Author’s address (for correspondence): Tiziana A.L. Brevini, Laboratory of Biomedical Embryology – UniStem, Center for Stem Cell Research, Universitá degli Studi di Milano, Faculty of Veterinary Medicine, via Celoria 10, 20133 Milan, Italy. Tel: 0039-02-50317980, Fax: 0039-02-50317989;. E-mail: email@example.com
Huge amounts of work have been dedicated to the establishment of embryonic stem cell lines from farm animal species since the successful isolation of embryonic stem cells from the mouse and from the human. However, no conclusive results have been obtained so far, and validated lines have yet to be established in domestic animals. Many limiting factors have been suggested and need to be studied further to isolate truly pluripotent cell lines from livestock. In this review, we will discuss the difficulties in deriving and maintaining embryonic stem cell lines from farm animal embryos and how can this lack of success be explained. We will summarize results obtained in our laboratory regarding derivation of pluripotent cells in the pigs. Problems related to the identification of standard methods for derivation, maintenance and characterization of cell lines will also be examined. We will focus our attention on the need for appropriate stemness-related marker molecules that can be used to reliably investigate pluripotency in domestic species. Finally, we will review data presently available on functional key pluripotency-maintaining pathways in farm animals.
The derivation of embryonic stem cells (ESCs) in farm animal species could represent a great improvement in developmental biology as well as applied biomedical research. In particular, farm animal ESCs could provide a powerful tool for genetic engineering aimed at improved production traits and products, for disease resistance and for biopharming (Keefer et al. 2007). These cells could also represent an excellent experimental model in pre-clinical trials, where the feasibility and the clinical potential of stem cell therapies could be studied. The close morphological and physiological resemblance to humans could be greatly advantageous for creation of biomedical models of human diseases and cell transplantation therapies. Although huge amounts of work have been dedicated to the establishment of ESC lines from farm animals over the past 20 years, no conclusive results have been obtained so far. Indeed, although progress towards the derivation of ESCs from porcine, bovine, caprine, ovine and equine species has been made and despite the many peer-reviewed journal articles describing ungulate ES or ES-like cell lines over the past 20 years, validated lines have yet to be established in farm animal species.
Why is it so difficult to derive and maintain ESCs from farm animal embryos? How can this lack of success be explained? Several concomitant factors have been suggested that hamper the process and still need to be elucidated. In this review, we will discuss some of the issues to be addressed and the hurdles to be overcome, related to the identification of standard methods for the derivation, maintenance and characterization of cell lines, the use of appropriate markers and elucidation of functional key pluripotency-maintaining pathways (Fig. 1). In particular, we will focus on the information available for the pig, which is a desirable species to create pluripotent cell lines, because of its striking similarities in terms of anatomy, physiology, metabolism and organ development with primates, and may therefore represent an important pre-clinical model for cell therapy approaches (Yang et al. 2000; Aleem Khan et al. 2006; Brevini et al. 2007a; Keefer et al. 2007).
Timing of Derivation
The first differentiation event in the mouse occurs at the late morula stage, when cells forming the outer compartment adopt an epithelial fate and the first two lineages, trophectoderm and inner cell mass (ICM), can be appreciated. Further differentiation of the ICM leads to formation of the epiblast and the primitive endoderm. In the interval between days 3.5 and 4.5, the hypoblast begins its evolution towards the extra embryonic endoderm and the yolk sac, while the epiblast develops into the embryo proper. The timing of this process is different in other species, and in particular, the interval between fertilization and the formation of the three early embryonic lineages is more protracted in porcine and bovine embryos, where epiblast formation begins at hatching and is complete towards day 12 (Vejlsted et al. 2005, 2006). This implies that no defined epiblast is present in these species before hatching. However, once formed, it will persist until day 13, when the major part of the epiblast differentiates into neural ectoderm and forms the neural plate. This process is accompanied by a gradual down-regulation of OCT4 and the onset of beta-tubulin III expression, which is correlated with neural differentiation, but, at the same time, indicates that embryos at this stage are no longer suitable for ESC derivation. Altogether, these observations indicate that farm animal species have a longer interval during which the epiblast is accessible and suggest that ICMs obtained from farm animal pre-hatching blastocysts may not be exactly equivalent to mouse epiblasts, usually isolated for ESC derivation.
What is the optimal time for isolation and establishment of porcine ESC lines? The available data are controversial. Chen et al. (1999) showed that recently hatched blastocysts had a higher success rate for establishment of pig ES-like cell cultures when compared with late-hatched blastocysts. Very poor outgrowth formation ability was detected in 5- to 6-day embryos, whereas 10- to 11-day embryos led to establishment of ES-like cell lines (Hochereau-de Reviers and Perreau 1993). Pluripotent epiblast cell cultures were obtained from early porcine blastocysts as well as from later-stage embryonic discs. Furthermore, no difference in the failure to inhibit spontaneous differentiation of the epiblast cells or in the inability to propagate epiblast cells was noted across these time points (Talbot et al. 1993a; Keefer et al. 2007).
Experiments conducted in our laboratory confirmed the possibility of establishing stable pluripotent cell lines using 6- to 7-day blastocysts. This was demonstrated both using in vivo- and in vitro-derived blastocysts, as well as parthenogenetic embryos that enabled successful establishment of cultures (Brevini et al. 2007b, 2010).
Cell-to-cell dissociation is a complicating factor in the culture of mammalian ESCs, with the exception of the mouse. Generally, in mice, dissociation protocols involve the use of trypsin–EDTA that does not seem to affect cell plating efficiencies (Robertson 1987). In contrast, enzymatic and chemical dissociation of primate ESCs give re-plating efficiencies of <1% (Thomson and Marshall 1998) and seem to render cells more prone to chromosome abnormalities (Cowan et al. 2004; Draper et al. 2004). This sensitivity seems to be more evident in ESCs derived from domestic animals. Primary cultures of undifferentiated ungulate epiblast cells are extremely sensitive to lysis, by either physical manipulation, withdrawal of calcium or exposure to trypsin–EDTA (Talbot et al. 1993a, 1995; Talbot and Garrett 2001). A short exposure of pig epiblast cell cultures to dissociating agents has a very damaging effect, with complete lysis in <10 min. In particular, exposure to Ca+2/Mg+2-free PBS for only 5 min results in cell rupture and lysis and complete cell disintegration after 30–60 min (Talbot and Garrett 2001). Interestingly, it has been reported that pig epiblast cells can be viably dissociated from each other if saline is used instead of PBS and if a rapid reattachment of the dissociated cells to a solid substrate is fostered (Keefer et al. 2007). This inability to resist dissociation from one another is a critical point for porcine ESCs passaging protocols. In line with this requirement, we mechanically disaggregate our pig ICMs and passage our pig cell lines using micro-loops and mechanical pipetting. We ensure not to disaggregate to single cell suspensions and always keep cell clumps with an average of 500–600 cells. This allows a more efficient re-plating of cells that can be maintained in culture for extended period of time, with a low apoptotic rate (11%) and rare differentiating events.
There are minimal data regarding appropriate culture conditions needed by ESCs derived from domestic species. On the contrary, much information is available for mouse and human ESCs, and these conditions have been frequently tested on putative ESCs in other species. It is well known that murine ESCs need a feeder layer to keep them pluripotent. Consistent with this, several authors demonstrated that for the survival of pig and bovine epiblast cells, it is necessary to have a murine feeder layer, for example, STO cells or mouse embryonic fibroblasts (Talbot et al. 1993a, 1995). Without these supporting cells, cultures of primary pig epiblast cells failed to grow, and instead, senesced and died over a 10- to 14-day interval (Keefer et al. 2007). In agreement with this, in our experience, porcine ICMs cultured in the absence of STO fibroblasts or on gelatin alone grew slowly and tended to differentiate and increase in size. Similar results were reported with feeder-free, short-term, primary cultures of pig ICMs, with or without the addition of leukaemia inhibitory factor (LIF) to the medium (Moore and Piedrahita 1997), indicating that this growth factor cannot substitute feeder cells for pig ESC line establishment. However, several studies demonstrated that the use of porcine embryonic fibroblasts does not appear to have any beneficial effect on the establishment of porcine ESCs (Hochereau-de Reviers and Perreau 1993; Wianny et al. 1997; Li et al. 2004). Porcine ICMs plated and grown on embryonic cells derived from the same species, in fact, were unable to be maintained for longer than 11 passages. Until now, the reason for these differences remains unknown, but murine feeder-layer requirement does not seem to be related only to the release of specific factors by the feeder-cell populations, but its presence appeared essential to ensure good culture conditions.
Another fundamental aspect related to the ICM ability to attach on feeder layer seems to be expression of adhesion molecules. We recently demonstrated a higher ability of parthenogenetic-derived ICMs to form outgrowth, compared with their bi-parental counterpart (Brevini et al. 2010). Although, further studies are needed to better elucidate this aspect, our preliminary data indicate that one possible explanation of this trend may be found in the up-regulation of beta integrin-1 and vitronectin gene expression detected in parthenogenetic ICMs (Brevini et al. 2010).
Consistent with these data, previous studies carried out in mouse ESCs have shown that altered expression of molecules involved in the implantation cascade that stabilize cell-to-cell adhesion leads to a reduced ability of the cells to adhere and to maintain a stable association with fibroblast feeder layers (Fassler et al. 1995).
A major goal at present is to develop better culture formulations to obtain homogenous pluripotent outgrowths from ungulate embryos and identify the best in vitro environment that would facilitate derivation of stable ESC culture. It is clear from earlier studies that culture conditions needed to sustain porcine ESCs have not been well defined.
At present, most laboratories use protocols similar to those described in the mouse, and ungulate ESCs are grown on a feeder layer, in medium supplemented with various other nutrients or components like basic fibroblast growth factor (bFGF) (Strelchenko 1996; Yadav et al. 2005), LIF (Strelchenko 1996; Saito et al. 2003), epidermal growth factor and stem cell factor (SCF) (Saito et al. 2003). These additions to the culture medium are performed without evidence of the presence on pig ESCs of the related receptors and are mainly based on previous experience in the murine species. At present, there is continued debate in the literature as to whether LIF and bFGF are required for sustained putative porcine ESC cultures (Fig. 2). Basic fibroblast growth factor exerts its effects, binding to one or more of its several specific receptors (Sperger et al. 2003; Ginis et al. 2004). We recently demonstrated the expression of FGFR-2 in porcine pluripotent cell lines (Brevini et al. 2010). In contrast, we found that porcine pluripotent cell lines do not express LIF receptors. This seems to indicate that the addition of LIF into the culture medium is not essential for the maintenance of pluripotency. However, its presence appears to inhibit the differentiation process (Brevini et al. 2008) as pig pluripotent cells were not able to differentiate, aggregate and give rise to embryoid bodies (EB) when LIF was added to the standard medium routinely used to induce EB formation, (Brevini et al. 2010). This suggests that the cytokine is unlikely to act through the canonic LIF receptor signalling pathway, but rather via alternative cascades (see chapter ‘Pluripotency signalling pathways in the pig’).
Lack of Molecular Markers of Pluripotency
Until a few years ago, criteria used to define ESCs were mainly morphological. Most colonies were accepted as pluripotent when cells ‘were small and rounded and had a large nucleus with one or two prominent nucleoli’ (Piedrahita et al. 1990). Alkaline phosphatase (AP) was the first demonstrated molecule to be used as a reliable marker for undifferentiated ESCs in pig and sheep (Talbot et al. 1993b). This first attempt was followed by the demonstration of stage-specific embryonic antigen-1 (SSEA-1) expression (Wianny et al. 1997) and by the absence of laminin and intermediate filaments, which are not present in the ICM and epiblast, but detectable once differentiation occurs (Piedrahita et al. 1990; Moore and Piedrahita 1997; Wianny et al. 1997).
At present, several pluripotency markers are expressed both in embryos and derived cells by most mammalian species, including OCT4, REX1, NANOG and SOX2 (Chen et al. 1999). However, several markers related with stemness have difference in expression according to the species considered. A clear example of this is represented by the OCT4 gene. In mouse and human ESCs, OCT4 is considered the main hallmark of pluripotency, as its expression is restricted to the ICM (Kirchhof et al. 2000; Mitalipov et al. 2003). A much different scenario has been described in farm animal species. In cattle, the gene was detected both in ICM and trophoblast cells; therefore, its expression is not limited to pluripotent cells of the early embryo (van Eijk et al. 1999). This report by van Eijk, who firstly described a non-ICM-restricted distribution of OCT4, was quickly followed by data collected in the pig, where the OCT4 protein is present in both the ICM and trophectoderm, and only at hatching was the signal confined to the epiblast component of the embryonic disc (Kirchhof et al. 2000; Vejlsted et al. 2006). We recently derived pluripotent pig lines from both in vivo (Brevini et al. 2006) and in vitro (Brevini et al. 2010) embryos and followed OCT4 expression during the life of the cell lines. Our results indicate that OCT4 is expressed at the time of attachment and during the first passages for a maximum of seven when its expression is down-regulated. However, despite the decrease in OCT4 expression, cells could be cultured for several months without any change in morphology and without up-regulation of specific differentiation markers. This observation was recently confirmed by Wolf et al. (2011), who reported gradual down-regulation of OCT4 expression, in spite of maintained ESC-like morphology in porcine cells. This suggests that other factors may be responsible for the maintenance of self-renewal of these cells in an undifferentiated status. Indeed, induced pluripotent porcine stem cells were recently obtained without OCT4 over-expression and by ectopic expression of only SOX2, KLF4 and c-MYC (Montserrat et al. 2011). These authors reported that not only OCT4 was completely dispensable, but its overexpression negatively affected the pluripotent potential of their reprogrammed cells that, nevertheless, followed the standard criteria for pluripotency, differentiated in vitro along multiple tissue lineages, had a stable karyotype and expressed endogenous pluripotency-related genes. The need for careful interpretation when OCT4 is used as a marker for pluripotency is further suggested by the observation that detectable levels of this gene have been reported in pig fibroblasts (Carlin et al. 2006), reinforcing the hypothesis of a broader role for OCT4 in pig development and cell biology compared with other species.
Pluripotency Signalling Pathways in the Pig
Specific cell signalling pathways have been shown to govern pluripotency in the mouse and primates. Mouse ESCs are pluripotent and can self-renew indefinitely in culture. They depend on the cytokines LIF and bone morphogenetic protein 4 (BMP-4) to maintain their undifferentiated state (Smith et al. 1988; Ying et al. 2003; Alberio et al. 2010). Conversely, human ESCs need bFGF and activin A for pluripotency and self-renewal, whereas LIF is dispensable (Dahéron et al. 2004; Vallier et al. 2005). These observations suggest that, although similar regulatory pathways are likely to exist and be conserved among species, species-related differences in the mechanisms controlling pluripotency are evident. An understanding of these aspects would indeed help to identify conditions adequate to support self-renewal requirements in farm animal species. Pig epiblast has been shown to depend on activin/nodal signalling for self-renewal, as previously shown for human ESCs, indicating that maintenance of pluripotency by this signalling mechanism is conserved in the two species (Alberio et al. 2010). We recently observed that both LIF and bFGF are needed to establish outgrowths and generate stable cell lines in pigs. At the same time, we detected expression of STAT3, but the absence of gp130 and LIFR transcripts, two specific subunits of LIF receptors. In agreement with this, a previous report demonstrated inconsistent expression of LIFR in porcine epiblast cells cultured for 24 hr after ICM isolation (Blomberg et al. 2008), and a further one could not detect LIFR in the epiblast cells of early embryos (Hall et al. 2009). Altogether, these observations would suggest the dispensability of LIF in supporting pluripotency in the porcine species and would seem to disagree with our results indicating LIF as an important factor, supporting both attachment and self-renewal. Further molecular characterization, however, indicated the possibility that this cytokine is unlikely to act through the gp130/LIFR/STAT3 signalling pathway, but rather via an alternative cascade involving phosphoinositide-3 kinase (PI3K), serine/threonine protein kinase (AKT) (a key effector in the PI3K pathway) and phosphatase and tensin homolog deleted on chromosome 10 (PTEN) (a negative regulator of the same pathway) (Fig. 3), known to be responsive to LIF and has been previously shown to trigger the expression of NANOG and to facilitate efficient proliferation and survival of murine ESCs (Welham et al. 2007).
As mentioned previously, pig outgrowth attachment and colony formation were also encouraged by bFGF, which has been reported to be necessary for proliferation and pluripotency (Mummery et al. 1993; Levenstein et al. 2006) and to fully replace LIF for maintenance of human ESCs self-renewal (Xu et al. 2005b). The concomitant presence of its receptor indicates the ability of porcine cells to bind bFGF and suggests a possible role of the FGF signalling pathway in self-renewal of porcine cells. Furthermore, it has been reported that bFGF can also bind and activate the PI3K/AKT cascade (Jirmanova et al. 2002; Xu et al. 2005a), and it cannot be ruled out that bFGF may synergistically act with LIF to exert its pluripotency-related effect through this pathway as well through its canonical one.
Notably, the PI3K/AKT cascade is known to modulate several regulatory pathways and to exert a powerful triggering effect on NANOG expression which, in contrast to OCT4, is constantly transcribed and steadily detectable in pig pluripotent cells. Altogether these observations lead us to hypothesize that NANOG may be able to maintain pig pluripotent cells in an undifferentiated state, also in the absence of the simultaneous expression of OCT4, and may thus represent a key molecule for self-renewal in the porcine species.
Since the successful isolation of ESCs from mouse in the early 1980s, significant efforts have been addressed to establish ESC lines from farm animal species. However, the active research has not yet lead to the derivation of bona fide ESC lines. The persistent lack of standard methods for the isolation, maintenance and characterization are the main limiting factors that may explain why no validated lines are available so far. Many reports have described cell lines in domestic species, which presented several, although not all of important features typical of ESCs. These observations suggest that similar regulatory pathways are likely to be present among different species and control the main mechanisms leading to outgrowth formation and cell line derivation and maintenance. However, it is also clear that species-related differences couple these shared mechanisms with a high specificity driving embryo development timing, a controlled concentration and expression of pluripotency-related molecules and the specific need for a defined micro-environment.
The research was funded by PRIN 2007 and FIRST 2007. GP was supported by Carraresi Foundation.
Conflicts of interest
None of the authors have any conflicts of interest to declare.