differentiation in development When human embryonic stem cell lines were first analyzed it became clear that they were naturally prone to differentiate in certain ways. Because they are capable of giving rise to any type of body cell, there was the possibility of artificially directing cellular differentiation in vitro. To direct differentiation toward a desired cell type, genes encoding suitable lineage-specific transcription factors could be transfected into the cells and overexpressed. A new era of regenerative medicine was envisaged where cells could be instructed to change into other cells, both in vitro and in vivo.
The idea of forcibly directing differentiation by overexpressing an appropriate transcription factor dates back to pioneering work by Harold Weintraub and colleagues in the late 1980s. They transfected mouse cultured fibroblasts with a cDNA encoding MYOD, and by overexpressing just this one transcription factor, were able to convert the fibroblasts into myoblasts. That result, although attracting considerable attention, still appeared unusual: there was considerable resistance to the idea that the epigenetic settings that determined the identity of a cell could easily be reset. Even by 2006, 10 years after the birth of Dolly the cloned sheep, the discovery that induced pluripotent stem cells could be created by overexpressing just four transcription factors was met with widespread amazement.
Since then there has been a paradigm shift: the idea that irreversible epigenetic marks are laid down as cells travel along the pathways of cell specialization is one whose time has come and gone. In hindsight, we should not have been too surprised: there are extraordinary examples of tissue regeneration in some species, and even in humans natural epigenetic reprogramming occasionally causes cells to change identity (Figure 1).
Fig1. Modes of epigenetic reprogramming to change cell identity. Red arrows signify naturally occurring modes of epigenetic reprogramming; blue arrows indicate artificial epigenetic reprogramming. Metaplasia—conversion of a differentiated cell to another of a different type—is common in some organisms but also occurs naturally in humans when cells are subject to extended physiological or pathological stress. Prolonged exposure to cigarette smoke, for example, can convert pseudostratified columnar epithelial cells of the airways into squamous epithelial cells, and gastroesophageal reflux causes squamous epithelial cells of the esophagus to convert to columnar epithelial cells. Artificial transdifferentiation can make larger changes in mammals, for example convertibility of fibroblasts and neurons. Dysplasia entails expansion of immature cells at the expense of more differentiated cells and is common in cancer where epigenetic changes cause cells to revert to undifferentiated states. (Adapted from Cherry A & Daley G [2012] Cell 148:1110–1122; PMID 22424223. With permission from Elsevier.)
Directed transdifferentiation
If artificial epigenetic reprogramming were to be used for therapeutic purposes, then inducing dedifferentiation to form a pluripotent stem cell followed by differentiation of the pluripotent cells to a suitable tissue progenitor cell might not seem to be the most efficient route. To replace insulin-producing pancreatic β cells, for example, it might be simpler to convert other pancreatic cells in vivo, or one might try to convert patient fibroblasts in vitro to some tissue progenitor cells that might simply then be directed toward the desired cell type in vivo.
Significant effort has therefore gone into changing the identity of a differentiated cell toward another desired cell type, a process known as transdifferentiation. The example of converting fibroblasts to myoblasts given above required a single transcription factor, but a variety of different transdifferentiations have been carried out and often require two or more transcription factors. And transdifferentiation has been possible between the three germ layers—ectoderm, endoderm, and mesoderm—as well as between cell types belonging to one germ layer (see Figure 2 for some examples). We consider the potential therapeutic applications in Chapter 22.
Fig2. Transdifferentiation by overexpression of transcription factors. (A) Examples of successful transdifferentiation of human and mouse cells, including conversion of cells belonging to different germ layers and to cells belonging to the same germ layer. (B) Examples of how mammalian fibroblasts can be programmed by combinations of transcription factors to give different cell types. As described in the following article, various miRNAs have also been overexpressed to direct transdifferentiation. (A, From Ladewig J et al. [2013] Nat Rev Mol Cell Biol 14:225–236; PMID 23847783. With permission from Springer Nature. Copyright © 2013; B, Adapted from Wang H et al. [2015] Differentiation 90:69–76; PMID 26525508. With permission from Elsevier.)
