The role of the stem cell epigenome in normal aging and rejuvenative therapy

Abstract

Adult stem cells are ultimately responsible for the lifelong maintenance of regenerating of tissues during both homeostasis and following injury. Hence, the functional attrition of adult stem cells is thought to be an important driving factor behind the progressive functional decline of tissues and organs that is observed during aging. The mechanistic cause underlying this age-associated exhaustion of functional stem cells is likely to be complex and multifactorial. However, it is clear that progressive remodeling of the epigenome and the resulting deregulation of gene expression programs can be considered a hallmark of aging, and is likely a key factor in mediating altered biological function of aged stem cells. In this review, we outline cell intrinsic and extrinsic mediators of epigenome remodeling during aging; discuss how such changes can impact on stem cell function; and describe how resetting the aged epigenome may rejuvenate some of the biological characteristics of stem cells.

Introduction

The human body undergoes multiple changes as it transitions into old age. These changes are often progressive and frequently detrimental, leading to an overall decline in health status with advancing age. Such detrimental changes can ultimately limit the lifespan of an individual and commonly impact negatively on quality of life. The drive towards promoting healthy aging calls for a better understanding of the mechanisms that drive this process, both at an organismal and a cellular level. Along these lines, specific biological hallmarks of aging have been proposed to better define this process. These hallmarks include cellular senescence, genomic instability, loss of proteostasis, mitochondrial dysfunction, epigenetic alterations and the depletion of somatic, tissue specific adult stem cells (1). Adult stem cells play the important role of ensuring the lifelong generation of the mature cells of the tissue to which they belong, both under homeostatic conditions and during the course of tissue regeneration in response to injury or infection. Any depletion of stem cell reserves will therefore result in the gradual functional attrition of regenerative tissues. Some well-known and characterized somatic stem cells in adults include hematopoietic stem cells (HSCs), neural stem cells (NSCs), intestinal stem cells (ISCs), mesenchymal stem cells (MSCs), muscle stem cells (MuSCs) and skin epithelial stem cells (2,3). As well as being capable of differentiating into multiple cell types in the tissue of origin, they are also thought to be capable of extensive self-renewal divisions, a process that would act to maintain the stem cell pool in the resident tissue throughout aging. However, the fact that many adult stem cell populations become depleted during aging suggests that self-renewal may be an imperfect process or that it cannot be sustained over long periods of time. Indeed, this frequently cited but poorly defined phenomenon is referred to as stem cell exhaustion. Even though stem cell exhaustion in itself is considered one of the hallmarks of aging, the other hallmarks likely comprise underlying mechanisms that drive the progressive functional attrition of stem cells. In this context, progressive remodeling of the epigenome is another hallmark of aging that has been proposed as a driver of stem cell dysfunction. The structure of heterochromatin is stabilized by multiple epigenetic marks, such as DNA methylation and histone modifications (4). Age-related DNA methylation changes have been associated with a global loss of methylation around heterochromatin regions and a focal hypermethylation in CpG islands of specific genes (5–8). This loss of heterochromatin state is also characterized by loss of repressive histone marks such as H3K27me3 and a gain of activating histone marks such as H3K4me3 (9). The epigenetic landscape and its regulators are known to play a vital role in normal stem cell maintenance through self-renewal and differentiation (10). Thus any age-related aberrant changes to the epigenome would likely have a significant impact on stem cell functionality with age. However, whether the observed changes in the epigenome is a cause or consequence of aging is a long standing debate. In this review, we will discuss the role of the epigenome in modulating stem cell aging; age-related disorders that originate from stem cells, such as cancer; and the prospect of cellular reprogramming as a route to rejuvenate aged stem cells.

Intrinsic and extrinsic factors that impact on the stem cell epigenome during aging

Cell-intrinsic factors

Metabolism

Metabolic dysfunction and genomic instability are two of the hallmarks of aging that are intrinsic to stem cells and which are also linked to epigenome status (Fig. 1). Stem cells such as HSCs (11), NSCs (12) and MuSCs (3) have been shown to normally exist in a quiescent state under homeostasis and are activated upon demand in the resident tissue. The metabolic state of a stem cell is closely linked to whether it is present in a quiescent or activated state. For example, quiescent stem cells have been shown to primarily utilize the glycolytic pathway for their metabolic needs, while proliferating stem cells and their differentiated progeny are more skewed towards reliance upon oxidative phosphorylation (OxPhos) (13,14). Importantly, TCA cycle intermediates such as acetyl-coA and α-ketoglutarate (α-KG) are important co-factors for epigenetic regulators such as histone acetyltransferases and histone demethylases, respectively. Another product of active mitochondria, s-adenosylmethionine (SAM), is required for the methylation of CpG residues by DNA methyltransferases (15). Hence the metabolic state of the cell can have a direct effect on its own epigenetic landscape via modulating the abundance of molecular entities that are required for remodeling. It has been shown in HSCs that treatment with α-KG or SAM has a direct consequence on HSC differentiation. Addition of α-KG was found to increase myeloid differentiation and the addition of SAM was found to preserve stem cell function of HSCs in culture (16). It has also been shown in MuSCs, that the levels of mitochondrial NAD⁺, which is a co-factor for NAD-dependent protein deacetylase Sirtuin 1 (SIRT1), affects MuSC self-renewal (17). These findings suggest that the levels of these metabolites in the cell, which can modulate the epigenetic state, can also have a direct influence on stem cell function and self-renewal. Aberrations in mitochondrial function increases with age and can hence also lead to a shift in the glycolysis-OxPhos balance in stem cells. This shift can in turn modulate the levels of different metabolites that are involved in stem cell function and maintenance. With epigenetic changes linked tightly to changes in level of different metabolites, mitochondrial dysfunction plays an important role in the aging stem cell epigenome and vice versa. Age-related mitochondrial dysfunction also leads to elevated generation of Reactive Oxygen Species (ROS), which in turn increases the formation of DNA adducts such as 7,8-Dihydro-8-oxoguanine (8-oxoG) (18). The presence of 8-oxoG has been shown to prevent methylation of cytosine residues that are immediately adjacent to it, by interfering with the binding of DNA methyltransferases. Elevated ROS in stem cells during aging could hence have a direct effect on epigenetic modifications. Elevated ROS in cells also plays a role in accumulation of nuclear DNA damage, causing an overall state of genomic instability, which is another factor in stem cell aging (19,20).

Figure 1

Factors affecting the epigenetic landscape of aging stem cells. In contrast to a young epigenome, an aged epigenome is defined by an overall loss of heterochromatin characterized by DNA hypomethylation and gain of activating histone marks. This altered epigenome makes aged stem cells pervasive to increased expression of lineage-specification genes and pro-inflammatory genes, with a repression of self-renewal genes. These changes in the epigenetic landscape during aging can trigger and be triggered by several cell-intrinsic (1a and 1b) and cell-extrinsic (2a, 2b and 2c) factors. (A) Level of cellular metabolites is one common effector of the epigenome that is regulated by these factors. Under normal mitochondrial function in young stem cells, the metabolite levels are controlled by the balance between glycolysis and oxidative phosphorylation (OxPhos) in quiescent and active stem cells respectively. However due to mitochondrial dysfunction (1a) in an aged-cell there is a Glycolysis: OxPhos imbalance that affects the level of various metabolites such as S-adenosylmethionine (SAM), α-ketoglutarate (α-KG) and acetyl-CoA. These metabolites normally act as co-factors for various epigenetic regulators such as DNA methyltransferases (DNMTs), histone acetylases and histone deacetylases. The change in availability of these co-factors during aging due to dysfunctional mitochondria, deregulates these epigenome editors, thereby driving the aberrant epigenetic landscape of an aging stem cell. The level of metabolites and the resultant epigenetic changes can also be impacted by cell-extrinsic factors such as the microbiome composition (2a) and the stem cell niche (2b), both of which are also altered during aging. (B) Another common mediator of the epigenome that links the different extrinsic and intrinsic factors is DNA damage response (DDR). In young cells, an efficient DDR mediated DNA repair involves epigenetic modifiers such as DNMTs that are recruited by the DDR complex. However, there is a marked reduction in expression of DDR proteins and DNA repair efficiency during aging. Since epigenetic alterations are tightly linked to efficient DNA repair, a deregulated repair mechanism can lead to deleterious epigenetic changes in aged stem cells. Additionally, the reduction in heterochromatin during aging is a cell-intrinsic mechanism that sensitizes the chromatin to acquire DNA damage (1b). Some cell-extrinsic factors such as a reduction in beneficial microbiome (2a) and an altered stem cell niche (2b) during aging also impact on DNA damage and the resultant epigenetic alterations, through mechanisms such as altered growth factor expression, replication stress and inflammation (2c).

DNA Damage

A major cause of DNA damage in stem cells has been shown to be DNA replication stress. Driving quiescent HSCs into active proliferation has been shown to induce DNA damage and reduction in HSC function (21). DNA damage has also been shown to accumulate with age and is proposed to act as one of the major causes of compromised HSC function (22). Other stem cells such as mammary stem cells, MSCs, skin epithelial stem cells and MuSCs have all been shown to acquire DNA damage, but have very efficient DNA repair processes under homeostasis (23–25). However, the process of aging is often accompanied by an overall decline or defects in DNA repair efficiency (26). For example, defects in the DNA damage response (DDR) have been shown to be the cause of premature aging in progeroid mouse models (27). While genetic DDR defects can lead to progeria, the DDR can become compromised during normal aging, leading to genomic instability, as has been shown in ISCs (28,29) and HSCs (30). In this context, the epigenetic state of a cell plays a key role in determining sensitivity towards DNA-damaging agents and the accessibility of the chromatin to proteins that mediate DDR. A compacted heterochromatin state makes the DNA less susceptible to DNA damage induction (31). In addition, epigenetic regulators such as histone deacetylases and acetylases have been shown to localize to damaged DNA residues and subsequently recruit further DDR related proteins. Further deposition of repressive histone marks at these sites enabled transcriptional repression of genes within the DNA damage foci to enable efficient repair (32). However, with age there is significant loss of heterochromatin (5,9). This would potentially act on two fronts during aging. Firstly, to make the less condensed chromatin more sensitive to DNA damage acquisition, and secondly to restrict efficient repair, causing DNA damage accumulation. As stem cells are known to accumulate DNA mutations with age, the epigenome would play a vital role in how efficiently and accurately the genome is repaired. In addition to the epigenome affecting DNA damage, DNA damage can also drive epigenome remodeling at the sites of damage. SIRT1 localizes to heterochromatin regions at sites of damage and facilitates the subsequent recruitment of DDR proteins as well as other epigenetic regulators such as DNA methyltransferases—DNMT1 and DNMT3B, leading to methylation of damaged gene promoters (33,34). The deposition of repressive histone marks are also observed at these sites. This would normally enable efficient repair. However, under repeated DNA damage accumulation during aging, this can also lead to aberrant DNA methylation and histone mark patterns in these focal regions of damage. In this way, the duration of DNA repair process, which is in turn determined by the efficiency and expression level of DDR proteins, can reshape the epigenetic landscape. Conversely, the epigenetic state of the genes involved in DDR might also play a role in their expression during stem cell aging. However, there is no direct evidence yet for epigenetic regulation of DDR genes in stem cells during normal aging.

Cell-extrinsic factors

Inflammation

While metabolic dysfunction and DNA damage can affect stem cell biology, multiple factors that are cell extrinsic are responsible for driving these intrinsic factors (Fig. 1). One such extrinsic factor that is being extensively studied in the context of age related disorders is inflammation. Inflammation is an evolutionary immunological defense mechanism towards pathogens, allergens and tissue damage. This is normally an acute process within the body, with the resolution of the underlying cause and anti-inflammatory factors keeping it in check. However, with the immune dysfunction that accompanies aging (35), the resolution of the inflammatory state can become compromised. The term ‘inflamm-aging’ was first coined by Franceschi et al. to describe this significant involvement of inflammation in the process of aging (36). Inflammation is mediated by complex interactions, encompassing many cellular processes, through mediators such as systemic factors like cytokines and chemokines. These systemic factors have been shown to have a direct effect on different stem cells and their function. For instance, chronic infections and chronic inflammation have been shown to drive quiescent HSCs into cycle and such inflammation-mediated repetitive cycling of HSCs leads to their functional exhaustion (21,37,38). Such inflammation induced effects on proliferation and differentiation have also been observed in tissue regeneration response of ISCs, NSCs and skin epithelial stem cells (39–41). As the immune function becomes deregulated with age, these processes can become even more compromised. This could potentially lead to promiscuous or deregulated stem cell activation, which can in turn affect its functional regenerative capacity. Although not well documented, this could have effects on the epigenome of the cells, since inflammation is regulated through complex transcriptional networks (42). A recent study showed that HSCs in mice that have been treated with lipopolysaccharide (LPS) mimicking a bacterial infection, retain an epigenetic memory of the encounter. It was further observed that the HSCs from these mice were able to elicit a stronger myeloid response upon secondary LPS treatments, facilitated by an open chromatin at enhancers of myeloid genes. (43) This finding suggests that infection or inflammatory signals can have a direct influence on the chromatin landscape surrounding inflammatory response genes and that HSCs can retain this epigenetic state. As an individual encounters multiple infections throughout their life, the epigenome could become increasingly primed towards the induction of inflammatory responses. Studies on other cell types such as intestinal epithelial cells and bronchial epithelial cells also point to the epigenetic landscape as a player in regulating infection response. A state of promoter hypomethylation in Toll-like-receptor (TLR) genes in these cells have been shown to affect their sensitivity to bacterial peptidoglycans in terms of pro-inflammatory responses (44,45). Furthermore, inhibition of DNMTs has been shown to induce expression of human endogenous retroviral elements, resulting in the up-regulation of inflammatory interferon response genes in some cell lines (46). Therefore, hypomethylation of the promoters of these endogenous retroviral elements as a result of age-associated mutations in DNMTs could be a mechanistic driver of a sustained basal inflammatory signal. Together these findings suggest that immune priming of stem cells, DNA methylation state of infection response genes, and expression of endogenous retroviral elements might be some of the epigenetic mechanisms which drive chronic inflammation that often accompanies aging. However, further studies are required to evaluate whether the epigenetic alterations induced by inflammation have a direct consequence on stem cell function during aging.

Microbiome

There is increasing evidence that the status of the gut microbiome can have a dramatic impact on the overall health of an individual. The gut microbiome release multiple systemic factors and can affect not only the intestine, but also very distal organs such as the brain (47). Disruption of the gut microbiome has been implicated in a range of age-associated conditions such as diabetes, metabolic syndrome and cancers (48–50). This suggests that the composition of gut microbiota plays an important role in the process of aging. One way that the microbiome composition has been shown to be significant is in the secretion of specific metabolites and inflammatory stimuli, which are induced by specific microbiota. Some of the metabolites that are secreted by the microbiota are known factors that can directly affect epigenetic modifiers. For example, common probiotic bacteria such as Lactobacillus and Bifidobacterium produce folate, which is directly involved in the process of DNA methylation (51,52). Furthermore, fermentation of dietary fibers by the gut microbiome produce metabolites such as acetate, propionate and butyrate, which have all been linked to epigenetic modifications (53–55). Specifically butyrates are produced by a wide range of gut microbiota such as Coprococcus, Clostridium, Eubacterium, Faecalibacterium, Fusobacterium, Megasphaera, Odoribacter, Roseburia Peptoniphilus and Subdoligranulum (56). At high levels, butyrates are known Histone Deacetylase (HDAC) inhibitors and can hence directly modulate histone acetylation marks (57,58). If the microbiome composition changes to a butyrate-high or butyrate-low microbiota, then this could directly affect histone modifications. Even though there is no direct evidence, these metabolites and the epigenetic changes that they are capable of inducing could potentially be responsible for the observed changes in stem cell function with age. Another mechanism through which the composition of the microbiota can mediate its effect is through modulation of host immunity and the accompanied inflammatory responses, which can clearly impact on the epigenetic landscape, as discussed in the previous section (59). There is increasing evidence which show that the genera composition of the microbiota is very different between young and aged individuals. With age there is an overall reduction in the beneficial microbes with a concomitant increase in microbiota normally associated with inflammation (60), reiterating the causal role of microbiome composition in aging-associated inflammation. Another indirect effect the microbiome has on aging is through DNA damage. For instance, deoxycholate is a microbiota-mediated metabolite that has been shown to induce senescence of gastric cardia stem cells in the gut through DNA damage (61,62). All in all, the microbiota composition, the plethora of factors secreted by them and the state of the host immunity, all play a role in determining stem cell behavior and the underlying epigenetic remodeling during aging (Fig. 1).

Stem-cell Niche

Systemic factors such as inflammatory cytokines, growth factors and metabolites are all also secreted by the cells residing in the stem cell niche. The cells of the stem cell niche are known to affect and control resident stem cell self-renewal and differentiation. This has been observed in multiple stem cell niches such as those of HSCs, MuSCs, NSCs and ISCs. (63–66) The cellular composition and organization of niche cells changes during aging and can cause deregulation of this niche-mediated stem cell maintenance, as has been observed in HSC and NSC-niches (67,68). In addition to changes in the niche-cell composition, the sensitivity of adjacent stem cells to the paracrine factors secreted by the niche-cells can also be affected. For instance, expression of the chemokine receptor CXCR4 has been shown to be reduced during aging in murine MSCs. The resulting lack of responsiveness to the CXCR4 ligand, CXCL12, compromises the function of MSCs by inducing elevated ROS, DNA damage and reduced proliferation (69). This compromised functionality of the HSC niche, in turn drives aging of HSCs (69). In addition to being an example for niche-mediated aging of stem cells, this also serves as an example of one stem cell population directly affecting the aging of another. Along these lines, MSCs and other stromal cells secrete various cytokines that impact HSC maintenance through paracrine signaling (70). Another example for this is Paneth cells within the intestine, that are known to secrete important factors such as Wnt3a and epidermal growth factors which modulate ISC differentiation (71). Since transcription of these secreted factors is regulated by the niche-cell epigenome, the epigenetic state of aged niche cells can in turn contribute to the aging of the adjacent stem cells. In terms of spatial organization of the stem cell niche, it has been found that specific locations in their niche can protect a subset of stem cells from aging-associated phenotypes. Quiescent HSCs that reside in perisinusoidal niches in mice have been observed to be protected from functional defects during aging. These HSCs also maintain their H4K27ac epigenetic polarity, which is normally disrupted with aging. (72) This shows the influence niche organization has on stem cell aging and its epigenetic integrity. However if such well-organized niche structures are disrupted, it might also expose stem cells to stress-mediated functional damage. For example, injury-mediated stress responses in aged MuSCs has been found to affect their functionality through acute epigenetic upregulation of Hox-genes (73). Even though there is a lack of studies on niche-cells affecting stem cell epigenome during aging, there is some sparse data on this topic using DNA methylation-based age prediction tools. This technique, known as an epigenetic clock, has been extensively studied by multiple groups and can be used to predict the chronological age of a cell (74). Recently, this tool was used in a 17 year follow up study performed on patients who received bone marrow transplantation. The DNA-methylation age of blood cells in these recipients were found to resemble that of the donor’s chronological age rather than that of the recipient’s chronological age. (75) This data seems to indicate that the epigenetic age of stem cells is independent of its niche, at least at the level of the DNA methylome clock. However, further studies in this area are required to determine if the niche can indeed play a role in modulating the epigenome of stem cells.

Stem cell potency, epigenetic drift, and malignant transformation of stem cells

Embryonic stem cells (ESC) are characterized by their multipotency, in terms of their capacity to differentiate into cell types comprising the three different germ-layers. The state of multipotency is strongly associated with the role of epigenetic regulators in enforcing a specific epigenetic state. For example, studies have shown that chromatin modifiers such as SWI/SNF complex and PRC2 complex, activate genes involved in self-renewal and repress developmental or differentiation genes in ESCs, respectively (76,77). During the course of development, there is however a sharp decline in stem cell multipotency (Fig. 2) as the resultant adult stem cells are committed to the role of tissue-specific maintenance. This lineage-restricted multi- or oligo-potency of adult stem cells can apparently also become more restricted with increasing age. This is exemplified in the aging of the hematopoietic system, where there is a significant skewing towards a myeloid/megakaryocytic output of HSCs, termed as ‘myeloid-bias’ (78,79). A similar finding of reduced multipotency during aging has also been observed in aged bone marrow MSCs isolated from rats (80). These observations indicate that there is an overall decrease of stem-cell mutipotency leading to a reduced diversity of cellular output of stem cells during development and aging.

Figure 2

Clonal evolution of stem cells during development, aging and malignant transformation. (A) During development, there is an organized clonal diversification of cells starting from the embryonic stem cell. The resultant clonal diversity (represented here with each color denoting an individual stem cell clone) might facilitate the formation of well-defined cellular hierarchies in the organism. However, along the process of development there is a sharp decline in multipotency (1) of the stem cells to differentiate into cells of different germ-layers, as the adult stem cells take over the role of tissue-maintenance. (B) Multipotency continues to reduce further during aging, leading to further diversification of the adult stem cells in terms of differentiation potential. This generation of stem cell clones with reduced diversity of cellular output has been attributed to clonal selection and clonal expansion (2), where certain stem cell clones exhibit increased clonal dominance due to a competitive advantage. These steps of clonal diversification and later expansion with age, can be linked to the observed stochastic divergence of the epigenetic landscape among different cells, known as an epigenetic drift. Aging is accompanied by an ever increasing epigenetic drift (3). (C) Increase in epigenetic drift (3) during aging and the resultant epigenetic instability can also lead to the introduction of epi-mutations in certain stem cell clones. Mutations in epigenetic modifiers have been shown to drive aberrant expansion of specific pre-malignant stem cell clones. (D) Pre-malignant stem cell clones are thought to be responsible for cancer initiation through further epigenetic and genomic instability, resulting in an expansion of a highly malignant stem cell clone. As the tumor develops, the cancer stem cells further continue to diversify due to increasing epigenetic plasticity and generate a high intra-tumor clonal heterogeneity. However, the process of clonal diversification and expansion here are much more unorganized and deregulated compared to that during development.

One reason for the loss of multipotency observed during development and aging might be clonal diversification. Clonal sub-lines of ESCs generated through limiting dilution from progenies of a single mouse ESC have been observed to be highly diverse in their transcriptome and also differentiation potential (81). This would indicate that clonal diversification could start as early as the first divisions of ESCs. For example, expansion of certain cardiac progenitor cells are observed during mouse heart development (82). This form of clonal expansion might be a necessary factor for the establishment of clear cellular hierarchies during development. Like all processes associated with development, these clonal events are also likely to be well organized and tightly regulated through various molecular mechanisms. The process of clonal diversification and expansion are not only restricted to the process of embryonic development, but has also been observed in adults during aging (Fig. 2). Age-associated lineage-bias found in HSCs during aging serve an example of clonal diversification (78). Another example of this is in the intestine, where each colon crypt is functionally taken over by a single ISC clone during aging, referred to as intra-crypt clonal dominance (83). This can also serve as a prime example for clonal expansion, as intra-crypt clonal dominance is a result of expansion of a single ISC clone that has a competitive advantage over the other ISCs in the space-limited crypt niche (84). Another instance of clonal expansion and clonal selection of stem cells is the aforementioned myeloid-bias in aged hematopoiesis. This bias has also been linked to age-related epigenetic alterations such a reduction in the expression of the histone acetyltransferase KAT6B with age (85). Hence, these events of clonal diversification and expansion might have a strong link to diversification of the stem cell epigenome with age.

Similar to stochastic genomic mutations being acquired during aging, the epigenome also undergoes stochastic changes with age. This leads to a divergence of the epigenetic landscape between different cells in the body, which is referred to as epigenetic drift. The phenomenon of epigenetic drift is a defining factor of aging that leads to epigenetic mosaicism within different tissues of the body and within different stem cells. This process of epigenetic drift is also known to be exacerbated by other age-associated conditions such as chronic inflammation. (86) One example of epigenetic drift as a determinant of clonal diversification is the observed cell–cell DNA methylation heterogeneity with age in MuSCs. This epigenetic drift in MuSCs have also been shown to be contextual, with CpGs around LINE1 elements being more homogeneous and around H3K27me3 regions being more heterogeneous (87).

With clonal expansion, comes a predisposition for selection of highly proliferative but functionally-compromised stem cell clones that can invasively populate a tissue, leading to age-related disorders such as cancer (Fig. 2). The phenomenon of clonal hematopoiesis is a well-studied example of this. HSC clones with certain mutations, mostly in epigenetic regulators such as DNMT3A, TET2, EZH2 and ASXL1, have been shown to expand due to clonal selection in individuals that do not have any underlying hematological diseases. Clonal hematopoiesis predisposes these individuals to developing malignancies such as acute myeloid leukemia (AML) (88). It is thought that the expanded mutant clones act as a pool from which the leukemia cell of origin emerges, eventually giving rise to leukemic stem cells which can then propagate the disease. The over-representation of mutations in proteins that encode epigenetic modifiers, suggests that clonal hematopoiesis can represent a state of epigenetic instability, where cellular transcription programs become more plastic and therefore more divergent across different sub-clones. Thus, epigenetic plasticity results in the generation of a diverse pool of rapidly evolving pre-malignant cells on which clonal selection can act to drive malignant transformation. The malignant transformation of clonal stem cells with epi-mutations have also been reported in other cancers such as glioblastoma with H3K3A and IDH1 mutations (89). As a tumor develops, intratumor heterogeneity continue to increase (90–92). This mechanism of increased clonality and potential clonal expansion is likely a major mechanism that is responsible for therapy-resistance of cancers, due to the formation of more resistant-subclones (93). While this phenomenon has largely been studied in the context of progressive acquisition of DNA mutations, it is likely that epipolymorphisms may account for a much broader range of biologically relevant intra-tumour heterogeneity during the evolution of the disease. Polycomb complex proteins such as EZH2 (94–96), PRC1.1 (97), BMI1 (98) and DNA methyltransferases such as DNMT3A (99) are all key epigenetic regulators that play a role in modeling the cancer stem cell epigenome. Cancer is also driven by the reactivation of key developmental signaling pathways such as Wnt, Notch and Hedgehog signaling. These pathways are also epigenetically regulated (100), and switching-on these developmental epigenetic and transcription programs facilitates the maintenance of the cancer stem cell pool in many cancer entities (101). This can be viewed as a vicious strategy of cellular reprogramming of more differentiated tumor cells towards a developmentally younger and more aggressive stem cell-like stage. If cancer stem cells can undergo epigenetic changes to revert to a younger stem cell like epigenome, it begs the question whether it is possible to reprogram old stem cells to a younger epigenetic state, without leading to the deleterious outcome of transformation.

Cellular reprogramming as a method of stem cell rejuvenation

Reprogramming of adult somatic cells into an embryonic stem cell state, called induced pluripotent stem cells (iPSCs), was first described over a decade ago by the work of Takahashi and Yamanaka (102). They achieved this by overexpressing four transcription factors—Oct4, Sox2, Klf4 and c-Myc, in mouse fibroblasts. These four factors, now commonly referred to as the Yamanaka factors or OSKM, are sufficient to reprogram somatic cells into a stem-cell like form capable of differentiating into cells of all three germ layers. OSKM induced cellular reprogramming has also been observed to almost completely remodel the epigenetic landscape of cells (103). This includes a reestablishment of DNA methylation and repressive histone marks at lineage specifying genes and activating chromatin marks at pluripotency genes (104). The epigenome of the resultant iPSCs have also been shown to be similar to that of an embryonic stem cell (ESC) (105). Thus, one might consider that the process of OSKM-mediated reprogramming reverses the process of aging at the epigenetic level.

A recent study involving young adults, supercentenarians and progeroid individuals showed that iPSCs can be generated from donor cells from all these sources, regardless of age. The resultant iPSCs were very similar in their transcriptome and functional competence (106). This suggests that cellular reprogramming can completely reset cells at a functional and epigenetic level, irrespective of the chronological age or whether they have undergone ‘accelerated aging’ as a result of progeroid disease. Another important study demonstrated that old myeloid biased HSCs can be reprogrammed with OSKM successfully to iPSCs. The resultant iPSCs from these myeloid-biased HSCs were now capable of generating HSCs that can now repopulate all hematopoietic lineages in adult mice that result from blastocyst complementation with these iPSCs (107). Hence, not only is the reprogramming sufficient to induce molecular changes but also capable of reversing the functional defects associated with aging (Fig. 3). However, the process of complete reprogramming results in a pluripotent cell that may be difficult to efficiently differentiate into a functionally normal adult stem cell ex vivo, which carries that additional risk that residual pluripotent cells are able to initiate teratomas. Furthermore, all four factors used in cellular reprogramming are known oncogenes which may become reactivated (108–111), making full reprogramming a non-practical option for aged stem cell rejuvenation. The process of partial reprogramming of somatic cells has recently gained traction as a strategy to facilitate the rejuvenation of aged cells, which at least reduces the risk of loss of cell identity.

Figure 3

Complete reprogramming versus partial reprogramming in rejuvenation of aged cells. The process of complete cellular reprogramming (A) of aged somatic cells into induced-pluripotent stem cells (iPSCs) has been achieved through stable expression of the transcription factors Oct4, Sox2, Klf4 and c-Myc, known as Yamanaka factors or OSKM. This form of reprogramming leads to a reset of the functional and molecular consequences of aging such as genomic instability (1), mitochondrial dysfunction (2) and deregulated epigenetic landscape (3), to a young-cell-like state. However, complete reprogramming is also accompanied by an apparent loss of cellular identity, making it a non-viable option for age rejuvenation. Partial reprogramming (B), on the other hand works through a cyclic or transient expression of OSKM. This type of reprogramming can lead to a similar functional and molecular reversal of aging achieved through complete reprogramming, but without the deleterious loss of cell identity. Hence, a partial reprogramming approach seems to be a more capable and better means for age rejuvenation and rectifying certain age-related conditions such as stem cell exhaustion.

In 2016, Ocampo et al. were able to show that partial reprogramming can ameliorate signs of aging without completely losing cellular identity. This partial reprogramming strategy in mouse fibroblasts showed a significant reduction in mitochondrial ROS production, senescence and stress associated systemic factors such as IL-6, and DNA damage indicators such as γH2AX and 53BP1 foci (112) (Fig. 3). All of these are known players of stem cell aging and age-related epigenome remodeling. They were also able to show that cyclic expression of the Yamanaka factors in mice that model Hutchinson-Gilford progeria syndrome, led to an overall increase in lifespan of these mice. These partially reprogrammed mice further showed restoration of the number of MuSCs and of hair follicle stem cells to normal levels, suggesting its critical impact on aged stem cells. They could additionally show that cyclic injection of OSKM into injured muscle of aged mice, mediated increased muscle regeneration through activation of MuSCs, compared to untreated aged mice. These effects were also observed to be dependent on epigenetic remodeling (112). Their work shows that partial programming can be applied in vivo with very efficient modulation of aged stem cells in terms of epigenetic remodeling and rejuvenation of functional potency. Another recent piece of work by Sarkar et al. has also extended this idea of partial reprogramming to human cells. Transient non-integrative expression of OCT4, SOX2, KLF4, c-MYC, NANOG and LIN28 (OSKMNL) in aged human fibroblasts and endothelial cells was found to reset their transcriptome and methylome clock to resemble that of young cells (113). By expressing OSKMNL in chondrocytes from aged Osteoarthritis patients, they were also able to reduce age-related inflammatory response in these cells. Human MuSCs from aged individuals were further shown to have increased engraftment in mice after expression of OSKMNL compared to untreated MuSCs (113). This suggests that human stem cell regenerative capacity can also be positively modulated through partial reprogramming. All in all, partial reprogramming seems to be formally capable of reversing certain age-associated phenotypes including stem cell exhaustion and epigenetic deregulation. However, direct translation of this methodology to a clinical setting is clearly not yet feasible.

Conclusions and future perspectives

The study of compromised stem cell populations as an underlying driver of aging has received much attention, since the regenerative capacity of many tissues depends on the functional potency of their respective stem cell pool. Many studies have furthered our understanding of the different factors involved in this deterioration of stem cell function. With the advent of technologies to better interrogate the epigenome, numerous studies have been devoted to characterize and understand the changes that occur in the epigenetic landscape with age. The role of epigenome remodeling in functional decline of stem cells with age has been suggested by many, but conclusive proof of a direct causal relationship is yet to be presented. Nonetheless, various cell-extrinsic and intrinsic factors clearly play a role in a complex network that modulates the epigenome during aging, and in turn this is highly likely to impact upon stem cell maintenance and differentiation. However, much still remains to be elucidated regarding the precise mechanism via which such stimuli remodel the epigenome in stem cells. Indeed, the phenomena of epigenome plasticity, clonal heterogeneity and clonal outgrowth adds another layer of complication into the study of stem cell aging, which may be more amenable to study given the explosion in single cell ‘omics’ technologies.

From a translational perspective, young stem cells have higher potency and demonstrate fewer disease associated defects when compared to old stem cells. Therefore devising a method to safely and efficiently reverse stem cell aging has been postulated as an ideal solution towards the treatment or prevention of age associated conditions such as cancer, immune deficiency or metabolic syndrome. The potential involvement of epigenetic drift as a driving force behind stem-cell aging, has led to the idea of resetting the epigenetic clock of stem cells as a route towards cellular rejuvenation. Cellular reprogramming, in particular partial reprogramming, has been used in proof of concept studies where the epigenetic clock has been reset. However, this approach is not without caveats, most notably with regards to the capacity to safely reprogram somatic cells and safely introduce these back into the patient. In addition, such partial reprogramming of the epigenome will not be capable of erasing all biological consequences of aging, such as somatic mutations. In order to address these caveats, it will clearly be necessary to further development our understanding of the role that the epigenome plays in stem cell aging, and continue to refine technologies that can be used in resetting the epigenetic clock in a controlled or targeted manner.

Funding

MDM was supported by funding from the German Research Foundation (DFG) SFB873 and FOR2674; the Helmholtz Zukunftsthema Aging and Metabolic Programming (AMPro) ZT-0026; and the Dietmar Hopp Stiftung.