OTX2 impedes self–renewal of porcine iPS cells through downregulation of NANOG expression

The transcription factor Otx2 acts as a negative switch in the regulation of transition from naive to primed pluripotency in mouse pluripotent stem cells. However, the molecular features and function of porcine OTX2 have not been well elucidated in porcine-induced pluripotent stem cells (piPSCs). By studying high-throughput transcriptome sequencing and interfering endogenous OTX2 expression, we demonstrate that OTX2 is able to downgrade the self-renewal of piPSCs. OTX2 is highly expressed in porcine brain, reproductive tissues, and preimplantation embryos, but is undetectable in fibroblasts and most somatic tissues. However, the known piPSC lines reported previously produced different levels of OTX2 depending on the induction procedures and culture conditions. Overexpression of porcine OTX2 can reduce the percentage of alkaline phosphatase-positive colonies and downregulate NANOG and OCT4 expression. In contrast, knockdown of OTX2 can significantly increase endogenous expressions of NANOG, OCT4, and ESRRB, and stabilize the pluripotent state of piPSCs. On the other hand, NANOG can directly bind to the OTX2 promoter as shown in ChIP-seq data and repress OTX2 promoter activity in a dose-dependent manner. These observations indicate that OTX2 and NANOG can form a negative feedback circuitry to regulate the pluripotency of porcine iPS cells.


INTRODUCTION
Pluripotent stem cells have two pluripotency states, naive and primed, depending on the origin of cells and the culture conditions. 1 In mouse, embryonic stem cells (ESCs) derived from the inner cell mass in the early stage of the blastocyst retain naive pluripotency, whereas epiblast stem cells derived from the epiblast in the late stage of the blastocyst retain primed pluripotency. [1][2][3] Clear distinctions between the two pluripotent states have been defined on the basis of cell morphology, cytokine supplementation, signaling pathways, formation of chimera and germline transmission in vivo, and transcriptomics. Naive pluripotent stem cells show compact and dome-like colonies, depend on leukemia inhibitory factor (LIF), and are able to generate germline transmitting chimeras. 1,2,4,5 In contrast, primed pluripotent stem cells form flattened colonies, depend on basic fibroblast growth factor (bFGF), and show very limited capacity to generate chimeric offspring. 2,6,7 Furthermore, a set of transcription factors (TF), such as Nanog, Oct4, Esrrb, and Otx2, has been shown to regulate the transaction between naive and primed pluripotent states. 1,8 The TF Otx2, which plays an essential role in the regulation of brain and sense organ development and neuronal differentiation, [9][10][11] is a crucial regulator in the transition of murine naive ESCs into primed epiblast stem cells. 12 Increasing the activity of Otx2 causes a significant enrichment of FGF protein expression and reduces the generation of chimeric embryos. Conversely, absence of Otx2 leads to an increase of LIF/STAT3 signaling activity. 12 Genome-wide mapping of enhancer activity and protein-DNA interaction profiles show that Otx2, as a cell-statespecific regulator, can interact with Oct4 and combine with primed-dominant enhancers to drive the reorganization of enhancer usages during differentiation. 13 Therefore, Otx2/Oct4bound enhancers are important for maintaining cellular identity and leading pluripotent stem cells to exit from naive state pluripotency. 14 In addition, knockdown of Otx2 leads the primed state cells to flip to naive ESCs by increasing the expression of Esrrb that is directly regulated by Nanog, and Esrrb can form a binding partner with Oct4 to bind to naive-dominant enhancers. 14,15 Thus, Otx2 may have a mutually exclusive effect on Nanog, which is well studied as a critical factor on blocking the differentiation of pluripotent stem cells 16,17 and maintaining the pluripotent state of stem cells. 12 Porcine-induced pluripotent stem cells (piPSCs) have been reported by several laboratories worldwide. [18][19][20][21][22][23][24] Some piPSC lines showed the primed state with bFGF-dependence and mouse epiblast stem cells-like morphology; [18][19][20][21] other piPSC lines showed the naive-like state with LIF-dependence and mouse ESC-like morphology. 23 Besides, piPSCs cultured with both LIF and bFGF represented the metastable state. 24 So far, porcine ESCs and naive state piPSCs are difficult to generate. The underlying problems are improper culture conditions used to generate the piPSCs and unclear cell-state-specific regulatory circuitries. Thus, naive piPSC generation will benefit from an understanding of the genetic and epigenetic mechanisms that control the self-renewal and differentiation of piPSCs.
In our previous studies, mRNA expression profiles showed clear differences in the expression status of TFs in LIF-dependent, 23 bFGF-dependent, 18 and LIF/bFGF-dependent piPSCs. 24 Interestingly, OTX2 expression in LIF-dependent piPSCs was distinguished from that in bFGF-dependent piPSCs, suggesting that OTX2 might act as a molecular marker to classify the different pluripotent states of pig iPSCs. In this study, we explored whether OTX2 was functionally relevant to the pluripotency of piPSCs. Also, we dissected the relationship between OTX2 and NANOG in such regulation.

RESULTS
Porcine OTX2 expression pattern In porcine tissues, including testis, ovary, and brain, OTX2 was highly expressed (Figure 1a). The qRT-PCR analysis further confirmed that OTX2 expression in testis, ovary, and brain was 8-17-fold higher than that in other tissues (Figure 1b). Since OTX2 was reported to be relevant to embryo development, 25,26 we then investigated the OTX2 expression in porcine embryos. Results of reverse transcription polymerase chain reaction (RT-PCR) and quantitative RT-PCR (qRT-PCR) showed high OTX2 expression in oocytes, indicating that OTX2 is a maternal factor; OTX2 expression increased in the preimplantation embryos during the zygotic transition starting from the eight-cell stage (Figures 1c and d). In cell-based experiments, OTX2 was expressed in porcine LIF/bFGFdependent iPS cells-piPS-g 24 and PS23 22 cells, but not in somatic porcine embryonic fibroblasts (PEF; Figure 1e). This observation allowed us to use these cell lines to explore the function of OTX2. RNA-seq data (E-MTAB-2634) showed that OTX2 was highly expressed in bFGF-dependent iPS cells (piPS-w), 18 but OTX2 expression was rather low in LIF-dependent iPS cells (piPS-j). 23 In LIF/bFGF-dependent iPS cells (piPS-g), OTX2 expression fell somewhere between piPS-w and piPS-j ( Figure 1f). Expression patterns of endogenous OTX2 in piPSCs were confirmed by qRT-PCR ( Figure 1g).
Characterization of porcine OTX2 Phylogenetic analysis showed that pig OTX2 was evolutionarily related to cattle OTX2, whereas human, baboon, marmoset, rat, and mouse shared the most Otx2 sequence identity (Figure 2a). Porcine OTX2 CDS was amplified from porcine brain tissue in which OTX2 is an essential factor for brain development, 12 confirmed by DNA sequencing, and submitted to GenBank (Accession No. KP779653). BLAST research found that the porcine OTX2 gene was located on the complementary strand of chromosome 1 and contained three exons (Figure 2b). Alignment of the cloned OTX2 CDS with the pig genome sequence showed 99% identity (Supplementary Figure S1). The 2 kb OTX2 promoter fragment and an 892 bp fragment of 3′ untranslated region (3′ UTR) were also cloned from porcine brain tissue, and confirmed by DNA sequencing (Supplementary Figure S2). The porcine OTX2 promoter sequence was submitted to GenBank (Accession No. KR135411).
The protein sequence of OTX2 has 288 amino acids (aa) and exhibits high homology with human and mouse Otx2 (Figure 2b). On the basis of the Conserved Domain Database, 27,28 OTX2 belongs to the OTX TF super family (Accession No. pfam03529). OTX2 has a homeodomain sequence with 56 amino acids (Accession No. cd00086; Figure 2c), which is predicted to be involved in DNA binding and homodimer or heterodimer formation in a sequence-specific manner.
To investigate the function of OTX2 in eukaryotic cells, we constructed the expression vectors pEGFP-OTX2 and pcDNA-OTX2, which were confirmed by enzyme digestion (Figure 2d). The pEGFP-OTX2 and a control pEGFP-C1 were transfected into HEK-293 T cells for 36 h, and the presence of EGFP-OTX2 fusion protein was confirmed by western blotting (Figure 2e). In a cell-based assay, EGFP-positive and EGFP-OTX2-positive cells were observed with fluorescence microscopy (Figure 2f). EGFP-OTX2 fusion protein was translocated in nuclei, whereas EGFP protein was displayed in whole cells (Figure 2f, inset). Thus, the cloned OTX2 CDS can be transcribed into functional protein and be used for further research.
OTX2 influences morphology and expression of pluripotent genes in piPSCs To investigate the regulatory function of OTX2 in piPSCs, the piPS PS23 cells were transfected by pcDNA-OTX2. Western blotting confirmed that cells transfected with pcDNA-OTX2 expressed high-level OTX2 protein (Figure 3a). Alkaline phosphatase (AP) staining showed that OTX2 + piPSC colonies were partially reduced the AP staining compared with the control group, which displayed a more compact morphology with uniform AP staining ( Figure 3a). The proportion of AP + colonies in OTX2 + piPSCs was significantly lower than that in the control group (Figure 3b), indicating that the pluripotent state of piPSCs was disturbed by ectopic OTX2 overexpression. To further uncover changes in self-renewal, we then examined endogenous pluripotent gene expression in OTX2 + piPSCs. The level of OTX2 in OTX2 + piPSCs was significantly increased comparing with the endogenous level of OTX2 in piPSCs ( Figure 3c). However, the expression level of NANOG was reduced more than 70% and the expression levels of OCT4 and ESRRB were reduced to 50% in OTX2 + piPSCs. Conversely, the expression of PITX2, a differentiation-related gene, was significantly increased ( Figure 3c). Interference in the expression of other core pluripotent genes, such as SOX2 and KLF4, by OTX2 overexpression was observed, but was not statistically significant ( Figure 3c). This observation proved the concept similar to the recently reported results that overexpression of OTX2 significantly downregulated expression of pluripotent genes NANOG, OCT4, SOX2, and so on. and decreased the self-renewal, survival and cell migration in human ES cells. 29 The RNA interfering assay showed that expression of OTX2 in both DNA level and protein level was significantly knocked down by all three siRNAs, in which siR-543 and siR-1115 caused 95% reduction in OTX2 expression ( Figure 4a). Besides, the AP staining showed that piPSCs treated with siRNAs exhibited much more compact clones compared with the control group ( Figure 4b). Moreover, statistics data showed that the percentage of fine AP-stained colonies in siR-543 treatment was significantly higher than that in a control group (Figure 4c). Additionally, we found that the expressions of endogenous NANOG, OCT4, and ESRRB significantly increased in piPSCs, and PITX2, a factor that regulates cell differentiation, was remarkably downregulated in the presence of reduced OTX2 expression ( Figure 4d). Interestingly, SOX2 expression was reduced significantly when OTX2 was knocked down, suggesting that further investigation of SOX2 function might reveal its role in regulating the pluripotent state in piPSCs. These observations indicated that OTX2 knockdown could stabilize the self-renewal and reduce the differentiation potential of piPSCs.
OTX2 downregulates NANOG expression It has been reported in mouse embryos that loss of Otx2 could severely affect Nanog expression. 12 RNA-seq data generated from different stages of porcine embryos showed that endogenous NANOG was highly expressed in eight-cell embryos, whereas OTX2 was highly expressed at the morula stage ( Figure 5a). 30 Transcriptome analysis, confirmed with qRT-PCR, showed that OTX2 expression increased significantly in the morula stage, and NANOG expression at the eight-cell stage was twofold higher than that at four-cell and morula stages (Figure 5b). In previous experiments (Figure 3c), we noticed that endogenous NANOG expression was significantly downregulated in OTX2 + piPSCs, suggesting that OTX2 might play a role in negatively regulating NANOG expression. To determine the effect of OTX2 on NANOG expression, the pcDNA-OTX2 and reporter vector pGL-NANOG were transiently cotransfected into HEK-293 T cells in which both OTX2 and NANOG were not expressed (Figure 5c). Luciferase assays showed that the activity of the NANOG promoter could be remarkably downregulated by OTX2 in both time-dependent and dose-dependent manners (Figure 5d). To further investigate activation of the NANOG promoter, expression vectors carrying OTX2, OCT4, and SOX2 were cotransfected with reporter pGL-NANOG into HEK-293 T cells, respectively. Results showed that addition of OCT4 and SOX2 significantly enhanced NANOG promoter activity, similar to a previous study. 31 However, addition of OTX2 alone significantly reduced NANOG promoter activity (Figure 5e). To investigate the synergistic effect of combining OTX2 with OCT4/SOX2, we applied OTX2 with OCT4 and SOX2. NANOG promoter activity increased slightly compared with the treatment lacking addition of OTX2, but the increase was not statistically significant ( Figure 5e). This result suggests that OTX2 and OCT4/SOX2 do not have a synergistic effect on the activity of porcine NANOG; however, the exclusive application of OTX2 does impede NANOG promoter activation.
To further investigate OTX2 regulation of NANOG expression, we constructed a series of reporter constructs of porcine NANOG, which retained the truncated promoter sequence reported previously. 31 Within the promoter sequence, seven putative OTX2 binding sites and multiple putative OCT4 and SOX2 binding sites were found on the basis of the online JASPAR software (Figure 5f, left). Double enzyme digestions confirmed the accuracy of these recombinant vectors (Figure 5f, right). NANOG reporter pGL-NANOG and three truncated constructs pGL-N1, pGL-N2, and pGL-N3 were transiently cotransfected with pcDNA-OTX2 into HEK-293 T cells, respectively. The promoter activity was significantly repressed in cells transfected with pGL-N3 and pGL-NANOG, but not in cells with pGL-N1 and pGL-N2 (Figure 5g). These results indicated that OTX2 could block activation of the NANOG promoter and its binding sites were probably located in the distal region of the NANOG promoter. In the future study, the gel shift assay and ChIP-seq experiment might reveal whether OTX2 binds directly or indirectly to the NANOG promoter.
NANOG directly regulates OTX2 expression ChIP-seq datasets from mouse, human, and pig showed that Nanog could directly bind to the mouse and human Otx2 promoter region. However, pig ChIP-seq profiles exhibited numerous noisy due to the quality of the anti-NANOG antibody that lacks pig specificity (Figure 6a). 32 To investigate whether NANOG could regulate porcine OTX2 expression, we cloned the 2 kb porcine OTX2 promoter and constructed OTX2 reporter vectors pOTX2-GFP and pGL-OTX2. The pOTX2-GFP and pEGFP-C1 were cotransfected into PEFs, a porcine epithelial cell line (PK15), and a porcine iPS cell line (PS23). Results showed that the OTX2 promoter was activated in PK15 and PS23 cells, but not in PEF cells, indicating that the cloned OTX2 promoter retains cell-type specificity (Figure 6b). NANOG overexpression in piPSCs significantly reduced OTX2 expression (Figure 6c). The dose-and timedependent assays showed that the activity of the porcine OTX2 promoter could be repressed by NANOG in a dose-dependent manner (Figure 6d). To further monitor the NANOG binding region, the truncated OTX2 reporter constructs were made and used for luciferase assays (Figure 6e and Supplementary Figure S3). Within the 2 kb promoter region, there are multiple predicted binding sites of NANOG, OCT4, and SOX2. After removing NANOG binding sites in distal region of OTX2 promoter, OTX2 activity was not repressed by NANOG (Figure 6f). These observations indicate that NANOG can negatively regulate activity of porcine OTX2 promoter. Thus, the binding of NANOG to the OTX2 promoter could form negative feedback regulatory circuitry to regulate the self-renewal of porcine iPS cells.

DISCUSSION
As a homeodomain-containing TF in mammals, OTX2 regulation is essential for the normal development of brain and the genesis of photoreceptors. [9][10][11]33 Also, Otx2 is described as the marker of the anterior central nervous system. 34,35 The known Otx2 in several mammals is highly conserved and shares a conserved homeodomain region. High-level expression of Otx2 was detected in mouse/human brain. 11,36,37 However, in pig, OTX2 is highly expressed not only in brain but also in reproductive tissues, as observed in this study. We found that porcine OTX2 transcripts initially existed in maternal components in oocytes, but were absent at two-cell and four-cell stages; the OTX2 level increased significantly after the eight-cell stage. In both mouse and human, 25,26 Otx2 expression was upregulated along with embryonic development, indicating that OTX2 is an intrinsic TF activated in the stage of zygotic genome activation and determines cell fate in early embryonic development. 33 We also found that in porcine iPS cell lines, OTX2 expression was fourfold higher in primed piPSCs than in naive-like piPSCs (Figure 1g). Furthermore, the OTX2 expression level in metastable state cells was between the levels in naive-like and primed cells. These results indicate that OTX2 can be used as a marker to evaluate the pluripotent state of piPSCs. However, more precise determinations of cell type and culture condition must be applied to verify this hypothesis in future work.
We found that OTX2 and OCT4/SOX2 did not have the synergistic effect on promoting NANOG expression. Alternatively, overexpression of OTX2 alone could repress the expression of pluripotent genes in piPSCs. OTX2 has been shown to bind to the human tenascin-C promoter and transrepress tenascin activity. 38,39 Investigations of regulatory circuitries of mouse Otx2 and other pluripotent factors have shown that Otx2 can interact with Oct4 and the two proteins cobind to primed-dominant enhancers to activate genes that promote exit from naive pluripotency. 14,40 As a cell-state-specific regulator, Otx2 dynamically binds to enhancers, H3K4me1 and H3K27ac, that have low histone signals, and opens the previous inaccessible chromatin sites in which Oct4 and other TFs can cobind to these sites and promote enhancer-promoter interactions to elevate the expression of the set of genes that are related to lineage differentiation in primed cells. 13 Thus, OTX2 is a negative regulator to influence iPS cell pluripotency. We assume that switching off porcine OTX2 expression during cellular reprogramming may help to gain the naive state piPSCs.
Of note, an abundance or lack of OTX2 in porcine iPS cells repressed and activated the NANOG promoter, respectively, in a time-and dose-dependent manner ( Figure 5). Studies of singlecell expression profiling under different chemical and genetic perturbations showed that the expression of polycomb target genes, including Otx2, existed in a repressed yet poised state with a unique chromatin signature in iPSCs and negatively correlated with Nanog expression, representing regulators governing initial steps in lineage commitment. 41 Similar observations were reported in mouse, in which constitutive and ubiquitous expression of Otx2 leads to a substantial reduction in Nanog expression; conversely, reduction of Otx2 causes strong activation of Nanog. 12 In mouse Otx2 − / − ESCs, cell colonies exhibited a sphere-like morphology, uniform AP staining, ubiquitous distribution of Nanog and Klf4, and a higher abundance Rex1. 12 In piPSCs, We found that knockdown of OTX2 with RNA-mediated interference also improved cell morphology and increased the expression of NANOG, OCT4, and ESRRB; however, KLF4 expression was unchanged and was independent of OTX2 activation, suggesting that other KLF family members may functionally overlap with KLF4, as reported previously, 24 to regulate porcine cell reprogramming.
The ChIP-seq result in the pig sample displayed severe noise due to the low quality of the anti-NANOG antibody. 32 ChIP-seq data from mouse and human showed that NANOG could directly bind to the Otx2 promoter region, which were also associated with Oct4 binding. A dose-dependent luciferase assay confirmed that NANOG repressed OTX2 activation ( Figure 6). Our findings suggest that OTX2 and NANOG may form negative feedback regulatory circuitry to maintain pluripotent states in piPSCs.
We found that, unlike OCT4 and NANOG expression, SOX2 expression was positively correlated with OTX2 expression. Overexpression of OTX2 slightly elevated SOX2 expression, however, knockdown of OTX2 expression significantly reduced SOX2 expression (Figure 4), suggesting that the SOX2 protein, which also functions to specify neural lineage, is a determinant of cellular reprogramming potential and is required for epiblast maintenance. 42,43 A previous study of Otx2 and Sox2 coregulation of Rax gene expression in fog showed that Otx2 overexpression was accompanied by increased expression of Sox2, in which the Otx2 protein directly bound to multiple cis-regulatory elements (N-2, AGATTA; N-3, GGATTA) that spatiotemporally control Sox2 expression. 44,45 Thus, the regulatory function of the SOX2 protein in piPSCs may be different from that in mouse pluripotent stem cells. It would be interesting to investigate the transactivation role of SOX2 during the transition of pluripotent states in piPSCs.
In summary, we have identified OTX2 as an important cell-statespecific regulator of the fate and pluripotency of piPSCs, through inhibition of pluripotent gene expressions. We also showed that OTX2 and NANOG exhibited negative feedback circuitry to balance the pluripotency of piPSCs. Further studies will determine the functional relationship between OTX2 and other pluripotent factors and dissect the molecular mechanisms that maintain pluripotency in piPSCs.
To construct the OTX2 reporter vector, a 2 kb OTX2 promoter fragment was amplified by PCR and cloned into pGEM-T Easy vector (A1360, Promega, Madison, WI, USA) and confirmed by DNA sequencing. The porcine OTX2 promoter sequence was showed in Supplementary Figure S2 and also submitted to GenBank (Accession No. KR135411). The OTX2 promoter fragment was then subcloned into Xhol/HindIII sites of pEGFP-1 (#6086-1, Clontech) and pGL3-basic (U47295, Promega), respectively. The recombinant vectors were named as pOTX2-GFP and pGL-OTX2. To dissect the core regulatory region of the OTX2 promoter, we amplified four DNA fragments on the basis of 2 kb OTX2 promoter sequence and constructed four reporter vectors that retain the truncated OTX2 promoter.

Cell transfection
To evaluate OTX2 expression, cells were seeded in culture dishes 24 h before transfection. After reaching 80% confluence, 4.0 μg pEGFP-OTX2 was transiently transfected into HEK-293 T cells plated on six-well plate using TurboFect transfection reagents (R0532, Thermo, Fair Lawn, NJ, USA) according to the manufacturer's instruction. After 36 h, GFP-positive cells were examined and collected for western blotting. To determine the tissue specificity of the OTX2 promoter, the vector pOTX2-GFP was transfected into PEF, PK15, and PS23 cells, respectively, with Lipofectamine 2000 Regent (#11668-019, Invitrogen) for 36 h.
To overexpress OTX2 in porcine pluripotent stem cells, PS23 cells plated on a six-well plate were transiently transfected with 3.5 μg pcDNA-OTX2 and pEGFP-OTX2, respectively, using Lipofectamine 2000 Regent for 36 h. The transfection efficiency of pEGFP-OTX2 was over 35%. To knockdown OTX2 expression in piPSCs, three small interfering RNAs (siRNAs; siR-543, siR-863, and siR-1115) and a negative control (siR-Ctrl) were synthesized (GenePharma, Shanghai, China), and the sequence of siRNAs are listed in Table 1. The 200 nM siRNA siR-543, siR-863, siR-1115, and siR-Ctrl were transiently transfected into PS23 cells, respectively, with Lipofectamine 2000 Regent for 36 h. The transfection efficiency of siRNA determined with a fluorescent-labeled control siRNA (GenePharma) was more than 60%. To increase siRNA interference efficiency, a second transfection of siRNA was conducted at 12 h after the first transfection, and samples were analyzed at 24 h after the second transfection.

Alkaline phosphatase staining
To perform AP staining, PS23 cells were washed twice using ice-cold phosphate buffered saline (PBS, pH 7.4), fixed with 4% paraformaldehyde in PBS for 10 min at room temperature, and washed three times using icecold PBS. Cells were then incubated at room temperature in 0. Data sources and bioinformatics analysis