Results

Regenerative responses of adult muscle stem cells and hESCs are dominantly inhibited by the aged systemic milieu
Previous work established that the upregulation of repair-specific molecular signaling mechanisms, such as Notch, and successful engagement of resident muscle stem cells in tissue repair are largely determined by the age of the systemic milieu, rather than by the cell-autonomous age of muscle cells, or by the differences in their numbers (Conboy & Rando, 2005; Conboy et al., 2005). Intriguingly, these experiments also hinted at a small but persistent inhibitory effect of the aged systemic milieu on the performance of young stem cells. Exploring this further, we found that young serum permits satellite cells to be myogenic, while old serum inhibits the satellite cell regenerative potential not only alone, but also when mixed with young serum, suggesting a dominant over-riding of ‘young’ serum factors (Fig. 1). Myofiber cultures, in which satellite cells have been activated by injury in vivo, were established from young (2–3 months) and old (22–24 months) C57-BL/6 male mice, as previously described (Conboy & Rando, 2002; Conboy et al., 2005). As previously shown, this method is well suited for the assessment of satellite cell regenerative myogenic capacity (Conboy & Rando, 2002; Wagers & Conboy, 2005). Isolated myofiber explants with associated satellite cells were cultured overnight in the presence of young or old serum (alone at 5% and 10%, and mixed at 5% young + 5% old); bromodeoxyuridine (BrdU) was added for the last 2 h of culture to measure the rate of cell proliferation. The effects of heterochronic systemic milieu on myogenic potential were examined as generation of proliferating myoblasts that express desmin and Myf5, and that spontaneously form multinucleated nascent myotubes. As shown in Fig. 1A and quantified in Fig. 1B, the age of sera clearly determined satellite cell regenerative potential and old serum strongly inhibited the myogenic potential of young satellite cells either when present alone, or when mixed with young sera. Similar data was obtained by using another myogenic marker, Pax7 (Supplementary Fig. S1). Additionally, there were two to three times fewer total cells generated in the presence of aged serum (not shown).
Fig. 1  The age of sera determined the regenerative potential of satellite cells. (A) Young satellite cells were cultured either in 5% or 10% young (Young), 10% old (Old), or in a 5%+ 5% mouse sera combination (young + old). Cells were analyzed by immunofluorescence microscopy, using anti-BrdU (red), antidesmin (green) or anti-Myf5 antibodies (green, small panels). Similar results are shown for Pax7 immunodetection (Supplementary Fig. S1). Hoechst (blue) labeled nuclei. (B) Three independent experiments were quantified [300 young myofibers per experiment] as percentage of desmin+/Myf5+/BrdU+ de novo generated cells for each age and culture condition. On average, two to three fewer cells were generated when cultured in the presence of old. Shown are identical microscope fields at ×40 magnification. At least three independent experiments produced similar results. (*) indicates P≤ 0.001 as compared to young sera.   Importantly, it was not simply the dilution of young serum factors that resulted in diminished myogenic capacity when young and old sera were mixed, because young sera promoted robust myogenesis both at 10% and 5%. Thus, old serum factors dominantly inhibited the myogenic capacity of young satellite cells even in the presence of young serum. This observation suggests that satellite cells of young mice engage in efficient myogenic responses, in part, because the inhibitory influence of old circulatory milieu is absent.
These data reveal that the regenerative potential of young muscle stem cells is determined by the age of the systemic milieu, prompting us to investigate whether hESCs would similarly succumb to inhibitory factors present in the aged circulation.
To determine the effects of aged serum on stem cell self-renewal/pluripotency, we analyzed hESC expression of Oct4 and studied the rate of hESC proliferation, by assessing BrdU incorporation (Fig. 2) and Ki67 expression (Supplementary Fig. S2). Specifically, these determinants of hESC regenerative potential were examined in the presence of heterochronic (young vs. old) mouse sera added to typical hESC medium, e.g., MEF-conditioned medium (MCM). Oct4 is expressed by self-renewing, pluripotent ESCs in culture, by the totipotent inner cell mass of the blastocyst and by the germ cells (Nichols et al., 1998; Pesce et al., 1999). Most cells in control cultures or young conditions expressed high levels of this marker of ‘stemness’, and maintained their normal phenotype and morphology throughout the various co-culture experiments performed in this study (see below).
Fig. 2  The regenerative potential of embryonic stem cells was negatively affected by aged mouse sera. (A) hESCs were cultured in MCM with 10% young (young) or old (old) mouse serum, or in three control media: MCM without mouse sera; GM (myoblast medium of Ham's F10 with 20% FBS) and DMEM/FBS (hESC differentiation medium of DMEM with 10% FBS). BrdU was added for the last 2 h of culture to measure the rate of cell proliferation. Immunodetection assays were performed for BrdU (red), Oct4 (red), and Ki67 (Supplementary Fig. S2). Hoechst (blue) labels nuclei. A high rate of hESC proliferation and Oct4 expression is displayed in all control media and in the presence of young mouse serum. In contrast, hESC proliferation and Oct4 expression are inhibited in the presence of old mouse serum, either alone or when mixed with young serum. MCM with mouse sera at 5% gave results similar to those observed with 10% young mouse sera or in control media (Supplementary Fig. S3). (B) Three independent experiments yielded similar results and were quantified as percentage of BrdU+ and Oct4+ cells for each culture condition. * indicates P < 0.001 as compared to young serum.   Importantly, at 10% aged serum dramatically inhibited the self-renewal and proliferative potential of hESCs, as judged by highly diminished Oct4 expression and a lack of BrdU incorporation. Again, the inhibitory factors in the aged milieu were dominant over the young, as evidenced by a decline in Oct4 expression, the low rate of BrdU incorporation, and Ki67 expression in young and old mixed environments (5% young + 5% old sera in MCM). Similar to the data shown for adult stem cells (ASCs) (Fig. 1), it was not simply a dilution of young serum factors as hESCs robustly proliferated and expressed high levels of Oct4 when cultured with 5% young sera in MCM (Supplementary Fig. S3). Quantification of multiple independent experiments has demonstrated that hESC expression of Oct4 and BrdU incorporation have been reduced by two- to threefold in the aged milieu (Fig. 2B).
As expected, hESCs cultured in control media, including MCM alone that does not contain either young or old serum, also displayed a high rate of proliferation and Oct4 expression (Fig. 2, control medium). Additionally, in this experimental set-up there was no general inhibitory effect of sera per se on hESC proliferation and Oct4 expression, as 10% young mouse sera (young) and 10–20% of FBS (growth medium and DMEM/FBS) allowed for a high rate of cell proliferation and for uniformly high Oct4 levels (Fig. 2).
When instead of immediate exposure to aged mouse serum, hESCs were first cultured overnight in MCM, these cells were no longer susceptible to the negative effects of old systemic milieu (Fig. 3), suggesting that hESC-produced factors established an embryonic microniche that may provide temporary protection from the aged environment. It appears that satellite cells do not have such anti-aging ability, because despite an initial activation in entirely young environments, e.g., after muscle injury to young muscle, isolated satellite cells remain susceptible to inhibition by the old mouse serum (Figs 1 and 4C). Similarly, culturing satellite cells isolated from noninjured muscle in growth-promoting medium for 1–2 days does not protect against the inhibitory affects of aged systemic milieu (not shown).
Fig. 3  Embryonic stem cells produce youthful microniche in culture. (A) As opposed to immediate exposure to old mouse serum after passaging (10% old), preculturing of hESCs for 24 h in feeder-free conditions, e.g., Matrigel™ + MCM, prior to replacing MCM with MCM + 10% old mouse sera, resulted in continuously high BrdU incorporation and Oct4 expression (embryonic microniche + 10% old). BrdU was added for the last 2 h of culture to measure the rate of cell proliferation. Immunodetection of BrdU and Oct4 (both in red) was performed as described in Experimental procedures. Hoechst (blue) labels nuclei. (B) Three independent experiments yielded similar results and were quantified as percentage of BrdU+/Oct4+ for each condition. * indicates P < 0.001 as compared to ‘old + MCM’.
Fig. 4  Aged muscle niche inhibits the regenerative potential of hESCs and satellite cells. (A) Immunodetection of a mouse-specific M-cadherin (green) or desmin (red; both human and mouse proteins are detected) revealed that hESCs underwent muscle lineage differentiation when co-cultured with young, but not old myofibers. The myogenic progeny of hESCs appears M-cadherin−/desmin+ (white arrow in young), as opposed to M-cadherin−/desmin− hESCs that lack myogenic commitment (white arrow in old). M-cadherin+/desmin+ cells are the myogenic progeny of mouse satellite cells (yellow arrows). To assess the effects of secreted factors produced by young vs. old myofibers on the rate of hESC proliferation, transient, 2 h BrdU incorporation was examined in hESCs cultured for 48 h with supernatants produced by heterochronic myofiber explants (See Experimental procedures for details). As compared to young myofiber-derived supernatants (young myofiber supernant), exposure to old myofiber-derived supernatants (old myofiber supernant) inhibited hESCs proliferation, as judged by BrdU immunodetection (red). As expected, the rate of hESCs proliferation was high in control media (shown in Fig. 2). Hoechst (blue) labels nuclei in all experiments. Quantification of desmin+/BrdU+ hESCs in direct myofiber cocultures, or with muscle supernatants, is shown in (B). * indicates P≤ 0.001 as compared to young. (C) Transwell co-cultures between purified young satellite cells and myofibers isolated from uninjured young (young myofiber) and old (old myofiber) muscle demonstrated that satellite cell regenerative myogenic capacity was inhibited by the aged differentiated muscle. Myogenic potential was determined by the ability of satellite cells to generate proliferating desmin+ myoblasts (immunodetection shown in green) and by rate of proliferation (2 h BrdU incorporation; immunodetection shown in red). (D) Satellite cell regenerative potential was quantified as percentage of desmin+/BrdU+ cells for transwell co-cultures with young or old uninjured myofibers (i.e., RM, resting muscle). n = 3; * indicates P≤ 0.05 as compared to young.   Comprehensively, these data establish that the inhibition of stem cell regenerative potential by the aged systemic milieu is conserved between species (mouse vs. human) and cell types (adult vs. embryonic stem cells). As summarized in Table 1, aged mouse sera similarly affected the expression of key molecular identifiers of both embryonic and adult stem cells, e.g., Oct4 in hESCs and Myf5 in mouse ASCs. As expected, adult mouse stem cells did not express Oct4, and hESCs did not express Myf5 in these experimental conditions (not shown). Moreover, aged systemic milieu had similar inhibitory effects on proliferation of hESCs and ASCs, suggesting that not only the regenerative capacity, but also the presence and expansion of stem cells will be significantly restricted in aged organs. Intriguingly, prolonged culturing of hESCs in their preferred in vitro conditions enables generation of an embryonic microniche that antagonizes the inhibitory influences of aged circulatory factors.
Table 1  Conservation of stem cell aging in the systemic environment   Quantified results from Figs 1, 2 are summarized and presented as mean percentages from experimental replicates ± SE. Rate of proliferation (BrdU) and cell-fate identifier (Oct4 or Myf5) are shown for both ESCs and ASCs cultured in heterochronic systemic conditions of 10% young (young), 10% old (old) or in 5%+ 5% mouse sera combination (young + old). Results for 5% young mouse sera are very similar to those for 10% young mouse sera and are shown in Fig. 1 (ASCs) and Supplementary Fig. S3 (hESCs).

The regenerative potential of hESCs and ASCs is inhibited by aged differentiated muscle
After establishing that the aged systemic niche negatively affects the regenerative capacity of hESCs and of ASCs, we then assessed whether myogenic potential and the rate of cell proliferation would be inhibited in hESCs and ASCs by the aged local muscle niche. Myofibers with associated satellite cells were isolated from young and old injured muscle, and were directly co-cultured with hESCs in typical hESC differentiation medium (DMEM/FBS). Similar to Fig. 1, the myogenic potential in these co-cultures was assayed by the expression of desmin, which is present in both fusion-competent myoblasts and newly formed myotubes. To analyze whether hESCs, mouse myogenic progenitor cells or both could express desmin in direct co-cultures, we costained these cells with a mouse-specific antibody to a myogenic marker, M-cadherin, which does not react with human protein, and a desmin-specific antibody that recognizes both mouse and human proteins. As shown in Fig. 4A, hESCs underwent myogenic differentiation in co-cultures with young myofibers (M-cadherin−/desmin+ mononucleated cells, white arrow in young). These myogenic progeny of hESCs in co-cultures with young myofibers could be of skeletal, smooth or cardiac muscle lineages (Debus et al., 1983; Fischman & Danto, 1985; Schultz & McCormick, 1994). As expected, the young mouse muscle progenitor cells (M-cadherin+/desmin+) were more advanced in their degree of myogenic differentiation, which was of skeletal muscle lineage, as judged by the formation of large, multinucleated de novo myotubes (yellow arrow in young). In addition to the myogenically differentiated human cells, co-cultures with young myofiber explants also contained some small undifferentiated hESC colonies, as determined by immunoreactivity to a human-specific antibody to the nuclear mitotic apparatus protein, NuMA and Oct4 expression (Supplementary Fig. S4).
In contrast, when co-cultured with the aged mouse myofibers, only mouse cells appeared desmin+ (Fig. 4A, yellow arrow in old). These aged myogenic cells were of skeletal muscle lineage, based on spontaneous generation of multinucleated myotubes (see Fig. 5B) and based on induced differentiation into myotubes in DMEM + 2% horse serum (not shown). Importantly, the myogenic differentiation of hESCs failed in the aged co-cultures (Fig. 4A, white arrow in old). Furthermore, colonies of hESCs in co-cultures with aged myofibers typically differentiated into cells with fibroblast morphology, which lacked Oct4 expression (not shown). Spontaneous production of desmin+ myogenic cells in control hESC cultures without myofibers, or with young/old mouse sera was less than 0.1% (not shown).
Fig. 5  In vitro co-culture with hESCs enhanced myogenesis of mouse cells. (A) 1 × 105 hESCs or control hMSCs were co-cultured with 5 × 106 primary mouse myoblasts. hESCs expressing Oct4 (immunodetection shown in red) dramatically enhanced myotube formation of co-cultured mouse myoblasts (immunodetection of eMyHC is shown in green), as compared to co-cultures between mouse myoblasts and human mesenchymal stem cells (Mb + hMSCs) or myoblasts alone (Mb alone). Experiments were carried out in myoblast differentiation medium. Hoechst (blue) labels nuclei throughout this figure. (B) 1 × 105 hESCs or control hMSCs were co-cultured with young or old myofiber-associated satellite cells, as described in Experimental procedures. Co-culture with hESCs (myofiber + hESC), but not hMSCs (myofiber + hMSC) or control medium (DMEM/FBS), greatly enhanced the myogenic potential of both young and old myofiber-associated satellite cells, based on immunodetection of percentage of desmin+ de novo generated myoblasts and multinucleated myotubes. These experiments were carried out in GM. Shown are myogenic responses of mouse cells only, judged by lack of immunoreactivity to human-specific/hESC-specific antigens, such as NuMA and Oct4; and presence of mouse-specific immunoreactivity, e.g., M-cadherin (not shown). Both young and old myofiber associated satellite cells exhibited considerable myogenic improvement over control conditions. n = 3.   In concert with the conservation of inhibitory affects of aged systemic niche, the negative influence of local muscle niche was also found to be conserved in its inhibition of hESC and ASC regenerative responses. Specifically, the myogenic capacity (generation of desmin+ myoblasts) was inhibited in young satellite cells co-cultured in a transwell system with aged myofibers (Fig. 4B). In addition, hESC and ASC proliferation (BrdU incorporation) was also inhibited by aged differentiated muscle (Fig. 4A,C). These data suggest that not only systemic but also local organ niches would inhibit key stem cell properties, e.g., myogenic capacity and the rate of proliferation in the aged organism. The conserved inhibitory influences of the differentiated muscle niche on hESC and ASC regenerative responses are summarized in Table 2.
Table 2  Conservation of stem cell aging in the local organ niche   Quantified results from Fig. 4 are summarized and presented as mean percentages from experimental replicates ± SE. Rate of proliferation (BrdU) and myogenic differentiation (desmin) are shown for both ESCs and ASCs, in the presence of young vs. old differentiated muscle environments (young myofiber or old myofiber).

hESCs indirectly enhance and rejuvenate the regeneration of skeletal muscle
While hESC properties were inhibited by aged differentiated muscle, the myogenic potential of aged satellite cells seemed to be enhanced by co-cultures with hESCs (Fig. 4A). Therefore, we further explored the enhancing and rejuvenating effects of hESCs on myogenic potential in vitro and in vivo, using human mesenchymal stem cells (hMSCs) as a negative control. First, we examined the effects of hESCs on myotube generation by co-culture with primary myoblasts freshly derived from activated-by-injury satellite cells (Conboy et al., 2003). As shown in Fig. 5A (Mb + hESC), primary myoblasts underwent very rapid and robust nascent myotube formation, when co-cultured with hESCs for 48 h in myoblast differentiation medium. Namely, remarkably large fused myotubes containing approximately 50–70 nuclei formed around hESCs colonies (Fig. 5A). In contrast, when co-cultured with hMSCs, myotube formation was no greater than in myoblast cultures alone (Fig. 5A, Mb + hMSC and Mb alone). Encouraged by these data, we analyzed the myogenic potential of young and old satellite cells co-cultured with hESCs for 48 h. As shown in Fig. 5B, hESCs conferred a much-enhanced myogenic capacity on both young and, importantly, old myofiber-associated satellite cells (rapid formation of desmin+ myogenic cells, many of which formed de novo multinucleated myotubes). Control co-cultures of these satellite cells with hMSCs displayed no enhanced myogenicity. In summary, while the myogenic potential (production of desmin+ fusion-competent cells) was more pronounced in young vs. old myofiber-associated satellite cells under all experimental conditions, a finding that is consistent with previous data (Conboy et al., 2003), a clear increase in myogenic potential of old satellite cells was noted in co-cultures with hESCs, as compared to control cultures devoid of hESCs (Fig. 4A,B).
Interestingly, in addition to the rejuvenating effects of direct co-cultures shown in Fig. 5, soluble factors present in hESC-conditioned culture supernatants were also able to enhance myogenesis of aged satellite cells (Supplementary Fig. S5). Thus, in agreement with the notion that an established embryonic microniche antagonizes the inhibitory effects of the aged environment on stem cell responses (Fig. 3), the hESC-produced factors enhanced myogenic capacity of even old mouse satellite cells.
Establishing that hESC-produced factors enhance adult myogenesis and rejuvenate the regenerative capacity of even aged satellite cells in vitro prompted us to examine whether the regeneration of old injured muscle will be improved by hESC transplantation in vivo. Additionally, based on the data shown above, we speculated that even if the host's repair capacity is improved, hESCs themselves will not be efficiently maintained or expanded in the context of old systemic and local organ environments, and will not directly contribute to the repair of aged skeletal muscle. To test these hypotheses, we injected 5 × 105 hESCs or control hMSCs into the tibialis anterior (TA) and gastrocnemius muscles of young and old mice at 24 h after cardiotoxin-induced injury, when activation/proliferation of endogenous satellite cells normally begins (Conboy et al., 2003, 2005; Wagers & Conboy, 2005). To avoid immune response against hESC antigens, mice were immunosuppressed using FK506 (Ito & Tanaka, 1997; Dumont, 2000). Muscle was isolated 5 days post-injury, when nascent differentiated myofibers normally replace the damaged tissue (Conboy et al., 2003), and 10 µm cryosections were analyzed for the success in tissue repair using hematoxylin and eosin (H&E) histochemistry and eMyHC immunodetection. H&E analysis reveals newly formed myofibers, based on their smaller size and centrally located nuclei. Additionally, de novo myofibers in the damaged area appear positive for eMyHC, while undamaged myofibers remain negative. As shown in Fig. 6A and quantified in 6B, injection of hESCs significantly enhanced regeneration of skeletal muscle. Remarkably, this positive embryonic effect was especially pronounced in old tissue.
Fig. 6  Skeletal muscle regeneration following hESC transplantation is a balance between the inhibitory influence of aged niches and the rejuvenating effects of hESCs. Young and old tibialis anterior and gastrocnemius muscles were injured by cardiotoxin injection. hESCs or hMSCs were transplanted at the site of injury and were analyzed by cryosectioning at Day 5 after injury (as described in Experimental procedures). (A) Newly regenerated myofibers were detected using eMyHC-specific antibody (green) and staining with H&E. In H&E staining, newly regenerated areas contain smaller, immature myofibers with centrally located nuclei. Uninjured myofibers are much larger, by comparison, with peripherally restricted nuclei. Poorly regenerated areas lack new myofibers and contain areas of fibrosis and inflammation. eMyHC immunodetection is specific for regenerating areas of muscle only. Both assays showed dramatic enhancement of muscle regeneration in ‘old + hESC’ vs. ‘old + hMSC’. Regeneration improvement was also seen in young + hESC, as compared to young + hMSC. (B) Quantification of muscle regeneration was performed by analyzing the density of newly formed myofibers per mm2 of injury site, which is the volume that typically covers the whole injured area. Multiple, 10 µm H&E sections were examined through the entire volume of injury in multiple, independently injured muscles. n = 20; * indicates P < 0.001 (‘old + hMSC’ compared to young + hMSC and ‘old + hMSC’ compared to ‘old + hESC’. (C) H&E and immunofluoresence staining for Oct4, and a human-specific antibody to NuMA, revealed the failure of hESCs to expand or persist in old, but the presence of hESCs in young muscle at 5 days post-transplantation. Hoechst (blue) labels nuclei.   Importantly, such enhanced and rejuvenated muscle repair stems from an indirect induction, as hESCs themselves (or control hMSCs) did not physically contribute to the mouse myofibers, as judged by near absence (less than 0.1%) of human-specific NuMA+ nuclei in de novo desmin+ myofibers, analyzed through multiple injury sites. An example of one regenerated myofiber from young muscle injected with hESCs, with NuMA+ nucleus in a field of NuMA−/desmin+ mouse myofibers, is shown in Supplementary Fig. S6. No such NuMA+/desmin+ myofibers were detected in aged regenerated muscle (not shown).
In agreement with the in vitro data, establishing that aged systemic and local niches inhibit hESC proliferation and Oct4 expression (Figs 2 and 4 and Supplementary Fig. S2), hESCs failed to expand or even persist in old muscle, as judged by the absence of NuMA+/Oct4+ hESC-derived cells in the aged tissue. In contrast, colonies of Numa+/Oct4+ hESC-derived cells that did not undergo myogenic differentiation were easily detected in young regenerating muscle (Fig. 6C). This finding validates several technical aspects of these experiments, and confirms the contrasting effects of young and old systemic and local organ niches on hESC self-renewal.
These data further confirm and extrapolate our findings and demonstrate that when exposed to both aged systemic and local organ niches, hESCs fail to persist and do not contribute to tissue repair directly. At the same time, these embryonic cells indirectly but significantly improve the repair of aged injured muscle in vivo.