Regeneration competent satellite cell niches in rat engineered skeletal muscle

Abstract Satellite cells reside in defined niches and are activated upon skeletal muscle injury to facilitate regeneration. Mechanistic studies of skeletal muscle regeneration are hampered by the inability to faithfully simulate satellite cell biology in vitro. We sought to overcome this limitation by developing tissue engineered skeletal muscle (ESM) with (1) satellite cell niches and (2) the capacity to regenerate after injury. ESMs contained quiescent Pax7‐positive satellite cells in morphologically defined niches. Satellite cells could be activated to repair (i) cardiotoxin and (ii) mechanical crush injuries. Activation of the Wnt‐pathway was essential for muscle regeneration. Finally, muscle progenitors from the engineered niche developed de novo ESM in vitro and regenerated skeletal muscle after cardiotoxin‐induced injury in vivo. We conclude that ESM with functional progenitor niches reminiscent of the in vivo satellite cell niches can be engineered in vitro. ESM may ultimately be exploited in disease modeling, drug screening, or muscle regeneration.

to the formation of new muscle fibers, which are characterized by expression of immature myosin isoforms. [4][5][6] Given these regenerative properties, satellite cells have been suggested as a potential cell therapeutic agent in severe skeletal muscle diseases. 7 However, implantation of in vitro propagated "satellite cells" has not been overly successful in treating muscular dystrophies. 8,9 The lack of regeneration described in these studies may be in part attributed to the irreversible activation of satellite cells under standard culture conditions, leading to differentiation into myoblasts and their subsequent fusion into multinucleated myotubes. 10,11 Only when satellite cells in their muscle fiber niche were transplanted, repopulation of the endogenous niche and widespread muscle regeneration were observed in a mouse model of muscular dystrophy. 1 In an attempt to better retain the satellite cell phenotype under defined in vitro culture conditions, pliant tissue culture substrates have been introduced to create an artificial niche. [12][13][14] Data from these studies suggest that biomechanical cues provided by the extracellular environment are important regulators of satellite cell biology.
Tissue engineering adds a third dimension to classical monolayer culture models with the aim to offer more "physiological" conditions and to enhance maturation toward an adult state. [15][16][17][18] Tissue engineering of skeletal muscle from rodent and human cells has been introduced roughly 20 years ago 19 with implications in drug screening, gene therapy, and tissue replacement therapy. [20][21][22][23][24] Ideal tissue engineered skeletal muscle should display all characteristic morphological (ie for example fusion of myotubes to form a true syncytium) and functional (ie for example tetanic contractions upon high frequency stimulation) properties of native skeletal muscle. [24][25][26] It should moreover contain functional satellite cell niches capable of muscle regeneration if activated by injury as this would allow to study muscle regeneration in the dish. 27 Here, we demonstrate that satellite cells from engineered skeletal muscle (ESM) are capable of regenerating muscle in vitro and in vivo. Satellite cells respond to pharmacological interventions demonstrating the principal utility of ESM in screens for regeneration inducing drugs in vitro.
Connective tissue and tendons were removed and the muscle was minced to a fine slurry. This was digested for 90 min at 37°C using 2.5 mg/ml Collagenase II (Biochrom) and 20 µg/ ml DNAse I (Calbiochem) in basal medium. The tissue suspension was subsequently passed through filters with gradually decreasing mesh size (250 100 40 µm, BD Biosciences). The cell suspension was then preplated for 1 hour in basal medium on cell culture dishes to reduce the number of fibroblasts. The nonattached cells were resuspended in proliferation medium (basal medium supplemented with 10 ng/ml bFGF; PeproTech), 10 ng/ ml EGF (PeproTech), Insulin-Transferrin-Selenite (ITS) + X (100x; Invitrogen) and expanded for 5 days on uncoated flasks without further passaging. For implantation experiments transgenic Lewis rats with ubiquitous expression of eGFP were used for skeletal myocyte preparation. 28

| Myoblast culture
For 2D culture primary myogenic cells were plated on the ESM matrix mixture (Collagen type I plus Matrigel ® ; diluted 1:50 in PBS) and cultured in parallel for 12 days in ESM culture medium (see above).

| Measurement of elastic modulus
The Young's modulus (E) of ESM was measured at 37°C in Tyrode's solution in thermostatted organ baths. Stress-strain curves were derived by repeatedly increasing the length of the ESM starting from slack length. Passive force was recorded after it had reached a steady-state. The stress (kPa) was calculated from the passive force (mN) and the ESM cross-sectional area (mm 2 ). Strain was calculated by dividing ESM length change (delta) by the ESM slack length. The Young's modulus was derived from the slope of the stress-strain curve at the range of physiological strain (0%-10% length change).

| Immunocytochemistry
ESMs were fixed in 4% formaldehyde or 70% ice cold ethanol overnight at 4°C. Primary skeletal muscle was fixed overnight in 4% formaldehyde. For cryosectioning ESM/skeletal muscle were immersed in 30% sucrose in PBS and embedded in OCT compound. Sections were cut at 10 µm thickness using a Leica Cryotome (CM3050 S). For whole mount staining of ESM, permeabilization and blocking was achieved by overnight incubation in phosphate buffered saline (PBS), 5% goat serum, 1% bovine serum albumin (BSA), and 0.1% Triton-X. The same buffer was used for the primary and secondary antibody incubation. Labelling of Pax7 (Developmental Studies Hybridoma Bank, DSHB), Ki67 (Pierce), embryonic myosin heavy chain (Vector Laboratories), tropomyosin and laminin (both Sigma), m-cadherin and caveolin-1 (both Santa Cruz), caveolin-3 (BD Biosciences), ED1/ ED2 (AbD Serotec), and prolyl-4-hydroxylase (Millipore), and GFP (Invitrogen) was performed overnight at 4°C followed by several washes with PBS. Incubation with appropriate Alexa fluorochrome labelled secondary antibodies (Invitrogen, 1:1,000) was performed overnight at 4°C. Where indicated labelling of f-actin and nuclei was performed using Alexa 488-conjugated phalloidin 1:60 (Invitrogen) and 4',6-diamidino-2-phenylindole (DAPI; 1 µg/ml) in parallel to secondary antibodies. Plated cells were fixed in 4% formaldehyde for 10 minutes at room temperature. After blocking for 30 minutes, primary antibodies were added overnight at 4°C. After three washes with PBS, appropriate secondary antibodies were applied for 1 hour at room temperature. Cells were washed and mounted in Fluoromount-G (Southern Biotech). All stainings were imaged on a Zeiss LSM 710/NLO confocal microscope. Quantification of labelled cells was performed by imaging at least three randomly chosen focal planes per ESM of at least three different series. Cell counting was done using ImageJ cell counter tool.

| Flow cytometry
Cells were fixed in 4% formaldehyde for 10 minutes. Staining for Pax7 (DSHB), desmin (Sigma), and ED1/ ED2 (AbD Serotec) was performed at 4°C for 45 minutes. Appropriate secondary antibodies were applied for 30 minutes at room temperature. Isotype controls were prepared in parallel. Samples were run on a LSRII cytometer and analysed with Facs Diva software (BD Biosciences).

| Quantitative real time PCR
Snap-frozen ESMs were lysed in RNA extraction buffer (Trizol, Invitrogen) using the Tissue lyser device (Quiagen, 25 strokes/second). RNA was extracted according to manufacturer's recommendations followed by reverse transcription of 1 µg RNA with Superscript II kit (Invitrogen). Quantitative PCR was performed using 50 ng of cDNA as template, TaqPolymerase Master Mix (Applied Biosystems), TaqMan ® primer/probe sets, and a 7900 HT Applied Biosystems Cycler. The following primer/probe sets were used:

| ESM injury models
To induce skeletal muscle injury, cardiotoxin (CTX: 25 µg/ ml, Latoxan) was applied for 24 hours. Mechanical ("crush") injury was induced with fine anatomical forceps (tip width: 1 mm). 3 short, full squeezes (one second each) were applied to one arm of the ESM yielding a reproducible local injury.
To irreversibly inactivate satellite-like cells, ESMs were taken off the holders and subjected to a single dose of 30 Gy (applied in ~ 13 minutes) irradiation using an STS Biobeam 8000 (Germany) gamma irradiator.

| Measurement of adenylate kinase (AK) activity
Cell lysis was assessed by measuring the release of adenylate kinase into the culture medium. Activity of AK was determined in culture medium supernatant after 24 hours of cardiotoxin treatment utilizing the Toxilight assay (Lonza) and a multiwell plate reader (FlexStation 3, Molecular devices). Values were calculated relative to full lysis induced by ToxiLight 100% lysis reagent.

| In vivo regeneration model
Animal experiments were approved by the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit (LAVES: AZ 11/0581). The hind legs of RNU nude rats (Charles River) were irradiated with 40 Gy using a RS 225 X-Ray Research System (Gulmay Medical Systems, UK). 24-72 hours after irradiation, the right TA muscle was injected with 100 µl of 20 µmol/L CTX to induce muscle damage. 72 hours after CTX-induced damage, the myogenic cell suspension from 2D or ESM culture was injected into the injured muscle (5x10 5 cells in 100 µl DMEM/F12 medium) using a 27 G needle.

| Statistical analysis
Data are presented as mean ± standard error of the mean (SEM). Statistical differences were determined using unpaired Student's t test or non-parametric Mann-Whitney test (two-tailed) or ANOVA followed by appropriate post hoc test as indicated. A p value < 0.05 was considered statistically significant.

| Engineered skeletal muscle with differentiated morphology and functionality
We isolated skeletal muscle cells from adult Wistar rat hind limb muscle and cultured them in vitro for maximally 5 days to expand cell numbers, but minimize the impact of in vitro culture on cellular properties. The primary muscle cell isolates contained 9 ± 2% Pax7-positive, 31 ± 3% MyoD-positive, 29 ± 4% Myogenin-positive, and 37 ± 4% desmin-positive myogenic cells (n = 3-4 preparations), with the remaining non-muscle cells being predominately prolyl-4-hydroxylase-positive fibroblasts ( Figure 1A, Figure S1). Engineered skeletal muscle (ESM) was generated from these primary muscle cell isolates, by mixing with solubilized collagen type 1 and Matrigel ™ . This reconstitution mixture was cast into circular molds, which facilitated condensation into mechanically stable circular tissue constructs within 5 days ( Figure 1B). We subsequently transferred ESMs onto custom made holders for additional 7 days to maintain them under a defined load ( Figure 1C). This procedure yielded contractile ( Figure 1D, Movie S1) ESM with morphologically well-differentiated actin and tropomyosin-positive muscle fibers lined by a Laminin-positive basal lamina  Figure 2D). When normalized to mean muscle cross-sectional area (CSA) the tetanic force corresponded to a specific force of 21 ± 1 kN/ m2 (n = 13). This is about 10% of the specific force of native fast skeletal muscle, 33 indicating high but not fully functional maturation of muscle fibers within ESM. Enhanced expression of mature myosin heavy chain transcripts ( Figure  2E and 2F) in ESM suggested advanced maturation compared to conventional 2D skeletal muscle cell culture. The fast twitch characteristics (time to peak: 43 ± 1 ms, time to 50% relaxation: 47 ± 2 ms, n = 20 ESM) and the absence of slow myosin transcripts provided functional and molecular evidence for the development of fast-type muscle rather than slow-type muscle phenotype in ESM and are comparable to day 4 neonatal rat fast muscle. 34

| Engineered skeletal muscle contains muscle progenitor niches
ESMs are composed of differentiated and functionally mature myofibers arranged as true multinucleated muscle syncytia. We next investigated if ESM contain a stem cell pool similar to native muscle. Interestingly, we identified mononucleated cells expressing the satellite cell transcription factor Pax7 in ESM, 77 ± 5% (n = 3 ESM) of which were positioned adjacent to differentiated skeletal myofibers, reminiscent of their typical anatomical location in vivo ( Figure 1A; arrow). In addition to Pax7, these cells also expressed other characteristic satellite cell markers, including caveolin-1 and m-cadherin ( Figure 3A, 3) and were situated beneath a collagen IV ( Figure 3C) positive basal lamina, a characteristic of the in vivo satellite cell niche. As further evidence for an in vivo satellite cell phenotype, 88 ± 7% (n = 3 ESM) of these Pax7 cells were negative for the cell cycle marker Ki67, indicative of cell cycle quiescence ( Figure 3D). Ki67 staining also indicated that fused myotubes were post-mitotic ( Figure 3D; arrows). Our model therefore seems to provide the appropriate cues to facilitate both, (1) fusion and differentiation of muscle cells into postmitotic myotubes and (2) development of a muscle progenitor cell niches reminiscent of the physiological satellite cell niches.
To scrutinize which conditions favor the formation of the niche in ESM we compared Pax7 expression in 3D (ESM) and parallel 2D cultures. Pax7 transcript and protein was clearly detectable in ESM, but not in 2D indicating the maintenance of bona fide skeletal muscle progenitor cells in 3D ( Figure 3E and 3F). We hypothesized that the main difference between 3D and 2D was the substrate stiffness. Interestingly, muscle syncytia in 12 day old ESM exhibited passive forces of 9.5 ± 1.4 kPa (elastic modulus [E], n = 10), which is similar to the elastic modulus reported for tibialis anterior muscle and which was reported to favor satellite cell "stemness". 12

| Cardiotoxin-induced muscle injury and regeneration
We next investigated the response of ESM to muscle injury to scrutinize the functionality of the engineered satellite cell niches. This was firstly tested by inducing muscle damage with cardiotoxin (CTX 25 µg/ml; for 24 hours) on ESM culture day 12 ( Figure 4A). On ESM culture day 14 (ie after an additional 24 hours of CTX washout) we noted severe muscle damage with almost complete destruction of skeletal muscle cells in ESM as determined by adenylate kinase release ( Figure 4B) as well as f-actin (to identify myotubes) * and propidium iodide (PI; to identify nuclei of dead or dying cells) labeling ( Figure 4C; arrows). This massive muscle cell lysis was paralleled by a complete loss of characteristic skeletal muscle transcripts, including embryonic myosin and fast adult myosin ( Figure 4D and 4E). Interestingly, Pax7 transcripts remained largely unaffected, suggesting survival of satellite cells under these conditions ( Figure 4F). Enhanced MyoD transcript expression (culture day 14; Figure 4G) provided first molecular evidence for the activation of the satellite cell niche in damaged ESM. This notion was further substantiated by histological analyses demonstrating marked cell cycle activation (Ki67-positivity) in Pax7 positive cells two days after injury (culture day 14, Figure 4H) and a subsequent increase in Pax7 positive cells at culture day 21 (ie 9 days after injury; Figure 4I). This finding is also in agreement with the concurrent up-regulation of Pax7 transcripts, observed in ESM on culture day 21 ( Figure  4F). As an apparent consequence of satellite cell activation, we observed re-expression of embryonic and adult myosin genes ( Figure 4D and 4E).
Importantly, the net effect of this regenerative response could be robustly assayed by tetanic force measurements. Here we found, consistent with the transcript and protein data, a partial recovery of tetanic force (to 32 ± 5% of initial force, n = 16) 7 days following CTX-induced complete contractile failure ( Figure 4J). The functional regeneration correlated with the morphological evidence of newly formed eMHC-positive muscle fibers ( Figure 4J, inset). Persistent activation of the regenerative potential in ESM resulted in a further normalization of contractile performance after additional 7 days of regeneration (to 56 ± 13% of initial force, n = 4, Figure 4J).
To test whether regeneration of contractile force was indeed related to the activation of satellite cells, we irradiated ESMs with 30 Gy before CTX injury ( Figure 4K). This experimental model is widely used to inhibit cell cycle activity and thus endogenous repair capacity in skeletal muscle in situ. 35,36 In line with the findings from respective experimental animal models, ESMs were not significantly affected by irradiation alone ( Figure 4K). However, functional and morphological regeneration 9 days after CTX injury (culture day 21) was completely abolished in irradiated ESMs ( Figure 4K, Figure S2A). This data suggests that cell division of satellite cells was essential for the muscle regeneration in non-irradiated ESM.
Considering the important role of macrophages in skeletal muscle regeneration in vivo and in vitro, 37,38 we asked whether ESM contain a macrophage population. Indeed, we identified a population of ED1 (CD68) and ED2 (CD163)positive macrophages that were present in the ESM input cells ( Figure S1) and that are likely instrumental in the robust regenerative response seen in ESM ( Figure S2B).

| Regeneration after mechanical injury
As CTX induced severe muscle injury, we asked whether a less dramatic and likely physiologically more realistic local mechanical injury would also activate satellite cells in ESM to facilitate skeletal muscle regeneration. By applying repeated mechanical compressions (3 times, one second long) with a pair of fine forceps restricted to one arm of the ESM at culture day 12 we induced a reproducible "crush injury" (Figure 5A and 5B). As opposed to the full contractile failure in response to the CTX injury protocol, we observed only a mild but reproducible deterioration of contractile force 2 days after injury (ie ESM culture day 14) and full recovery 7 days following injury ( Figure 5C). Notably, the molecular consequences of the regenerative response to the crush injury model were similar as in the cardiotoxin injury model with up-regulation of Pax7 (4 ± 0.8 fold; Figure 5D) 2 days after injury, followed by MyoD (2.4 ± 0.5 fold Figure 5E) and embryonic myosin (2.7 ± 0.6 fold; Figure 5F) up-regulation

| Pharmacological modulation of muscle regeneration in vitro
We next tested if muscle regeneration can be modulated in vitro by pharmacological treatment. ESM injured by CTX were treated with substances targeting signaling pathways implicated in skeletal muscle regeneration. Also the effect of the drug was tested in uninjured ESM thereby measuring the effect on muscle homeostasis ( Figure 6A). Inhibition of Notch signaling by a γ-secretase inhibitor (GSI) reduced satellite cell expansion and muscle regeneration ( Figure 6B, Figure S3) without negative effect on muscle baseline function (muscle homeostasis, Figure 6C). Similarly, inhibition of canonical Wnt signaling (by IWR-1) had a negative impact on regeneration without effects on muscle homeostasis ( Figure 6B,C). These results are consistent with the role of Notch and Wnt signaling described in myogenic regeneration in vivo. 10,[39][40][41] Confirming the essential role of Wnt signaling in muscle regeneration an activator of canonical Wnt signaling (CHIR-99021) enhanced muscle regeneration ( Figure  6B), but muscle function was depressed under basal conditions ( Figure 6C). Both Angiotensin-II and Oxytocin did not have a significant effect on muscle regeneration in ESM ( Figure 6B), Angiotensin-II negatively affected basal muscle homeostasis ( Figure 6C). Collectively, these data suggest that ESM can be used to screen for compounds with skeletal muscle regeneration enhancing properties.

| ESM-derived Pax7 + satellite cells repopulate the skeletal muscle stem cell niche in vitro and in vivo
To fully assess the autonomous regenerative potential of the myogenic cells in the ESM, we assessed whether functional satellite cells could be isolated from ESM using a protocol similar to that for the isolation of satellite cells from skeletal muscle. 42 To unequivocally identify the progeny of ESMderived satellite cells in vitro and in vivo we constructed ESM from eGFP-transgenic rats. 28 To isolate the satellite cells, we dispersed the ESMs enzymatically and separated small, round cells from remaining myofibers by differential straining ( Figure 7A). The isolated mononucleated cells contained 17 ± 5% (n = 4) Pax7 + cells without evidence of any remaining myofibers ( Figure 7A). Interestingly, ESMisolated cells could be utilized to construct "2nd generation ESM" with similar functional performance as the parental "1st generation ESM" ( Figure 7B) indicating high myogenic potential of the isolates. The observation of Pax7 + adjacent to the newly formed skeletal muscle provided evidence for the reconstruction of satellite cell niches also in the "2nd generation ESM" with similar numbers of Pax7 + cells (1st generation: 0.18 ± 0.03 Pax7 + nuclei per myonucleus, 2nd generation: 0.22 ± 0.02 Pax7 + nuclei per myonucleus, n = 3; Figure 7C).
To investigate the potential of ESM-derived satellite cells in vivo we established a CTX injury model in immunocompromised RNU (nude) rats. In this model we first blocked endogenous regeneration by irradiation. As anticipated, muscle regeneration after CTX-injury was greatly reduced in irradiated muscle ( Figure S3). Three days after injury eGFP + ESM-derived cells were injected and could be traced by their epifluorescence even macroscopically ( Figure 7D). In parallel we injected cells from age-matched 2D cultures (14 days; 2D F I G U R E 6 Muscle regeneration enhancing drugs and underlying mode of action identified in ESM. A, Schematic of experimental protocol: cardiotoxin (CTX; 25 µg/ml) was applied on ESM culture day 12 for 24 hours followed by drug or vehicle treatment from culture day 14-21 ("Muscle regeneration"). In parallel the drug effects were tested on uninjured ESM ("Muscle homeostasis"). B-C, 80 Hz tetanic stimulation induced twitch tension (TT) at day 21 of injured B, and uninjured C, ESM treated with indicated compounds and concentrations (GSI: γ-secretase inhibitor, n = 5/6; IWR-1: canonical Wnt inhibitor, n = 12/6; CHIR: Wnt activator, n = 5/3; AT-II: Angiotensin-II, n = 3/3; OT: Oxytocin, n = 3/3). TT was normalized to vehicle treated ESM (n = 16/12; mean ± SEM marked by dashed lines). *P < 0.05 vs. vehicle treated control ESM by unpaired Student's t-test (two-tailed) and ESM were from the same muscle cell isolations). Despite the presence of eGFP + -cells derived from 2D culture in the muscle after 30 days, we did not observe muscle regeneration ( Figure 7E). Conversely, we observed substantial remuscularization -on average 17 ± 2 eGFP + donor muscle fibers -in rats (n = 5) injected with ESM-derived cells ( Figure 7F, G). Staining for Pax7 identified eGFP + donor cells in a satellite cell position after implantation ( Figure 7H).

| DISCUSSION
Satellite cells are the regenerative components in skeletal muscle and typically located in morphologically defined satellite cell niches. 4,5 Here, we describe a tissue engineering modality capable of retaining satellite cell "anatomy" and functionality, both could not be readily anticipated given the commonly observed rapid loss of muscle stem cells properties in standard 2D cultures. 10,11,43 In addition, we demonstrate that ESMs exhibit many properties of native skeletal muscle, including (i) morphological features of a syncytium composed of fused myocytes, (ii) basal contractile performance reminiscent of skeletal muscle (positive force-frequency-and length-force-relationships, tetanic contractions), and (iii) responsiveness to pharmacological agents targeting the nicotinergic-acetylcholine receptors. The observation that satellite cells with true muscle stem cell properties 1,11,[44][45][46][47] can be preserved in ESM appears particularly intriguing and supports earlier studies on the regeneration potential of tissue engineered muscle. 25 These properties include expression of characteristic proteins (Pax7, caveolin-1, m-cadherin), localization in a typical anatomical niche underneath the basal lamina and adjacent to well differentiated myotubes in a non-proliferating state with the propensity for cell cycle activation, myogenic differentiation and functional regeneration upon injury. We have confirmed the regenerative capacity of the engineered satellite cell niche in the following in vitro and in vivo models: (1) severe muscle damage by cardiotoxin (CTX) administration, (2) mild mechanical ("crush") muscle damage, (3) re-isolation of functional satellite cells from "1st generation" ESMs for the generation of "2nd generation" ESMs, and (4) regeneration of CTX-induced skeletal muscle damage in nude rats with impaired endogenous regeneration capacity.
Intramuscular CTX-injection is an established method for the induction of skeletal muscle damage in vivo. 48,49 We adapted this model to induce ESM damage and found that CTX at 25 µg/ml for 24 hours induced marked myotube damage resulting in full contractile failure of ESM. An interesting observation was that this protocol did not destroy the reengineered satellite cell niche in ESM as evidenced by largely unaltered Pax7 + transcript abundance and a reduced, but not eliminated Pax7 + cell count. This data is on the one hand in line with the reported high susceptibility of differentiated muscle cells to CTX 6 and suggests on the other hand that satellite cells entail an intrinsic resistance to CTX. Based on these observations, we hypothesized that the unaffected Pax7-positive satellite cells are activated upon injury to regenerate ESM. We tested this hypothesis by making use of an irradiation protocol typically applied in vivo to deplete/ inactivate the satellite stem cell pool 35,36 and could indeed observe that the endogenous regenerative capacity of ESM in response to CTX-injury was abolished by irradiation, which advances earlier studies by demonstrating that muscle regeneration in vitro is indeed driven by satellite cells reentering the cell cycle. 25 In line with a recent report 38 and findings in other organ models, 50 macrophages may also play a crucial role in skeletal muscle regeneration. Given the muscle origin of the cells used for ESM generation and the presence of macrophages in most, if not all primary cell isolates (refer for example to primary cells isolated from the heart and their use in tissue engineering 30 ), we anticipated and finally confirmed the presence of macrophages in ESM. Future studies will have to address the heterologous cell content dynamics of ESM under homeostatic and injury conditions in more detail to better understand the contribution of specific cell types. The observation that the regenerative response in ESM can be enhanced or decreased by pharmacological modulation is in our view exciting as it provides early proof-of-concept for a utilization in phenotypic drug screens.
We substantiated the role of Pax7 + cells in muscle repair by making use of a novel in vitro model of spatially defined "crush injury". In this series of experiments locally confined activation of Pax7 + muscle progenitors/satellite cells provided evidence for the specificity of the proposed muscle repair mechanism. Collectively, our findings are in agreement F I G U R E 7 Repopulation of the satellite cell niche in vitro and in vivo. A, Myofibers with attached Pax7-positive satellite cells (lower left panel) were isolated from eGFP + ESM ("1st generation ESM"); after enzymatic dispersion and differential straining only mononucleated cells were retrieved (middle panel) and utilized to either generate "2nd generation ESM" or as implant into CTX-damaged skeletal muscle of regeneration incompetent (by γ-irradiation) nude rats. Scale bars: 1 mm in full ESM images, 20 µm in lower left, and 100 µm in middle panel. B, 80 Hz tetanic stimulus induced witch tension (TT) of parental 1st and 2nd generation ESM (12 days old, n = 3/group). C, Immunostaining for Pax7-positive satellite cells in 1st and 2nd generation ESM (Pax7: white, f-actin: green, nuclei: blue). Scale bars: 10 µm. D, Macroscopic view of explanted nude rat TA muscle 33 days after CTX-injury (upper panel) and 30 days after implantation of myogenic cells from eGFP + -ESM (lower panel). Note that both images were taken under similar epifluoresence excitation at 488 nm. Scale bars: 1 mm. Immunostaining of explanted TA muscle 30 days after injection of E, 2D culture-derived or F, ESM-derived myogenic cells; eGFP: green epifluorescence, anti-GFP: white, f-actin: magenta, nuclei: blue; graft area marked by dotted line is denoted by eGFP + -cells Scale bars: 100 µm. G, Quantification of eGFP + donor myofibers per muscle section in injured area (n = 3 [2D], n = 5 [ESM] animals from two experimental series). *P < .05 by Mann-Whitney test. H, Representative images of a eGFP + Pax7 + cell in satellite cell position (Arrow) 30 days after ESM-derived cell injection. eGFP: green epifluorescence, Pax7: white, f-actin: magenta, nuclei: blue. Scale bar: 50 µm with the concept that ESMs resemble bona fide skeletal muscle surrogates, which encompasses the capacity for autonomous structural and functional regeneration via satellite cell activation.
Finally, we were able to re-isolate functional satellite cells from ESM in a similar way as demonstrated for satellite cell isolation from native muscle. 42 These data suggest that ESMs may be further exploited as a in vitro satellite cell expansion/preservation tool for potential therapeutic applications. This assumption is supported by implantation experiments whereby the injection of ESM-derived isolates into injured and by irradiation regeneration-incompetent muscle resulted in considerable remuscularization. As the ESM-derived isolates also contained non-muscle cells we cannot exclude that those supported engraftment of muscle stem cells. The observation that injections of 2D cultured muscle cells, similarly isolated as the cells for the ESM cultures, did not result in any palpable regeneration corroborates that the ESM culture format provided a unique environment for the maintenance of a functional satellite cell niche in vitro.
Notably, the majority of satellite cells were quiescent in ESM under basal conditions. This is remarkable because during isolation of primary satellite cells they are typically activated to proliferate and progress along the default myogenic differentiation pathway. 51 The observation of predominantly quiescent Pax7 + cells in ESM may be explained by the presence of a population of Pax7 + /MyoD + "reserve cells" [52][53][54][55][56][57] in the original ESM reconstitution mixture. Alternatively, there may have been a selection bias for specific stem/progenitor subpopulations with the capacity for symmetric and asymmetric self-renewal in the primary skeletal muscle cell isolate. 48 Further studies utilizing transgenic mouse models (eg Myf5-Cre/ROSA-YFP, NES-GFP/Myf5 nLacZ/+ 48,56,57 ) may be instructive to address whether the engineered satellite cell niche in ESM is indeed populated by a distinct myogenic cell subpopulation.
In vivo, the satellite cell niche is influenced by a variety of factors including growth factors, extracellular matrix (ECM), cell-cell interactions, and likely other as of yet unidentified biochemical and biophysical cues. 4 We aimed to control for the influence of extracellular matrix and growth factors on satellite cell function by parallel culture of the primary skeletal muscle cell isolate under conventional 2D conditions using the same extracellular matrix (ie rat tail collagen type 1 and Matrigel ® ) as adhesive growth substrate and medium as in 3D culture (ESM). The observation that these matrix and medium factors alone did not suffice to preserve Pax7 + cells suggests that other factors may be more relevant. Tissue elasticity was discovered to control stem cell properties in satellite cells. 12 In this study, growth of satellite cells on a substrate with tissue-like elasticity (12 kPa) was sufficient to retain a high degree of the cells regenerative potential.
Interestingly, the elastic modulus of a day 12 ESM is in a similar range (9.5 kPa) and thus it is tempting to speculate that the tissue-like biomechanical properties of ESM contributed to preservation of satellite cell function in our model. However, further dissection of biochemical and mechanical cues as demonstrated by a recent elegant study from Quarta et al is required to fully define and control the satellite cell niche in ESM. 13 We conclude that the ESM technology enables robust engineering of skeletal muscle with functional satellite cell niches in vitro. Importantly, regenerative potential could be documented morphologically, but also by very sensitive means of isometric force measurements. We feel that partial and full recovery of tissue functionality after injury, as observed in the CTX and the novel "crush" injury model, as well as complete loss of regeneration after irradiation provide strong evidence for the "physiological" functionality of the engineered satellite cell niche in ESM. Given these findings we propose that the ESM model may represent a versatile tool to (1) further dissect satellite cell biology, (2) establish skeletal muscle disease models in vitro for drug development, and (3) provide potential therapeutic satellite cells for cellbased skeletal muscle repair. The availability of human pluripotent stem cells and methods for their directed myoblast differentiation 58,59 will facilitate the translation of the reported rat to a human model. In fact, alternative models of engineered human skeletal muscle using the forced expression of Pax7 60 or directed multi-lineage differentiation in induced pluripotent stem cells 61 have recently been reported. These advances will be important for applications in screens for regeneration inducing compounds and studies of cell based muscle repair. A challenge will be to establish the cellular complexity of bona fide skeletal muscle, which may be achieved by a further refinement of multi-lineage stem cell differentiation or the optimization of primary cell isolation from human skeletal muscle.