Epigenetic reprogramming and induced pluripotency

Abstract

The cloning of animals from adult cells has demonstrated that the developmental state of adult cells can be reprogrammed into that of embryonic cells by uncharacterized factors within the oocyte. More recently, transcription factors have been identified that can induce pluripotency in somatic cells without the use of oocytes, generating induced pluripotent stem (iPS) cells. iPS cells provide a unique platform to dissect the molecular mechanisms that underlie epigenetic reprogramming. Moreover, iPS cells can teach us about principles of normal development and disease, and might ultimately facilitate the treatment of patients by custom-tailored cell therapy.

Fig. 1.

The developmental potential and epigenetic states of cells at different stages of development. A modification of C. H. Waddington's epigenetic landscape model, showing cell populations with different developmental potentials (left) and their respective epigenetic states (right). Developmental restrictions can be illustrated as marbles rolling down a landscape into one of several valleys (cell fates). Colored marbles correspond to different differentiation states (purple, totipotent; blue, pluripotent; red, multipotent; green, unipotent). Examples of reprogramming processes are shown by dashed arrows. Adapted, with permission, from Waddington (Waddington, 1957).

Fig. 2.

Examples of transcription factor-mediated reprogramming. Hierarchy of cell populations (blue shading) that appear during normal development and their relationship to each other (green lines). Dashed red lines illustrate examples of transcription factor-induced reprogramming. The bracketed cMyc gene indicates that this factor is dispensable for reprogramming. ICM, inner cell mass; ES, embryonic stem cell; iPS, induced pluripotent stem cell.

Fig. 3.

Steps involved in direct reprogramming to pluripotency. The starting, intermediate and end stages of reprogramming to pluripotency that can be identified during the generation of iPS cells are shown. `Intermediate cells' appear only transiently before converting into iPS cells, whereas `partially reprogrammed cells' can be stably propagated and converted into iPS cells upon treatment with DNA demethylating agents and knockdown of lineage-specific genes. Although not proven, it is assumed that partially reprogrammed cells originate from transient intermediate cells. The defining molecular and cellular characteristics are shown above and below each cell population.

Fig. 4.

ES cell transcription factor network and implications for reprogramming. (A) The reprogramming factors Oct4, Sox2 and Klf4 (light blue) often co-bind promoter regions with other transcription factors, including Nanog, Nr0b1 (nuclear receptor subfamily 0, group B, member 1), Esrrb (estrogen-related receptor, beta), Zfp281 (zinc finger protein 281) and Nac1 (nucleus accumbens associated 1; all of which have been purified in large protein complexes with Oct4 or Nanog), as well as with Stat3 and Smad1 (transcription factors downstream of the Bmp4 and Lif signaling pathways that maintain ES cell self-renewal and pluripotency) (Chen et al., 2008; Kim et al., 2008a; Wang et al., 2006). The recruitment of co-activators, such as the histone acetyltransferase (HAT) p300 is often observed (yellow) (Chen et al., 2008). This binding pattern is found in transcriptionally active genes in ES cells. ES cell target groups and implications for reprogramming are also indicated. (B) In ES cells, genes bound by either Oct4, Sox2 or Klf4 are often repressed, potentially through the recruitment of Polycomb group (PcG) proteins or histone deacetylases (HDACs), but become activated upon differentiation (Liang et al., 2008; Lee et al., 2006). (C) cMyc is proposed to bind and activate largely different sets of genes to Oct4, Klf4 and Sox2, but in collaboration with other transcription factors (Kim et al., 2008a; Chen et al., 2008).

Fig. 5.

Pathways to DNA demethylation of key pluripotency genes. (A) The establishment of symmetric DNA methylation patterns could be prevented passively during replication by the steric hindrance of Dnmt1 due to the stochastic binding of the reprogramming factors to target sites or by inhibiting Dnmt1 function indirectly. Hemimethylation of the DNA would result in a progressive loss of methylation upon further rounds of cell division. (B) Alternatively, DNA methylation could be actively removed by the recruitment of a demethylating enzyme.