Epigenetic modifications such as histone acetylation and DNA methylation play a paramount role in regulating gene expression and exhibit unique changes during aging and age-related disease (Fraga et al., 2007; Johnson et al., 2012). Modifications to epigenetic machinery can directly impact longevity (Lin et al., 2005) and health (Klein et al., 2011) as well as prevent differentiation of stem cells into somatic tissues (Bröske et al., 2009), highlighting the importance of a well-functioning epigenome. Emerging studies suggest that iPSCs may harbor a higher number of genetic and epigenetic abnormalities than both ESCs and the somatic cells that they originate from (Pera, 2011). Moreover, there are mixed data regarding the epigenetic memory of iPSCs and whether this memory affects the differentiation potential of reprogrammed cells (Fig. 1).
Figure 1. Epigenetic memory and reprogramming. There are controversial data regarding the epigenetic memory of induced pluripotent stem cell (iPSC)s and whether or not this memory affects the differentiation potential of reprogrammed cells. iPS cells have been reported to feature incomplete epigenetic reprogramming compared to ESCs, retaining residual methylation signatures characteristic of their tissue of origin that favor differentiation into lineages related to the donor cell.
Download figure to PowerPoint
It was recently shown that low-passage iPSCs can feature incomplete epigenetic reprogramming compared to ESCs, retaining residual DNA methylation signatures that are characteristic of their tissue of origin and favor differentiation into lineages related to the donor cell (Fig. 1). iPSCs derived from mouse neural progenitors, for example, contained methylomic signatures at loci important for hematopoietic differentiation, resulting in a decreased propensity for differentiating into hematopoietic cell types. Treatment with chromatin-modifying compounds reduced DNA methylation at these loci and increased the blood-forming potential of the low-passage iPSCs, suggesting that the effects of these epigenetic marks can be attenuated via pharmaceutical intervention (Kim et al., 2010). Conflicting data exist regarding the retention of these methylation signatures with passage number. Some iPSC clones derived from human neonatal keratinocytes and umbilical cord blood cells were documented to maintain tissue-specific methylation memory at high passage numbers (Kim et al., 2011), while iPSCs derived from mouse myogenic cells, fibroblasts, and hematopoietic cells reportedly lost their epigenetic memory with continued passage in culture (Polo et al., 2010). More recently, genetically matched human iPSC clones from dermal fibroblasts and bone marrow stromal cells of the same donor were generated and differentiated into osteogenic and chondrogenic lineages. The authors found that, although the iPSCs exhibited an epigenetic memory characteristic of the donor tissue used, the clones varied in their differentiation propensity. Moreover, no correlation was found between the cell type of origin and the propensity of an iPSC clone to differentiate into bone and cartilage (Nasu et al., 2013). Further work is required to determine whether this epigenetic memory affects the pluripotency of iPSCs and whether this influence declines with passage number or varies with the donor tissue or iPSC line used.
Unlike stem cells, somatic cells have a limited division capacity and senesce in vitro. The inability to further replicate is termed replicative senescence and can be induced by a plethora of factors, such as short and uncapped telomeres, oxidative stress, and mitochondrial DNA damage (Fig. 2) (Chen et al., 2007). The enzyme telomerase, which maintains telomere length and long-term self-renewal potential of stem cells, is strongly expressed in ESCs and is inactive in most somatic cells. As such, telomere length gradually decreases with every cell division in a typical somatic cell, eventually resulting in replicative senescence (Harley et al., 1990). The lifespan of normal human fibroblasts can be extended in vitro by exogenous introduction of plasmids expressing the catalytic subunit of telomerase hTERT, resulting in an increased telomerase activity (Bodnar et al., 1998). Donor cells that are difficult to reprogram can be more efficiently induced into the pluripotent state by adding hTERT and SV40 large T antigen to the reprogramming factors originally used by Yamanaka's laboratory (Park et al., 2008). Furthermore, telomerase-deficient mice exhibit a sharp reduction in reprogramming efficiency that can be restored by the reintroduction of telomerase (Marion et al., 2009), highlighting the imperative role telomerase plays in iPS reprogramming.
Figure 2. Aspects of aging and reprogramming. There are currently conflicting data regarding the ability of reprogramming to fully rejuvenate an aged somatic cell and reverse age-related changes such as DNA damage, shortened telomeres, and dysfunctional mitochondria. Moreover, contentious data exist suggesting that cells derived from induced pluripotent stem cells may be subject to premature senescence.
Download figure to PowerPoint
Through the use of factor-based reprogramming, Yu et al. reported that human iPSCs display levels of telomerase activity characteristic of ESCs (Yu et al., 2007). In a recent report using mouse iPSCs, it was found that, at passage eight, iPSCs have shorter telomeres than ESCs, yet longer telomeres than the embryonic fibroblasts they were derived from. It was only after subsequent cell divisions and further passaging that telomeres were restored to the lengths found in ESCs (Marion et al., 2009). In six of seven iPS cell lines derived from human skin fibroblasts, telomeres were substantially elongated compared to their parental cells and had telomere lengths comparable to ESCs at passage five (Suhr et al., 2009) (Table S1). At passage 25, several iPSC clones derived from human neonatal foreskin fibroblasts had significantly longer telomeres than their parental donors (Yehezkel et al., 2011). iPSCs reprogrammed from senescent and centenarian human cells exhibited telomere lengths that were identical to or greater than the lengths observed in ESCs and did not shorten with increased passage number (Lapasset et al., 2011) (Table S1).
A study conducted by Vaziri et al. compared telomerase activity and telomere length in iPSCs to hESCs. The authors reported that five well-studied human iPS cell lines had significantly shorter telomeres than three commonly used ES cell lines as well as reduced levels of telomerase activity. Six iPS cell clones were then generated from the ESC-derived cell line EN13 and, after further culturing, telomeres in five of the six lines shortened to lengths comparable to those observed in the widely used iPS cell lines. One clone, however, expressed high levels of telomerase and displayed a progressive increase in telomere length for over 60 days, procuring a length equivalent to that of the parental ES cell line (Vaziri et al., 2010). Mathew et al. also reported the variability in telomere length and telomerase expression in human iPSC clones. Only clones with the highest level of telomerase expression were observed to have telomere lengths comparable to ESCs (Mathew et al., 2010). In iPSCs derived from patients with dyskeratosis congenita, a disease characterized by shortened telomeres and defects in telomerase function, reprogramming was capable of restoring telomere length in some iPSCs (Agarwal et al., 2010), but not others (Batista et al., 2011) (see Table S1 for a detailed overview of telomere length data). For those iPSCs where telomere length was elongated, telomere length was observed to increase with further passaging (Agarwal et al., 2010).
When telomere chromatin is transcribed, long noncoding RNA transcripts referred to as telomeric-repeat-containing RNA (TERRAs) are generated. TERRA levels are positively correlated with telomere length and may act as regulators of telomerase expression (Schoeftner & Blasco, 2008). In mouse iPSCs, TERRAs were observed to be increased compared to differentiated cells, yet lower compared to ESCs at passage eight (Marion et al., 2009). At passage 25, human iPSCs had higher TERRA levels than their parental source, yet these levels were found to vary from clone to clone (Yehezkel et al., 2011).
These results indicate that, although telomerase activity is clearly reconstituted during the reprogramming process, there is a significant variability in telomere length among various iPS cell lines. This length is sensitive to passage number and can even vary among cell lines derived from the same tissue or parental cell type. This is not unique, however, as it was recently shown that substantial variability in telomere length and telomerase expression exists for ESCs as well. For both iPSCs and ESCs, telomere length was found to be highly correlated with proliferation efficiency and developmental pluripotency (Huang et al., 2011). As such, caution should be made when drawing comparisons for telomere length, ensuring that passage number is controlled for and that appropriate controls (e.g., numerous ES cell lines and donor tissue) are used.