Geroconversion of aged muscle stem cells under regenerative pressure

Abstract

Regeneration of skeletal muscle relies on a population of quiescent stem cells (satellite cells) and is impaired in very old (geriatric) individuals undergoing sarcopenia. Stem cell function is essential for organismal homeostasis, providing a renewable source of cells to repair damaged tissues. In adult organisms, age-dependent loss-of-function of tissue-specific stem cells is causally related with a decline in regenerative potential. Although environmental manipulations have shown good promise in the reversal of these conditions, recently we demonstrated that muscle stem cell aging is, in fact, a progressive process that results in persistent and irreversible changes in stem cell intrinsic properties. Global gene expression analyses uncovered an induction of p16^INK4a in satellite cells of physiologically aged geriatric and progeric mice that inhibits satellite cell-dependent muscle regeneration. Aged satellite cells lose the repression of the INK4a locus, which switches stem cell reversible quiescence into a pre-senescent state; upon regenerative or proliferative pressure, these cells undergo accelerated senescence (geroconversion), through Rb-mediated repression of E2F target genes. p16^INK4a silencing rejuvenated satellite cells, restoring regeneration in geriatric and progeric muscles. Thus, p16^INK4a/Rb-driven stem cell senescence is causally implicated in the intrinsic defective regeneration of sarcopenic muscle. Here we discuss on how cellular senescence may be a common mechanism of stem cell aging at the organism level and show that induction of p16^INK4a in young muscle stem cells through deletion of the Polycomb complex protein Bmi1 recapitulates the geriatric phenotype.

Keywords:

• Senescence
• Stem cells
• Skeletal muscle
• Aging
• Geroconversion
• Quiescence
• Satellite cells

Stem Cell Aging

Adult or somatic stem cells represent a small population of long-lived progenitor cells which reside in an undifferentiated state within most or all fully differentiated adult tissues and whose function is fundamental for maintaining tissue homeostasis and for mediating growth and regeneration after damage or in disease. It has been postulated that aging of an organism is in great part a direct result of stem cell loss-of-function. With age there is both a progressive decline in the stem cell self-renewal ability, causing in some cases a depletion of the stem cell pool, and a loss of activation/proliferative potential, resulting in impairment of the regenerative process due to an inability to generate the appropriate progeny.^1,2 Tissues vary in their homeostatic and regenerative requirements and that is reflected on the turnover of their associated stem cell populations.^3,4 The heterogeneity in stem cell differentiation and proliferative potential among different tissues underlay different susceptibilities to age-associated stressors throughout the organism lifespan. Stem cells of tissues with high cellular turnover, such as the intestine or the haematopoietic system, will be subjected to genomic instability that is associated with high rates of DNA replication. By contrast, low cellular turnover tissues, such as the skeletal muscle and the cerebral cortex, contain stem cell populations that are for the most part maintained in a quiescent state and only induced to proliferate upon injury or stress. These non-dividing cells undergo a progressive decline in function with chronological aging, although the mechanism behind this process is poorly understood.^5-7 Cellular senescence corresponds to a state of irreversible cellular arrest and can be triggered in primary cells by different stimuli, including expression of activated oncogenes (oncogene-induced senescence), or serial passaging (replicative senescence). Specifically, accelerated cellular senescence (i.e. geroconversion) is believed to occur when cellular growth is promoted while cell cycle progression is inhibited.^8,9 The concept of cellular senescence as a driving force for age related pathologies was first introduced by Hayflick and Moorhead in 1961 and the hypothesis has sustained until the present without a definitive prove.¹⁰ Meanwhile, senescent cells have indeed been observed during organismal aging^11-18 as well as in other different physiological contexts, including embryonic development,^19,20 tissue repair^21,22 and tumor suppression.²³ The physiological role of senescence in these contexts has also been addressed and it involves functions that range from tissue remodeling during development to limiting excessive tissue fibrosis during wound healing and tissue repair.^24,25 However, how do senescent cells promote age-related tissue dysfunction is a question that remains to be answered. One hypothesis is that senescence acts on stem cells to limit tissues regenerative potential during aging. Some hints of such mechanism have been observed in skeletal muscle and fat tissue from BubR1 progeroid mice,²⁶ where senescence markers co-localized with resident stem cells.

The resident adult stem cells of skeletal muscle are termed satellite cells since they are located between the myofibers and the basal lamina. Satellite cells express the paired box transcription factor Pax7 and normally reside in a quiescent state until activated by environmental signals (e.g., injury or stress) or in pathological environments (e.g., in degenerative muscle diseases).^27-29 Following activation, satellite cells originate 2 types of progeny: self-renewed satellite cells which maintain the stem cell pool, and committed myogenic cells, which form new muscle fibers where necessary to sustain postnatal muscle repair and growth.³⁰ Satellite cell aging is characterized by a loss of regenerative potential,^31,32 which is particularly pronounced in sarcopenic muscles.³³ This is mostly a consequence of impairments in activation and proliferative potential in response to injury, which account for the primary age-specific defects in muscle repair. The regenerative defects are also associated with a delayed inflammation and an increase in tissue fibrosis and aged satellite cells display a tendency to enter alternative differentiation programs by adopting fibroblastic and adipogenic fates.^34-36

Given the negative impact of satellite cell aging on muscle regeneration, there has been an extensive research effort in characterizing the origin of the aging phenotypes, focusing on strategies of rejuvenation. The prevalent view has been that satellite-cell-dependent muscle regeneration decline with aging is mostly due to extrinsic factors and can therefore be reversed by exposure to an external youthful environment.² This view originates in heterochronic parabiosis and hetero-grafting studies between young and old (non-geriatric) mice. Several labs have shown that the regenerative capacity of injured muscle tissue from old mice can be enhanced to match the efficiency of young mice when their blood circulation is joined.^34,37,38 This is mostly a consequence of satellite cell rejuvenation by exposure to young serum, leading to the re-establishment of proliferative and myogenic potential necessary to regenerate the injured muscle.³⁸ These data supported earlier hetero-grafting studies^39-43 and suggest that there are factors that accumulate either systemically or in the niche of old animals, which are deleterious to the satellite cell and regenerative process. More importantly, it suggests that the old serum/niche is depleted of factors that are present in the blood of young animals, which could be therefore sufficient to enhance regenerative potential of the injured skeletal muscle and can rejuvenate old (non-geriatric) satellite cells. Several niche-derived and systemic factors responsible for the regenerative decline in old animals have been identified, including Transforming growth factor β (TGFβ) and Fibroblast Growth Factor 2 (FGF2) from the niche,^32,44,45 as well as systemically derived Wnt signaling.³⁴ These studies focused on pathways which are over-activated in aging and proposed rejuvenation strategies based on inhibition of these signaling cascades. Two recent studies, however, have identified systemic derived factors capable of reversing age-related impairment of muscle regeneration and rejuvenating old satellite cells.^46,47 Growth differentiation factor 11 (GDF11) and Oxytocin are systemic factors depleted in the serum from old animals, and its supplementation was sufficient to restore muscle regeneration capacity and improve myogenic potential and genomic integrity in old satellite cells.^46,47 Of note, GDF11 is a TGFβ-family member, as Myostatin,⁴⁸ and given the anti-myogenic and pro-atrophying roles of this growth factor family, the pro-regenerative function of GFD11 came as unexpected.

Together, these studies support the idea that there is in fact an important contribution of the environment to the process of satellite cell aging, and that many of the changes that occur during muscle stem cell aging can be reversed. While this model holds true for classically classified old animals (around 20 months of age), one would expect that if age-associated senescence is a consequence of combined cellular stresses as recently proposed,²⁵ they would take a longer period of time to accumulate and senescent cells may not be observed until later stages in life. Here, we present and discuss our recent data⁴⁹ showing that in “healthy” geriatric animals (beyond 28 months), satellite cells enter an irreversible pre-senescent state that affects their intrinsic regenerative and self-renewal ability. We propose that maintenance of quiescence in adult life depends on the active repression of senescent pathways and that the molecular mechanisms behind satellite cell geroconversion in geriatric animals relay on the de-repression of the p16^INK4a locus. Because senescent cells have also been observed in patients with several age-related pathologies, associated with the affected tissues,^24,50,51 including cancer, which also increases with age,⁵² we propose the geroconversion into a senescent state might be common to stem cells in different tissues and the underlying cause of several age-related dysfunctions.

Old vs Geriatric Satellite Cells: An Irreversible Switch

Although old animals show features consistent with the onset of muscle atrophy it is not until geriatric ages that a full sarcopenic phenotype can be detected. Consistent with this, regenerative capacity after muscle injury, which is already decreased in old animals is further impaired in geriatric animals, correlating with a progressive decrease in activation rate of satellite cells in response to injury. Interestingly, the worsening of the phenotype is not due to a loss of satellite cells. Satellite cell numbers decrease in old animals compared to adult (5-6 months old) mice,^32,53 but this decline in number is no further accentuated after 24 months of age.⁴⁹ This observation suggested to us that there may be intrinsic changes that occur in geriatric satellite cells that underlie the marked drop in regenerative ability and satellite cell activation potential observed in geriatric animals. To address this question, we have isolated satellite cells from donor mice of different ages—adult, old and geriatric—and compared their regenerative and self-renewal capabilities. Consistent with previous studies (see above), satellite cells isolated from old mice and transplanted into injured muscles of young hosts were able to regenerate the muscle with an efficiency similar to adult satellite cells. In agreement, old satellite cell activation potential was not significantly distinct to that of adult satellite cells after transplant into young hosts. These observations confirm that old satellite cells can be rejuvenated by exposure to a young environment and their loss of regenerative ability and activation potential can be rescued. Importantly, satellite cells isolated from geriatric mice and transplanted into injured muscles of young hosts remained unable to regenerate the damaged muscle, indicating that the impairment in geriatric satellite cell function cannot be reversed by exposure to a youthful environment.⁴⁹ Supporting this idea is the observation that geriatric satellite cell activation remained impaired after transplant into young hosts, even after a 21 d adaptation period. These observations strongly suggested that there were intrinsic changes that occur specifically in geriatric satellite cells that cannot be reversed by youthful environmental cues.

We have used global gene expression analysis of satellite cells freshly isolated from resting skeletal muscle to identify transcriptional programs that may be responsible for the intrinsic changes described above.⁵⁴ Interestingly, a gene set enrichment analysis (GSEA)⁵⁵ of the significantly upregulated genes in geriatric satellite cells was enriched in gene sets controlled by Polycomb repressive complexes (PRC1 and PRC2), which are known mediators of cellular senescence.⁵⁶ Furthermore, comparison with a comprehensive list of genes implicated in cellular senescence^57-59 showed specific up-regulation of several genes, including the master regulator of senescence Cdkn2a (p16^INK4a).^{15,17,49,60,61}

Satellite Cell Geroconversion

The switch into a pre-senescent state anticipated by the transcriptome profile led to the hypothesis that in geriatric animals a quiescent to senescent conversion in muscle stem cells may be the underlying cause for an irreversible impairment in activation/proliferative potential, required both for self-renewal and for the proliferative expansion that precedes myogenesis. Indeed, we observed that after successive rounds of muscle injury, geriatric but not old satellite cells showed a progressive and cumulative loss of self-renewal ability. Consistently, the proliferative capacity of geriatric satellite cells induced by regeneration in vivo or growth conditions in vitro was markedly reduced. Confirming the switch from a quiescent to a pre-senescent state, satellite cells from geriatric animals placed under this proliferative pressure acquired characteristics of full senescence, such as senescence-associated β-galactosidase activity and expression of several senescence markers. These features were observed both in vivo and in vitro and could not be reversed by adaptation to a youthful environment confirming that the quiescence-to-senescence switch of geriatric satellite cells is irreversible and leads to accelerated senescence (i.e., geroconversion). Interestingly, these features of cellular senescence were also observed in human satellite cells from sarcopenic geriatric individuals, which express p16^INK4a and other senescence markers. This is the first time that senescent stem cells are observed in vivo during skeletal muscle's normal aging and supports the previously proposed hypothesis that senescence may indeed be a driving force for age-associated tissue dysfunction and regenerative decline.^25,26,49

p16^INK4a/Rb/E2F Axis in Satellite Cell Geroconversion

INK4a refers to a chromosomal locus encoding for 2 alternative reading frame proteins described as tumor suppressors: p16^INK4a and ARF (p19^ARF in mouse and p14^ARF in human).⁶² Each of these proteins has been shown to have a critical role in senescence, cell cycle control and tumor suppression, and this locus is often functionally inactivated in human tumors. p16^INK4a is a cyclin D-dependent kinase (cdk) inhibitor acting upstream of retinoblastoma protein (pRb), preventing its phosphorylation and functional inactivation by cdks.⁶³

p16^INK4a gene expression increases with age in a variety of tissues⁶⁴ and induction of p16^INK4a expression can cause senescence of a variety of cell types in culture and in vivo.⁶³ Although a universal marker that is exclusively expressed in senescent cells has not been identified, most senescent cells seem to express p16^INK4a, making it a generally accepted biomarker of senescence. Removal of senescent cells using a strategy of inducible elimination of p16^INK4a-expressing cells upon drug administration delays the onset of some of the most common age-related phenotypes in adipose tissue, skeletal muscle and eye of BubR1 progeroid mice,¹⁷ reinforcing the idea that p16^INK4a associated senescence plays a central role in age-dependent tissue dysfunction.

Because p16^INK4a was the only senescence-associated gene identified by our global gene expression profile to be up-regulated in different sarcopenia-associated satellite cell signatures, we sought to investigate a role for p16^INK4a in satellite cell geroconversion. p16^INK4a overexpression in young satellite cells was sufficient to prevent efficient activation and to promote geroconversion under proliferative pressure both during regeneration and in vitro growth conditions, suggesting that p16^INK4a is a central molecular effector of quiescence-to-senescence switch of satellite cells (Fig. 1). Because p16^INK4a acts upstream of retinoblastoma/E2F mediated repression of proliferation-promoting transcription factors we sought to investigate the involvement of this pathway in satellite cell geroconversion. Indeed, geriatric satellite cells which express high levels of p16^INK4a under proliferative pressure also show a reduction in the phosphorylated form of retinoblastoma and a consequent repression of Rb/E2F target genes required for cell cycle progression. Importantly, the activation of this Rb/E2F axis could not be reversed by environmental cues, when geriatric satellite cells were transplanted into injured young muscles. Consistent with the senescent phenotype observed in human geriatric satellite cells, Rb phosphorylation and Rb/E2F target genes are also repressed in satellite cells from geriatric sarcopenic individuals.⁴⁹ In agreement with our findings, a coetaneous publication of the Blau lab,⁶⁵ reported that the activation of the p38 MAPK cascade in aged satellite cells, as it has been demonstrated in different somatic cell types,^66-70 could underlie p16^INK4a, and that inhibition of the p38 MAPK pathway rejuvenated the satellite cells.⁶⁵

Figure 1. Satellite cell geroconversion disrupts muscle regeneration. Satellite cells stay quiescent in young and adult mice under normal conditions. During aging, geriatric satellite cells of sarcopenic muscles undergo derepression of p16^INK4a, a regulator of cellular senescence, losing as a consequence their reversible quiescent state. Instead, they adopt a senescent-like state (becoming pre-senescent cells), which impairs the regeneration process by specifically preventing stem cell activation, proliferation and self-renewal.

Epigenetic Regulation of the p16^INK4a Locus

Given the central role of the p16^INK4a on satellite cell geroconversion, it is important to understand the mechanisms involved in the regulation of its locus in the context of satellite cell aging. Epigenetic regulation through Polycomb repressive complexes is a key component of p16^INK4a silencing and is achieved through the sequential establishment of the H3K27me3 chromatin mark and the mono-ubiquitination of histone H2A, by PRC2 and PRC1 complexes, respectively.^71-73 Bmi1 (a member of the PRC1 complex) is found strongly enriched at the INK4a promoter, together with PRC2 members and the associated H3K27me3 mark, in different tissues.^74-76 The critical role of PRCs in regulating the INK4A locus is further supported by studies showing that over-expression of PRC1 and PRC2 components is sufficient to bypass senescence, while cells lacking PRC1 components, including Bmi1, show aberrant expression of the p16^INK4a gene.^15,75,77 In agreement with the extensive data on the regulatory functions of PRC proteins, we found that also in young satellite cells the INK4a locus (both at the RD domain and at exon1a/2 regions,^78,79 was marked with the PRC1 repressive mark H2AUb and that this mark was significantly reduced in satellite cells from geriatric mice.⁴⁹ This pattern of locus occupancy is consistent with p16^INK4a transcriptional repression and induction in young and geriatric satellite cells, respectively.

To explore the relevance of INK4a locus derepression and associated p16INKa expression in satellite cell function, we used Bmi1-deficient mice. Bmi1^−/− mice display a number of defects including a progressive decrease in the number of haematopoietic cells, severe neurological abnormalities and markedly shortened lifespan.⁸⁰ Young adult Bmi1-deficient mice were considerably smaller than the age-matched WT littermates and displayed a corresponding reduction in muscle size (Fig. 2A). Evaluation of myofiber cross-sectional area showed that these differences were not significant at early post-natal stages but dramatically increased thereafter, suggesting premature muscle wasting (Fig. 2A), and in agreement with a previous report.⁸¹ Several studies indicate that Bmi1 is required for the self-renewal and post-natal maintenance of the stem cell population of haematopoietic, neural, gastrointestinal and breast tissues.^82-85 In most cases, the defect in self-renewal in the absence of Bmi1 was associated with increased p16^INK4a expression and INK4a deficiency partially rescues haematopoietic and neural stem cell self-renewal of Bmi1^−/− mice.^86,87 We reasoned that loss of repression of the INK4a locus and the high expression of p16^INK4a might also provoke defective self-renewal in Bmi1 null satellite cells (Fig. 2B). We forced these cell to undergo cycles of self-renewal through sequential rounds of quiescence, activation, proliferation and return to quiescence, by performing successive rounds of muscle injury. We quantified self-renewal efficiency by the number of quiescent (Pax7⁺/Ki67⁻) satellite cells associated to newly formed myofibers. 21 days after CTX injury, the number of Pax7⁺/Ki67⁻ satellite cells was significantly reduced in Bmi1^−/− regenerated fibers compared to the WT (Fig. 2C). The regenerated muscle was re-injured to trigger the reactivation of satellite cells from the quiescent state, new round of myogenesis and self-renewal. Importantly, the number of self-renewed satellite cells was further reduced in Bmi1-deficient mice correlating with a regeneration defect and an upregulation of p16^INK4a (Fig. 2B). These results indicate that a deregulation of the Bmi1/p16^INK4a axis leads to an intrinsic defect in satellite cell self-renewal, that mimics what is observed in geriatric animals.⁴⁹

Figure 2. Atrophic muscles and pre-senescent satellite cells in Bmi1-deficient mice. (A) Right: Representative images of Hematoxilin/Eosin (H/E) stained cryosections of tibialis anterior (TA) muscle from Bmi1^+/+ and Bm1^−/− animals at 30 d after birth (p30). Left: Boxplot represent the quantification of myofiber size, evaluated by the cross-sectional area at the indicated days after birth. n = 10 mice, 500 fibers were measured per mice. (B) Expression of p16^INK4a, evaluated by RT-qPCR, in satellite cells isolated from Bmi1^+/+ and Bm1^−/− mice by FACS as described.⁴⁹ n = 7 mice. (C) Self-renewed satellite cells (Pax7⁺/Ki67⁻) identified by immunostaining in cryosections from regenerating TA muscle of Bmi1^+/+ and Bm1^−/− mice analyzed 21 d after being injury-induced by cardiotoxin injection (single muscle injury). Twenty one days after the first injury, the muscles were re-injured and satellite cell self-renewal was analyzed 21 d after the second injury (double muscle injury). n = 7 mice. Scale bars, 50 μm. Data are mean ± s.e.m. Two-sided Mann-Whitney U test was used to assess statistical significance. P values are indicated. n.s. not significant. Boxplot center lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software;⁹³ whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles.

Genetic Rejuvenation of Geriatric Satellite Cells

Our data suggest that there is an active protective mechanism repressing senescence pathways in young satellite cells, and that its deregulation in geriatric individuals leads to intrinsic changes to the satellite cells that cannot be reversed by a youthful environment. However, this does not mean that senescence cannot be overcome and geriatric satellite cells rejuvenated to restore muscle regeneration in sarcopenic individuals. On the contrary, our results show that p16^INK4a inhibition is an effective rejuvenation strategy for geriatric satellite cells. p16^INK4a silencing with short hairpin specifically in isolated geriatric satellite cells before transplantation into young muscles was sufficient to restore activation potential. p16^INK4a silencing also restored self-renewal capacity of satellite cells both in vivo and in vitro and improved regeneration efficiency. Importantly, both in mouse and human satellite cells, p16^INK4a silencing alone limited satellite cell geroconversion, as shown by enhanced proliferative ability and reduced senescence makers.⁴⁹ This suggests that genetic rejuvenation of satellite cells by preventing p16^INK4a expression may be an effective way to prevent poor regeneration phenotypes in humans undergoing sarcopenia, though some concerns apply, first whether such therapeutic intervention will be feasible and second whether the benefits of a muscle stem cell rejuvenation outweigh the possible risks of a reduction of a tumor-suppressor in the context of aging.

Understanding how intrinsic aging is triggered may be the key to bypass these concerns. In that sense, as mentioned above, Cosgrove et al. have recently demonstrated that the p38 MAPK cascade is part of the signaling pathway leading to p16^INK4a expression and that its pharmacological inhibition is able to rescue the poor proliferation phenotype of old satellite cells.⁶⁵ The environmental or niche extrinsic factors that may be implicated in the activation of the p38 MAPK pathway remain mostly unknown, though it has been shown that FGF-2 signals may be upstream of p38 MAPK pathway in the committed cell during asymmetric division of young satellite cells,⁸⁸ and that in an aging environment FGF-2 levels increase, asymmetric division is impaired and symmetric division with p38 MAPK activation in both cells is induced.⁸⁹ Inhibition of the p38 MAPK pathway restores asymmetric division and aged satellite cell function.⁸⁹ Though inhibition of p38 MAPK is increasingly considered as a therapeutic strategy -mostly because of its role as a master regulator of inflammatory responses-, the facts that p38 MAPKs are crucial stress-responsive kinases expressed in every cell and the discovery of p38 MAPK as a main actor in the development of tumorigenesis⁹⁰ have generated many concerns over its practical use in long term therapies. While this paper was being reviewed, two new papers reported the involvement of the JAK/STAT signaling pathway in the reduced expansion of satellite cells with aging, and the rescue of this defect by JAK/STAT inhibitors.^91,92

Concluding Remarks

In vivo stem cell senescence is an emerging concept that may help understand and design strategies to treat several age related pathologies. Our data shows that the maintenance of satellite cell quiescence throughout adult life requires the active repression of genes associated with senescence and that the loss of epigenetic marks regulating these genes in advanced aging results in a quiescence-to-senescence cellular switch. Because the pathways involved in cellular senescence are conserved, it is likely that cellular senescence occurs in other tissues during physiological aging and that this may be a common mechanism that coordinates tissue aging at the organism level. Understanding the underlying mechanism involved in the de-repression of these pathways with advanced aging, and particularly in aging stem cells, may help design strategies to delay the aging process and the age-associated tissue regenerative decline. However, because senescence has been shown to have other physiological roles beyond aging, strategies involving the repression of senescence pathways should be looked at with care and be considered in an age- and cell/tissue-dependent manner.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We are indebted to AL Serrano, L García-Prat, S Gutarra, L Ortet and members of the Cell Biology Group for their contributions to this study.

Funding

The authors acknowledge funding from MINECO-Spain (SAF2012–38547, FIS-PI13/025, PLE2009–0124), DuchennePP-NL, E-Rare, Fundació Marató TV3, MDA, EU-FP7 (Myoage, Optistem and Endostem) and AFM.