Pluripotent Stem Cells for Transgenesis in the Rabbit: A Utopia?

Featured Application: Rabbit pluripotent stem cells represent valuable tools for creating human disease models in vivo via the production of transgenic animals and in vitro by generating organoids. They are also of interest for the creation of bioreactors, i.e., transgenic rabbits producing pharmacological molecules in their milk. Abstract: Pluripotent stem cells (PSCs) possess the following two main properties: self-renewal and pluripotency. Self-renewal is deﬁned as the ability to proliferate in an undi ﬀ erentiated state and pluripotency as the capacity to di ﬀ erentiate into cells of the three germ layers, i.e., ectoderm, mesoderm, and endoderm. PSCs are derived from early embryos as embryonic stem cells (ESCs) or are produced by reprogramming somatic cells into induced pluripotent stem cells (iPSCs). In mice, PSCs can be stabilized into two states of pluripotency, namely naive and primed. Naive and primed PSCs notably di ﬀ er by their ability to colonize a host blastocyst to produce germline-competent chimeras; hence, naive PSCs are valuable for transgenesis, whereas primed PSCs are not. Thanks to its physiological and developmental peculiarities similar to those of primates, the rabbit is an interesting animal model for studying human diseases and early embryonic development. Both ESCs and iPSCs have been described in rabbits. They self-renew in the primed state of pluripotency and, therefore, cannot be used for transgenesis. This review presents the available data on the pluripotent state and the chimeric ability of these rabbit PSCs. It also examines the potential barriers that compromise their intended use as producers of germline-competent chimeras and proposes possible alternatives to exploit them for transgenesis.


Introduction
Pluripotent stem cells (PSCs) were first derived from the inner cell mass (ICM) of mouse embryos in 1981 [1,2]. These cells, named embryonic stem cells (ESCs), were shown to be able to colonize the epiblast of host blastocysts and, consequently, to produce chimeric mice in 1984 [3]. This ability of mouse ESCs (mESCs) has made it possible to develop transgenesis techniques in this species, facilitating substantial progress in functional genetics. Later on, PSCs were derived from the late epiblast of mouse post-implantation embryos [4,5]. These cells, named epiblast stem cells (EpiSCs), display several different features from mESCs, especially with respect to their transcriptome and epigenome, and are notably not able to colonize host blastocysts. At the same time, mouse PSCs were obtained by reprogramming differentiated somatic cells by overexpression of four pluripotency factors, namely Oct4, Sox2, Klf4, and cMyc [6]. These cells, named induced pluripotent stem cells (iPSCs), can exhibit features of mESCs or EpiSCs depending on the medium used during their reprogramming [7,8]. Therefore, mESCs and EpiSCs epitomize two states of pluripotency existing in vivo during early embryonic development-the naive state, corresponding to early epiblast cells from preimplantation embryos, and the primed state, corresponding to late epiblast cells from post-implantation embryos, respectively [9]. In vitro naive PSCs are sustained by the leukemia inhibitory factor (LIF)/gp130/STAT3 and bone morphogenetic protein 4 (BMP4)/ALK/ SMAD1-5-8 signaling pathways, whereas primed PSCs are supported by the fibroblast growth factor 2 (FGF2) and activin A/transforming growth factor beta (TGFβ)/SMAD2-3 signaling pathways [10].
To date, all PSC lines, ESCs and iPSCs, obtained in non-rodent mammalian species display features of primed pluripotency [11,12]. Therefore, it seems that non-rodent PSCs are only able to stabilize in culture in the primed state of pluripotency [13]. For example, primate PSCs present a transcriptome very different from that of human early blastocyst epiblast cells [14], but is rather closer to that of post-implantation late blastocyst epiblast cells according to an analysis in monkeys [15]. The actual research challenge in the PSC field is to reprogram primed cells toward the naive state of pluripotency, in order to obtain and culture more genetically stable cells, which are easier to handle by single-cell dissociation [16,17] and are more useful for cellular therapies or the production of disease models [18,19]. The possibility to revert primed PSCs into naive PSCs was first demonstrated in mice by overexpressing pluripotency genes such as Klf4, Nanog, Stat3, Tfcp2l1, or Prmd14, either alone or in combination [20][21][22][23][24], and by strengthening the signaling pathways sustaining the naive state using inhibitors of mitogen-activated extracellular regulated kinase (MEK) and glycogen synthase kinase 3 beta (GSK3β) in a medium called 2iLIF [25,26]. Such strategies for reprogramming primed PSCs have been extensively studied for human PSCs and have produced naive-like cells with a heterogenous reconfiguration of their transcriptome and epigenome [27][28][29][30] as well as a variable capacity to produce interspecies chimeras after microinjection into mouse blastocysts [31,32]. These variations in the molecular and functional characteristics of primate PSCs show that embryonic cells could be stabilized in vitro at different stages along a continuum of pluripotency, the ends of which are epitomized by the naive and primed states, respectively [33].
Lagomorphs and primates share many similarities in their embryonic development [34], in particular in the timing of the embryonic genome activation at the 8/16-cell stages [35], the timing of the waves of DNA demethylation and methylation [36,37], and the timing of the random inactivation of the second X chromosome [38]. Like their human counterparts, rabbit embryos develop as a flat disc on the surface of the conceptus [39], and present the advantage of implanting very late (at E6.75) due to a mechanism similar to that of human embryos [40]. Above all, the gastrulation of rabbit embryos begins before implantation (at E6.0) so that the epiblast remains more easily accessible for experimentation than in rodents [41]. These similarities and particularities make rabbits an interesting model not only for the study of the biology of PSCs, but also to be used to create transgenic animal models of human development and diseases and to improve interspecies chimerism tests.
In this review, we provide a state-of-the-art discussion of the pluripotent state and chimeric ability of rabbit PSCs. We discuss the potential barriers to their use in the formation of chimeras and propose possible alternatives to exploit them for transgenesis in rabbit.

Rabbit Embryonic Stem Cells (rbESCs)
Rabbit embryonic cell cultures were described by American teams in 1993 [42,43], but the first lines of rbESCs were not published until 15 years later by two teams from China and Japan [44,45]. RbESCs are derived from ICM cells of early blastocysts (E3.5-E4.0). They are cultured on feeder cells and form flat colonies ( Figure 1). The self-renewal of rbESCs depends on the activin A/TGFβ/SMAD2-3 and FGF2 pathways [46,47]. FGF2 appears indispensable by inducing the PI3K/AKT and MAPK pathways [48,49], while the WNT/β-catenin pathway may also be indirectly activated by FGF2 [50]. It is possible to derive cells without any growth factor in the medium if they are cultured on feeder cells, showing that LIF is not necessary for the maintenance of rbESCs [51]. However, the addition of LIF to the culture medium of rbESCs has often been used [48,49,52], and several studies have described the effect of LIF on the derivation of rbESCs and the induction of LIF-receptor expression [52,53]. Unlike the derivation of rodent ESCs stabilized in the naive state [54], MEK and GSK3β inhibitors do not enhance epiblast cell differentiation in vivo [55] or rbESC derivation in vitro [51]. In addition, the maintenance of dome-shaped naive-like rbESC colonies in the presence of these two inhibitors requires both FGF2 and feeder cells [56].
Appl. Sci. 2020, 10, x FOR PEER REVIEW 3 of 18 [52,53]. Unlike the derivation of rodent ESCs stabilized in the naive state [54], MEK and GSK3β inhibitors do not enhance epiblast cell differentiation in vivo [55] or rbESC derivation in vitro [51]. In addition, the maintenance of dome-shaped naive-like rbESC colonies in the presence of these two inhibitors requires both FGF2 and feeder cells [56]. RbESCs are unstable and their enzymatic single-cell dissociation induces higher rates of cell proliferation than clump passages [51], but generates more chromosomal abnormalities as described for human ESCs [57]. The inhibitor of the rho kinase (ROCKi), which acts by blocking the apoptotic RbESCs are unstable and their enzymatic single-cell dissociation induces higher rates of cell proliferation than clump passages [51], but generates more chromosomal abnormalities as described for human ESCs [57]. The inhibitor of the rho kinase (ROCKi), which acts by blocking the apoptotic response induced by cell dissociation [58], does not increase the clonogenicity of rbESCs, unlike human ESCs [59], but leads to their arrest of proliferation and differentiation (our own unpublished data). This difference can be explained by the high expression levels of ROCK1 and ROCK2 genes in rbESCs [60], which may be essential to maintain cell pluripotency by eliminating cells engaged into differentiation.

Rabbit-Induced Pluripotent Stem Cells (rbiPSCs)
RbiPSCs were produced by reprogramming rabbit somatic cells (adult liver and stomach cells or adult and embryonic fibroblasts) using the classical combination of the four human genes (hOCT4, hSOX2, hKLF4, hc-MYC) overexpressed by murine leukemia virus (MuLV)-based retroviral vectors [47,[61][62][63]. RbiPSC lines show reactivation of endogenous genes of the pluripotency core, OCT4, SOX2, and NANOG, as well as silencing of all four transgenes; they are therefore fully reprogrammed [47]. Like rbESC lines, they form flat colonies ( Figure 1) and self-renew in the presence of FGF2, knockout serum replacement (KOSR), and mitomycin-treated mouse embryonic fibroblasts (MEFs) producing activin A, as feeder cells. The use of serum and LIF-based medium did not allow the reprogramming of adult ear fibroblasts [47]. Attempts to directly reprogram rabbit fibroblasts into the naive state used, in addition to the four usual human transgenes, LIF, KOSR, MEK, and GSK3β inhibitors as well as forskolin, a protein kinase C (PKC) agonist shown to reinforce KLF2 and KLF4 expression in human PSCs [27]. The resulting female rbiPSC lines displayed reactivation of the second X chromosome, formed dome-shaped colonies, exhibited improved growth and clonogenicity rate, and required both LIF/STAT3 and PI3K/AKT pathways to self-renew [64]. However, these lines were not fully reprogrammed since their naive-like state depended on the expression of the four transgenes only maintained in presence of doxycycline. This obstacle, posed by the inability of iPSCs to silence exogenous reprogramming factors and to depend on them for self-renewal, is often observed in domesticated animals [12].

Pluripotent State of rbPSCs
Both rbESC and rbiPSC lines express the cardinal markers of PSCs [47]. They are positive for alkaline phosphatase activity [44]. They express the pluripotency-associated OCT4 and NANOG transcription factors, as well as the SSEA-1, SSEA-4, TRA1-60, and E-cadherin cell surface markers [52,61]. They also display a normal karyotype (42XX or 42XY) [44,47], produce embryoid bodies in vitro [45,63], and form teratoma containing tissues of ectodermal, mesodermal and endodermal origin upon injection under the kidney capsule in severe combined immunodeficient (SCID) mice [47,61]. However, rbPSCs can only be stabilized in the primed state of pluripotency. The suppression of FGF2 and/or culturing without feeder cells lead to their differentiation. Thus, these cells can be cultured only on a synthetic matrix, with either medium conditioned on mitomycin-treated MEF or addition of activin A [51,65]. In the same way, rbPSCs differentiate in the presence of inhibitors of SMAD2-3, TGFβ-receptor, or FGF-receptor [46,50].
Transcriptomic comparisons of rbPSCs with rabbit epiblast cells from E4-and E6-stage embryos show overexpression of 17 genes involved in the WNT/β-catenin signaling pathway, including TCF4, LEF1, and WNT5A [60]. This result is consistent with the decrease in the rate of rbESC multiplication after inhibition of the WNT pathway by a Frizzled-1 antagonist or an anti-Wnt3a antibody [50]. This effect is associated with an increase in phosphorylation of β-catenin and SMAD1-5-8, as well as a reduction in expression of SSEA-4 pluripotency marker, indicating cell commitment to differentiation. The WNT pathway is involved in maintaining the balance between self-renewal and differentiation of both mouse and human ESCs [66][67][68]. It is therefore also important in sustaining the pluripotency of rbPSCs, although the 2iLIF medium allowing the self-renewal of rodent PSCs and containing the GSK3β inhibitor CHIR99021 did not allow the stabilization of rbESCs.
RbPSCs form flat colonies characteristic of their primed pluripotent state ( Figure 1). They express 173 genes involved in cytoskeletal organization and function compared to rabbit epiblast cells at the E4 and E6 stages [60]. Among them are gene-encoding filamins (FLNA and FNLC) and vinculin (VCL), which function to stabilize the actin cytoskeleton and its anchoring to the plasma membrane. These molecules promote cell-to-matrix interactions rather than cell-to-cell adhesions, and therefore may interfere with colony morphology. These proteins are also more expressed in EpiSCs than in mESCs [69,70].
Interestingly, a comparison of rbESC and rbiPSC lines produced from the same breed of New Zealand white rabbit showed different features that bring rbiPSC lines closer to the naive state of pluripotency than rbESC lines [47]. These characteristics mainly concern their proliferation rate, their resistance to unicellular dissociation, their global transcriptome, and their expression of markers specific to the primed and naive pluripotency states [71]. The variation of these characteristics allows the classification of rbPSC lines on a graduated scale of primed pluripotency with rbESCs at one end (most primed state) and rbiPSCs at the other end (closer to the naive state) (Figure 1).

Reprogramming of rbPSCs Toward the Naive State of Pluripotency
As in humans, the reprogramming of primed rbPSCs toward the naive state was addressed by mimicking the protocols used in rodents and was based on the overexpression of pluripotency genes associated with the use of media sustaining the naive state. To date, two studies have been published in this area. The first used overexpression of the hOCT4 gene and the 2iLIF medium supplemented by KOSR, forskolin, and kenpaullone, a potent inhibitor of several cell-cycle complexes, such as CDK/cyclin [72]. In the second, we made use of the hKLF2-hKLF4 gene duo with a serum/LIF-based medium to produce the cells named rbEKA, for rabbit enhanced KLF activity [65] ( Figure 1). Both studies showed an improvement in some properties of the reprogrammed rbPSCs, such as increased differentiation capacity [73] or reprogramming of the transcriptome, epigenome, and miRNome [65]. As previously seen with the methods used to directly produce naive-like rbESCs [56] and rbiPSCs [64], the techniques tested improve the pluripotency of primed rabbit cells, but without reaching a stable naive state, nor the capacity to produce germline-competent chimeras. In addition, all of these rbPSC lines present highly variable properties due to the heterogeneity of the rabbit strains used, the variability in the production techniques and media used, and the characterization criteria tested. Therefore, how rbPSCs can be stabilized in the naive state of pluripotency remains a question open to new research. However, the molecular pathways responsible for pluripotency in this species, although closer to those of primates than rodents, present some idiosyncrasies that have yet to be adequately characterized [74].

Definition and Benefits of Chimeras
Chimeras are biological entities made up of cells of different genetic origins. Depending on whether the donor cells are involved only in the development of the recipient's somatic tissues or in the development of all of the recipient's organs, including the germline, chimeras can be referred to as somatic or germline chimeras, respectively. Chimera production is commonly used in developmental biology to study the differentiation potential of cells transplanted into a genetically different host [75]. In the field of PSCs, it is used to confirm their ability to differentiate in all embryonic lineages, including the germline, by transferring the donor genome to the next generation. It thus remains the reference technique, at least in rodents, where intraspecific syngeneic or allogeneic chimeras are easily achievable. It is more questionable in the case of xenogenic interspecific chimeras, which may involve factors other than the mere differentiation capacity of the donor cells [76].

Techniques Used in Mice
In mice, germinal chimeras can be obtained by different techniques, namely microinjection of mPSCs into the blastocoel of recipient blastocysts or into 4/8-cell stage embryos [77], aggregation of mESCs to morula [78], or by tetraploid complementation, i.e., injection of pluripotent cells into host embryos rendered tetraploid by electrofusion [79]. The latter technique makes it possible to obtain a mouse from only microinjected cells, since the tetraploid cells of the recipient embryo can only produce extraembryonic tissues [80]. Only naive mouse cells cultured in LIF/serum or in 2iLIF media are capable of producing germline chimeras with these techniques [81][82][83]. In contrast, primed EpiSCs are not capable of colonizing a host blastocyst [5], unless they are injected into post-implantation embryos [84]. Similarly, primed PSCs from cynomolgus monkeys become capable of colonizing a recipient embryo only when they have been reprogrammed in a naive-like state [85].

Chimeras Produced in Rabbit
So far, the chimeric rabbit pups or fetuses reported in the literature have been obtained with freshly isolated embryonic cells and one of four different microinjection techniques [86]. The microinjection of an isolated ICM into an early blastocyst has produced 10% and 6% chimeric animals in the absence of or after three days of ICM culture, respectively [42,87]. Another study found that the aggregation of two embryos within the same E3-stage morula had given rise to about 5% chimeric fetuses [88,89]. Microinjection of 1 to 3 blastomeres into an embryo at the 8/16-cell stage had yielded upwards of 2% of germline chimeras [90], and microinjection of dissociated gonadal cells from an E20-stage male fetus into an 8/16-cell embryo produced 2.5% chimeric fetuses [91]. In contrast, the longer-cultured rbPSCs produced at best only 10% of blastocysts with colonized ICM and 3% of E10-stage fetuses that were very weakly chimeric [65]. Only two publications have reported the successful use of rbESCs to produce some coat color chimeras (4%) [92] or one cloned rabbit (1%) [93]. However, these observations remain isolated and apparently were not reproducible even in the laboratories concerned. In any case, these percentages are far below those obtained in mice, where the proportions of chimeric animals resulting from the microinjection of mPSCs into embryos ranged from 20% to 30%, and the rates of chimerism obtained reached nearly 100% [81,82].

Colonization and Chimerism
It is important to distinguish embryo colonization, which corresponds to the presence of donor cells in preimplantation embryos, from chimerism, which represents the participation of donor cells to embryonic development in parallel with host cells. The mere presence of donor cells in the ICM of recipient blastocyst does not presage their participation in embryonic development and may only reflect their autonomous development [94]. Thus, a chimeric fetus must be made up of a homogeneous mosaic of cells of both origins; this was the case for cynomolgus chimeras obtained with naive-like monkey PSCs, even if the rate of chimerism was still low [85]. Similar to primed primate PSCs, primed rbESCs were unable to colonize the ICM of host blastocysts after microinjection at the morula stage [45,47,61]. However, primed rbiPSCs yielded colonized blastocysts after the aggregation of morula with cells expressing LacZ or green fluorescent protein (GFP) marker genes [47] (Figure 2). Similarly, rbESCs derived in the presence of serum and in the absence of FGF2 produced early blastocysts with GFP-positive cells in the ICM, with 2% or 16% colonized embryos, respectively, depending on the lineage [51]. These percentages are comparable to those obtained with reprogrammed rbiPSCs cultured in 2iLIF medium containing KOSR, forskolin, and kenpaullone [72]. In vitro, reprogrammed rbEKA cells allowed a higher rate of colonized embryos to be obtained, up to 33% at the E3.5 stage and 25% at the E4.5 stage, but the percentage of embryos with microinjected cells into the ICM was only 10% on average [65]. In the same study, after the transfer of rbEKA-treated embryos into surrogate females, six E6-stage pre-gastrula (1.4%) and two E10-stage fetuses (3%) were found colonized by rbEKA cells [65] (Figure 2). The number and the localization of GFP-expressing rbEKA cells showed the amplification and the participation of the injected cells to the formation of the host epiblast. Pre-gastrula showed a fairly uniform distribution of fluorescent cells throughout the embryonic disc and the only gastrula obtained showed GFP-positive cells in the primitive streak, and therefore potentially entering into the process of gastrulation. However, the percentage of colonized embryos decreased as embryonic development progressed, and above all, the rate of colonization continually declined. Even though most ICMs appeared to be massively colonized at the E3.5 and E4.5 stages, the rate of chimerism gradually decreased to 1% in fetuses according to a semi-quantitative PCR estimate [65]. The injected cells observed in these colonized embryos were far from being in the majority, as is the case in murine chimeras produced with a single donor cell [95]. These low colonization rates are consistent with those of interspecific chimeras generated by microinjection of naive-like reprogrammed primate PSCs into mouse or pig embryos [96][97][98].

Barriers to Overcome
Recently, cell competition has been shown to be a fundamental process underlying early embryonic development in mice [99]. Cell competition is a short-range intercellular communication,

Barriers to Overcome
Recently, cell competition has been shown to be a fundamental process underlying early embryonic development in mice [99]. Cell competition is a short-range intercellular communication, in which neighboring cells compare their fitness and winner cells eliminate loser cells [100,101]. The production of chimeras also requires close interaction between donor and host cells and depends on the result of these cell competitions [102]. The transplanted cells must be able to survive, proliferate, and differentiate at least at the same rates as the host embryo cells in order to participate in the development. Rodent naive PSCs behave as winner cells, whereas primed rbPSCs, like primed primate PSCs, behave more like loser cells after injection into host blastocysts. Several barriers should be overcome to convert loser PSCs into winner cells in this context.

The Pluripotency Barrier
As previously seen, even if all PSCs can differentiate into the three embryonic lineages within embryoid bodies or teratoma, only naive mPSCs can create germline chimeras after microinjection into blastocysts, whereas primed EpiSCs cannot [5]. Classically, naive mPSCs are involved in the development of the embryo proper, but not in the development of extraembryonic tissues, except for a subpopulation of so-called 2C cells. These 2C cells were identified among mPSCs cultured in 2iLIF medium, and express endogenous retrovirus activity. They also show features similar to 2-cell stage embryos, and present dual developmental ability [103,104]. Modification of the potentiality of the mPSCs was also possible by using different cocktails of small molecules that allow chromatin remodeling [95] or differentiation blockade [105]. The resulting cells, called extended pluripotent stem cells, showed the ability to participate in the formation of extraembryonic tissues without modifying their self-renewal [106].

The Developmental Barrier
The developmental stage of the donor cells and that of the recipient embryo must match in order to produce chimeras [107]. EpiSCs are unable to colonize the ICM of preimplantation blastocysts, but implant and develop perfectly in the epiblast of post-implantation embryos [84]. Similarly, human PSCs can only participate in the post-implantation development of mouse embryos [94,108]. This concordance is not necessarily respected if the donor cells provide a selective, even vital, benefit to the host, such as mESCs expressing Pdx1 gene injected into mouse a-pancreatic embryos [109]. Modification or alteration of the development of the recipient embryo may also promote the integration of the donor cells. Activin A activated trophectoderm differentiation to the detriment of the epiblast of mouse embryos, which were then more easily colonized by microinjected mPSCs [110]. Similarly, FGF4 added to the mouse embryo culture medium promoted the differentiation of ICM cells into hypoblasts and, consequently, the formation of epiblasts by the injected mPSCs [111].

The Apoptotic Barrier
A recent study showed that most mESCs cultured in serum/LIF-based medium and microinjected into an 8-cell stage embryo entered apoptosis within 10 h after microinjection [112]. In this medium, the pluripotency state of mESCs is heterogeneous and only transiently Rex1-positive cells were able to colonize the embryo, whereas transiently negative cells entered apoptosis after microinjection. In contrast, the level of apoptosis is extremely low for mESCs cultured in 2iLIF medium, which explains their increased colonization efficiency. Similarly, overexpression of the anti-apoptotic Bcl2 gene in EpiSCs [32] or human ESCs [113] promoted the survival of cells microinjected into preimplantation mouse embryos. Interesting observations have been made in the naked mole rat, a rodent characterized by a high longevity and a low incidence of spontaneous tumors. In this experimental model, the reprogramming of fibroblasts into iPSCs required, in addition to the four classical Oct4, Sox2, Klf4, and cMyc genes, the overexpression of SV40-T, whose role is to inhibit the tumor-suppressive protein retinoblastoma [114]. The resulting iPSCs have contributed to the formation of embryonic and extraembryonic tissues of mouse fetuses [115].

The Cell Division Barrier
Competition between donor and host cells may be based on the speed of cell division. The injected PSCs must multiply at least at the same rate as the host cells to prevent progressive elimination. This is particularly true at the peak of cell division, which precedes and marks the onset of gastrulation [116,117]. Mouse PSCs showed a particularly fast cell cycle with a very short G1 phase, and a close interaction was demonstrated between the factors that maintain the cell cycle and pluripotency [118][119][120]. Among them, the c-Myc gene plays a role both in the regulation of the G1 phase [121] and in the acquisition and maintenance of pluripotency [122]. The G1 phase of mESCs is independent of cyclin D [123], but is controlled by cyclin E [124]. Modulation of cyclin E expression increased or decreased the length of the G1 phase, and chimeras produced with mESCs lacking cyclins D and E, which then showed a G1 phase elongation, displayed implantation of these cells only in the placenta and in neural structures of the host [125]. In addition, the level of c-Myc gene expression was shown to be the direct driver of cell competition in mouse embryos, with higher and lower expressions of c-Myc corresponding to winner and loser cells, respectively [126,127].

The Epigenetic Barrier
Epigenetic modifications of the paternal and maternal genomes during the early stages of embryonic development are crucial [128], and often hinder the reprogramming of the transferred nuclei during embryonic cloning [129,130]. Similarly, the epigenomes of PSCs vary according to their pluripotency states [131]. In mice, the derivation of ESCs with two successive cocktails of inhibitors and specific growth factors led to advanced pluripotent stem cells with increased potential for differentiation in embryonic and extraembryonic lineages [132]. These cells displayed a hypermethylated genome relative to the hypomethylated genome found in ICM cells or mESCs, and also showed adequate imprinting gene configuration. In addition, naked mole rat cells showed very high levels of the repressive mark H3K27me3 and very low levels of the activator mark H3K27ac, which may confer high resistance to tumor transformation [114]. To be reprogrammed into iPSCs, these cells required the intervention of the SV40-T gene, which altered these epigenetic marks and may likely permit the formation of chimeras with mouse embryos [115].

The Adhesion Barrier
Injected PSCs need to interact directly with host cells in order to participate efficiently in embryonic development. Differences in the level of adhesion proteins have been found between naive and primed PSC membranes [133]. Transient overexpression of the Cdh1 gene, which encodes the E-cadherin adhesion molecule, conferred on EpiSCs the ability to colonize murine preimplantation blastocysts and to produce 5% chimeric mice [134]. In contrast, low density cultures of mESCs in the absence of LIF caused the methylation of the Cdh1 gene and the loss of E-cadherin expression. These cells, called ENPSC M , displayed the same characteristics as EpiSCs and lost their ability to form chimeras [135]. Similarly, rbPSCs express E-cadherin at variable levels, and these seem to be correlated with their capacity for embryonic colonization [51,65].

The Metabolic Barrier
By acting at the interface between the environment and the epigenome, energy metabolism and mitochondrial activity are major factors affecting pluripotency and the differentiation of embryonic cells [136]. In vivo, reduced access to oxygen limits the embryo to a glycolytic mechanism based on the transformation of glucose into pyruvate and lactate, until compaction. Glucose consumption then increases, resulting in a difference between the trophectoderm, which produces more ATP by oxidative phosphorylation (OxPhos) of glucose, and the ICM, which favors the glycolytic mechanism (glycolysis) [137]. The role of metabolism on the loss of pluripotency of embryonic cells is reinforced by access to oxygen and changes in the composition of the embryo's microenvironment during peri-implantation development [138]. Similarly, metabolism is a central element in the physiology of PSCs since they exhibit different states of pluripotency depending on their culture conditions [139]. Naive PSCs exhibit a bivalent metabolism using both glycolytic and OxPhos mechanisms, whereas primed PSCs preferentially use glycolysis [140]. In the same way, somatic cell reprogramming requires a metabolic switching between OxPhos and glycolytic mechanisms to produce iPSCs [141]. Pluripotency genes are also involved in the activation of metabolic pathways. The naive factor ESRRB activates mitochondrial OxPhos and STAT3 binds to the mitochondrial genome to enhance oxidative metabolism, whereas the primed factor LIN28 coordinates with LET-7 to enhance glycolysis and repress OxPhos [142]. Therefore, microinjected PSCs must be able to adjust their energy metabolism to match that of the ICM cells in order to participate in embryo development.

The Germinal Competency and the Euploid Karyotype
A third state on the pluripotency scale has been termed the formative state and has been described in mice [143]. This state corresponds to epiblast cells of peri-implantation mouse embryos at the E5.5 stage that are committed to the primed state but are still able to differentiate into germline cells. The formative cells express specific early post-implantation genes, such as Otx2, Sox3, and Oct6, and have recently been established in vitro in mice [144,145]. In order to produce germline chimeras, the cells used must be in the naive or formative state, since cells in the primed state have lost the capacity to differentiate into germline cells. Finally, the transplanted cells must have an euploid karyotype, as the accumulation of chromosomal abnormalities during long-term culture can greatly reduce their ability to form chimeras [57].

Conclusions
To date, rbESCs derived from early epiblast of the E3.5-stage blastocysts and rbiPSCs resulting from somatic cell reprogramming show most of the features characteristic of primed pluripotency and display a transcriptome equivalent to that of late epiblast of the E6.0-stage pre-gastrula. None of the methods used in mice and primates to reprogram primed PSCs toward the naive state produced rabbit cells capable of efficiency colonizing a host blastocyst. The technique required to create bona fide naive rbPSCs and, therefore, to use them to produce germline chimeras and transgenic rabbits must focus on the signaling pathways that can support the core of pluripotency genes in this species, since these pathways appear to vary between mammals. Nevertheless, current rbPSC lines may be useful for transgenesis, but this will require developing alternative methods such as their differentiation into female or male germ cells and their exploitation as nuclear donor cells for cloning. These methods have shown to be effective in mice [146][147][148][149]. Although these techniques require further testing in rabbits, they could make an important contribution to improved cloning, which is inefficient in this species [150]. In any case, rbPSCs remain valuable tools for transgenesis, as they can be genetically modified in vitro and especially verified for mutations created before producing the transgenic rabbit strains. Moreover, iPSC technology is also of interest for the preservation of species biodiversity [151], by allowing easy sampling from valuable animals without any breeding steps. This would consequently enlarge the number of frozen samples with somatic tissues. Moreover, PSC lines are useful to develop tissue-specific cell lines or 3D organoids and to create in vitro disease models [152]. Finally, the study of the biology of rbPSCs makes it possible to understand the maintenance of pluripotency in a model species other than rodents and primates, rabbits being suitable for this purpose due to the accessibility of their early embryonic development.