Lost in the debate about ethical and moral dilemmas surrounding human embryonic stem cell research are the considerable practical difficulties with using such cells as a tool in regenerative medicine. In 2006 Yamanaka and his colleagues reported that fully developed somatic cells can be induced to form pluripotent stem cells (iPSCs) that are functionally similar to embryonic stem cells. This groundbreaking discovery offered a possible solution to the controversial use of human embryos, but also the potential of deriving autologous cells from the patients who would be the ultimate recipients and beneficiaries of the regenerative products of their own iPSCs, reducing the problem of tissue rejection.
Capture, amplification, and differentiation of lineage-committed progenitor cells may ultimately prove more practical [than iPSCs] for tissue regeneration.
Progress since that initial discovery has been rapid, with major advances in increasing reprogramming efficiency and safety, as well as a much better understanding of the reprogramming process itself. However, the finding that inactivation of p53, the “guardian of the genome,” increased the efficiency of the reprogramming process was a harbinger of later studies reporting evidence that major deleterious genomic changes were associated with the reprogramming process. This damage included increased frequencies of copy number variations (CNVs) as well as point mutations. Culturing of cells reduced the number of cells with CNVs, suggesting that many of these variations were deleterious, at least under competitive growth conditions in vitro.
What might be the source of this damage? The epigenetic modifications of DNA that occur during embryonic development and are associated with lineage commitment usually involve 5-methylation of cytosine in CpG-rich sequences associated with the regulatory region of most genes. Methylated cytosine recruits repressor proteins that suppress transcription of proximal genetic elements. A well-characterized family of DNA methyltransferases (DNMTs) is responsible for de novo methylation and its maintenance, resulting in relative stability of DNA methyl “marks” in somatic cells. Very early in embryonic development, the expression of key genes associated with the pluripotent state is rapidly extinguished by methylation. An essential element of cellular reprogramming is wide-scale erasure of DNA methylation—notably demethylation of endogenous pluripotency genes like nanog and Oct4. While many cellular modifications, such as phosphorylation and protein methylation, are readily and directly reversible, demethylation of DNA is not so straightforward.
Reprogramming of fully differentiated somatic cells to form induced pluripotent stem cells (iPSCs) is achieved by transduction of four reprogramming genes—Oct4/c-Myc/Sox2/Klf4—and subsequent culturing over a period of ~30 days. During this process, some of the cells successfully de-differentiate due to erasure of methylated marks on endogenous pluripotency genes such as Oct4. There are several variations on the method of gene transduction and on the composition of the exogenous genes.
DNA demethylation can be characterized as passive or active. Passive demethylation generally occurs when DNA is replicated and DNMT1 (which has a preference for hemi-methylated DNA) is inhibited from methylating the newly synthesized strand. If inhibition of DNMT1 persists through another round of DNA replication and cell division, cells will contain regions of unmethylated DNA. Active demethylation may be achieved by several mechanisms, most of which share the requirement for DNA repair, typically base-excision repair. (For an excellent review of demethylation processes, see Wu & Zhang, Nat Rev Mol Cell Biol, 11:607-20, 2010.) In mammals, it appears likely that removal of 5-methyl cytosine residues is brought about through conversion of the methylated cytosine to an intermediate nucleotide that must then be excised and replaced by a new cytosine. While DNA repair is an extraordinarily faithful process, it is not entirely foolproof, and rare mistakes may contribute to the genetic changes observed when comparing iPSCs with the cell source from which they are derived.
The discovery of significant numbers of genetic changes and mutations in iPSCs raises concerns about the safety of their use in regenerative medicine and about their authenticity as tools for modeling diseases. Several questions naturally emerge. For example, could damage incidence be reduced by increasing the fidelity of base repair during reprogramming? Secondly, is DNA damage an inevitable corollary of reprogramming, and if so, supposing it were possible to eliminate the risk of damage, could demethylation be achieved at all?
The first question might be addressed by inducing the expression of repair genes to increase the cellular capacity for and fidelity of repair, although this is unlikely to be practical for human therapeutic applications. The second might be approached by promoting passive demethylation through temporary inactivation of DNMT1 during the period when the four exogenously introduced reprogramming genes—Oct4/c-Myc/Sox2/Klf4—are expressed. The activity of both DNMT1 and its relative DNMT3 is essential for development, and mice lacking a functional gene for either one are embryonic lethal; hence the need for reversible repression. Suppression of maintenance methylation during the generation of iPSCs would allow accumulation of demethylated DNA, a natural process that occurs during the first few days after fertilization. However, since active demethylation will still occur, a positive result of this test would be a reduction in mutational frequency. If this is found to be the case, additional means to promote the ratio of passive to active demethylation might be sought, including lengthening the period for reprogramming (since this has been found to help select against certain genetic abnormalities).
The overall importance of completely preventing DNA damage during reprogramming remains an open question. Ideally, pristine iPSCs should be the goal, but this is an issue of calculating the risk-benefit ratio, which is dependent upon the unique situation of each recipient patient. Nonetheless, minimization of mutations will be required for approval of clinical trials.
These concerns may become moot if reprogramming can be avoided in the first place, as an even larger hurdle is efficient generation of therapeutically useful, differentiated cell types from iPSCs. In this respect, capture, amplification, and differentiation of lineage-committed progenitor cells may ultimately prove more practical for tissue regeneration. Indeed, inactivation of DNMT1 in pancreatic ß cells helps convert them to glucagon-producing ß-like cells, suggesting that the inhibition of maintenance methylation may be a means to manipulate progenitor cell fate (Dev Cell, 20:419-29, 2011).
Jim Woodgett is director of research at the Samuel Lunenfeld Research Institute at Mount Sinai Hospital in Toronto, Canada. His interests include understanding of signaling systems, such as the Wnt pathway, that modulate cell fate, pluripotency, and differentiation.
- S. Wu, Y. Zhang, “Active DNA demethylation: many roads lead to Rome,” Nat Rev Mol Cell Biol, 11:607-20, 2010.
- Y. Mayshar, et al., “Identification and classification of chromosomal aberrations in human induced pluripotent stem cells”, Cell Stem Cell, 7:521-31, 2010.
- S.M. Hussein, et al., “Copy number variation and selection during reprogramming to pluripotency,” Nature, 471:58-62.
- A. Gore, et al., “Somatic coding mutations in human induced pluripotent stem cells,” Nature, 471:63-7, 2011.
- L.C. Laurent, et al., “Dynamic changes in the copy number of plurippotecy and cell proliferation genes in human ESCs and iPSCs during reprogramming and time in culture,” Cell Stem Cell, 8:106-18, 2011.
- S. Dhawan et al., “Pancreatic b cell identity is maintained by DNA methylation-mediated repression of Arx,” Dev Cell, 20:419-29, 2011.