Abstract
The successful production of animals and embryonic stem cells using somatic cell nuclear transfer (SCNT) has demonstrated the unmatched nuclear reprogramming capacity of the oocyte and helped prove the degree of plasticity of differentiated cells. The introduction of transcription factors to generate induced pluripotent stem cells (iPSCs) displaced SCNT and, due to its ease of implementation, became the method of choice for cell reprogramming. Nonetheless, iPSC derivation remains inefficient and stochastic. This review article focuses on using the oocyte as a source of reprogramming factors, comparing the SCNT and iPSC mechanisms for remodeling chromatin and acquiring pluripotency.
Introduction
M
Despite the overwhelming number of alternative methods and molecules used, iPSC derivation efficiency has remained low. Reprogramming rates vary from 10 to 0.0001% [3 –7], depending mostly on the combination of factors and the delivery method used. The kinetics of iPSC derivation also varies greatly, ranging from 2 to 8 weeks from the time a somatic cell is exposed to the reprogramming factors until an iPSC colony is observed [8]. Such variability is difficult to explain, underscoring the need for the independent validation of protocols using an assortment of donor somatic cell types.
Somatic cell nuclear transfer (SCNT) has successfully produced live cloned animals in more than 20 different mammalian species [9 –12]. SCNT has also proven effective for generating ESCs in mice, nonhuman primates, and humans [13]. Its reported efficiency varies from 1 to 50% depending on the species, tissue type, and sex, among other factors. The kinetics, however, are always consistent with the developmental program of each species; reprogramming happens within hours after nuclear transfer and concludes with the formation of the inner cell mass at the blastocyst stage [13 –16].
Both technologies seem to have issues with variability and stochasticity in the reprogramming process; however, the generation of ESCs using SCNT (NT-ESC) in mice and humans is 10- to 100-fold more efficient, as well as less variable, faster, and is considered less stochastic than iPSC derivation [17 –20]. However, besides the challenging technical optimization for efficient oocyte enucleation, donor cell fusion, and cytoplast activation, one limitation for human SCNT is the availability of high-quality oocytes. Ovarian stimulation protocol, oocyte collection timing, and donor-associated factors such as age seem to play crucial role on developmental and reprogramming capacity of human oocytes after SCNT. More studies are needed to properly optimize human SCNT for the generation of pluripotent cells [13,21]. Nonetheless, many laboratories have shown the equivalency of NT-ESCs to in vitro fertilized (IVF) ESCs [13 –15,22]; they have shown that SCNT is transcriptionally and epigenetically more similar to IVF-ESCs [23] and with lower somatic mutation load [24] than iPSCs of syngeneic background. However, a recent comparison of SCNT and iPSC technologies found that both can cause molecular defects and have no significant differences in gene expression, methylation profile, or de novo coding mutation rates [25]. These different findings may arise from the techniques used for iPSC generation, differences among somatic cell lines, and the bioinformatics analysis [26]. Nonetheless, a recent report showed that once a pluripotent cell line has been successfully passaged multiple times in vitro under the same culture conditions, the molecular differences among them tend to disappear [27].
The process of iPSC derivation could, as described by Buganim et al., be divided into three phases. In the first phase, called the “initiation or early phase,” the reprogramming factors introduced into the target cells result in an increase of proliferation, cell death, changes in metabolism, modification of histone residues, activation of RNA processing, and DNA repair. During the second phase, which can last up to several weeks, a stochastic reactivation of pluripotency genes and glycolysis occurs. The final phase, called “maturation and stabilization,” establishes the core pluripotency as well as an ESC-like epigenetic landscape [17,28].
Over the last 4 years, we have made great progress at unveiling the SCNT mechanism that lets an oocyte erase the epigenetic memory of a somatic cell and faithfully establish a pattern of gene expression that more closely resembles that of a preimplantation embryo.
We hypothesize that oocyte reprogramming factors (ORP) bear responsibility for the exceptional reprogramming capacity of SCNT. In what follows, we will describe known ORPs, and propose likely new candidates that could improve our understanding of SCNT, while making the process of iPSC generation more efficient.
Maternal Histones and Chaperones
During SCNT, shortly after exposure of the somatic cell's nucleus to the oocyte cytosol, its chromatin incorporates maternal histones, a well-understood process that also occurs during fertilization [29 –32].
The most studied of such histones are H1FOO, H3.3, and TH2A/TH2B [33], with H2A.X now emerging as a fourth histone variant. These maternal histones also play a critical role in nuclear reprogramming [34] (Table 1).
iPSC, induced pluripotent stem cell; SCNT, somatic cell nuclear transfer.
Overexpression of H1FOO in mouse ESCs favors the maintenance of pluripotency, impairing the normal differentiation of embryoid bodies [35]. Although its overexpression along with the four Yamanaka factors (OSKM) does not appear to increase the efficiency of reprogramming [34,36], it does improve pluripotency acquisition when using only three factors (OSK) [36].
Another maternal histone variant, H3.3—unlike its family members H3.1 and H3.2—can be deposited into the zygotic genome independent of DNA replication, using its committed chaperone HIRA [37]. Donor nuclei must incorporate H3.3 for successful SCNT in Xenopus and mouse [30,34,38,39], and the mouse zygote requires HIRA for the first cell cleavage/division [40]. It remains undetermined whether overexpression of H3.3 and HIRA could increase the efficiency of iPSC derivation.
TH2A and TH2B are highly expressed in mammalian oocytes [34,41]. Shinagawa et al. showed that TH2A/TH2B double mutant oocytes cannot develop normally after fertilization, because they fail to reactivate the paternal genome, suggesting a potential role during acquisition of pluripotency [42]. These results were confirmed when efficient iPSC reprogramming was shown to require TH2A/TH2B, which is likely because these variants increase chromatin accessibility [34]. Their overexpression, together with their phosphorylated chaperone nucleoplasmin (NPM), facilitates the derivation of iPSCs. Research has demonstrated the involvement of TH2A/TH2B and NPM-2 during iPSC derivation using Oct4 and Klf4 alone [34,43].
Another variant of histone H2A present in the oocyte, H2A.X, is incorporated to chromatin after fertilization in both male and female pronuclei and in preimplantation embryos up to the blastocyst stage. H2A.X is also highly expressed in mouse ESCs [44,45]. Phosphorylation of the Ser139 residue of H2A.X (γH2A.X) has been associated with DNA damage response and used to measure DNA damage during SCNT [46]. Although H2A.X incorporation in mice does not require it [45], phosphorylation increases significantly throughout preimplantation development in the absence of DNA damage [47] and during bovine and porcine IVF and SCNT [46,48]. During iPSC reprogramming, the levels of γH2A.X increase and then later decreases during cell differentiation [49,50]. H2A.X depletion reduces the efficiency of iPSC derivation [51], compromising self-renewal activity in ESCs [50], although the exact mechanism remains elusive.
In our laboratory, we have recently shown that ASF1A, a chaperone for histones H3 and H4 required for H3K56 acetylation and specifically enriched in the metaphase II human oocyte [52], is necessary for reprogramming human somatic cells into iPSCs [53] (Table 2). Overexpression of ASF1A and OCT4 in human adult fibroblasts exposed to the oocyte-specific paracrine growth factor GDF9 can reprogram human adult dermal fibroblasts into pluripotent cells. Others have shown, using an in vitro model, that the histone chaperone CAF-1 is required for the incorporation of histones bearing this mark (H3K56 acetylation) in humans [54]. However, the most studied role of the CAF-1 complex is its function promoting histone H3 and H4 deposition onto newly synthesized DNA during replication or DNA repair and establishing repressive histone marks, including H4K20me3 and H3K9me3, at specific DNA regions [55,56]. Supporting this repressive role on chromatin, Cheloufi et al. showed that modulating the downregulation of the p150 and/or p60 subunits of CAF-1 and overexpression of the OSKM factors increased the efficient generation of mouse iPSCs—while decreasing the amount of time needed [57]. However, heterochromatin spatial organization and epigenetic marking of heterochromatin (H3K9me3 and H4K20me3) in mouse pluripotent embryonic stem cells (mESCs) and early embryos are crucial for proper development. A loss of p150CAF-1 in homozygous mutants leads to developmental arrest at the 16-cell stage and a decrease of heterochromatin organization in mESCs [55]. Further study of CAF-1, both its behavior and role during SCNT, could help us understand this apparent discord in the role of CAF-1 on pluripotency.
Regulation of Histone Modifications in Mammalian Oocytes and iPSCs
Histone acetylation
Histone acetyltransferases (HATs) and histone deacetylases (HDACs) control the steady-state level of histone acetylation. More than 17 isoforms of mammalian HDACs have been identified in mammals. They are generally classified into five groups: class I (HDACs 1, 2, 3, and 8), class IIa (HDACs 4, 5, 7, and 9), class IIb (HDACs 6 and 10), class III (SIRTs 1, 2, 3, 4, 5, 6, and 7), and class IV (HDAC 11) [58,59].
The mRNAs of HDACs 1, 2, 3, 4, 6, 8, 9, and SIRT1 are present during mouse oocyte maturation [60,61]; those of HDACs 1, 2, 3, and 7 are present in bovine oocytes [62]. In human MII oocytes, we have found mRNAs for HDACs 1, 2, 3, 6, 7, and 9, as well as for SIRT 1, 2, 6, and 7 [52] (Table 3).
HDAC, histone deacetylases.
Ongoing studies are seeking to determine whether any of these HATs and HDAC enzymes present in the oocyte play a role during reprogramming.
While evidence shows that deacetylation of H3K9, H3K14, and H4K16 is crucial for proper nuclei reprogramming during mouse SCNT [63], this phenomenon seems to be associated only with these specific histone residues and not with general histone deacetylation, as acetylation marks of other lysine residues on core histones, for example, H4K8 and H4K12, persisted in the genome of cloned embryos after SCNT.
Nonspecific HDAC inhibitors (HDACi) such as suberoylanilide hydroxamic acid (SAHA), oxamflatin, valproic acid (VPA), trichostatin (TSA), and sodium butyrate seem capable of increasing the efficiency of nuclear reprogramming [64 –68] (Table 4).
SAHA, suberoylanilide hydroxamic acid; NT-ESC, nuclear transfer-derived embryonic stem cell; VPA, valproic acid; TSA, trichostatin.
In the context of SCNT, SAHA and oxamflatin (HDAC I and IIa/b inhibitors) can improve the development of cloned mice [69,70]. VPA also improves the in vitro development of both bovine and miniature pig SCNT embryos by enhancing nuclear reprogramming, judging by the cleavage and blastocyst formation rates of SCNT embryos and by the decrease of apoptosis in the treated blastocysts [67,71,72].
Perhaps the most striking results using HDACi were reported by Kishigami et al.; they used TSA during SCNT to derive cloned mouse pups and nuclear transfer-derived ESCs (NT-ESCs) [64,65]. They attributed both the observed fivefold increase in the success rate of mouse cloning from cumulus cells and the threefold improvement in the establishment of NT-ESCs from cloned blastocysts to the effects of TSA.
The findings that HDACs (VPA, TSA, sodium butyrate, and SAHA) increase SCNT reprogramming were also used to increase the efficiency of iPSC derivation [25,73 –76] (reviewed in Ref. [77]).
Supporting the use of HDACi, either VPA exposure or Hdac2 downregulation, proved necessary for mouse iPSC generation using miR302/367 reprogramming without the expression of OSKM factors [6].
So far, we have described examples of genome-wide histone acetylation during reprogramming. However, histone acetylation of specific residues is also crucial for cell reprogramming. The presence of acetylated histone H3K56 (H3K56ac) reflects the epigenetic differences between hESCs and somatic cells more accurately than the other active histone marks, such as acetylated H3K4me3 and H3K9 [78]. In human ESCs, H3K56ac correlates positively with the binding of NANOG, SOX2, and OCT4 transcription factors at their target gene promoters, playing a central role in promoting hESC pluripotency [53,78,79]. As mentioned before, the importance of this histone modification in reprogramming was supported by data from our laboratory showing that the oocyte-enriched histone chaperone ASF1A is a key factor for iPSC generation, at least, in part, through H3K56 acetylation of specific promoters [53]. It remains to be determined whether H3K56 HATs (such as p300, CBP, and Gcn5/KAT2A) or specific chemical compounds can affect iPSC generation efficiency.
Recent studies of the localization patterns of key acetylated residues in the core of histone H3 (H3K64ac, H3K122ac, and H3K56ac) in mouse preimplantation embryos underscore the importance of these H3 modification dynamics as potential regulators of cell plasticity [80], making them appealing test targets for their potential to augment iPSC derivation efficiency.
Histone methylation
During SCNT, active histone methylation changes of the somatic genome have been observed as early as the one-cell embryonic stage. This includes the acquisition of such active posttranscriptional histone modifications as H3K4me2/3 and H3K79me3, which are predominantly associated with chromatin accessibility. These residues are also localized in large regions of the oocyte and the ESC genome [81 –84].
Rapid genome-wide changes of H3K4me2/3 distribution are among the earliest events observed in both the initial phase of SCNT reprogramming and after transcription factor reprogramming [85]. Wdr5, a key component of the Set/MLL complex responsible for H3K4 methylation in oocytes, is required for vertebrate (Xenopus) development [86,87]. Wdr5 interacts with Oct4 and has proven essential for mouse embryonic stem cells (ESC) self-renewal and for efficient mouse iPSC reprogramming [88]. It remains undetermined whether its overexpression can improve iPSC reprogramming efficiency.
Erasure and remodeling of repressive histone modifications
H3K4me3, described as a mechanism by which somatic genes of donor cells can retain their epigenetic memory, limits proper reprogramming during nuclear reprogramming and embryonic development [89]. Overexpression of the H3K4-specific demethylase Kdm5b in donor cells improved transcriptional reprogramming in Xenopus NT embryos, making it a potential candidate to improve mammalian iPSC reprogramming. Supporting this idea, Kdm5b reactivation during mouse SCNT proved crucial for proper development [90]; however, total demethylation of H3K4 min IVF embryos impedes zygotic genome activation (ZGA), leading to embryonic lethality and demonstrating the necessity of some levels of methylation for proper development of fertilized embryos [91] (Table 4). Kdm5b is involved in the regulation of both cell pluripotency and differentiation. Kdm5b knockdown leads to a loss of pluripotency [92], whereas its overexpression reduces cell differentiation and increases stem cell self-renewal [93]. Others have shown that kdm5b is dispensable for ESC self-renewal, but essential for ESC differentiation along the neural lineage [94] and that its downregulation favors the self-renewal of mESCs and improves iPSC derivation [95]. While some of these findings seem contradictory, it is clear that a better understanding of the regulatory process controlling H3K9 methylation should help to clarify the cell's transition to and from a pluripotent state (Table 5).
H3K9me3-marked heterochromatin regions are considered epigenetic barriers for both SCNT and transcription factor reprogramming [96 –101]. However, the role of H3K9me in reprogramming seems to depend on context—it can either facilitate or block pluripotency depending on its genome localization. It regulates such key trophectoderm determinant genes as CDX2 and has also proven important for silencing lineage-specific genes [100,102 –104].
The role of H3K9 demethylation during reprogramming was provided by experiments using BIX-01294, a small molecule inhibitor of the G9a histone methyltransferase—also known as KDM4D, which increases the efficiency of transcription factor reprogramming [105,106] (reviewed in Ref. [77] and [107]). Recent SCNT studies show the importance of H3K9 m3 demethylation in improving the efficiency of mouse, monkey, bovine, pig, and human SCNT [90,97,98,108 –111]. Overexpression of KDM4D, and in combination with TSA, was proven crucial for achieving the cloning of macaque monkeys from SCNT embryos [112]. Cloned mouse, human, and monkey blastocyst SCNT embryos treated with Kdm4d and TSA were transcriptionally closer to the IVF embryos used as controls. Nonetheless, a significant percentage of somatic cell genes remained unchanged.
H3K27me3, when catalyzed by polycomb repressive complex-2 (PRC2), silences large regions of chromatin containing developmentally regulated genes. Loss of this mark promotes a transiently open chromatin state, which has been observed in mature oocytes and during the earliest stages of development [113,114], in both SCNT and iPSC reprogramming [115,116]. Utx, a demethylase also known as Kdm6a, mediates the removal of this repressive histone mark [117 –119]. While Utx overexpression does not increase the efficiency of iPSC formation, it is required for the acquisition of ground state pluripotency. Interestingly, mESCs lacking UTX can self-renew and properly differentiate [120].
Many other chromatin modifiers contribute to resetting the epigenomes of reprogrammable cells. Of interest, the methyltransferases KDM2A/B and DOTL1 are responsible for the repressive marks H3K36me2/3 and H3K79m2, and they seem to play a role in reprogramming, as their downregulation significantly enhances the derivation of iPSCs [102,121] (Table 6).
Predominantly associated with accessible chromatin or active promoters.
BMI1, another chromatin-remodeling complex that affects the efficiency of mouse iPSC derivation [122], is a major component of the polycomb group complex 1. Peak expression of BMI1, which is expressed in human and mouse oocytes [52,123,124], occurs at around the one-cell stage [124]. It acts as an essential epigenetic repressor of multiple regulatory genes involved in embryonic development and is required for self-renewal of tissue-specific somatic stem cells. BMI1, in combination with OCT4, can reprogram mouse fibroblasts, functionally replacing KLF4 and c-MYC and increasing reprogramming efficiency [122,125 –127].
DNA methylation
In mammals, 5-methylcytosine (5mC) demethylation is required for early embryonic development and acquisition of pluripotency [128]. In fact, global DNA demethylation appears to be a shared attribute of reprogramming events [22,25,129,130]. At the transcriptome level, early passage ESCs and SCNT-derived pluripotent cells are quite similar at the transcriptome level compared to iPSCs. What sets iPSCs apart is the level of DNA methylation of differentially expressed genes [22]. Adding an inhibitor of DNA methylation, such as 5′ azacitidine, to the cell culture media not only increases the efficiency of reprogramming of somatic cells to iPSCs but also converts a percentage of pre-iPSCs into iPSCs [131]. Similar results were obtained by suppressing the expression of the methyltransferase DNMT1 with short hairpin RNAs [131].
The role of DNA-methyltransferase enzymes during cell differentiation has been known for some time; the mechanism of 5mC demethylation, however, was revealed only recently. Briefly, the 10–11 translocation proteins (TET1, TET2, and TET3), which belong to the family of oxygenases, catalyze 5mC into 5-hydroxymethylcytosine, a first step toward complete demethylation by passive dilution or active excision of oxidized bases [132].
Oocytes have high levels of TET3, presumably for sperm demethylation during normal development. TET3 has also proven responsible for reactivating the OCT4 promoter during SCNT experiments [133]. TET1 and TET2 appear to affect the demethylation and reactivation of genes and regulatory regions important for acquisition of pluripotency during iPSC reprogramming, where interaction with NANOG seems crucial [134 –136]. During reprogramming using the Yamanaka factors, TET1 can replace OCT4 without compromising the derivation of fully pluripotent mouse iPSCs, partly by promoting endogenous OCT4 gene demethylation [136,137]. Whether the absence of TET proteins prevents iPSC reprogramming remains unclear. Experiments using shRNA for TET mRNA support the previous statement [137], others have shown that mouse fibroblasts in which TET1 and TET2 have been knocked out can be efficiently reprogrammed into iPSCs [138].
Other proteins that influence reprogramming through DNA methylation modulation are PARP1, PRG1, and BAF155. The poly (ADP-ribose) polymerase-1 (PARP1) plays a complementary role in establishing early epigenetic marks during somatic cell reprogramming by regulating 5mC modification [135]. BRG1 and BAF155, two components of the SWI/SNF (BAF) chromatin remodeling complex, enhance reprogramming when used with OSKM factors. They facilitate the formation of euchromatin and induce specific demethylation and binding of the reprogramming factors to the regulatory region of key pluripotency-related gene promoters [139].
Roles of DNA Replication and Transcription in Reprogramming Somatic Chromatin
While strong evidence supports the conclusion that RNA synthesis is a key cellular process during nuclear reprogramming using several different methods, that is, cell fusion, SCNT, or iPSC reprogramming, only some methods require DNA replication, while remaining dispensable in others. SCNT and cell fusion experiments have shown that, without DNA synthesis, donor nuclei suffer morphological changes (such as volume enlargement, appearance of diffused/dispersed chromatin, and acidophilic) and increase RNA synthesis soon after transplantation. DNA replication is important for gaining pluripotency during cell fusion and transcription factor-driven reprogramming, as shown by experiments inhibiting DNA polymerase activity that effectively blocks acquisition of pluripotency [140 –148].
Some have proposed that the chromatin changes associated with DNA synthesis provide a window of opportunity for key transcription factors to bind to their response elements [149 –151]. Such oocyte factors as histone B4 [152], NPM [153], nuclear actin, WAVE1 [154,155], and H3.3 (see “Maternal Histones and Chaperones” section) [29] could bear responsibility, at least in part, for modifying the somatic chromatin and making it accessible to the oocyte transcriptional machinery, thus enabling unusually high amounts of Pol II loading.
During SCNT, the oocyte transcription factor TBP2 and the oocyte Pol II subunit RPB1 replace their somatic equivalents in the transplanted nuclei, suggesting that reprogramming requires a shift in the basal transcriptional machinery [142]. Several lines of evidence support this notion. Bui et al. (2010) demonstrated that they could reprogram mouse SCNT embryos more efficiently by incubating them for a few hours with TSA before the first cleavage. They showed that TSA induced high levels of newly synthesized RNA, implying that a transplanted nucleus can be reprogrammed more successfully by switching its transcriptional profile more quickly [156]. Jullien et al. injected mouse somatic cell nuclei into Germinal Vesicle (GV) stage Xenopus oocytes. Their results showed that somatic cells can switch their transcriptional machinery to an oocyte pattern in as few as 6 h, and can complete reprogramming within 48 h [142]. The authors also showed that the transplanted somatic cells expressed the oocyte-specific transcription factors TBP2 and TAF4b, which likely help to establish a new transcription profile in somatic cells [157 –160] (Table 5).
New oocyte-specific transcription factors
So far, we have focused on the oocyte's unique ability to force genome-wide epigenetic changes in somatic cells, which will, in turn, reprogram their nuclei into pluripotency. Nonetheless, the oocyte expresses transcription factors capable of single-handedly inducing nuclear reprogramming. A case in point is the zinc finger protein GLIS1—a nonspecific transcription factor found to be enriched in unfertilized eggs and in embryos at the one-cell stage—that also plays an important role during ZGA and preimplantation development of mouse and bovine embryos [161,162]. GLIS1 can replace c-myc from the OSKM formula in mouse and human fibroblasts. It directly interacts with OCT4, SOX2, and KLF4 in a p53-independent manner, promoting multiple pathways that facilitate the reprogramming and upregulating of other transcription factors, such as NANOG, GATA4, FOXA2, and NKX2 [162] (Table 7). GLIS1 also upregulates the expression of LIN28, an RNA binding protein that prevents the production of active Let-7 microRNA, a known repressor of reprogramming genes. LIN28 reaches its peak of expression at the zygote stage in mice and at the eight-cell stage in humans [163 –166]. Together with OCT4, SOX2, and NANOG, LIN28 is one of the key genes first described for reprogramming human and mouse fibroblasts. The oocyte-specific homeobox 6 (Obox6) transcription factor reportedly improves OSKM reprogramming in mouse cells [167]. Obox6 was first described as a mouse oocyte-specific transcription factor [168]. Others then showed that mESCs [169] and early mouse preimplantation embryos overexpressed it; its highest level of expression occurs from the two-cell stage to the morula stage [170]. A related member of the same gene family, Obox4, is also overexpressed in mESCs, a redundancy among the Obox family members that could explain why Obox6-null mice develop normally, lacking an obvious phenotype normally [169]. Overexpression of Obox4 in mESCs upregulates histone family genes related to chromatin packaging and remodeling, suggesting a role in controlling chromatin structures as well [169]. The potential role of Obox6 in increasing SCNT efficiency remains untested, and its exact cell function is currently under investigation.
Noncoding RNAs
Small noncoding RNAs
Some of the mechanisms governing ZGA include posttranscriptional repression, which is mainly mediated by small RNAs, including the three main small RNA pathways (miRNA, siRNA, and piRNA) [171 –173].
MicroRNAs are transcribed as precursor RNAs that are transported to the cytoplasm and cleaved by Dicer to produce miRNAs that are ∼22 nt long. They are then loaded on Argonaute (AGO) proteins that mediate the silencing effects. The miRNA pathway is the dominant mammalian small RNA pathway. However, RNA sequencing experiments show that maternal mouse miRNAs constitute a relatively minor population during oocyte maturation, contributing to the characteristic mRNA stability of the mature oocyte. MicroRNAs become the dominant small RNA class toward the end of the preimplantation period [174,175]. The maternal miRNAs present in GV oocytes are eliminated and, during meiotic maturation and after fertilization, substituted with“zygotic” miRNAs [175,176]. Among them, the miR-290 cluster, highly abundant in pluripotent stem cells, counteracts the role of miRNA Let-7 as a pluripotency gene repressor [177].
It is crucial to continue to characterize species-specific miRNAs in the oocytes of other vertebrates, since most miRNA preimplantation studies have used mouse and rat models [176,178 –182]. In rodents, miRNAs are functionally blocked in the oocyte [183 –186], and they are relatively independent from the miRNA biogenesis factor DGCR8 for developmental competence [187]. Mice also have the more active oocyte-specific Dicer isoform (DicerO), with its preferential binding affinity for long double-stranded RNA (dsRNA), creating small interfering RNAs (siRNAs) [188] that become the predominant regulatory RNA molecule in mouse oocytes.
MicroRNAs are important regulators of ESC pluripotency, exerting their function by targeting the 3′UTR or the coding region of pluripotency factors OCT2, SOX2, or KLF4, modulating the differentiation capacity of ESCs [189,190] or targeting regions necessary for pluripotency maintenance. ESC-specific miRNAs are also called ESC-specific cell cycle-regulating (ESCC) miRNAs [191,192].
Analysis of the miRNA profile of ESCs and iPSCs led to the discovery of miRNAs capable of affecting the reprogramming efficiency of iPSC generation. The miR-290 and miR-302 clusters can enhance OSKM-based mouse iPSC reprogramming [193]. Similar findings were observed in human cells using the human miRNA orthologs, that is, human miR-302 and miR-372 clusters [5]. Also, when analyzing ESCC miRNAs upregulated during reprogramming (as miR-17-92, miR-106b-25, and miR-106a-363 clusters), miR-93 and miR-106b miRNAs increased OSK MEF reprogramming [194].
In the absence of OSKM or any other exogenous transcription factor, miRNAs can reprogram mouse and human somatic cells into iPSCs. Two different combinations of miRNAs have proven effective: the combination of the miR-200c, miR-302, and miR-369 family [195] and the combined miR-302 and miR-367 clusters [6].
MicroRNA-125b, highly expressed in mature oocytes and early embryos [178,196], has been described as crucial for bovine SCNT reprogramming [196]. Two other strong candidates, miR-21 and miR-130a—due to their high expression in MII oocytes and preimplantation embryos between the two- and eight-cell states, should be further tested in the context of both SCNT and iPSC reprogramming.
The description of the miRNA profile in human oocytes and cumulus cells [182] suggested new miRNA reprogramming candidates for testing: miR-184, miR-100, miR-10A, miR-29a, miR-30d, miR-93, miR-125a, and miR-320a. Of these, only miR-93 was previously tested, and it proved effective at enhancing iPS reprogramming [194], whereas others (miR-184, miR10a, miR100, and miR125a) seem to play roles in blocking pluripotency, as their upregulation in iPSCs or ESCs promotes cell lineage-specific differentiation [197 –200]. Their downregulation favors pluripotency and the capacity for self-renewal (miR-29a, miR-21, miR-30d, and miR-320a) [201 –203].
The miRNA interaction network and function are complex, species specific, and highly temporally regulated during development. Against this backdrop, continuing testing oocyte-specific miRNA candidates is a priority (Table 8).
ESC, embryonic stem cell.
siRNAs are generated from long dsRNA by the RNAse Dicer, provoking enzymatic, sequence-specific mRNA degradation. The amount of endogenous siRNA in most mammalian cells is almost irrelevant [204] and not well understood. The prediction would be that it seems fair to predict that siRNA activity in most mammals is relatively low and nonessential for meiotic progression and ZGA [205]. Mouse oocytes remain an exception, as mentioned; they contain abundant siRNAs, in part, because of the higher activity of their more active specific Dicer isoform (Dicero).
Piwi-interacting RNAs (piRNAs) are longer than miRNAs and siRNAs, and their complex biogenesis involves PIWI proteins (a different class of the AGO protein family expressed during gametogenesis), but not Dicer [206]. While piRNAs are necessary for normal spermatogenesis [207 –209], data from mouse oocytes show that they are not essential for development [207], although comparative analysis with bovine and primate oocytes suggests that this pathway could be more active in other mammalian oocytes [210].
Long noncoding RNAs
Although no consensus exists for an exact definition, the term long noncoding RNAs (lncRNAs) refers to a heterogeneous class of transcripts that are longer than 200 nucleotides and have no evident protein-coding capacity [211 –213]. LncRNAs are usually spliced from multiexonic RNA precursors, processed as mRNA (capped and polyadenylated) and located in both the nucleus and cytoplasm. Although their sequence is less conserved than mRNA or small noncoding RNAs, their structure and function are evolutionarily preserved [214,215]. Information regarding genomics, evolution, and mechanisms have been reviewed extensively and is not the scope of this article [211,213]. We can briefly divide lncRNA functions into four main categories that are not mutually exclusive [212]: (i) molecular scaffolds, normally in the nucleus, that interact with RNA, DNA, and/or proteins to form and regulate these complexes and regulate their function in gene expression as chromatin remodeling complexes—for example, PRC2, WDR5, and MLL, and specific DNA methyltransferases; (ii) guides binding proteins to specific genome sequences to exert their repressing or activating function; (iii) enhancer RNAs (eRNAs) that bind to promoters and induce their transcription activity; and (iv) competing endogenous RNAs (ceRNAs) bind to other cellular components—mainly RNA, miRNA, or proteins—thereby preventing their activity.
Several recent studies point out the importance of lncRNAs in maintaining the pluripotency of ESCs and in reprogramming iPSCs. Guttman et al. identified an lncRNA signature in mESCs that is lost during differentiation [216,217]. This was confirmed by studies showing that many of these lncRNAs regulate, and are regulated by, pluripotency-associated factors [218,219].
Many studies have identified lncRNAs in the oocytes and preimplantation embryos of vertebrates [220 –222], including humans [223]. Interestingly, the lncRNA species reach the lowest levels in oocytes, where maternal mRNA is the most abundant RNA. LncRNA peaks in the early stages of development, at the time ZGA occurs or soon thereafter, becoming the predominant category of transcripts. When analyzing the identity of the lncRNAs of human preimplantation embryos, the majority of them are developmental-stage specific, confirming their subtle regulation and function [224]. This kind of meta-analysis of RNA expression at different stages, including the mature oocyte, provides a novel list of lncRNAs likely to play crucial roles not only during development but also during reprogramming. The LncRNA BCAR4 (breast cancer anti-estrogen resistance 4) is specifically expressed in the mature human oocyte until ZGA occurs, while others, such as AFAP1-AS1 (AFAP1 antisense RNA 1) and AACSP1 (acetoacetyl-CoA synthetase pseudogene 1), extend their expression to the eight-cell embryo and are also activated in certain cancer samples.
Based on experimental evidence and the growing list of lncRNAs expressed during preimplantation development (Table 9), we anticipate that lncRNAs will prove to have a central role during the acquisition and maintenance of pluripotency [224].
lncRNA, long noncoding RNA; PRC2, polycomb repressive complex-2.
Conclusions
The evidence cited in this article supports the notion that the study of genes and gene products present in the oocyte (more specifically the unfertilized metaphase II oocyte) can help us understand how somatic cells acquire pluripotency.
With the advent of new and affordable genomic and proteomic tools, we can now revisit some old questions about the basic mechanisms (and their key regulators) by which an oocyte cytosol exerts its influence on a given nucleus. This is an exciting new era to study the process of cellular reprogramming in a temporally controlled manner and with high level of resolution. Soon we will be able to answer our most urgent questions, while opening new avenues of research in such seemingly unrelated fields as aging, the developmental origins of disease, and cancer.
Footnotes
Author Disclosure Statement
No competing financial interests exist.