivalent to the uninjured contralateral TA (Fig. 1I). Quantification of the myofibers with centrally located nuclei to indicate the number of new or regenerating myofibers revealed that there was a slight (not statistically significant) increase in the hMP TEMG-treated VML compared to the no treatment VML, and both had significantly more than the uninjured contralateral TA (Fig. 1J).

Figure 1: hMP TEMGs Increase Retention of MHC after VML Injury.

(A) Schematic of hMP TEMGs in vitro fabrication, culture, and assessment. (B) Representative images of fibers stained for DAPI, MHC, and Desmin at each time point, respectively (Scale Bar = 250 µm). (C) Quantification of MHC+ area as a percentage fiber area at each time point (n = 3). (D) Quantification of Desmin+ area as a percentage of fiber area at each time point (n = 3). (E) Schematic of hMP TEMGs in vivo implantation and assessment. (F) Representative images of TA cross-sections at 28 DPI stained for DAPI, MHC, and Desmin for each group (Scale Bar = 1000 µm). (G) Quantification of TAs MHC+ area as a percentage of TA cross-section area from each group (n = 4). (H) Quantification of TAs Demin+ area as a percentage of TA cross-section area from each group (n = 4). (I) Quantification of muscle cross section area in mm² (n = 4). (J) Quantification of the number of myofibers with centrally located nuclei for each group (n = 4). Significance was determined using a one-way ANOVA with multiple comparisons and Tukey post-test. Error bars represent standard error of the mean (SEM). *: p < 0.05; **: p < 0.01; ***: p < 0.001.

hASCs have high levels of PDGFRα expression and exhibit a maximal threshold for promoting myogenesis in hMP-hASC TEMGs

To confirm that our hASCs contained a large population of cells that expressed high levels of PDGFRα, hASCs were expanded for 2 passages then evaluated via flow cytometry for surface markers CD31, CD45, CD34, and CD140a (PDGFRα). We found that 98.8% of the cells were CD31-/CD45- and 87.7% of the CD31-/CD45- cells were PDGFRα+ (Fig. S3A & B). To evaluate the effect of hASC co-culture with hMP TEMGs on myogenesis in a concentration dependent manner, hMP TEMGs were seeded with hASCs at 0% (hMP TEMG), 0.5% (4,500 hASCs), 5% (45,000 hASCs), and 50% (450,000 hASCs) of hMPs concentration on day 7 post-hMP seeding (Fig. S3C). These concentrations were chosen to determine the impact of dosing across a broad range. The hMP-hASC TEMGs were cultured for total of 21 days post-hMP seeding (Fig. S3D). We observed that increasing the concentration of hASCs increased the pro-myogenic effects by both increasing myogenic proliferation and maturation of hMPs, however, with the addition of 50% hASCs these pro-myogenic effects were no longer observed. The normalized area of MHC+ myofibers showed an increase in myogenic maturity with concentration up to 5%, then a decrease in myogenic maturity at 50% (Fig. S3E). The normalized area of Desmin+ hMPs exhibited an increase in fiber coverage that plateaued at 0.5% and 5%, then a significant decrease fiber coverage at 50% (Fig. S3F). These results suggest that at higher concentrations, the proliferating hASCs significantly outnumber the hMPs on the fiber.

Delayed seeding of hASCs onto hMP TEMGs promotes their pro-myogenic effects

Given the positive effects of 5% hASCs, we evaluated whether delaying their seeding onto the TEMGs would further enhance the beneficial effects by allowing the hMPs time to adhere to the microfiber bundles, elongate, and initiate their differentiation into multinucleated myotubes. The hASCs were seeded at 5% of the hMPs concentration on days 0 (D0), 3 (D3), or 7 (D7) (Fig. 2A). These hMP-hASC TEMGs were cultured until 21 days post-hMP seeding (Fig. 2B). We observed that after seeding hMPs on fibrin microfiber bundles, longer delays prior to co-culture with hASCs led to an increase in the pro-myogenic effects of the co-culture. The normalized area of MHC+ myofibers and Desmin+ hMPs showed an increase in maturity and fiber coverage with delayed co-culture in vitro (Fig. 2C, D). This supported our hypothesis that myogenic cells benefited from time to stabilize in vitro prior to co-culture. We next evaluated this formulation (with delayed hASC-seeding) of the hMP-hASC TEMGs for their capacity to regenerate VML injuries.

Figure 2: Delayed Seeding of hASCs to hMP TEMGs Enhanced Their Pro-Myogenic Effects.

(A) Schematic of hMP-hASC TEMGs delayed co-culture testing and assessment. (B) Representative images from the 5% hASC co-cultured on hMP TEMGs stained for DAPI, MHC, and Desmin at each day of seeding, respectively (Scale Bar = 250 µm). (C & D) Quantification of MHC and Desmin expression on hMP TEMGs co-cultured with hASCs at each day of hASC seeding (n = 4). *: p < 0.05; **: p < 0.01; ***: p < 0.001.

hASCs co-cultured on hMPs TEMGs promote excessive adipose tissue growth when implanted in a murine model of TA VML in vivo

To evaluate the effectiveness of hMP-hASC TEMGs for treatment of VML, hMPs were cultured for 21 days in vitro, either alone or with the addition of 5% hASCs on D7, before implantation in a TA VML injury for 28 days and compared to VML that was not treated at 28 DPI. After 28 DPI the skin superficial to the TA shaved, nair applied, surgically incised, and the TA was harvested (Fig. 3A). At 28 DPI, we observed that instead of increasing regeneration such that the tissue would more closely resemble uninjured skeletal muscle, the hMP-hASC TEMGs promoted the growth of a large tissue mass at the site of injury comprised majorly of adipose tissue. FAPs have been shown to be the primary source of fibrosis and fatty degeneration in skeletal muscle. However, Masson’s Trichrome did not indicate that the mass was caused by excessive fibrosis (Fig. 3B) but instead indicated that the mass was comprised largely of adipose tissue (which is seen as the white space where the oil droplets were lost during processing). Harvested TAs were weighed and normalized to the weight of the contralateral uninjured TA (Fig. 3C).

Figure 3: hASCs co-cultured on hMPs TEMGs promote excessive adipose tissue growth when implanted in a murine model of TA VML.

(A) Representative images after 28 DPI of the skin superficial to the TA shaved and naired, opened, and the TA harvested, respectively. (B) Representative images of TA cross-sections at 28 DPI stained with Masson’s Trichrome for each group (Scale Bar = 2 mm). (C) Quantification of the weight of the harvested TA at 28 DPI divided by the weight of the contralateral harvested TA on the same mouse (n = 4). Significance was determined using a one-way ANOVA with multiple comparisons and Tukey post-test. Error bars represent standard error of the mean (SEM). *: p < 0.05; **: p < 0.01; ***: p < 0.001.

hASCs co-cultured on hMPs TEMGs promote cell survival and Collagen IV deposition when implanted in a murine model of TA VML in vivo

After 28 DPI harvested TAs were sectioned and stained with Collagen IV (Col IV), human Lamin A + C (hLamin AC used to distinguish the implanted human cells (ASCs and MPs)), and DAPI, to provide a visualization of the implanted human cells and the collagen IV deposition (Fig. 4A, Fig. S4). While the survival of the implanted cells increased with the hMP-hASC TEMGs, hLamin AC staining indicated that the mass was majorly comprised of host tissue. Additionally, we observed that the hMP-hASC TEMGs increased ECM deposition in the form of Col IV. Quantification of the number of hLamin AC cells showed an increase in survival of human cells at 28 DPI when treated with TEMGs that were supplemented with hASCs compared to those without hASCs (Fig. 4B). The area of Col IV signal was divided by the cross-section area of the TA to provide a % Coverage, which showed an increase Col IV % coverage at 28 DPI when treated with TEMGs that were supplemented with hASCs compared to those without hASCs (Fig. 4C).

Figure 4: Co-culture of hASCs with hMPs on Fibrin Microfiber Bundles to Create TEMGs Promotes Increased Human Cell Survival Post-Implantation and Increased Col IV Deposition.

(A) Representative images of TA cross-sections at 28 DPI stained for DAPI, Col IV, and hLamin AC for each group (Scale Bar = 1000 µm) and cropped images of representative regions (Scale Bar = 250 µm). (B) Quantification of the number of hLamin AC+ cells in a cross-section from each group (n = 4). (C) Quantification of TAs Col IV+ area as a percentage of each TA cross-section area from each group (n = 4). Significance was determined using a one-way ANOVA with multiple comparisons and Tukey post-test. Error bars represent standard error of the mean (SEM). *: p < 0.05; **: p < 0.01; ***: p < 0.001.

hASCs co-cultured on hMPs TEMGs skew host macrophages towards the M2 phenotype when implanted in a murine model of TA VML in vivo

After 28 DPI harvested TAs were sectioned and stained with F4/80, iNOS, CD206, and DAPI, to assess the host macrophages phenotype and whether this changed in response to the different treatments (Fig. 5A, Fig. S5). F4/80 (or EGF-like module-containing mucin-like hormone receptor-like 1) is a cell surface glycoprotein that is expressed in murine macrophages. To identify the polarization of macrophages F4/80 is combined with a pro-regenerative marker like CD206 or a pro-inflammatory marker like inducible Nitric Oxide Synthases (iNOS) which is expressed on inflammatory macrophages and lymphocytes. We observed a skewing towards a more pro-regenerative macrophage phenotype when treated with hMP-hASC TEMGs. F4/80+/CD206+ signal was employed as a measure of the M2 response and F4/80+/iNOS+ as M1. The M2/M1 increased at 28 DPI when treated with hMP-hASC TEMGs that compared to hMP TEMGs without hASCs (Fig. 5B).

Figure 5: Co-culture of hASCs with hMPs on Fibrin Microfiber Bundles to Create TEMGs Promotes a Skew to the M2 Phenotype for Host Macrophages.

(A) Representative images of TA cross-sections at 28 DPI stained for DAPI, F4/80, CD206, and iNOS for each group (Scale Bar = 1000 µm) and cropped images of representative regions (Scale Bar = 250 µm). (B) Quantification of the M2/M1 ration in a cross-section from each group (n = 4). Significance was determined using a one-way ANOVA with multiple comparisons and Tukey post-test. Error bars represent standard error of the mean (SEM). *: p < 0.05; **: p < 0.01; ***: p < 0.001.

TEMGs promote restoration of vasculature and nerves in a murine (NSG) model of TA VML

After 28 DPI harvested TAs were sectioned and stained with CD31, β III tubulin (β3T), and DAPI, to provide a visualization of the vasculature and nerves (Fig. 6A, Fig. S6). Quantification of the CD31 and β3T stained muscle cross-sections revealed an increase in the number of CD31+ vessels between the no treatment VML and TEMG-treated VML (although only the hMP-hASC TEMG showed significance) and confirmed that the number of CD31+ vessels in the TEMG treated VML was statistically equivalent to the uninjured contralateral TA (Fig. 6B). Quantification of the number of β3T+ nerves showed that there was a slight (not statistically significant) increase in the TEMG treated VML compared to the no treatment VML, and TEMG treated VML was statistically equivalent to the uninjured contralateral TA (Fig. 6C). When are of these vessels and nerves were measured and normalized to the size of the muscle cross-section, it showed an increase in vessel coverage in the no treatment group compared to the other groups (although only when compared to hMP-hASC TEMGs was it significant), and a trend in increased nerve coverage with TEMG treatment (Fig. 6D and Fig. 6E, respectively).

Figure 6: hMP-hASC TEMGs Increase in Vascular Number and Decrease in Vascular Density Compared to VML with No Treatment at Day 28 Post-VML Injury.

(A) Representative images of TA cross-sections stained with CD31 and β3T for each group (Scale Bar = 1000 µm). (B) Quantification of CD31+ vessels as a number per TA cross-section for each group (n = 4). (C) Quantification of β3T+ nerves as a number per TA cross-section for each group (n = 4). (D) Quantification of CD31+ vessels area normalized to TA cross-section area for each group (n = 4). (D) Quantification of β3T+ nerves area normalized to TA cross-section area for each group (n = 4). Significance was determined using a one-way ANOVA with multiple comparisons and Tukey post-test. Error bars represent standard error of the mean (SEM). *: p < 0.05; **: p < 0.01; ***: p < 0.001.

To investigate the effects on acetylcholine receptor (AchR) clusters and neuromuscular junction (NMJs) formation, the sectioned TAs were stained with α-Bungarotoxin (αBTX) to mark regions of αBTX and β3T across the entire muscle cross-section (Fig. 7A, Fig. S7). Quantification of the αBTX stained muscle cross-sections revealed that hMP-hASC TEMG treatment significantly decreased the number (Fig. 7B) and area (Fig. 7C) of AchR clusters, while hMP TEMG treatment showed a similar number of AchR clusters as the uninjured contralateral TA. The percentage of AchR clusters co-localized with nerves forming NMJs (Fig. 7D) was similar between groups except for the significant decrease after hMP-hASC TEMG treatment.

Figure 7: hMP TEMGs maintain a similar number of NMJs as Uninjured muscle at Day 28 Post-VML Injury.

(A) Representative images of TA cross-sections stained with αBTX and β3T for each group (Scale Bar = 1000 µm). (B) Quantification of αBTX+ staining per TA cross-section for each group (n = 4). (C) Quantification of αBTX+ staining an area normalized to TA cross-section area for each group (n = 4). (D) Quantification of the percent of αBTX+ clusters co-localized with β3T+ stains per TA cross-section for each group (n = 4). Significance was determined using a one-way ANOVA with multiple comparisons and Tukey post-test. Error bars represent standard error of the mean (SEM). *: p < 0.05; **: p < 0.01; ***: p < 0.001.

hASCs promote proliferation and maturation of hMPs while hMPs decrease inflammatory signaling and apoptosis in hASCs

To elucidate the interactions between hASCs and hMPs, an indirect co-culture was performed in which hMP TEMGs and monolayer hASCs were separated by an insert with 6 µm pores and compared to individual cultures in which the hMP-TEMGs and the monolayer hASCs were cultured separately. After 21 days of culture, RNA was collected from the cells and qPCR was run on a series of myogenic (Fig. 8A), cell cycle (Fig. 8B & D), and FAP transcripts (Fig. 8C & E). The hMP TEMGs indirectly co-cultured with hASCs showed few statistically significant changes in myogenic genes, and no statistically significant changes in non-myogenic genes. However, these changes overall indicate an increase in maturity (decrease in immature myogenic transcripts and large increase in MYH7) and an increase in proliferation (MKI67) of hMPs due to the presence of hASCs. The hASCs indirectly co-cultured with hMP TEMGs showed few statistically significant changes as well. These changes reflected an increase in maturity with a slight skew towards adipogenesis, a decrease in immune signaling, an increase in cell survival, and decreased proliferation. Specifically, the hMP TEMGs that were indirectly co-cultured with hASCs expressed decreased MYOG (as well as a trend towards decreased MYF5, MYF6, and MYOD1), decreased MYH1, and increased MYH7. The hMP TEMGs that were indirectly co-cultured with hASCs also expressed higher MKI67 (p=0.1845). Conversely, the hASCs that were indirectly co-cultured with hMPs expressed decreased CCL2 (an immune recruitment factor) and CCN4 (fibroblast marker) (as well as decreased CD34 and PDGFRA; progenitor markers). hASCs that were indirectly co-cultured with hMP TEMGs expressed decreased CASP3 and TP53 (cell death markers) (as well as a non-statistically significant decrease in MKI67; proliferation marker) and increased PPARGC1A (adipogenic marker) although not statistically significant.

Figure 8: In co-culture hASCs promoted proliferation and maturation of hMPs while hMPs promoted decreased inflammatory signaling and increased cell survival in hASCs.

(A) Schematic of the experimental design (B) qPCR quantification of fibers for myogenic transcripts (n = 3–4). (C) qPCR quantification of fibers for cell cycle transcripts (n = 3–4). (D) qPCR quantification of plates for FAP transcripts (n = 3–4). (E) qPCR quantification of plates for cell cycle transcripts (n = 3–4). Significance was determined for qPCR using unpaired two-tailed t-tests for each individual transcript. Error bars represent standard error of the mean (SEM). *: p < 0.05; **: p < 0.01; ***: p < 0.001.

Discussion

Favorable interactions between cells and biomaterial scaffolds are essential to creating TEMGs with high regenerative efficacy. In our murine model, roughly 40% of the TA muscle is removed to create a VML injury. Yet, the hMP TEMGs constructed with fibrin microfiber bundles elicited close to 100% restoration of muscle cross-sectional area and roughly 90% weight recovery with high MHC expression in the defect region. The regeneration observed compares favorably with similar prior studies in which ECM was used with hMPs.^10,21 The electrospun fibrin microfiber bundles in our prior works have been shown to stimulate the proliferation and maturation of myogenic cells¹⁷ and enhance myogenesis in engineered hTEMGs.^9,16,18–20 However, acellular fibrin microfiber bundles do not stimulate regeneration as well as when seeded with pro-myogenic cell populations, which promote robust muscle regeneration and angiogenic ingrowth while minimizing scarring. Together, these data suggest that the scaffold is an essential component, but not sufficient to promote regeneration on its own. The successful regeneration outcomes observed with the hMP TEMGs were much greater than in our previous studies in which hASCs or iPSC-derived myogenic progenitor cells were used to create TEMGs^9,16,18–20, and suggest that they represent a promising strategy for creating autologous therapies for VML even in the absence of supporting cell populations.

In spite of the unprecedented regenerative outcomes, we sought to assess whether the addition of ASCs could further enhance neurovascular infiltration into our tissues. In vivo studies confirmed that the number of vessels and nerves increased, however the density of these vessels and nerves were similar to that of uninjured muscle. It is likely that the increase in vascular density of the no treatment group is due to chronic inflammation that occurred in the untreated injury. ASCs promote proliferation and differentiation of MPs in 2D co-culture.^43–45 This remains true even if only ASC extract is used to supplement MP media.⁴⁶ Thus, indirect signaling by hASCs may be sufficient to further enhance the pro-survival and pro-regenerative outcomes of hMPs, which was confirmed in the in vitro indirect co-culture study. Based on this promising in vitro data, we assessed the direct contribution of hASCs to VML healing in conjunction with hMPs. We hypothesized that inclusion of hASCs at the initiation of 3D culture could be detrimental to muscle tissue formation if hASC proliferation impeded direct hMP-hMP contact and the potential of the hMPs to fuse as they matured. Similarly, we reasoned that above a critical concentration threshold, ASCs may impede the pro-regenerative capacities of hMPs, as our prior studies using hASCs exclusively to treat VML did not induce improvements in myogenesis. In vitro studies validated both hypotheses. Overall, the importance of timing and dosage of hASCs to promote hMP myogenesis is demonstrated in vitro. Further fine-tuning is needed to translate this increased myogenesis from hASCs seen in vitro to treatment of VML in vivo, such that the TEMGs continue to promote M2 macrophage polarization and ECM deposition while avoiding the overgrowth of adipose tissue.

Our previous work with hASC TEMGs did not result in excessive adipose tissue at the site of injury,⁹ and the hMP TEMGs did not promote adipose tissue growth. Neither did the hMP-hASC TEMGs exhibit large growths of adipocytes in vitro. We speculate that the adipose tissue growth might be a result of the signaling interaction between the transplanted hMPs and hASCs and the endogenous FAPs within the context of the injury microenvironment in immunocompromised NSG mice resulting in dysregulated FAP proliferation and adipogenesis. In the regions of adipose deposition, while there are human cells, the majority of cells are of non-human origin. Activation and regulation of FAPs is primarily controlled by the immune response in skeletal muscle injury. Anti-inflammatory (or pro-regenerative) macrophages promote FAP expansion via TGF-β⁴⁷ and adipogenesis via cytokines like IL-4^48,49 and we observed an increase in polarization towards M2 macrophages in the groups that had the large adipose tissue growths. However, M1 macrophages are also inducing FAPs to undergo apoptosis.⁴⁷ While these immune cells are present in our NSG mice, NSG mice lack T cells, B cells, and NK cells that coordinate with innate immune cells and thus the innate immune cells are not fully functional. We speculate that in immunocompetent animals, the regulation of FAPs apoptosis would be restored which might inhibit excessive adipose tissue formation. Future work in an immunocompetent animal model will be beneficial to confirm this. NSG immunocompromised mice were necessary to avoid rejection of the implanted human cells, however, in the clinical setting these cells would be isolated from the patient and culture expanded before re-implanting as an autologous TEMG, avoiding the need for immunosuppression. Implantation of syngeneic MPs and other muscle resident cells into C57BL/6J mice with TA defects has been performed previously without rejection⁵⁰ supporting the potential feasibility of this approach. However, greater care would be required in a clinical setting to ensure that cultivation methods for the expansion of autologous cells do not introduce immunogens.

The in vitro indirect co-culture was used to assess the potential signaling between the hMPs and hASCs that may have given rise to the lipoma-like tissue. CCL2 is known to be a pro-inflammatory (M1) stimulatory factor, while IL-10 is a pro-regenerative (M2) stimulatory factor.⁵¹ The main sources of IL-10 in regenerating skeletal muscle are macrophages and regulatory T cells (Tregs), however, it has also been demonstrated that FAPs increase IL-10 expression upon muscle damage-induced activation ⁵² IL-10 inhibits synthesis of TNF-α (which is anti-myogenic), and thus is likely secreted by FAPs to prevent TNF-α induced apoptosis.⁵³ IL-10 also promotes skewing of immune cells towards an anti-inflammatory phenotype, creating a more pro-regenerative niche. Decreased expression of CCL2 (and a trend towards increased expression of IL-10) in hASCs co-cultured with hMPs is a possible mechanism that supports the skewing towards M2 observed from the hMP-hASC TEMGs. Furthermore, FAPs are the main source of Wnt in skeletal muscle.⁵⁴ Through regulation of GSK/B-catenin, Wnt signaling modulates PPARγ expression leading to adipogenesis, as well as follistatin expression which promotes myogenic multinucleation.^54–57 The decrease in CCN4 (also known as WISP-1, which is induced by Wnt1) in hASCs co-cultured with hMPs indicates a decrease in Wnt signaling and likely decrease in β-catenin, which would result in increased adipogenesis and decreased promotion of myogenic multinucleation. This increased adipogenesis and decreased myogenesis could have participated in the formation of the large quantities of adipose tissue seen after hMP-hASC TEMG implantation.

In conclusion, the current study provides a hMP TEMG design in which electrospun fibrin microfiber bundles, when combined with hMPs, enabled unprecedented regeneration of muscle weight, cross-sectional area, and MHC expression. The use of hASCs as a myogenic support cell was demonstrated by the promotion of proliferation and differentiation in hMPs during in vitro co-culture, with timing and dosage playing an important role in this interaction. The hMP-hASC TEMGs showed increased cell survival post-implantation and skewed local macrophages towards an M2 phenotype but resulted in large adipose tissue growths in vivo. The changes in transcription due to interactions between hASCs and hMPs which led to these findings in vitro and in vivo were identified through indirect co-culture. Supporting in vitro outcomes, a decrease in MYOG (which is expected to decrease as myogenic cells mature and fuse), a decrease in MYH1 (fast-twitch, IIX), and a large increase in MYH2 (slow-twitch, I), as well as a trend in increased MKI67 (proliferation) were observed. Additionally, decreased CCL2 (M1 promotion), CCN4 (Wnt signaling), CASP3 (apoptosis) and TP53 (cell cycle/apoptosis), as well as a trend in increased PPARGC1A (PPARγ signaling, pro-adipogenesis), supported in vivo outcomes. Consequently, the potential for incorporating hASCs in a therapeutic setting remains questionable. Future studies might test immunocompetent mice and syngeneic cells to assess whether, in animals with an intact immune system, the in vivo signaling between transplanted ASCs and myogenic cells enhance the therapeutic outcomes.

Materials and Methods

Electrospinning Fibrin Scaffolds

Fibrin scaffolds were electrospun as described previously.^9,16,18,19 Briefly, fibrin microfiber bundles were created through a co-extrusion of syringes of 1% fibrinogen (Sigma-Aldrich, St. Louis, MO, USA) in 0.2% polyethylene oxide (average Mv ~4,000 kDa, Sigma-Aldrich) and 0.75% alginate (Sigma-Aldrich) in 0.2% polyethylene oxide (average Mv ~4,000 kDa, Sigma-Aldrich), combined through a y-syringe. The combined solutions extruded through a 27G needle-tip with ~3.5 kV of applied voltage. The resulting stream was collected on a rotating dish (~35 rpm), containing 50 mM CaCl₂ and 20 U/mL thrombin (Sigma-Aldrich) to crosslink the fiber for 5.75 min. The fiber was allowed to crosslink for an additional 3–5 minutes and was wrapped 4 times around a rectangular frame (15 mm x 30 mm) made from acrylonitrile-butadiene-styrene. Following this, scaffolds were incubated overnight in 250 mM sodium citrate (Sigma-Aldrich) to dissolve the alginate, washed in DI water to remove remaining dissolved alginate, and stored in DI water for up to 2 weeks before use.

Primary Human Adipose-Derived Stem Cell Isolation

The hASCs were isolated from lipoaspirate tissue as previously described (under an institutional review board-approved protocol).^9,58 Two female donor sources were used: a 26-year-old Asian and a 37-year-old African American. In vitro studies utilized hASCs from both donors while in vivo studies utilized hASCs from the second donor. Briefly, tissue was digested with Type I Collagenase (1 mg/mL; Worthington Biochemical Corp., Lakewood, NJ, USA) to isolate the stromal vascular fraction of cells. These cells were plated onto tissue culture plastic as “passage 0 ASCs” when they reached 80–90% confluence, then cryopreserved for future studies.

Primary Human Myoblast Cell Culture

The hMPs were purchased from Lonza (USA). Two Caucasian male donor sources were used: a 29-year-old and a 27-year-old. hMPs were expanded in tissue culture flasks up to passage 3 using hMP Media: Skeletal Muscle Cell Growth Basal Medium-2 (Lonza, USA) supplemented with the Skeletal Muscle Cell Growth Medium-2 BulletKit [gentamicin–amphotericin (GA-1000), human epidermal growth factor (hEGF), dexamethasone, L-Glutamine, and 10% FBS].

Primary Human Adipose-Derived Stem Cell Culture

The hASCs were thawed and expanded in Growth Medium: high-glucose DMEM (ThermoFisher Scientific, Waltham, MA, USA) with 10% fetal bovine serum (FBS; Atlanta Biologicals, Flowery Branch, GA, USA), 1% penicillin/streptomycin (P/S; ThermoFisher Scientific), and 1 ng/mL FGF-2 (PeproTech, Rocky Hill, NJ, USA). The hASCs were then trypsinized and used at passage 2 for all experiments. The phenotypic profile of the cells at this passage were examined via flow cytometry for CD31, CD34, CD45, and CD140a. Briefly, hASCs at passage 1 were thawed and expanded as described above. Passage 2 ASCs were then suspended in phosphate buffered saline (PBS) containing 2% FBS and incubated with monoclonal antibodies 30 minutes at 4°C. Cells were then analyzed with a flow cytometer (Attune NxT Flow Cytometer; ThermoFisher Scientific). Two female donor sources were used: a 26-year-old Asian and a 37-year-old African American. In vitro studies utilized hASCs from both donors while in vivo studies utilized hASCs from the 37-year-old donor.

Seeding hMPs on Scaffolds

For optimal cell adhesion to the fibrin microfiber bundles, hMPs were suspended in thrombin at 1.25 U/mL in hMP Media containing 12.5 mmol/L CaCl₂. Fibrinogen was diluted in high-glucose DMEM (ThermoFisher Scientific, Waltham, MA, USA) to 20 mg/mL and equal parts of fibrinogen and the thrombin cell mixture were mixed and immediately pipetted onto the scaffold surface in 15 µL total. Cells were seeded at a concentration of 900,000 cells/fiber (30 mm). Following seeding, constructs were incubated for 0.5 hours before the media was applied. The hMPs were cultured in hMP media with 30 µg/mL aprotinin (Affymetrix, Santa Clara, CA, USA). Media was exchanged every 2 days for all studies. In the concentration experiment hASCs were seeded in the same manner as the hMPs, however, this led to a decrease in hMP proliferation. Therefore, in all subsequent experiments hASCs were suspended in hMP media and seeded by pipetting onto the scaffold surface in 15 µL total. Construct were incubated for 1 hour before the hMP Media was applied. Halfway through this incubation, 15 µL of hMP media was added to maintain sample hydration.

VML Defect Model

All animal and surgical procedures were performed in accordance with the Institutional Animal Care and Use Committee at Johns Hopkins University School of Medicine. For all animal experiments male NSG immunodeficient mice (Jackson Lab, Bar Harbor, ME, USA) aged 8–10 weeks (n = 4) were used to allow for implantation of human cells. Critical-sized defects (~30–50% of TA muscle) were performed, as previously described.¹⁸ Briefly, following isoflurane anesthetization, the right TA was exposed, and part of the muscle was removed. Engineered muscle grafts were then placed in the defect site (4 skeletal muscle grafts per defect) and sutured to the remaining muscle using nonabsorbable sutures (6–0 Nylon, ETHICON, San Lorenzo, Puerto Rico, USA). Implanted scaffolds were grown 21 days prior to implantation. Sutures were also used to close the skin and OstiFen^™/Carprofen (MWI, Boise, ID, USA) (5 mg/kg) was injected subcutaneously after the surgery for pain management. The mice were sacrificed by isoflurane overdose and cervical dislocation 4 weeks after the surgery. Immediately following this, the TA muscle, including the implanted scaffolds, was removed and cryopreserved for sectioning.

Indirect Co-Culture

900,000 hMPs (passage 3; Lonza) were seeded on electrospun fibrin microfiber bundles to create hMP TEMGs. After 7 days, hMP TEMGs were co-cultured through a 40 µm filter with passage 2 hASCs seeded (at 5% the concentration of hMPs, 45,000) onto 6-well plates and cultured for 14 more days then compared to hMP TEMGs cultured for 21 days and hASCs cultured on 6-well plates for 14 days. hMP Media was used for the entirety of cultures.

Whole Mount Immunostaining

Samples were fixed in methanol-free PFA at 4 °C for 3 h, washed with PBS, then blocked and permeabilized with 10% donkey serum (Sigma-Aldrich) in PBS with 0.2% Triton-X for 3 hours at 4 °C. Samples were incubated with primary antibodies overnight at 4 °C (100 rpm) followed by three 1 h washes. Scaffolds were then incubated with secondary antibodies overnight at 4 C (100 rpm) followed by three 1 h washes with DAPI in the second wash (1:2000; Sigma). Samples were imaged with a confocal microscope (Zeiss LSM 510 and 710).

Histology

TA muscles were sectioned on a cryostat (Leica CM3050 S) at a thickness of 10 μm for cross-sections. Serial sections were collected starting from the bottom suture on 15 glass slides, with approximately 8–10 sections along the length of the TA collected per slide. Masson’s Trichrome and H & E stains were imaged on an inverted laser fluorescence microscope (Zeiss AxioObserver.A1). For immunohistochemistry, slides were fixed in cold 4% formaldehyde for 10 min and rinsed with PBS three times for 15 min each, then blocked in 10% normal donkey (Sigma) in PBS for 1 hour at room temperature (RT). Slides were incubated with antigen-specific primary antibodies in blocking solution overnight at 4 °C. Primary antibodies included rabbit anti-desmin (1:200; SCBT), mouse anti-myosin heavy chain (all isoforms) (1:200; DSHB), goat anti-CD31 (1:400; R&D Systems), rabbit anti-β tubulin III (1:200; Sigma), rabbit anti-lamin A + C (5 1:200; Abcam), rabbit anti-collagen IV (1:200; SCBT), mouse anti-iNOS (1:100; Abcamn), rat anti-F4/80 (1:200; Abcam), rabbit anti-CD206 (1:200; Abcam), mouse anti-Sarcomeric α-Actinin (1:200; Abcam), and rabbit anti-laminin (1:200; Sigma). After three 5 min washes with PBS, slides were incubated with secondary antibodies and DAPI (1:2000) diluted in blocking solution for 1 hour at RT. Slides were washed three times for 5 min then mounted with 50% glycerol and imaged on a confocal microscope (Zeiss LSM 710). One section was quantified from each slide per sample, from the middle location longitudinally within the TA defect region.

Statistical Analysis

All image quantification was performed using FIJI. Fiber and cross-section quantifications were performed by manually outlining and measuring the area of the fiber then quantifying the area of each positive signal and dividing by the total area then multiplying by 100% to get a percent coverage. Muscle cross-sectional areas and number of myofibers with centrally located nuclei were quantified from the H & E staining; the number of myofibers with centrally located nuclei was obtained by having someone manually outline the fibers with nuclei not residing along the outside of the muscle fibers who was blinded to the samples while quantifying (performed using Fiji). Statistical analysis was performed using GraphPad Prism 5 software. Statistical significance was determined by either a one-way ANOVA with multiple comparisons and Tukey post-test for image quantifications or unpaired two-tailed t-tests for qPCR. Error bars represent standard error of the mean (SEM).