Characterization and isolation of highly purified porcine satellite cells

Pig is an important food source and an excellent system to model human diseases. Careful characterization of the swine skeletal muscle stem cells (satellite cells) will shed lights on generation of swine skeletal muscle disease model and efficient production of porcine meat for the food industry. Paired box protein 7 (Pax7) is a highly conserved transcription factor shared by satellite cells from various species. However, the sequence of Pax7 has not been characterized in pig. The lack of method to isolate highly purified satellite cells hinders the thorough characterization of the swine satellite cells. Here we found molecular markers for swine satellite cells and revealed that the porcine satellite cells were heterogeneous in various pieces of skeletal muscle. We further developed a method to isolate highly purified satellite cells directly from porcine muscles using fluorescence-activated cell sorting. We next characterized the proliferation and differentiation abilities of isolated satellite cells in vitro; and found that long-term culturing of satellite cells in vitro led to stemness loss.


INTRODUCTION
Satellite cells are a heterogeneous population of adult stem cells located in skeletal muscles. [1][2][3][4] These cells reside between the muscle sarcolemma and the basal lamina of muscle fibers and remain quiescent under normal conditions. 1,2,5,6 Upon injury, quiescent satellite cells are activated to repair muscle injury while a subset of the activated satellite cells return to quiescent status after self-renewal. 1,2,7 Satellite cells are essential for muscle regeneration and mice lacking satellite cells display severe muscle regeneration defects. [8][9][10] Satellite cells impairment occurs in muscular diseases such as Duchenne muscular dystrophy (DMD) 11 and in aged muscles. 12 Pigs have many advantages in modeling human diseases due to their similar anatomic and physiological features to human beings. 13,14 For example, the swine DMD model recapitulates human symptoms better than mouse model. The severe progressive dystrophic changes of skeletal muscles, impaired mobility, muscle weakness, and a much shorter life span are common symptoms in human DMD patients. 15 These symptoms can only be recapitulated in pig DMD model, 16 but not mouse DMD model, 17 supporting the notion that pigs can model human diseases better than rodents. Characterization of satellite cells in pig will be a valuable addition to our understanding of the porcine model system.
Paired box protein 7 (Pax7) has been shown to be the critical regulator of satellite cell maintenance and proliferation in several species. [18][19][20] Pax7 is a highly conserved protein present in many mammalian species such as human, mouse, and cattle. In mouse, Pax7 can directly regulate MyoD and Myf5 to modulate satellite cell maintenance and proliferation. [21][22][23][24] The full-length CDS of Pax7 has been found in mouse (Accession #: NM_011039) and human (Accession #: NM_001135254) database. But there are only three short partial sequences for Pax7 in pig existing in the database currently.
Satellite cells frequency and functions are heterogeneous in both mouse and human skeletal muscles. 4,25,26 Mouse soleus and diaphragm muscles have more satellite cells per mm 3 compared to other limb muscles. 25 Human temporalis muscle has estimated higher satellite cell while other body and limb muscles being analyzed show similar satellite cell numbers. 26 But little is known about the satellite cell frequency heterogeneity in pig muscles.
Highly purified mouse satellite cell population can be obtained by fluorescence-activated cell sorting (FACS) sorting. But the method to purify porcine satellite cells is not well established. Several approaches have been reported to isolate the myogenic lineage cells from porcine muscle tissues, including Percoll gradient centrifugation and preplating. [27][28][29][30][31] However, the cell identity of the myogenic lineage population obtained by the above methods has not been carefully characterized yet. The purification efficiency of Percoll gradient centrifugation has big variation as reported by different groups. 27,32 Furthermore, in cells obtained by Percoll gradient centrifugation, only a portion (o60%) of them stain positive for neural cell adhesion molecule (N-Cam), which is considered to be a porcine satellite cell marker. 28 Consistently, only 65% of the cells obtained by this method are able to fuse and form myotubes. 33 Another widely used approach to isolate porcine myogenic lineage cells is the preplating method. 29,34,35 Pax7 immunofluorescent staining indicates that satellite cells only account 20-50% of the cell population obtained by this method. 29,35 Thus, it is necessary to develop a more efficient method to obtain highly purified satellite cells.
Previous studies in mouse have suggested that highly purified satellite cells can be obtained by FACS sorting. 2,36,37 The cell surface markers for FACS sorting have been well characterized in murine muscle tissues. The most commonly used cell surface markers in murine are CD34 and α7-integrin. 2,36,38 The cell surface markers to specify satellite cells in pig are not well characterized. It has been reported that CD56 was expressed in a myogenic cell population residing in human 26,39 and swine skeletal muscles. 40 Further analysis in human muscles showed that the cells expressing CD56 can be further divided to two populations based on CD34 expression. CD56 + CD34 − cells isolated from skeletal muscles displayed sole myogenic potency; while CD56 + CD34 + cells showed both adipogenic and myogenic potentials, 40 though the scenario in pig has not been well characterized. CD29 (β1-integrin) is one of the integrin family members that can form heterodimers to link the extracellular matrix to actin cytoskeleton within cells. 41 CD29deficient embryonic myoblasts displayed impaired fusion ability in vitro. 42 Satellite cells lacking CD29 fail to maintain quiescent stage and cannot sustain the expansion and self-renewal of this stem cell pool during regeneration. 43 In mouse, more than 95% of satellite cells are CD29 and Pax7 double positive; 44 in human, 96% of satellite cells are CD29 and Pax7 double positive. 26 These observations suggest that CD29 could be a surface marker to specify satellite cells together with other markers.
It has been reported that long-term cultivation of satellite cells in vitro resulted in loss of most of the regenerative potentials in vivo after transplantation in mouse, 36,45 canine, 46 and human. 47 Scattered reports have shown that the progenitor cells derived from swine skeletal muscles display multilineage differentiation capacities after being cultured in vitro for 3 months, 48 but the identity of these cells has not been characterized. Whether the lack of ability to be expanded in vitro is also true in pig satellite cells is not known yet.
Here we characterize the satellite cell frequency in variant pieces of porcine skeletal muscles. The cell surface markers specific for porcine satellite cells have been identified, and a FACS sorting method to isolate porcine satellite cells based on the surface markers has been established. We further demonstrate that similar to satellite cells in other species, porcine satellite cells lose their stem cell identities and differentiation potentials after long-term in vitro culturing. Shadow indicated the conserved domains. Mouse (Mus musculus, Ms), cattle (Bos taurus, Bt), chick (Gallus gallus, Gg) Pax7 proteins, and partial pig (Sus scrofa, Ss) Pax7 peptide sequences were aligned with human (Homo sapiens, Hs) Pax7 protein. Black indicated the partial porcine Pax7 cDNA sequences. (b) The nucleotide sequence of pig Pax7 and the predicted protein sequence. The conserved PAX, Homeobox, and Pax7 were boxed.

RESULTS
Clone of the full-length porcine Pax7 gene Pax7 is an important regulator of muscle satellite cells. [18][19][20] We set out to compare the similarity of Pax7 peptide sequence between pig and other species. However, we could not find the full-length Pax7 sequence in the NCBI database. There are only several partial porcine Pax7 cDNA sequences in NCBI nucleotide database (Accession #: AY653213, XM_013992407, and XM_013991732).
These translated peptide sequences were aligned with Pax7 peptide sequences from several species (Figure 1a). Peptide sequence (Accession #: AAT72072) translated from AY653213 displayed 100% identity to Pax7 peptide sequences from multiple species (Figure 1a and Supplementary Figure S1). Pax3 belongs to the same family as Pax7 and share high similarity to Pax7. Sequence comparison indicated differences on two key amino acids between peptide sequence derived from AY653213 and Pax3 amino-acid sequence from various species (Supplementary Figure S1), suggesting that AY653213 could be a partial sequence from porcine Pax7.
To identify the complete cDNA sequence of Pax7 in pig, we performed 5′ and 3′ rapid amplification of cDNA ends (RACE) using AY653213 as the seed sequence with total RNA extracted from porcine myoblasts (Supplementary Figure S2a). A 400 bp fragment was generated by the 5′ RACE; and a 5000 bp fragment was generated by the 3′ RACE (Supplementary Figure S2b). The two fragments could be assembled into a 1512 bp CDS (Accession #: KX815346), which encoded a 503 amino-acid polypeptide ( Figure 1b). The assembled nucleic acid sequence showed 93% identity to human Pax7 and 90% identity to mouse Pax7 (Supplementary Figure S3a). The translated polypeptide displayed 99% identity to human Pax7 and 98% identity to mouse Pax7 protein (Supplementary Figure S3b). Paired box domain (PAX), homeobox domain (Homeobox), and Pax7 domain, the three highly conserved domains of Pax7 from various species, were all found in the amino-acid sequence translated from the full-length pig Pax7 CDS (Figure 1b).  Significance was analyzed by student's t-test. * indicated P o0.05, ** indicated P o0.01. *** indicated P o0.001. NS indicated no significant difference.
The satellite cell number varied at different age and location We tested several antibodies and found the antibody recognizing porcine Pax7 protein. This antibody was utilized to detect the endogenous porcine satellite cells. Consistent with observations in mouse 18 and human, 26 Pax7-positive satellite cells were detected underneath the basal membrane as indicated by Laminin immunofluorescent staining in both postnatal and adult pigs (Figures 2a and b). The satellite cell number is important for skeletal muscle homeostasis maintenance. It has been reported that the satellite cell number decreased in adult when compared to neonatal mice. 49 The number of satellite cells also differed between various pieces of skeletal muscles in the same individual. 25 We next checked the satellite cell number in semitendinosus from 1-and 25-week-old pig by Pax7 immunofluorescent staining. The number of satellite cells decreased in adult skeletal muscle (25 week old) compared to neonatal (1 week old) muscles in pig (Figures 2b and c). To compare the satellite cell numbers in different muscle pieces, we analyzed nine types of adult pig skeletal muscle from different locations. Peroneus tertius, psoas major, and extensor carpi radialis had high number of satellite cells; while intercopstal muscle and biceps femoris had low amount of satellite cells (Figures 2d and e). These results suggest that the distribution of satellite cells in different type of skeletal muscle is different.
Identification of the satellite cell-specific cell surface markers In order to set up FACS sorting method to isolate porcine satellite cells, we next characterized the specificity of the surface marker for porcine satellite cells. CD56 and CD29 have been reported to mark satellite cells in human. We performed immunofluorescent staining assays to verify whether these surface markers were also applicable to porcine satellite cells. Both CD56 and CD29 colocalized in the same cell with Pax7 in porcine skeletal muscles (Figure 3a), suggesting that they could be used as cell surface marker for isolation of satellite cells from pig skeletal muscles. Antibodies against CD31 (an endothelial cell marker) 50 and CD45 (a hematopoietic cell marker) 48 were utilized to distinguish the endothelial and hematopoietic cells present in the mononuclear cells isolated from porcine skeletal muscles. Based on these results, we isolated CD31 − CD45 − CD56 + CD29 + cells from mononuclear cell population obtained from porcine skeletal muscles by FACS sorting (Figure 3b). The Pax7 and Myf5 mRNA level in this cell population was compared with that in CD31 + CD45 + and CD31 − CD45 − CD56 − CD29 + populations by RT-qPCR.  Figure S4b), suggesting that the specific enrichment of Pax7 + cells in CD31 − CD45 − CD56 + CD29 + cell population isolated from porcine skeletal muscles. Taken together, satellite cells can be isolated from pig skeletal muscles by FACS sorting using CD31 and CD45 negative selection followed by CD56 and CD29 positive selection.
To analyze the proliferation abilities of porcine satellite cells in vitro, we cultured the satellite cells for several days in vitro. After being cultured in vitro for 96 h, the cell size started to increase and the cell shape became more spindle-like (Figures 4a-c and Supplementary Figure S5a). Previous reports have shown that the cell size increased when satellite cells being activated and differentiating to myoblasts. 51 To further characterize the identity of the cells isolated from pig skeletal muscles by FACS sorting, we performed immunofluorescent staining of MyoD and Pax7. In the round satellite cells which were cultured in vitro for less than 6 h, Pax7 protein was detected. MyoD protein level was relatively low (Figure 4c). While in cells being cultured for 96 h with bigger cell size, Pax7 was also detected (Figure 4c). MyoD protein displayed punctuated distribution and higher level in nuclei (Figure 4c). The Pax7 mRNA level was decreased during in vitro culturing (Figure 4d). These results suggest that consistent with the circumstance in mouse, 5,51 the porcine satellite cells being cultured in vitro for over 96 h are activated and started to differentiate to myoblasts.
We next assayed the differentiation potentials of the porcine satellite cells in vitro. Porcine satellite cells were differentiated in either 2% horse serum (HS) or 0.4% Ultroser G (UG) (Figures 4e  and f). The myosin heavy chain expression level was dramatically higher in UG-induced differentiation (Figure 4f). The differentiation efficiency was also enhanced by UG induction (Figure 4f).
We next examined whether cell density would affect pig satellite cell differentiation. We seeded 2 × 10 4 , 1 × 10 5 , and 5 × 10 5 pig satellite cells in 3.5 cm dish for 2 days (Supplementary Figure S5b) and induced differentiation with UG, respectively. When the density of the seeded pig satellite cells was low, the differentiation efficiency decreased; when the cell density reached the threshold, the differentiation efficiency plateaued (Figures 4g  and h). Higher seeding cell density yielded myotubes with elevated MyHC expression level and higher percentage of myotube nuclei (≥3 nuclei) in total nuclei (Figures 4i and j). Together, these results suggest that high seed cell density facilitates pig satellite cell differentiation.
We further examined the contractility of the differentiated myotubes. After being cultured in medium containing UG and Ca 2+ for 6 days, myotubes differentiated from swine satellite cells were able to contract (video was included in Supplementary File S7). These results suggest that swine satellite cells were able to generate myotubes with contractility.
Porcine satellite cells lost their stemness after long-term in vitro expansion Both mouse and human satellite cells have been reported to suffer the loss of abilities to repair muscle injury in vivo after days of in vitro culturing. 5,36,45,47,52 We therefore examined whether porcine satellite cells share the same feature. As described above, the cell size of porcine satellite cells increased and the cell shape became more spindle-like after 96 h of in vitro culturing (Figure 4a), suggesting these cells to be in a more differentiated status. The morphology of porcine satellite cells remained to be spindle-like after serial expansion in vitro (Figure 5a). These cells could be propagated logarithmly in vitro (Figure 5b). To further characterize the identity of the cells being serially expanded in vitro, CD56 and CD29 expression was analyzed by FACS. The expression level of both CD56 and CD29 decreased after passage 4 (Figure 5c), suggesting the gradual loss of satellite cell characters. Consistently, the percentage of Pax7 + cells also decreased dramatically after passage 4 and 5 (Figure 4d). The expression level of Pax7 echoed the morphology changes observed above. After 96 h of in vitro culturing, the Pax7 mRNA level dropped over 30 fold (Figure 5e). Serial expansion in vitro of these cells further drove the Pax7 mRNA level and the percentage of Pax7 + cells down (Figures 5d and e). After passage 4 and 5, the Pax7 expression level decreased to very low level (Figure 5e), suggesting that these cells have lost their satellite cell features. Consistently, the cells expanded in vitro for over 5 passages displayed significant lower differentiation potentials (Figures 5f  and g). Taken together, these results suggest that pig satellite cells quickly lose their stem cell features in vitro.
We further examined the ability of pig satellite cells to repair muscle injury in vivo by cell transplantation. Satellite cells were isolated from 1-week-old piglets by FACS sorting. TA muscle injury in immunodeficient recipient mice (NOD.Cg-Prkdc scid Il2rg tm1Wjl / SzJ) was induced by cardiotoxin injection. Various numbers of pig satellite cells were transplanted to the recipient mice. Four weeks after transplantation, TA muscles from recipient mice were collected to make cryo-sections and stained with antibody specifically recognizing pig Lamin A/C. Pig Lamin A/C can be detected in TA muscles of the recipient mice (Figure 6a and Supplementary Figure S6a). One week after transplantation, the porcine Lamin A/C could also be detected in newly formed myofibers (indicated by Myh3 staining; Figure 6b), suggesting that the transplanted pig satellite cells are able to engraft and contribute to muscle regeneration in recipient mice. We further examined whether the transplanted swine satellite cells were capable of homing to the right niche. Pax7-expressing satellite cells were detected in TA muscles of the recipient mice 4 weeks after transplantation (Supplementary Figure S6b). The Pax7expressing satellite cells of pig origin (indicated by Lamin A/C staining) were detected at the similar location where the endogenous satellite cells resided (Supplementary Figure S6c). These results suggest that the transplanted swine satellite cells were capable of homing to the niche. We next investigated whether in vitro expansion of porcine satellite cells affects their engraftment efficiency. The same number of freshly isolated pig satellite cells and porcine satellite cells expanded in vitro for different period were transplanted to immunodeficient recipient mice, respectively. The antibody specifically recognizing porcine Lamin A/C was used to perform immunofluorescent staining with muscle tissue sections 4 weeks after transplantation. The pig satellite cells expanded in vitro for 2 passages displayed similar engraftment efficiency with the freshly isolated pig satellite cells. However, few myofibers generated from pig satellite cells could be detected when these cells were expanded in vitro for 5 passages (Figures 6c and d), suggesting that pig satellite cells lost their stemness and abilities to repair muscle injury in vivo after prolonged in vitro expansion. These results are also consistent with the morphology and differentiation potential changes described above (Figures 4 and 5), suggesting that pig satellite cells lost their stem cell features after being expanded in vitro for 5 passages. Taken together, these results suggest that pig satellite cells lost their stemness and differentiation abilities after in vitro expansion, a feature shared by both mouse and human satellite cells.

DISCUSSION
Here we characterized pig muscle satellite cells and provided basic information for further utilization of these cells. We cloned the fulllength cDNA sequence of porcine Pax7 and characterized the numbers of satellite cells located at various piece of skeletal muscles. A FACS-sorting-based method was established to isolate highly purified pig satellite cells. Using this method, we isolated pig satellite cells and found that they lost their stem cell morphology, differentiation potential, and ability to repair muscle injury in vivo after prolonged in vitro culturing.
The neonatal mice have high numbers of satellite cells, while the number decreased in the adult (reviewed by Fu et al.). 53 Similar phenomenon has been observed in pig. The number of satellite cells was much higher in 1-week-old pig compared to that in 25-week-old pig. The number of satellite cells varied at different locations. In mouse, extraocular and diaphragm have high number of satellite cells, Gastrocnemius and TA have low number of satellite cells. 25 In human muscles, it is indicated that under normal homeostatic conditions, most adult human skeletal muscle contains a relatively homogeneous frequency of Pax7 satellite cells per fiber. 26 Both pig and human have larger fibers than mouse, but the relative homogeneous frequency of satellite cells in pig fibers still needs more studies.
The loss of abilities to repair muscle injury in vivo of satellite cells after being cultured in vitro (empty amplification) is a common problem shared by mouse and human satellite cells. 5,36,47,52 This problem hampers the application of satellite cell in clinic. Our results revealed that pig satellite cells have the same problem. It suggests that the factors required for stemness maintenance are lacking in the in vitro culturing system. Recently, Fu et al. found that cytokines secreted by activated T cells during the acute inflammation after muscle injury can help maintain the stemness of mouse satellite cells in vitro for over 2 months. 36 Charville et al. 47 showed that inhibition of p38 signaling is important to maintain the quiescent stage and stemness of human satellite cells. Whether these factors also work to maintain stemness of pig satellite cells is an interesting question to pursue.

Animals and pig muscle tissues
All animal procedures were performed according to the institutional and national guidelines and approved by the Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences in this study. Pig muscle tissues were collected from commercial pig 'large white' at the time of killing (asphyxiated by CO 2 of 1 week old, or stunned electrically and then killed by cutting the jugular vein of 25 week old). The NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ (NSG) mice were purchased from Jackson Laboratories (Bar Harbor, ME, USA).

Pig satellite cells isolation
Pig satellite cells were isolated from 1-or 25-week-old male pigs as previously described and adapted to pig tissues. 26,47 Briefly, freshly collected pig muscle was either immediately digested or kept in DMEM (Gibco, Grand Island, NY, USA) at 4°C. The tissues were dissected and dissociated with collagenase D (Roche 2 mg/ml, Indianapolis, IN, USA) and dispase II (Roche 1.07 U/ml) in DMEM supplement with 1% penicillinstreptomycin (P.S.) at 37°C for 1.5 h. The mixture was triturated with pipette once per 15-20 min. After digestion, the muscle was aspirated and ejected in and out of the syringe 10 times with a 20-ml syringe and a 20gauge needle. After centrifuging at 100 × g for 5 min, the supernatant was collected and centrifuged at 1000 × g for 5 min at 4°C. The cells were washed with 20% FBS in DMEM and filtered through a 100 μm cell strainer followed by a 40 μm cell strainer. The cells were then centrifuged at 1000 × g for 5 min at 4°C and incubated with the erythrocyte lysis buffer (ACK) buffer for 5 min on ice. Then the cells were washed with PBS twice and cell pellet was reconstituted with FACS buffer (1% BSA in PBS) or frozen in FBS supplement with 10% dimethyl sulfoxide (DMSO).
Satellite cell culture and differentiation Dishes (Corning, Corning, NY, USA) were coated with 0.05% collagen type I (Corning). FACS isolated pig satellite cells were cultured on collagencoated dishes in F10 medium (F10 medium (Gibco) containing 15% FBS, 5 ng/ml FGF (Invitrogen, Carlsbad, CA, USA) and 1% P.S.). Freshly isolated satellite cells were collected and plated in collagen-coated dish and cultured for less than 6 h to attach. Cells cultured at various hours were fixed with 4% paraformaldehyde (PFA) for immunofluorescent staining. For serial expansion, cells were serially passaged to maintain a density of o60% confluence and counted at each passage. Pig satellite cells differentiation were induced with DMEM (Invitrogen) with 2% horse serum (Hyclone, Logan, UT, USA) or 0.4% Ultroser G (Pall Corporation 15950-017; East Hills, NY, USA). Differentiation was induced for 4 days. The expanded and differentiated cells were fixed with 4% PFA for immunofluorescent staining.
To observe the myotube contraction, pig satellite cells were differentiated in DMEM containing 0.4% Ultroser G (Pall Corporation 15950-017) for 6 days, medium was changed at day 4.
For preplating method, isolated mononuclear cells were firstly cultured in F10 basal medium for 2 days. Cells were then treated with 0.25% trypsin for 5 min and stopped with F10 basal medium, centrifuged and re-plated. To separate satellite cells from fibroblast cells, the cells were incubated in an uncoated dish at 37°C for 60 min and then transferred to a new collagen-coated dish. This preplate technical was repeated three times to remove residual fibroblast cells. The bright field images were acquired by Olympus DP73 microscope.

Immunofluorescent analysis of cultured cells
Cultured satellite cells were fixed with ice-cold 4% PFA (in PBS) for 20 min, rinsed with PBS, and permeabilized in 0.5% Triton X-100 (in PBS) for 8 min.

Pig muscle analysis
The pig muscle tissues were cut into small pieces and then frozen in liquid nitrogen. Cross sections were cut at 8 μm and processed for immunofluorescent staining. The cross sections were fixed with 4% PFA at room temperature for 10 min, washed in PBS. The sections were permeabilized with 0.5% Triton X-100 for 6 min and washed with PBS again. Sections were then blocked for 1 h with 10% goat serum at room temperature. For Pax7 and Laminin co-staining, sections were incubated at 4°C overnight with the following primary antibodies: anti-Laminin antibody (1 : 50, Sigma, St Louis, MO, USA; Cat# L9393) or rat anti-Laminin (1 : 1000, Abcam, Cambridge, MA, USA; Cat# 11576), mouse anti-Pax7 (1 : 25, Developmental Studies Hybridoma Bank, Cat# PAX7). After PBS wash for 5 min three times, the sections were stained with Alexa 488-, 561-or 647-labeled anti-mouse or -rabbit or -rat antibodies (Invitrogen) and then mounted with VECTASHIELD mounting medium with DAPI (Vector Laboratories, Cat# H-1200).
For sections co-stained with Pax7, Laminin, and CD56/CD29, Pax7 and Laminin co-staining were first performed. After incubation with the secondary antibody above, sections were washed with PBS for 5 min three times and incubated with the Alexa Fluor 488 anti-human CD29 (1 : 25, BioLegend, Cat# 303016), APC anti-human CD56 (1 : 25, BioLegend, Cat# 304610) for 2-4 h at room temperature. After washing with PBS for three times, sections were mounted with VECTASHIELD mounting medium with DAPI (Vector Laboratories, Cat# H-1200). All images were acquired by Leica SP8 confocal microscope and processed with Adobe Photoshop CS5 to adjust brightness and contrast for publication.

Pig satellite cells transplantation
NSG mice were pretreated with 18 gamma (Gy) 1 day before transplantation limited to the hindlimb. Muscle injury was induced by the injection of CTX (Sigma) to TA muscle. Briefly, 15 μl of 10 μM CTX was injected at each TA muscle using 28 gauge needles. After 24 h, freshly isolated and cultured pig satellite cells (treated by trypsin) at passage 2 and 5 were counted, centrifuged, and resuspended in 15 μl PBS with 10 5 cells. Cells were injected into the TA muscle using 28 gauge needles. Recipient mice were treated with daily drinking water supplement with neomycin sulfate (0.25 mg/mg) for 2 weeks. The transplanted TA muscles were collected 1 week, 4 weeks after transplantation in liquid nitrogen. Serial 10 μm transverse sections of the TA muscle were stored in − 80°C fridge.
All glass slides were removed from − 80°C and warmed at room temperature for 10 min. Sections were rehydrated with PBS and fixed in 4% PFA for 10 min at room temperature. After washed with PBS for 5 min three times, the sections were permeabilized with 0.5% Triton X-100 for 6 min and again washed with PBS. Then sections were blocked for 1 h in 10% goat serum (in PBS) with mouse-on-mouse blocking reagent (Vector Laboratories). After blocking, sections were washed with PBS and incubated overnight with primary antibodies at 4°C. Next day, sections were then washed with PBS for 5 min three times and incubated with secondary antibodies 1 h at room temperature. Finally, sections were washed with PBS and mounted with VECTASHIELD mounting medium with DAPI (H-1200 Vector Laboratories). Antibodies used in the staining were mouse anti-Pax7 (1 : 20, Developmental Studies Hybridoma Bank, Cat# PAX7), mouse anti-Myh3 (1 : 100, Developmental Studies Hybridoma Bank), mouse anti-Lamin A/C (1 : 50, Santa Cruz Biotechnology, Santa Cruz, CA, USA; Cat# sc-7292), rabbit anti-Dystrophin antibody (1 : 50, Abcam, Cat# ab15277). The sections were next stained with Alexa 488-, 561-labeled antimouse or -rabbit antibodies (Invitrogen). All images were acquired by Leica SP8 confocal microscope and processed with Adobe Photoshop CS5 to adjust brightness and contrast for publication.
Gene expression analysis RNA was extracted from cells using the RNAprep pure cell/bacteria kit including DNase treatment (TIANGEN, Beijing, China) according to the manufacturer's instruction. One microgram of total RNA from each sample was reverse transcribed to cDNA using M-MLV transcriptase (Promega, Madison, WI, USA) according to the manufacturer's instruction. Relative gene expression was determined and performed in triplicate using a SYBR Green PCR master mix on an ABI 7500 fast real-time PCR system.
Cell size, Pax7 + cells population, differentiation, and engraftment efficiency measurement The cell sizes were measured by ImageJ software (NIH). Bright field images of satellite cells cultured at different time points were acquired using Olympus DP73 microscope. Immediately after isolation, freshly isolated satellite cells were first plated and allowed to adhere for less than 6 h at 37°C.
Pax7 + cells population was expressed as the number of nuclei co-stained with DAPI and Pax7 divided by the total number of nuclei in the same field. More than 400 nuclei from five randomly chosen fields were analyzed.
MyHC + cells population was expressed as the number of nuclei costained with MyHC + and DAPI divided by the total number of nuclei in the same field. More than 3000 nuclei from four randomly chosen fields were analyzed.
The percentage of myotube nuclei in total nuclei was expressed as the nuclei inside myotube (at least three nuclei) divided by total nuclei in the same field. More than 3000 nuclei from four randomly chosen fields were analyzed.

Statistics
Statistical analyses were performed using GraphPad Prism 5 (GraphPad Software). For comparisons of two treatment groups, a student's t-test was used. Results were means ± S.E.M. unless stated. Po0.05 was considered significant.