Hypoxia Increases Mouse Satellite Cell Clone Proliferation Maintaining both In Vitro and In Vivo Heterogeneity and Myogenic Potential

Satellite cells (SCs) are essential for postnatal muscle growth and regeneration, however, their expansion potential in vitro is limited. Recently, hypoxia has been used to enhance proliferative abilities in vitro of various primary cultures. Here, by isolating SCs from single mouse hindlimb skeletal myofibers, we were able to distinguish two subpopulations of clonally cultured SCs (Low Proliferative Clones - LPC - and High Proliferative Clones - HPC), which, as shown in rat skeletal muscle, were present at a fixed proportion. In addition, culturing LPC and HPC at a low level of oxygen we observed a two fold increased proliferation both for LPC and HPC. LPC showed higher myogenic regulatory factor (MRF) expression than HPC, particularly under the hypoxic condition. Notably, a different myogenic potential between LPC and HPC was retained in vivo: green fluorescent protein (GFP)+LPC transplantation in cardiotoxin-injured Tibialis Anterior led to a higher number of new GFP+muscle fibers per transplanted cell than GFP+HPC. Interestingly, the in vivo myogenic potential of a single cell from an LPC is similar if cultured both in normoxia and hypoxia. Therefore, starting from a single satellite cell, hypoxia allows a larger expansion of LPC than normal O2 conditions, obtaining a consistent amount of cells for transplantation, but maintaining their myogenic regeneration potential.


Introduction
Adult skeletal muscle exhibits a considerable capacity to regenerate and completely repair its structure and physiological function quickly after injury. During postnatal development, new myonuclei are supplied by muscle-specific stem cells, known as satellite cells (SCs), which divide to provide myoblasts for muscle growth and/or repair [1,2]. Anatomically identified as mononuclear cells that reside between the sarcolemma and the surrounding basal lamina of individual myofibers [3], these muscle-specific stem cells are normally in an undifferentiated mitotically-quiescent state within the sublaminal niche [1,4], but they are activated in response to stress, induced by growth or trauma. Activated SCs, called myogenic precursor cells, proliferate and differentiate to fuse with each other or with existing multinucleated myofibers [5]. Moreover, they can generate progeny that re-establish the quiescent SCs pool, a self-renewing process essential for sustaining the SC compartment [6].
The stem cell niche is a three dimensional informative structure that physically defines stem cells' localization and directs them to activation, differentiation and self-renewal. The balancing between signals required for stem cell quiescence or activation is the key homeostatic function of the stem cell niche. This dynamic microenvironment spatially and temporally controls stem cells by a mixture of chemokines, cytokines, extracellular matrix molecules and gases. The understanding and mimicking of this complex niche is extremely important for a more thorough stem cell characterisation and culture. Indeed, stem cell cultures usually show limited self-renewal or expansion, suggesting that classical stem cell culture conditions need to be optimized. Reproducing in vitro the stem cell microenvironment can be complex but the benefits would include a better understanding of stem cell biology and expansion for regenerative medicine prospective. Among the different factors influencing the niche, oxygen concentration plays a particularly relevant role in various tissues. Increasing evidence demonstrates that oxygen tension does not represent only a metabolic substrate, but also an important signalling molecule within the stem cell niche, where it regulates cell proliferation and plasticity [7][8][9][10]. Stem cell responses to different oxygen tensions have been studied in several cell types, such as neural stem cells, haematopoietic stem cells and, more recently, human SCs [8][9][10][11][12][13][14]. In these studies, hypoxic conditions produced an enhancement of stem cell proliferation and a modification of the cell fate when induced to differentiate. Physiological oxygen tension in the majority of tissues is limited and, generally, lower than classic culture conditions (20% O 2 ). In particular, in the adult skeletal muscle, oxygen levels in physiological conditions generally vary between 2% and 10%, depending on the analysed muscle [7,15]. Indeed, in vivo hypoxic condition (ischemic muscle damage) produces SC activation, proliferation, differentiation and maturation to muscle fibers, as well as maintaining the local stem cell pool [11,16,17]. Other then oxygen (low or high) stimulus, activation of SCs leading to myogenesis is regulated by transcriptional and epigenetical control of myogenic regulatory factors (MRFs), which comprise of myogenic factor 5 (Myf5) and myoblast determination protein (MyoD) [18][19][20]. While quiescent SCs express the paired/ homeodomain gene Pax7, satellite cell-derived myoblasts firstly proliferate and express both Pax7 and MyoD, then downregulate Pax7 maintaining MyoD expression throughout the myogenic differentiation process. A population of Pax7+MyoD+cells, however, withdraw from the cell cycle, downregulate MyoD and return to quiescence by sustaining Pax7 expression [21], through a selfrenewal process.
In vivo grafting experiments showed the ability of SCs to give rise to both new myofibers and new SCs maintaining the stem cell pool throughout life: for this reason, SCs are considered the natural source for muscle tissue-engineering. However, accounting for only 2-4% of the total myofiber nuclei [22], achievement of sufficient numbers of cells for transplantation is essential for clinical application. Consequently, more attention has been recently directed to improving the expansion of SCs before in vivo application, maintaining their myogenic potential and regenerative capacity. However, in order to obtain the best results, the in vitro culture of SCs should mimic the in vivo microenvironment, providing some stimuli that are present in the specific niche in which SCs normally reside.
Given the sensitivity of SCs to their physiological microenvironment and that they react to in vivo hypoxic muscle conditions with activation, proliferation and differentiation, also maintaining their stem cell reserve pool, we hypothesized that culturing SCs at a low level of oxygen, such as in their native niche, could give insight into cells' heterogeneity and, at the same time, more expansion potential. Here we aimed to compare the effect of hypoxic (2% O 2 ) vs. traditional culture conditions (normoxia -20% O 2 ) on the heterogeneity, clonogenicity, proliferation, marker expression, differentiation and in vivo potential of mouse SC clones from freshly isolated myofibers. As in our previous experiments in rats [23], after 5 days of culture it was possible to distinguish two subpopulations of mouse SC clones: low proliferative clones (LPC) and high proliferative clones (HPC) with distinct proliferation rates. Hypoxia increased the proliferation of all the clones, but maintained the distinction and the proportion between the two subpopulations. LPC also revealed higher amounts of MRFs and CD34 expression. Most importantly, single LPC transplanted in cardiotoxin-injured muscles revealed stronger engraftment and muscle regeneration than HPC. Notably, the Engraftment Index calculated by a single LPC cell was similar as it has been cultured under 2% or 20% of O 2 , proving that hypoxia led to the expansion of SCs, maintaining their intrinsic heterogeneity without loosing their myogenic capacity.

Animals
Three month-old male wild-type C57BL/6J mice and transgenic C57BL/6-(ACTB-EGFP)/J mice, with the expression of the enhanced GFP under the control of chicken beta actin promoter, were used in this study.
Animal care and experimental procedures were approved by the University of Padua's Animal care and Use Committee (CEASA, protocol #41056, 13/01/2010) and were communicated to the Ministry of Health and local authorities in accordance with Italian law.

Single Myofiber Isolation
Single muscle fibers with associated SCs were isolated from hindlimb skeletal muscles after enzymatic digestion with 0.2% (w/ v) type I-collagenase (Sigma-Aldrich), reconstituted in Dulbecco's Modified Eagle Medium (DMEM) low-glucose (GIBCO-Invitrogen) (supplemented with 1% penicillin-streptomycin, GIBCO-Invitrogen) for 2 hours at 37uC as previously described [23]. Briefly, following digestion, muscles were transferred to a 100 mm horse serum-coated plate containing 9 mL of plating medium (1 st dilution), composed of DMEM low-glucose, 10% Horse Serum (HS, GIBCO-Invitrogen), 1% penicillin-streptomycin (P/S), 0.5% chicken embryo extract (CEE, Seralab) and gently stirred with the same medium using a micropipette to release single myofibers. To minimize the contribution of non-myogenic cells, under phase contrast microscope, single fibers were carefully sucked up and serially transfer through two 10 mm HS-coated dishes containing 9 mL of plating medium (2 nd and 3 rd dilution).

Primary Satellite Cell Isolation
Single isolated myofibers were transferred from the 3 rd dilution into a vial containing 1 mL of DMEM low-glucose. They were then triturated 20 times using an 18 G needle mounted onto a 1 mL syringe, to disengage SCs. The resulting cell suspension was filtered through 20 mm and 50 mm cell strainer and centrifuged at 1200 g for 5 minutes. The supernatant was gently removed and cells (named 'stripped' SCs) were suspended in phosphate buffered saline 1X (PBS) for citospun deposition or in a proliferative medium, composed of DMEM low-glucose with 20% Fetal Bovine Serum (FBS, GIBCO-Invitrogen), 10% HS, 1% P/S and 0,5% CEE, for clonal cultures.
Alternatively, for some experiments SCs were isolated from skeletal muscles through a FACS-sorting method [24]. In brief, we carefully removed nerves, blood vessels, tendons and fat tissue from hindlimb muscles under a dissection microscope. Muscles were minced and then digested with dispase (BD), collagenase type I (Sigma) and CaCl 2 (Sigma) for 1 hour. Muscle slurries were filtered through a 70 mm and 100 mm cell strainer. The resulting freshly isolated mononuclear cell population was suspended in PBS containing 1% FBS and stained with anti SM/C-2.6 antibody (kindly provided by Fukada S.) for 30 minutes at 4uC [25]. Cy5-PerCp was used as the secondary antibody. Anti-CD45 PE, anti-CD31 PE and anti-Sca1 PE antibodies were used for lineage depletion. Cell sorting was performed on a FACSAria (BD Biosciences). The resulting sorted cell population was suspended in the proliferative medium for clonal cultures.

The Cytospun of Primary Satellite Cells
Freshly isolated SCs obtained with the stripping method were suspended in PBS and were induced to adhere to glass coverslips (100 cells/coverslip) by centrifugation with cytospin Centrifuge MPW-223-c at 700 rpm for 5 minutes. Cell spots were then airdried and fixed in 4% paraformaldehyde (PFA, Sigma-Aldrich) in PBS to be immunostained.

Primary Satellite Cell Clonal Cultures
For each experiment, stripped SCs or SM/C-2.6+cells suspended in proliferating medium were dispensed into 96-well plates with limiting dilution (0.5 cell/well) (0.1 mL/well). Dishes were split into either 20% or 2% O 2 conditions, both under 37.5uC and 5% CO 2 in a humidified culture incubator. Clones were observed at 5 days and the cells were counted with inverted-microscope analysis.

ATP Quantification
Cell viability quantification was performed using the CellTiter-GloH Luminescent Cell Viability Assay (Promega) based on quantification of the ATP present, following manufacturer's instructions. In brief, after 5 days in culture, clone plates were equilibrated at room temperature for 30 minutes and cell lysis was induced. By the addition of the specific CellTiter-GloH Reagent, a luminescent signal was generated proportional to the amount of ATP. Luminescence was recorded with a luminometer (Packard Instruments).

TUNEL Assay
Cell apoptosis was determined performing TUNEL assay as previously described [26] following manufacturer's instructions. In brief, cells were permeabilized with 0.1% Triton X-100 (Sigma) on ice, washed twice with PBS and then incubated with a mixture of terminal deoxynucleotidyl transferase and fluorescein-labeled nucleotides for 1 h at 37uC. The total number of nuclei was visualized by fluorescence microscopy after incubation of cultures with DAPI for 10 min at room temperature. The apoptotic rate (percent of TUNEL-positive cells) was estimated by counting random fields for each clone.

Immunocytochemistry on Freshly Isolated Satellite Cells
Citospuned stripped SCs were rinsed in PBS and permeabilized with Triton X-100 (Fluka) 0.5% in PBS for 5 minutes. Nonspecific interactions were blocked with 10% HS for 30 minutes. After washing, cells were incubated with primary antibodies 1 hour at 37uC or overnight at 4uC. They were then washed and incubated with labeled secondary antibodies for 1 hour at 37uC. Nuclei were counterstained with fluorescent mounting medium (DAKO) plus 100 ng/mL 49,6-diamidino-2-phenylindole (DAPI). The following primary antibodies were used: mouse anti-mouse Pax7 (

RNA Extraction and Amplification
Total RNA was extracted from single clones using RNeasy Micro Kit (Qiagen) following the manufacturer's instructions. In order to obtain a sufficient amount of cDNA for the RealTime PCR experiments, purified RNA samples were amplified using TransPlex Complete Whole Transcriptome Amplification Kit (Sigma-Aldrich) in accordance with the manufacturer's instructions [27]. Resulting cDNA has been quantified with a Nanodrop ND-2000 Spectrophotometer (Thermo Scientific) and we obtained an average of about 2 mg of cDNA from each clone.

RealTime PCR
Quantitative RealTime PCR reactions have been carried out in triplicate with Lc FastStart Dna Master PLUS SYBR (Roche) in a LightCycler II instrument (Roche), using 2 ng of amplified cDNA and 2 mL of a 3 mM mix of specific forward and reverse primers. Results have been expressed in arbitrary units as the ratio of target gene mRNA content to housekeeping gene mRNA content, further normalized for the number of cells in each considered clone, in order to obtain a ''single cell equivalent'' quantification. b2microglobulin was used as housekeeping gene. Primer sequences are reported in Table S1.

In vivo Transplantation of Single Clones
SC clones were transplanted in vivo in a murine muscle injury model. Tibialis anterior (TA) of C57BL/6J mice were injected locally with 30 mL of cardiotoxin (Naja nigricollis snake venom, Latoxan) 10 mM. Single GFP+LPC and HPC were detached at 5  Single clones were re-suspended in 30 mL of PBS and injected separately 48 hours after injury in TA damaged muscles. Transplanted animals were sacrificed after 4 weeks, TA muscles were collected and fixed in PFA 4% for 1 hour at 4uC. Muscles were dehydrated at 4uC with a sucrose gradient method: sucrose 10% in PBS for 30 minutes, 15% for 30 minutes and 30% overnight. Then muscles were frozen in isopentane cooled in liquid nitrogen, and subsequently processed using a cryostat (Leica) to produce transverse sections for histochemistry and immunostaining.
Histochemistry. Haematoxylin and Eosin staining (Bio-Optica for rapid frozen section) of transplanted muscle was performed. Muscle sections were put in reagent A for 45-60 seconds and rinsed in tap water, then were dipped in Buffer solution and rinsed again in tap water. Slides were put in reagent C for 30 seconds and dipped first in ethanol 95%, then in absolute ethanol and finally in xylene. Sections were mounted in synthetic mounting medium.
Immunohistochemistry. GFP+single clone-transplanted-TA sections were rinsed in PBS and permeabilized with Triton X-100 0.5% in PBS for 10 minutes. Non specific interactions were blocked with 10% HS for 15 minutes at room temperature. After washing, slides were incubated with primary antibody rabbit antimouse Laminin (Sigma-Aldrich, dil. 1:100) for 75 minutes at 37uC. They were then washed and incubated with labeled secondary Alexa Fluor chicken anti-rabbit 594 (Invitrogen, dil. 1:200) for 1 hour at 37uC. After washing, sections were incubated with rabbit anti-Green Fluorescent Protein Alexa Fluor 488 Conjugate (Invitrogen, dil. 1:200) for 1 hour to enhance GFP fluorescence signal. Nuclei were counterstained with fluorescent mounting medium plus 100 ng/mL DAPI.
For immunofluorescence and Engraftment Index (EI) analyses, 10 random fields for each analyzed muscle were acquired with an inverted microscope (Olympus IX71) at 20X magnification. Pictures were analyzed with Photoshop and ImageJ systems to count GFP+and centrally nucleated fibers. Random fields were collected from serial sections of the whole transplanted muscle to obtain wide representative sampling from different (apical, median and caudal) levels and areas. EI has been determined for each transplanted clone, dividing the corresponding percentage of GFP+fibers achieved for the number of GFP+transplanted cells (raw data in Table S2).
DNA extraction and PCR analysis. DNA from transplanted TA was extracted with a DNeasy Blood & Tissue kit (QIAGEN GmbH) and then quantified with a ND-1000 spectrophotometer (Thermo Scientific). Genomic DNA samples extracted from C57BL/6-(ACTB-EGFP)/J and WT mice organs were respectively used as positive and negative controls for the amplification of the GFP gene. PCR reactions for the TERT and GFP genes were carried out as previously described [28], primers sequences are reported in Table S1.

Statistical Analysis
A model based cluster analysis with normal mixture modelling was applied [29] to identify the groups with the number of cells per clone. The number of groups and their classification were estimated with the algorithm mclust using R software. quantification was performed using RealTime PCR. Graphs indicate the amount of MyoD, Myf5 and Hif-1a expression respectively at single cell level (n = 4, mean6SEM, ANOVA analysis determined a significant difference among groups for MyoD and Hif-1a expression; Student t-test: * p,0.05, *** p,0.001). doi:10.1371/journal.pone.0049860.g002 Values are reported as means 6 SEM. Significant differences were determined using one-way ANOVA and Student's t-test when required, with * = p,0.05, ** = p,0.01 and *** = p,0.001.

Hypoxia Increased Clonal SCs Proliferation
Individual myofibers were isolated from three month-old C57BL/6J mice hindlimb skeletal muscles after enzymatic digestion. Fiber-associated SCs were mechanically extracted (stripping) for clonal cultures (Fig. 1A). Each individual myofiber contains a few tens of SCs, and some hundreds of myofibers were collected to obtain several SC clone multiwell plates (a number between 12 and 20, 96-well plates were used for each experiment). Citospun were performed immediately after the stripping passage to investigate the myogenic phenotype of the isolated population (Fig. S1). Almost all the cells expressed Pax7 and a7integrin (94% and 96% respectively), which characterise quiescent SCs. Citospuned cells were also highly positive for CD34 membrane protein (88%). However, stripped SCs showed low percentages of MRFs expression: 13% of cells were positive for MyoD and 32% for Myf5. Finally, contaminant non-myogenic cells were not found in the freshly isolated population, as demonstrated by Sca1, CD31 and CD45 stainings.
Cells were subsequently seeded in proliferative medium and cultured in presence of high (20%-normoxia condition) or low (2% -hypoxia condition) oxygen levels. Initially, clones were observed and followed from day 2 to day 10 after seeding and a kinetic growth curve was made for each clone. The highest number of cell divisions was achieved at day 5 of culture for both oxygen conditions where clones reached the best ratio cells per clone. At day 10, indeed, cell proliferation tended to slow down with the formation of rare myotubes (data not shown). For this reason day 5 was chosen as the time point for further analyses. The number of clones after 5 days of culture was significantly higher in hypoxia condition (Fig. 1B), highlighting improved cell survival and growth at low oxygen. In addition, at this time of culture the number of cells per clone showed a bi-modal distribution. In fact, through a model based cluster analysis with normal mixture modelling of the distribution of cell number per clone, we identified two subpopulations with different proliferation rate as previously reported in rat [23]: Low Proliferative Clones (LPC) and High Proliferative Clones (HPC) (Fig. 1D). LPC and HPC showed i) homogeneous round and small spindle shape and ii) growth by uniform distribution without cluster formation (Fig. 1B) both at normoxia and hypoxia conditions. However, both LPC and HPC grew better at 2% than 20% O 2 and therefore the cut-off value between the two subgroups was always twice in the hypoxia in respect to the normoxia condition ( Fig. 1D and H). Furthermore, the significant difference between the number of cells per clone of LPC and HPC ( Fig. 1E and I) not only confirmed their differences in proliferative abilities, but also highlighted the proliferation enhancement by hypoxia. ATP quantification underlined similar differences, revealing higher ATP levels in HPC than LPC (Fig. 1G). The amount of ATP, which signals the presence of metabolically active cells, is directly proportional to the number of cells present in culture. Moreover, ATP detection, both in LPC and HPC, was also significatively higher in hypoxia in respect to normoxia (Fig. 1M). In theory, the lower number of cells displayed by LPC compared to HPC could be the consequence of both/ either cell death and/or proliferation. However, TUNEL assay performed after 5 days of culture was always negative, both in normoxia and hypoxia conditions for both clones (data not shown). Intriguingly, the two subgroups appeared in each experiment in similar proportions; in particular, about 30% of the total amount of clones were HPC, whereas LPC were more representative (about 70%) ( Fig. 1F and L). This fixed proportion was very similar between the normoxia and hypoxia culture conditions and confirmed the previous observation in rat SCs [23]. Having revealed above that 98% of freshly isolated cells were Pax7+, non-SC contamination would not explain the presence of 30% cells belonging to HPC. Nevertheless, as in control experiments, the clonal subpopulations' appearance was also validated by means of an alternative SC isolation protocol. Hindlimb skeletal muscles were subjected to enzymatic digestion. The resulting cell population FACS depleted for CD45, CD31 and Sca1, sorted for the expression of SM/C-2.6, (Fig. S2A), and cloned, revealed similar morphology and, most importantly, the same distribution in LPC and HPC as in stripped SC cultures ( Fig.  S2B and C). Together, these findings highlighted the higher proliferation profile of HPC than LPC, revealing proliferation heterogeneity within the SC population. These two subgroups were detectable both in normoxia and hypoxia, but with a clear proliferation enhancement in the latter. Thus, low oxygen increased SC clone proliferation maintaining their heterogeneity.

LPC and HPC Differences were Maintained at Different Oxygen Concentrations
At day 5 of culture, the expression of several markers was analyzed using immunocytochemistry to understand whether the two subpopulations showed different expression signatures and how hypoxia could affect it. Alpha7integrin marker was completely maintained in both LPC and HPC (100%) during the culture period, both under normoxia and hypoxia ( Fig. 2A and   Table S2). In both normoxia and hypoxia conditions cells from LPC possess stronger myogenic potential (mean6SEM, p,0.01 with ANOVA, * p,0.05, ** p,0.01 with Student t-test). (B) Cartoon emphasizing on one hand the same myogenic potential at single cell level (dotted line) for LPC and HPC respectively when cultured at different conditions and, on the other (continuous line), the greater muscle regeneration ability of LPC cultured under 2% O 2 . In this condition, LPC proliferation is enhanced and as a consequence the achievable myogenic engraftment is higher. doi:10.1371/journal.pone.0049860.g005 B). A higher proportion of cells expressed MyoD in LPC than in HPC and this difference was maintained at different oxygen concentrations ( Fig. 2C and D). Similar results were observed when MyoD gene expression was evaluated at single cell level using RealTime PCR (Fig. 2M). In accordance, expression of Myf5 was higher in LPC than in HPC, but the marker reached a significant difference only in hypoxia condition, both at protein and gene expression level (Fig. 2E, F and N). Moreover, CD34, a typical quiescence SC marker that rapidly decreases after cell activation, showed significantly higher expression both at 20% and 2% O 2 on LPC (56% and 55% respectively) when compared to HPC (41% and 40% respectively; Fig. 2G and H). Besides, CD31 was not present at all (0%) in all the conditions confirming the absence of cells from other origin (Fig. 2I and L). Evaluation of hif-1a expression at single cell level with RealTime PCR showed a higher expression of the factor in LPC compared to HPC, both under normoxia or hypoxia conditions (Fig. 2O). When cultured on proliferative medium, several desmin+myotubes were observed both in LPC and HPC clones. Intriguingly, low oxygen cultivation triggered myotube generation in all the clones (Fig. 3; ANOVA analysis showed p,0.05 among all the groups; Student t-test showed p,0.05 for both LPC and HPC hypoxia cultured clones compared to both normoxia cultured LPC and HPC). While, conversely, when adipogenic differentiation was considered, rare adipocytes were observed exclusively in HPC expanded in hypoxia and differentiated in adipogenic media in normoxia (Fig. S3).

Hypoxia Culture Condition Induced Expansion of LPC Maintaining Myogenic Engraftment Potential in vivo
To test whether LPC and HPC possessed different muscle regeneration ability, C57BL/6J cardiotoxin damaged mice were transplanted with single GFP+clones cultured for 5 days under normoxia or hypoxia ( Fig. 4A and Table S2). After 30 days the mice were sacrificed and transplanted TA were analysed for engraftment and muscle regeneration. Cell engraftment was confirmed by PCR (Fig. 4H). Single clones were able to generate GFP+fibers ( Fig. 4B and Table S2), and LPC showed the greatest myogenic potential at single cell level. GFP signal was determined through staining with anti-GFP antibody. GFP+and GFP-fibers were counted in each random field and the percentage of GFP+fibers in respect to the total number of fibers was calculated. The percentage of GFP+fibers per transplanted TA (Table S2) was determined as the average of the percentages of GFP+fibers per field (the variance plotted in Fig. 4D and 4E was calculated on the basis of results derived from all the transplanted animals). In particular, GFP+fibers from normoxia cultured single LPC showed a higher trend in forming GFP+fibers than HPC (Fig. 4D), whereas hypoxia cultured LPC treated muscles possessed approximately twice GFP+fibers than the HPC treated (Fig. 4E). A similar trend was maintained when the number of centrally nucleated fibers was evaluated ( Fig. 4F and G). Interestingly, hypoxia cultured clone transplantation led to a higher amount of GFP+centrally nucleated fibers (Fig. 4G right), highlighting a wider transplanted cell involvement in the regeneration process. However, quantifications of GFP+fibers and GFP+centrally nucleated fibers in transplanted muscles showed no statistically significant differences among the groups. This could be ascribed to the fact that fiber percentages do not take into account the number of transplanted cells per TA (Table  S2). Starting from this observation, the Engraftment Index (EI) calculated as the proportion of GFP+fibers per number of GFP+transplanted cells, revealed that LPC cultured in normoxia and hypoxia conditions gave a significantly higher (5 fold) EI in respect to HPC (Fig. 5A and B). Therefore, (i) LPC cultured 5 days at 2% O 2 grew efficiently allowing the transplantation of more cells per clone when compared to LPC normoxia cultured; and (ii) one single cell of LPC cultured both in normoxia and hypoxia possessed the same EI; in our setting, isolation and expansion of LPC at 2% O 2 was the most efficient way to culture mouse SCs (Fig. 5B).
Here, for the first time, we found that the best muscle regenerative potential can be obtained when low proliferative satellite cell clones (LPC) were cultured in hypoxia (2% O 2 ). Although they showed similar morphology, LPC and HPC were present in different proportions, with higher LPC development (,70%) in respect to HPC (,30%). This disproportion was fixed throughout all the experiments and reflected our previous work on rat SC clones [23]. We also excluded the possibility that intrinsic heterogeneity within SCs population was determined by the isolation method. Indeed, clone morphology and bi-modal distribution were confirmed among CD452/CD312/Sca12/ SM/C-2.6+cells. Several studies indicated that SCs constitute a population negative for the hematopoietic marker CD45, the endothelial marker CD31 and the Stem cell antigen 1 (Sca1), through flow cytometric analysis [25,41,55,56]. After this Lineage depletion, different specific markers have been identified to isolate quiescent SCs: a7integrin, b1integrin, CD34, M-cadherin, c-met, CXCR4, caveolin-1, and syndecan-3 and -4 [5,2,25,36,[45][46][47][48][49][50]57]. In particular, SM/C-2.6 antibody permits detection of mononuclear cells residing beneath the basal lamina of myofibers that are positive for Pax7, M-cadherin, CD34 and CTR, negative for MyoD and Ki67, and which are capable of self-renewal, making this antibody suitable for quiescent SCs isolation [25,[58][59][60].
Interestingly, we found that LPC and HPC proportions were maintained even when cells were cultured at different oxygen concentrations. Oxygen may be indeed an important component of stem cell niche: the low oxygen condition, in fact, leads to increased proliferation and maintenance of the naïve state of stem cells [7,10]. Here, the hypoxic conditions, while maintaining cell heterogeneity, displayed higher numbers of cells per clone and ATP quantity in respect to normoxia, as previously observed for other stem cells [7,[11][12][13][14]61]. This has important implication in SC expansion and confirms previous work done on human SCs [7,11]. Moreover, hypoxia influenced phenotype and differentiation capacity of SC clones. Hypoxia cultured HPC had the exclusive ability to differentiate into adipocytes when subsequently cultured in adipogenic conditions and normoxia. It still remains however, to understand if this could be an effect of a low oxygen environment [7] or whether HPC may represent in vivo a source of adipogenic cells within the skeletal muscle in pathological conditions [62]. We have shown that adipocytes can be derived from human SCs [46], while in rat, HPC, but not LPC, have been shown to be able to differentiate spontaneously into adipocytes [23].
It remains a matter of debate whether LPC or HPC may represent in vivo a distinct population with self-renewal characteristics. LPC have indeed shown a higher number of cells positive to Myf5 and CD34 which are commonly associated to quiescent or self-renewing SCs [63][64][65]. In particular, CD34 expression is extinguished shortly after the isolated fibers are placed in culture and become up-regulated only in SCs that revert back to a reserve state [66,67]. Moreover, recently described slow-dividing SC subpopulations [37,47], which show extensive contribution to muscle regeneration in vivo, resemble characteristics of LPC. On the other hand both Myf5 and CD34 are also expressed by HPC while LPC show higher presence of MyoD+cells, which is usually a characteristic of committed myogenic cells. Therefore, whether symmetric and asymmetric divisions, attributed to distinct subsets of SCs [37,46,47,64,68], belongs to LPC, HPC or both, still have to be investigated.
LPC and HPC differed also when expression of hypoxiainducible factor-1 (HIF-1) was investigated. HIF-1 plays an essential role in oxygen homeostasis, regulating several genes involved in angiogenesis, glucose uptake, cell proliferation and survival [69]. At normoxia, this factor is constitutively expressed, but it is essentially regulated by ubiquitinylation and proteolytic degradation [70]. However, under severe hypoxia, the HIF-1a protein level increases by relaxing its degradation, and HIF-1a can form an active complex with HIF-1b.
Since hypoxia increased SC proliferation we determined the HIF-1a gene expression involvement in clones cultured under normoxia and hypoxia. Intriguingly, LPC showed higher HIF-1a expression at a single cell level than HPC, but similar expression levels between low or high oxygen conditions. This is in accordance with other studies, demonstrating that, in contrast to other tissues, skeletal muscle expresses high amounts of HIF-1a in normoxia as well as in hypoxia [71,72]. It has been reported that the expression of HIF-1a synchronized with that of myogenic regulatory genes early during muscle regeneration. Furthermore, important observations from recent studies, described that HIF-1a protein expression decreases concomitantly with increased MyoD and myogenin protein expression after the induction of myogenic differentiation [73] and suggested that HIF-1a expression in vitro is regulated mainly at the protein level [74]. This fact could explain the absence of a statistical difference between normoxia and hypoxia cultured clones in terms of HIF-1a gene expression at 5 days of culture. Conversely, since HIF-1a was less expressed in HPC cells, the involvement of this hypoxia-inducible factor in satellite cell clone proliferation remains unclear and needs further investigation.
Similar to other studies [55,56,75] both mouse HPC and LPC can contribute in vivo to the SC pool when injected after injury. When single GFP+HPC and LPC clones were injected in C57BL/ 6J cardiotoxin damaged mice we observed skeletal muscle regeneration, which was more efficient with LPC cells. Despite the relatively low myotube formation displayed in vitro, single LPC transplantation effectively produced wider regenerating myofibers especially after hypoxic culture, supporting the idea that the population of slow-dividing SCs is capable of robust niche repopulation and more efficient muscle regeneration than fasterdividing population [37,47]. In particular, hypoxia increased both central-and peripheral-nucleated GFP+fibers in respect to normoxic conditions, suggesting that a low oxygen microenvironment could improve both SCs expansion and their in vivo active regeneration potential. We can speculate that culturing SCs in hypoxic conditions could improve the maintenance of a cell which may actively repopulate the in vivo SC niche and participate in new myofiber formation.
Extending these observations to a single cell level, the EI pointed out that hypoxia maintained the LPC single cell engraftment potential when transplanted in vivo. This implies that not only does a regenerative environment has a preferential effect for engraftment of slow-dividing cells [47], but this also confirms that the expansion achieved by low oxygen conditions do not impair the LPC in vivo potential.
In conclusion, our results show that 2% O 2 allows for a better expansion of SCs and maintains their intrinsic myogenic potential for more efficient cell transplantation. This indicates the significant role hypoxia plays as a key component for SCs expansion. In particular, hypoxia conditions promoted LPC growth, SCs heterogeneity maintenance and in vivo myogenic potential preservation. Finally, our results provide insights into the heterogeneity in marker expression within the clonal population, highlighting possible differences in the activation state and self-renewal process. Future studies are required to better understand LPC and HPC properties in muscle homeostasis and regeneration. Figure S1 Freshly isolated SC marker expression characterization. Freshly stripped SCs were characterized for their initial marker expression signature with immunofluorescence after citospun (in red the specific marker merged with DAPI; bar = 100 mm). Graphs indicated positive cell number averages (n = 8, mean6SEM). SCs displayed almost total positivity for Pax7, a7integrin and CD34, whilst they were negative for Sca1, CD31 and CD45. Respectively 13% and 32% of cells were positive for MyoD and Myf5. (TIF) Figure S2 SCs sorted by SM/C-2.6 marker and clonal cultured, compared with stripped SCs. SCs were isolated from hindlimb skeletal muscles after several enzymatic passages. After Lineage depletion (Lin-) for CD45, CD31 and Sca1 (A -left, 35.8% of total cell population was Lin-), SM/C-2.6+cells were sorted (A -right, 17.7% of Lin-cells, 2% of total cell population) and cultured at clonal density. After 5 days of culture, clone morphology appeared similar between SM/C-2.6+and stripped cells (B) and it was possible to distinguish between LPC and HPC in both the clonal cultures (B) (bar = 100 mm). The proportion of LPC versus HPC in SM/C-2.6+cell clones was similar to that obtained in stripped cells cultures (C -proportion between LPC and HPC in SM/C-2.6+clones, mean). (TIF) Figure S3 Adipogenic differentiation of hypoxia cultured clones. Clones cultured for 5 days in proliferative medium and subsequently for 10 days in adipogenic differentiation medium and normoxia. Adipocytes developed in HPC cultured initially in 2% O 2 (right) and not in LPC (left) cultures. Insets with higher magnification of a single adipocyte with lipid vacuoles positive for Oil-Red-O specific staining (red, bar = 75 mm). (TIF)