Keywords

8.1 Introduction

The prognosis of patients with severe heart failure is extremely poor, and heart transplantation is the only effective treatment (Lund et al. 2015). However, lack of donors is a major problem worldwide. Cardiac regenerative therapy using human pluripotent stem cells (PSCs) may represent an effective alternative treatment option for heart transplantation. Human induced PSCs (iPSCs) have the potential to differentiate into various types of cells, similar to human embryonic stem cells (ESCs) (Takahashi 2007; Thomson et al. 1998), and may have applications as a new cell source for regenerative medicine in the context of various diseases, including severe heart failure (Burridge et al. 2012; Passier et al. 2008).

Although cardiac differentiation protocols have dramatically improved (Laflamme et al. 2007; Burridge et al. 2014; Lian 2012; Zhang et al. 2012; Willems et al. 2011; Minami et al. 2012), it may be impossible to stably differentiate into only target cells because many factors, including the specific cell lines used, affect differentiation efficiency (Kattman et al. 2011; Elliott et al. 2011; Osafune et al. 2008). Moreover, cardiac regenerative medicine using human PSCs will require hundreds of millions of cardiomyocytes. The use of this many cells increases the risk of contamination with residual PSCs or noncardiac proliferating cells, which is a major cause of tumor formation (Hentze et al. 2009; Miura et al. 2009; Kawamura et al. 2016; Zhang et al. 2014). Thus, many technologies have been developed to prevent tumor formation in cardiac regenerative medicine, including selective elimination of residual PSCs (Fig. 8.1a) and complete purification of cardiomyocytes (Fig. 8.1b).

Fig. 8.1
figure 1

Strategies for prevention of tumor formation. (a) Selective elimination of residual PSCs. (b) Purification of only cardiomyocytes from PSC-derived mixed cell populations

Full size image

In this chapter, we introduce these two strategies and discuss the use of these approaches for safe cardiac regeneration.

8.2 Elimination of Residual Pluripotent Stem Cells

Many studies have described methods for selective elimination of residual PSCs that have the capacity for teratoma formation (Fig. 8.1a). This strategy could theoretically have applications in all fields and is discussed in more detail in the following sections.

8.2.1 Cell Sorting by Stem Cell Markers

Separation strategies based on cell sorting using fluorescent-activated cell sorting (FACS) or magnetic-activated cell sorting (MACS) have been reported to eliminate residual undifferentiated PSCs. In such cell sorting methods, human PSC surface markers, such as TRA1–60, SSEA-4, and SSEA-5, are used (Fong et al. 2009; Tang 2011). In addition, claudin 6, a tight-junction protein specific for human PSCs, is also a useful surface marker for selective elimination of residual human PSCs through FACS (Ben-David et al. 2013). While these strategies are simple, they are not suitable for large-scale culture because they require single-cell dissociation, which would be a time-consuming process when sorting a large number of cells.

8.2.2 Small Molecules or Toxins

Many studies have reported the elimination of undifferentiated human PSCs by utilization of toxins or small molecules. Of the toxins commonly used for this purpose, podocalyxin-like protein 1, a primary cytotoxic antibody for human PSCs, can eliminate residual PSCs (Choo 2008; Tan et al. 2009). In addition, Clostridium perfringens enterotoxin, which binds to claudin 6, has been reported to eliminate undifferentiated PSCs (Ben-David et al. 2013). Recently, Tateno et al. identified a human PSC-specific lectin (rBC2LCN) by glycome analysis and created a recombinant lectin-toxin fusion protein (rBC2LCN) using the catalytic domain of Pseudomonas aeruginosa exotoxin A (Tateno et al. 2015). This fusion protein could be specifically taken up into human PSCs and could kill residual PSCs.

Of the small molecules commonly used for elimination of undifferentiated PSCs, Bieberich et al. showed that ceramide analogs induce apoptosis and eliminate residual PSCs; indeed, treatment with ceramide analogs prevents teratoma formation after transplantation (Bieberich et al. 2004). Lee et al. targeted the hPSC-specific anti-apoptotic factor survivin and demonstrated that inhibition of survivin could selectively eliminate pluripotent stem cells with teratoma potential (Lee 2013). Using chemical screening, Dabir et al. identified a small molecule that inhibits the translocation of redox-regulated proteins to the mitochondria and showed that this small molecule induces apoptosis in human ESCs but not in differentiated cells (Dabir Deepa et al. 2013). Furthermore, Ben-David et al. identified a small molecule that inhibits the biosynthesis of oleic acid and specifically kills human PSCs using a screening library of more than 50,000 small molecules (Ben-David 2013).

These strategies using small molecules or toxins have many advantages because they are simple, efficient, and applicable for large numbers of cells and do not require single-cell dissociation. However, the cost of these strategies may be high owing to the need for large amounts of antibodies or small molecules. In addition, the components of these systems may affect other PSC-derived differentiated cells.

8.2.3 Metabolism

Improving our understanding of metabolic processes in human PSCs is necessary in order to remove undifferentiated tumor-forming cells by exploiting the metabolic environment (Fig. 8.2). Many studies have examined glucose metabolism in mouse and human PSCs (Kondoh et al. 2007; Panopoulos et al. 2011; Folmes Clifford et al. 2011). Folmes et al. reported that human iPSCs exhibit characteristics of elevated glucose utilization compared with mouse embryonic fibroblasts (MEFs) and that inhibition of glucose metabolism reduces the reprogramming efficiency of the cells (Folmes Clifford et al. 2011). Our group also showed that mouse and human PSCs mainly depend on activated glycolysis for ATP and biomass production and that glucose deprivation efficiently removes residual PSCs (Tohyama 2013).

Fig. 8.2
figure 2

Metabolism in human PSCs. Human PSCs depend on glucose and glutamine metabolism. Glycolysis contributes to ATP and biomass (amino acids and nucleotides) production. Glutamine metabolism contributes to not only ATP generation via OXPHOS but also to the maintenance of pluripotency via reduced glutathione synthesis. Methionine metabolism plays a role to produce S-adenosyl-methionine (SAM) that leads to maintain pluripotency via histone methylation. G6P glucose-6-phosphate, 3PG glycerate 3-phosphate, Gln glutamine, Glu glutamate, αKG α-ketoglutarate

Full size image

In contrast, few studies have examined the effects of amino acid metabolism on mouse or human PSCs. Shyh-Chang et al. reported that mouse ESCs are critically dependent on threonine catabolism, which is important for synthesis of S-adenosyl-methionine (SAM) and nucleotides (Shyh-Chang et al. 2013). Threonine starvation leads to decreased SAM levels, resulting in inhibition of histone H3K4 trimethylation and preventing mouse ESCs from maintaining pluripotency (Shyh-Chang et al. 2013; Wang et al. 2009). Additionally, Shiraki et al. evaluated the effects of essential amino acid deprivation on cell survival in human PSCs and found that methionine deprivation was the most effective inhibitor of human PSCs. They also reported that methionine is the main source of SAM production in human PSCs (Shiraki et al. 2014). Furthermore, Moussaieff et al. revealed that glucose-derived cytosolic acetyl-CoA contributes to the maintenance of pluripotency by induction of histone pan-acetylation (Moussaieff et al. 2015), and Carey et al. reported that naïve mouse ESCs utilize both glucose and glutamine catabolism to maintain a high level of intracellular α-ketoglutarate (αKG), which promotes histone and DNA demethylation and maintains pluripotency (Carey and Finley 2014). Recently, our group demonstrated that glutamine oxidation during the later steps of the tricarboxylic acid (TCA) cycle plays a key role in cell survival in human PSCs. In glucose-depleted conditions, glutaminolysis activation is increased, thereby promoting ATP production via oxidative phosphorylation (OXPHOS). Interestingly, human PSCs cannot utilize pyruvate efficiently because the expression levels of metabolic enzyme-related genes in the early steps of the TCA cycle are low, whereas those involved in the synthesis of cytosolic acetyl CoA are high. As a result, glucose deprivation and glutamine deprivation are most effective for elimination of residual human PSCs (Tohyama et al. 2016). Glutamine metabolism also contributes to synthesize reduced glutathione that plays a role in maintenance of pluripotency via prevention of OCT4 (Marsboom et al. 2016). Similar to approaches using small molecules and toxins, these approaches have many advantages, such as simplicity, efficiency, and suitability for large-scale culture, and do not require single-cell dissociation. Furthermore, these approaches are not expensive because they do not utilize antibodies or small molecules. However, these metabolic approaches also have the potential to cause damage to the other PSC-derived differentiated cells. Therefore, supplementation with alternative metabolites may be required to minimize the effects on other cells (Tohyama 2013).

8.3 Purification of Target Cells

Recently, both undifferentiated PSCs and other immature proliferating cells have been shown to have potential for tumor formation (Nori et al. 2015). In addition, contamination with noncardiomyocytes may induce arrhythmia after transplantation. Thus, complete purification of cardiomyocytes derived from human PSCs is necessary for safe realization of cardiac regenerative medicine (Fig. 8.1b).

8.3.1 Genetic Manipulation

Several studies have reported fluorescent protein expression-based purification of cardiomyocytes derived from mouse PSCs using various combinations of cardiomyocyte-specific promoters (e.g., αMHC, Mlc2v, Nkx2–5, and ANP) and reporters (e.g., green fluorescent protein [GFP]) (Gassanov et al. 2004; Anderson et al. 2007; Huber et al. 2007; van Laake et al. 2010). In humans, Elliott et al. introduced sequences encoding enhanced GFP (EGFP) into the NKX2–5 locus (NKX2–5-EGFP) by homologous recombination (Elliott et al. 2011). Furthermore, Ma et al. generated hiPSCs expressing a blasticidin-resistance gene under the control of the MYH6 promoter (MYH6-blasticidin) and obtain pure cardiomyocytes (Ma et al. 2011). While these methods are useful for basic research, they are not suitable for clinical application because they lack stability and safety. Therefore, it is necessary to establish nongenetic methods of purifying cardiomyocytes for clinical applications.

8.3.2 Nongenetic Cell Sorting

In nongenetic cell sorting strategies, some groups have attempted to obtain cardiac progenitor cells, whereas other groups, including ours, have attempted to isolate only cardiomyocytes. Yamashita et al. succeeded in obtaining mouse ESC-derived Flk-1-positive mesodermal cells, which could differentiate into several mesodermal lineages, including cardiomyocytes, smooth muscle cells, and endothelial cells (Yamashita et al. 2000). Hidaka et al. reported that prion protein and platelet-derived growth factor (PDGF) receptor α double-positive cells derived from mouse ESCs could differentiate into cardiomyocytes and smooth muscle cells (Hidaka et al. 2009).

On the other hand, it is difficult to isolate only cardiomyocytes because cardiomyocyte-specific surface markers have not yet been identified. Therefore, to isolate pure cardiomyocytes, we focused on the structural characteristics of cells rather than surface markers; using this approach, we succeeded in developing a nongenetic cardiomyocyte purification method (Hattori 2010). In short, because cardiomyocytes have many mature mitochondria and high mitochondrial membrane potential, we successfully purified cardiomyocytes (>99% purity) derived from mouse and human PSCs by a combination of FACS and the mitochondrial dye tetramethylrhodamine methyl ester perchlorate (TMRM). The fluorescence intensity of TMRM dye disappeared within 1 day, while that of other mitochondrial dyes was sustained over 5 days. Therefore, the effects of TMRM dye were suppressed. Other groups have also established nongenetic cardiomyocyte purification methods using FACS or MACS with antibodies against cell surface markers, including ALCAM (CD166) (Rust et al. 2009), signal-regulatory protein alpha (SIRPA) (Dubois et al. 2011), and vascular cell adhesion molecule 1 (VCAM1) (Uosaki 2011). Although cell sorting methods using antibodies or mitochondrial dyes are useful for the production of small numbers of cardiomyocytes, these methods are time consuming when using human PSC-derived mixed cell populations as the source cells (Fig. 8.3a). In addition, these methods require single-cell dissociation, which can damage the target cells, and transplantation of target cells with antibodies may result in immunogenicity. Therefore, further studies are needed to establish methods for scalable production of human PSC-derived pure cardiomyocytes for clinical applications.

Fig. 8.3
figure 3

Purification of only cardiomyocytes from PSC-derived mixed cell populations. (a), Purification of cardiomyocytes by a combination of FACS and antibodies or dyes. (b) Purification of cardiomyocytes by metabolic culture conditions

Full size image

8.3.3 Metabolic Selection

To establish an ideal method for scalable production of human cardiomyocytes for clinical applications, our group aimed to purify cardiomyocytes using specific metabolic culture conditions in which only cardiomyocytes and not residual PSCs can survive (Fig. 8.3b). To evaluate metabolic differences between PSCs and cardiomyocytes, we performed metabolome and transcriptome analyses. As mentioned above, we found that the PSCs mainly depended on activated glycolysis and that glucose deprivation could eliminate residual PSCs. However, because glucose-depleted conditions are also fatal for cardiomyocytes, supplementation with an alternative energy source is necessary for survival of cardiomyocytes. Interestingly, glucose and lactate are major energy substrates in fetal hearts, while fatty acids are major energy substrates in adult hearts based on the levels of energy substrates in the blood (Neely and Morgan 1974). Because PSC-derived cardiomyocytes show a fetal phenotype (Uosaki et al. 2015), we hypothesized that PSC-derived cardiomyocytes could efficiently utilize lactate for energy production and showed that mouse and human PSC-derived cardiomyocytes could survive under glucose-depleted and lactate-supplemented conditions. Moreover, because human PSC-derived noncardiac proliferating cells also depended on glycolysis like PSCs and cannot survive under these conditions, we were able to obtain pure cardiomyocytes (>95%) under these conditions (see Fig. 8.4).

Fig. 8.4
figure 4

Metabolic differences in human PSCs and PSC-derived cardiomyocytes. Under glucose- and glutamine-depleted conditions with pyruvate or lactate supplementation, human PSCs cannot utilize pyruvate efficiently because of low gene expression during the early steps in the TCA cycle. In contrast, human PSC-derived cardiomyocytes can efficiently use lactate-derived pyruvate because of high gene expression during the early steps in the TCA cycle

Full size image

As mentioned above, our group has recently reported that human PSCs depend on glycolysis and glutamine oxidation for ATP generation. Glucose and glutamine deprivation enabled complete removal of human PSCs in a much shorter period (Tohyama et al. 2016). Surprisingly, lactate supplementation could rescue only human PSC-derived cardiomyocytes because cardiomyocytes efficiently utilize lactate not only for ATP generation via OXPHOS but also for glutamate synthesis under glucose- and glutamine-depleted conditions. In short, lactate can compensate for the lack of intermediate metabolites and overcome the problem of cell damage in human PSC-derived cardiomyocytes, whereas residual human PSCs cannot utilize lactate-derived pyruvate, as mentioned above (Tohyama et al. 2016). In addition, most of the obtained pure cardiomyocytes were myosin light chain 2v (MLC2v)-positive ventricular cells. This metabolism-based method has the following advantages: (1) suitability for large-scale production of pure cardiomyocytes (Fonoudi et al. 2015; Hemmi et al. 2014), (2) simple procedure without specialized instrumentation, (3) low cost of culture medium, and (4) high yield of target cells (Aalto-Setala et al. 2015).

8.3.4 Other Nongenetic Methods

Xu succeeded in enriching human PSC-derived cardiomyocytes using a Percoll density gradient procedure (Xu et al. 2006), yielding cultures containing 35–66% cardiomyocytes. Nguyen et al. reported that the formation of cardiospheres derived from human PSCs enabled the enrichment of cardiomyocytes to over 80% (Nguyen Doan et al. 2014).

Ban et al. reported the purification of cardiomyocytes from mouse and human PSCs by a combination of FACS and molecular probes consisting of 15–30-bp dual-labeled oligonucleotides with a fluorophore and a quencher. In short, molecular probes could be used to identify and visualize cardiomyocyte-specific mRNA in live cells (Ban 2013). Recently, Miki et al. succeeded in establishing an efficient method for purifying cardiomyocytes based on endogenous microRNA (miRNA) activity (Miki et al. 2015). They utilized synthetic mRNAs encoding a fluorescent protein with sequences targeted by cardiomyocyte-specific miRNAs and purify cardiomyocytes with or without FACS.

Conclusions

To realize safe cardiac regenerative medicine using human PSCs, it is important to provide systems for producing target cells with high quality and sufficient quantity. Based on this requirement, metabolic selection systems may be an ideal method to efficiently obtain large numbers of cardiomyocytes derived from human PSCs (Tohyama 2013; Tohyama et al. 2016). This method using glucose-depleted media is also applicable to drug screening and elucidation of pathogenesis using patient-specific iPSCs (Burridge et al. 2016; Kodo et al. 2016; Matsa et al. 2016; Hinson et al. 2015; Dudek et al. 2015). At the same time, methods are needed to detect tumor-forming PSCs with higher sensitivity. Several studies have reported the detection of PSCs at a ratio of 0.001–0.01% (Tano et al. 2014; Kuroda et al. 2012). Further studies are needed to determine whether this sensitivity is sufficient for evaluation of the safety of techniques for regenerative medicine. Recent studies showed that human PSC-derived transplanted cardiomyocytes could electrically integrate with the host heart (Shiba et al. 2012; Gerbin et al. 2015) and mature over time (Hattori 2010; Chong et al. 2014; Funakoshi et al. 2016). Although the effectiveness of transplantation of human PSC-derived cardiomyocytes has been demonstrated in large animals (Chong et al. 2014; Ye et al. 2014; Kawamura et al. 2012), there is a risk of ventricular arrhythmia (Chong et al. 2014; Shiba et al. 2014). While the mechanism is unknown, there are two major possibilities: contamination with noncardiac cells derived from human PSCs and immaturity of PSC-derived cardiomyocytes. Large-scale purification methods for cardiomyocytes may yield solutions for overcoming both of these challenges in order to realize safe cardiac regenerative medicine.