Multiple Paths to Reprogramming

1. Introduction

For a long time, differentiation was considered a “one way process”; Conrad Waddington, in the 1950s, described cellular differentiation and development as a ball rolling towards different one-way ramified valleys, giving rise to specific cell fates, irreversibly [1]. However, in the last decades, a series of studies have shown that somatic cells and stem cells are more plastic than previously believed. Using different technical approaches, the epigenetic barriers imposed during development in differentiated cells can be erased, and cells can re-acquire pluripotency through a process, known as “reprogramming”.

The first evidence came at the end of the 1950’s from the pivotal experiments performed by J.B. Gordon in the zoology department at Oxford University [2]. At that time, embryologists, not aware of epigenetic regulation, i.e. the role of chromatin and its crucial modifications in cell fate determination, wondered whether development and cellular differentiation arise upon specific restriction of the genetic information contained in their nuclei. To answer this basic but intriguing question, Gordon used a technique, now known as somatic cellular nuclear transfer (SCNT) in Xenopus laevis laevis. For these experiments, nuclei from intestinal epithelial cells of albino tadpoles were transferred into unfertilized and enucleated wild-type frog oocytes. This resulted in the development of normal albino frogs, which in some cases were also fertile. They concluded from these studies that adult nuclei contain the genetic information necessary for the development of a frog. Moreover, cellular differentiation, during development, does not occur through loss of genetic information. These findings were exciting for the scientific community but at the same time controversial for two reasons: I) the efficiency with which a ‘cloned’ frog reached the adult stage was around 1% and II) the same technique did not work with mammalian cells [3].

For forty years, the scientific community was not able to use SCNT in other species. Finally, in 1997, by using the same technique, Ian Wilmut and colleagues, at the Roslin Institute in Edinburgh, Scotland, succeeded in generating the sheep named “Dolly” by SCNT, further confirming that genetic modifications, leading to cellular differentiation, are not irreversible. Two key improvements in his technical strategy led to the first cloning of a mammal: unfertilized oocytes were used as recipients and donor cells were induced to exit from the normal cellular cycle, by serum withdrawal [4, 5].

One year later, Wakayama and colleagues [6] reported that SCNT also allowed the cloning of the most used animal model, the mouse. Again, another technical advance led to this progress: the use of an enucleation pipette, which allowed for the removal of the nucleus from the oocytes. This advance also allowed the conclusion that reprogramming factors are not oocyte-specific, meaning that SCNT can be done also using zygotes and fertilized eggs [7], and that the molecules responsible for reprogramming were present in the cell cytoplasm. In general, nuclear transfer (see Figure 1) involves two steps: a) de-differentiation of a somatic donor cell to an embryonic state and the in vitromaturation till the blastocyst stage (also known as therapeutic cloning); b) the further development of the cloned blastocyst, after the implantation in the maternal uterus (reproductive cloning).

Therapeutic cloning permits the derivation of nuclear transfer derived embryonic stem cells (ntESCs). Recently, the efficiency of isolation of ntESCs drastically increased, at least in mice, from 1% to 20% [8]. It has also been possible to derive similar cells in cats, dogs, wolves, goats and monkeys. Although the isolation of human ntESCs has been reported, this paper has been retracted later on [9]. Thus, the possibility of therapeutic cloning with human cells needs to still be demonstrated. However, the therapeutic cloning remains a promising technology for regenerative medicine, considering that ntESCs, from other species, were able to differentiate into all the cell types of an adult body.

Reproductive cloning is technically more difficult than therapeutic cloning, as it involves the further development in vivoin a pseudopregnant female. Embryos, derived after SCNT, develop till blastocyst stage with a good efficiency (20-50% depending on the species), but most of them die in the post-implantation stage, without reaching birth (1-5% of survival rate to birth, depending on the species).

Analysis of the cloned animals also showed several abnormalities: increased telomere shortening (which may have caused the premature death of Dolly), altered gene expression during development, prolonged gestation, fetal or placental edema, increased risk of obesity and cancer. The reasons for these pathologies remain not fully understood. The defects may be due to infidelity of the reprogramming: residual epigenetic memory of the donor cell may be present an/or imprinting of important developmental genes may be altered. Nevertheless, reproductive cloning remains attractive and may have potential implications in agriculture and industrial biotechnology. However, as it relates to humans, cloning (also therapeutic) remains controversial as theoretically, it may allow the cloning of a human being.

Evidence that differentiation is reversible also comes from another technique, known as cell fusion [10]. In cell fusion experiments, two or more cells can be fused together (by using polyethylene glycol (PEG) or electrofusion) to generate a single cell, called heterokaryon or hybrid. The larger or more dividing cell type is the “dominant” one, and the “recessive” cell will convert its gene expression profile to the one imposed by the dominant cell type. Obviously, alteration in the ratio of the two cell types during fusion will affect the final fate of the fused cells [11].

A heterokaryon, produced by inhibiting cell division, is a fused cell that becomes multinucleated and survives only short-term. If the cell cycle is not blocked, the fused cells will form a hybrid, because upon the first division the two different nuclei will become a single nucleus, having 4n chromosomes (see Figure 2). Its karyotype can be: 1) euploid, when fused cells are from same species (the two cell types have the same number of chromosomes, thus, their segregation will be balanced); 2) aneuploid, when cells fused are from different species (the two cell types have a different number of chromosomes, thus, they will be lost and/or rearranged).

Cell fusion experiments advanced medical knowledge on cell plasticity. In 1969, Harris et al. fused tumor cells with normal cells and demonstrated that there are trans-acting oncosuppressor genes. Upon fusion, malignancy was suppressed and this was not due to the loss of an oncogene, as after prolonged in vitroculture malignancy re-emerged [12]. Fibroblasts can be induced to produce albumin or melanin if fused with hepatocytes [13, 14] or melanocytes [15, 16], respectively.

In 1983, Blau et al. [17] produced for the first time heterokaryons from diploid human amniocytes fused with differentiated mouse muscle cells. She demonstrated that the heterokaryons express many human muscle-related genes and that this activation was mediated by factors present in the cytoplasm (as non-dividing heterokaryons have distinct nuclei). Similar heterokaryons with muscle cells can be produced not only by fusing them with amniocytes but also with cells of the three embryonic lineages (mesoderm, ectoderm, endoderm) [18].

In 1997, Surani, Tada and colleagues demonstrated, by producing proliferative hybrids that cell fusion not only ”switches” the fate of different cell types but also “reprograms” them to a pluripotent state. Thymocytes from adult mice were fused to embryonic germ cells, pluripotent stem cells derived from primordial germ cells (PGCs). By using DNA sensitive restriction enzymes, they demonstrated that the genome of the somatic cell underwent a general demethylation, with reactivation of imprinted and non-imprinted genes, resembling the reprogramming events occurring in germ cell development [19].

They also fused female thymocytes, derived from Oct4-GFP mice, with mouse ESCs [17]. Two days after fusion, expression of GFP, from the thymocyte, was detected. The X chromosome, normally silenced in adult female cells, was reactivated. Moreover, hybrids had developmental potential, like ESCs, as they contributed to the three germ layers of chimeric animals, upon blastocyst aggregation [18]. Using the same approach, in 2005, Cowan succeeded in creating hybrids between human somatic cells and human ESCs [19].

This further elucidated that the differentiation state of cells is plastic and reversible; both SCNT and cell fusion experiments clearly demonstrated that it is possible to reset the epigenetic landscape of somatic cells. Despite all these studies were already present in the literature, the field of reprogramming only became jumpstarted in 2006 when Takahashi and Yamanaka [20] demonstrated that the overexpression of pluripotency-related transcription factors (TFs) can dedifferentiate adult fibroblasts to induced pluripotent stem cells (iPSCs), iPSCs strongly resemble ESCs. iPSC technology is an inefficient process, but differently from SCNT or cell fusion, may have in the near future therapeutic applications, including human disease modeling, drug screening and patient-specific cell therapy (see Figure 3).

After this publication [20], several studies demonstrated the potential of epigenetic reprogramming. Indeed, there is now evidence that use of different “cocktails” of TFs allows not only to redirect fibroblasts to an ESC-fate but also to a lineage-specific cell types/precursors, like cardiomyocytes, neuronal precursors, hepatocytes and blood cells, from a tissue different than the tissue from which the somatic cell was isolated [21-23].

Finally, It has also become clear that cell plasticity and reprogramming may be partially achieved or enhanced by the culture microenvironment. An increasing number of studies is showing how small molecules, including epigenetic modifiers and signaling pathway modulators, play a crucial role in cell-fate determination [24]. All together, these studies highlight that culture media influences the epigenetic-state of the cells in which they are cultured and thus their features [25]. In this chapter, we will discuss:

1. Reprogramming to the pluripotency-state, describing transcription factors used in mouse and human, different methodologies and potentiality of iPSC technolology

2. Lineage conversion, illustrating the differentiated cell types/precursors obtained and the differences of this approach with the iPSC technology

3. Culture mediated reprogramming, providing the published data which highlight the influence of culture media and small molecules on stem cells fate and features

2. Reprogramming to the pluripotency-state

In 1987, two key discoveries highlighted how crucial the role is of some “master” TFs in cell fate determination. During Drosophila Melanogasterdevelopment, the gene Antennopedia (Antp)specifies the formation of the thoracic segment, which will then form the legs of the adult fly. Heat-induced overexpression of Atnp,at specific larval stage, led to the formation of additional legs instead of antennae [26]. Similarly, ectopic expression of eyeless(the homologous of Pax6in mice) caused the development of functional eyes on the wings, antennae and legs of drosophila [27].

Twenty years after these pivotal experiments, Shinya Yamanaka’s group used the same TF-based technology to reprogram adult fibroblasts to pluripotent state [20]. Mouse embryonic and adult fibroblasts, transduced with retroviral vectors encoding for Oct4, Sox2, Klf4 and c-Myc and cultured in ESC-medium, erased their differentiated epigenetic state and reestablished the pluripotent state; these cells were named induced pluripotent stem cells (iPSCs). Murine iPSCs exhibited morphological and growth properties of ESCs, and expressed alkaline phosphatase and SSEA1. No differences, if compared with ESCs, could be detected in their methylation status, X activation status, embryoid body (EB) formation, in vitrodifferentiation capacity (ectodermal, mesodermal and endodermal), teratoma formation and in vivodevelopmental potential (contribution to the three germ layers of chimeric animals).

Noteworthy, pluripotent stem cells (PSCs) possess mechanisms that lead to the silencing of the integrated transgenes. Therefore, the expression of the four TFs is necessary for generating iPSCs but dispensable for maintenance of the iPSC fate; hence, pluripotency and selfrenewal capacity rely on the trans-activation of the endogenous genes, suggesting a true and complete reprogramming. Mouse iPSCs, like ESCs, were germline competent [28] and supported the development of a mice in tetraploid complementation assay [29]. In this assay, embryonic cells at the two cell-stage are fused together. This results in a tetraploid blastocyst in which just the extraembryonic tissues will further develop; by complementing the tetraploid embryo with normal diploid PSCs, it is possible to generate an individual, completely derived from the diploid PSCs. Interestingly, the same combination of TFs [30] or a somewhat different one (Oct4, Sox2, Nanog and Lin28) can be used for the reprogramming of human cells [31].

This discovery is groundbreaking because with iPSC technology, PSCs can now be induced/derived for autologous cell transplantation, avoiding immunological problems and ethical issues related to the use of human ESCs. In addition, iPSCs from patients carrying a disease can be derived and used to better understand the biological problem leading to the disease as well as for drug-screening.

3. Lineage conversion

The discovery of iPSCs, together with previous experiments involving SCNT and cell fusion, clearly showed that differentiation is a reversible process and that cells are more ‘plastic’ than previously believed. Therefore, a new field, called lineage reprogramming, emerged rapidly in the last five years. Recent attempts have demonstrated that, the forced overexpression of TFs can also convert one cell type to another of the same or of other somatic germ layers. Lineage reprogramming depends on the capacity of certain TFs to overcome the existing epigenetic barriers and to rapidly initiate the new identity-specific gene network [110-113].

Examples of direct lineage conversion were described already in 1986; Davis, Lassar et al. [114, 115] converted different fibroblast lines into myogenic cells by overexpression of MyoD, a basic helix-loop-helix transcription factor, in combination with the demethylating agent, 5-azacytidine. Subsequent studies confirmed that myogenic conversion, as shown by presence of desmin and myosin heavy chain, could be achieved in vitrostarting from a variety of cell types (adipose, melanoma, neuroblastoma and liver cell lines). However, the complete downregulation of the ‘original’ tissue-specific genes was only seen when starting with mesodermal cells and not with endodermal or ectodermal cell lines [116, 117].

A similar transdifferantiation was also seen in the blood system. The deletion of Pax5 in pro-B cells resulted in their switch into the T-cell lineage [118, 119]. Later on, the same group investigated this transdifferentiation more extensively; mature B cells were isolated from Pax5 knockout mice and transplanted back into a lymphocyte deficient recipient. Surprisingly, they could detect in the reconstituted mice donor pro B cells, which then gradually converted into T cells [120]. This demonstrated that lost of Pax5 led to a T cell phenotype through de-differentiation rather than direct transdifferentiation. Another example of direct conversion came from the work done by Graf and colleagues [121]; overexpression of C/EBPα orC/EBPβ,a basic leucine zipper TF binding CCAAT enhancers, induced a macrophage phenotype (as shown by Mac1 expression) in bone marrow or spleen-derived B cells. In the induced macrophages almost all the B cell genes analyzed were downregulated and cells acquired phagocytic function in vitro.

The above examples describe experimental conversions but there were also cases in which this conversion occurs naturally. Jarriault et al. [122], demonstrated that the epithelial rectal cell ‘Y’, migrates anterodorsally from the rectum to become a ‘PDA’ motor neuron. This conversion from Y to PDA is not direct but occurs through a de-differentiation state, in which the initial (Y-cell) and the final (PDA-cell) identity are not present [123]. In this section, we will describe the relevant cell types, recently, obtained by lineage reprogramming.

4. Culture mediated reprogramming

Reprogramming to the pluripotency-state and lineage conversion are achieved through the forced expression of TFs. However, in the last decade, several reports have highlighted how culture medium per se can be responsible for (partial) reprogramming. Moreover, there is an increasing amout of evidences showing that small molecules, including epigenetic modifiers and signaling pathway inhibitors, enhance the efficiency and kinetics of reprogramming.

Epiblast stem cells (EpiSCs) are isolated from post-implantation embryos between E5.5-E7.5. EpiSCs are the post-implantation equivalent of ESCs; they still express Oct4, Nanog and Sox2 but express lower levels of Stella and Rex1 [147]. ESCs and EpiSCs have also different culture requirements and features. While ESC selfrenewal is LIF dependent, EpiSC proliferation requires bFGF and Activin signaling. EpiSC female lines, but not ESC lines, have one of the X chromosome inactive. Importantly, EpiSCs, differently from ESCs, do not have the ability to contribute to chimeras in vivo, when aggregated into recipient morula/blastocysts.

In 2009, Bao et al. [148] demonstrated that established EpiSC lines could de-differentiate/revert into an ESC-like state by culturing EpiSCs in ESC medium (cointaining LIF) for 2-5 weeks. Once ’reverted’ cells lost all the features of the original EpiSCs and acquired all ESC-characteristics ( X was reactivated, growth was LIF-dependent and cells were capable to contribute to chimeras). This report showed that the simple manipulation of culture medium can dedifferentiate EpiSCs to a more primitive ESC-state but this is not the only case reported in literature.

In 2004, Kanatsu-Shinohara et al. [149] descibed that mouse germline stem cells (GSCs) isolated from neonatal testis reverted occasionally into cells with ESC-like colonies morphology within 4-7 weeks if cultured in LIF, epidermal growth factor (EGF), glial cell line-derived neurotrophic factor (GDNF) and fibroblast growth factor 2 (FGF2). The reverted cells were named multipotent germ stem cells (mGSs); they expressed not only Oct4 (already present in GSCs) but also Nanog and Sox2 at ESC-level. Analysis on mGSs showed the loss of spermatogonial properties (although the erasure of the androgenic imprinting was not complete) and the gain of ESC features (teratoma formation and contribution to chimeras with germline transmission). However, despite their siimilarity to ESCs, mGSs were not able to form offspring, after tetraploid complementation. Unipotent germline stem cells, but this time from adult testis, were converted into germline-derived pluripotent stem cells (gPSs) by Ko and colleagues [150]. Reprogrammed cells, like in the above case, were higly similar to ESCs but again, they could not form live animals in tetraploid complementation assay. The reason for this is most likely the residual persistence of androgenetic imprinting. The possibility to reprogram a germline stem cell into a cell with pluripotent features, even without the capacity of forming chimeric animal, is interesting because it might allow autologous cell therapy without embryo-manipulation. Similar conversions with mouse cells were also described by other laboratories [151, 152].

In 2008, Conrad and colleagues [153], showed that cells derived from human testis can be converted into cells with human ESC-like features. Cells isolated from human testis were cultured in GDNF-containing medium for 4 days and then selected based on the expression of CD49f and further selection on laminin matrix in medium cointaining LIF. 3-4 weeks later colonies with ESC-morphology appeared; human adult GSCs (haGSCs), like human ESCs, expressed SSEA4, TRA 1-60, TRA 1-81 and generated EBs and teratomas. However, a later report [154] questioned the previous finding of Conrad, arguing whether haGSCs really expressed Oct4, Nanog and Sox2; moreover, microarray data comparison further showed that haGSCs are similar to fibroblasts-derived from human testic biopsies but not to hESCs.

These studies strongly suggest that stem/progenitor cells derived from testis can to some extent be converted, by long-term culture, to cells with ESC-like properties, without any reprogramming factors; however, converted cells differ significantly from ESCs. The propensity of GSCs to be converted to ESC-like cells may depend on their Oct4 expression. Although gonads are the only place where Oct4 is functionally expressed in adult healthy-rodents [127], many reports described the isolation of Oct4⁺ cells from rodents [155-169]. It remains to be determined whether culture mediated reprogramming is responsible for the Oct4 re-activation in such cell lines.

In 2002, our group [170] isolated multipotent adult progenitor cells (MAPCs) from rodent bone marrow (BM), upon prolonged culture at low density in a medium cointaining LIF, PDGF and EGF. Murine MAPCs differentiated in vitrointo cells of the three germ layers and one murine line was also able to contribute to chimeric mice, although at low efficiency and without germline trasmission. Subsequently analysis on rodent MAPC lines [171, 172] showed a lineage marker profile (Oct4, Gata4, Gata6, Sox7 and SSEA1) found also in the nascent hypoblast of the blastocyst and in rat blastocyst-derived Extraembryonic Endoderm Precursor cells (XEN-Ps) [173]. Recently, we demonstrated that similar cells are not present in fresh BM but appear after prolonged in vitroculture. To ascertain whether the MAPC culture system reprograms BM cells to the equivalent of XEN-P, we, first, showed that rMAPC and XEN-P cells exhibit similar features under reciprocal culture conditions. Second, we reported, using the same MAPC medium, the quick and efficient isolation of new cell lines directly from blastocyst, which we termed Hypoblast Stem Cells (HypoSCs) and which strongly resemble XEN-P in features and developmental potential [174].

Moreover, specific culture media may also be responsible for the broader differentiation potential described for some adult stem cell types [175] and this should be more considered in stem cell research, especially when reaching clinical trials phases [176].

5. Conclusions

The importance and the impact on society of reprogramming has been recently recognised by the Nobel Assembly at the Karolinka Institute of Stockholm, which co-awarded John Gurdon and Shiniya Yamanaka with the Nobel Price in Medicine 2012. Their outstanding reports demonstrated that cellular fate is plastic and that differentiation is a reversible process. Epigenetic markers imposed by development can be erased through the multiple pathways to reprogramming. This means the epigenetic landscape as described by Waddington should be revised, as balls are capable of rolling back up and over the hill. The SCNT and the forced expression of TFs show that somatic cells can re-acquire all the features, lost upon their differentiation. Adult somatic cells can be redirected to the pluripotency-state or can be converted into cells of another lineage.

Although the precise mechanism via which the phenotype of all these cells can be changed remains to be fully elucidated, the iPSC technology is drastically changing and boosting the stem cell field; it allows one to obtain pluripotent stem cells for autologous therapy, avoiding the problems of immune rejection as well as the ethical issues related to the use of human embryo for scientific purposes. The possibility to also obtain precursors, with restricted differentiation potential, may be another alternative to reach the bedside, as it is likely associated with lower tumorigenicity. It is also clear that culture conditions have such a significant effect on cell fate, not only during reprogramming but also in establishing the potential of different adult stem cells, that this should be kept in mind when comparing studies across laboratories, and definitely when contemplating clinical trials with cultured stem cells.

Acknowledgments