Dental Pulp Stem Cells Differentiation Reveals New Insights in Oct4A Dynamics

Although the role played by the core transcription factor network, which includes c-Myc, Klf4, Nanog, and Oct4, in the maintenance of embryonic stem cell (ES) pluripotency and in the reprogramming of adult cells is well established, its persistence and function in adult stem cells are still debated. To verify its persistence and clarify the role played by these molecules in adult stem cell function, we investigated the expression pattern of embryonic and adult stem cell markers in undifferentiated and fully differentiated dental pulp stem cells (DPSC). A particular attention was devoted to the expression pattern and intracellular localization of the stemness-associated isoform A of Oct4 (Oct4A). Our data demonstrate that: Oct4, Nanog, Klf4 and c-Myc are expressed in adult stem cells and, with the exception of c-Myc, they are significantly down-regulated following differentiation. Cell differentiation was also associated with a significant reduction in the fraction of DPSC expressing the stem cell markers CD10, CD29 and CD117. Moreover, a nuclear to cytoplasm shuttling of Oct4A was identified in differentiated cells, which was associated with Oct4A phosphorylation. The present study would highlight the importance of the post-translational modifications in DPSC stemness maintenance, by which stem cells balance self-renewal versus differentiation. Understanding and controlling these mechanisms may be of great importance for stemness maintenance and stem cells clinical use, as well as for cancer research.


Introduction
In embryonic stem cell, Oct4 and its protein binding partners form complex autoregulatory circuits in which Oct4, Sox2, and Nanog proteins bind to each other's promoters, constituting an auto-feedback system that is necessary for embryonic stem cell selfrenewal and pluripotency maintenance, by preventing differentiation [1][2][3][4]. It is well established that Oct4 m-RNA and protein disappear relatively quickly following differentiation [1][2][3][4] (although re-appearance of Oct4 m-RNA [5] and protein [6] have been noted). Differentiation pattern is complemented by a decreased expression of the proteins involved in the autoregulatory circuit, such as Nanog, Sox2 [7], and by the expression of specific lineage markers typically expressed in differentiated tissues. The transcription factor Oct4 (POU5F1) is currently considered as a main regulator of ES pluripotency and self renewal abilities [1][2][3][4]. Such stemness properties are attributed to Oct4A, one of the two isoforms produced by the Oct4 gene. The function of the second Oct4 isoform, Oct4B, is still largely unknown [8]. Another important difference between the two Oct4 isoforms is their intracellular localization: while Oct4A behaves as a nuclear protein, Oct4B is a cytoplasmic protein [9]. Post-translational modifications have also been reported as a regulatory mechanism for Oct4, Nanog, and Sox2 in ES cells [2,[10][11][12]. Oct4 and Sox2 are both sumoylated [11,12], Baltus et al. [12] reported that Sox2 is acetylated [12] and also Saxe et al. [2] reported the presence of phosphorylation sites on Oct4 sequence. However, although Oct4 has been observed in adult stem cells, its functional role in this setting is still controversial [13]. In this study we used a human dental pulp stem cell (DPSC) population, whose stemness has been reported by many authors [14][15][16][17][18][19][20], to investigate and compare embryonic and adult stem cell markers in undifferentiated DPSC and afterward differentiation, with particular interest in Oct4A and its subcellular re-localization. Because of its subcellular relocalization in relation to the differentiation switch, we hypothesize here a functional role for Oct4A post-translational modifications in adult stem cells.

Cell isolation, culture and ethical statement
After written informed consent of donor parents and ethics approval from the Ethics Committee of the Medical Faculty of Udine, dental pulp was extracted, using a syringe needle, from human deciduous teeth of 5 to 7 year old children (n = 10) and was cultured as previously described by Ferro et al. [21]. The human serum used in this study was obtained after written informed consent of the donors from the ''Medicina Trasfusionale'' department of the ''Santa Maria della Misericordia'' hospital of Udine. Human embryonic carcinoma stem cells (Ntera2), purchased from ATCC (ATCC-LGC, Milan, IT), were used as positive control for embryonic stem markers as suggested by Liedtke et al. [22], and were cultured according to [23]. MCF7 breast cancer cell line, purchased from ATCC (ATCC-LGC, Milan, IT), was used as negative control for Oct4 expression [24]. Human osteoblast like cells, hOB, (ATCC-LGC, Milan, IT) were used as positive control for osteoblastic differentiation and were cultured by the method [25]. Human primary thyroid cells were cultured as already described by [26], and used as osteoblastic negative control. Human Supranumeral Teeth Buds (STB) were isolated, following same methods used for DPSC, cultured as reported by [27] and were used as odontoblastic positive control. HepG2, human hepato-cellular carcinoma, cell line (ATCC-LGC) was used as positive control for hepatic differentiation and was cultured as in [28]. Small fragments of cardiac tissue were used as positive control for cardiomyocyte (from discarded hearts with permission). Be2C, human neuro-blastoma (ATCC-LGC), cell line was used as positive control for neural differentiation and was cultured as reported in [29].

Osteoblastic differentiation
DPSCs at fifth passage in culture (P5-DPSC) were plated at a density of 4610 4 cells/cm 2 and osteo-induced in a medium already described by Ferro et al. [21,30]. One month old osteodifferentiated DPSC were used for characterization analyses except in X-Ray Diffraction (XRD) and Fourier Transform Infra Red (FTIR) spectroscopy. To demonstrate bone formation DPSC were allowed to aggregate by enzymatically detaching, approximately 5610 6 cells, and centrifuging them at 1610 3 rpm in 15 ml tube. Cells were then transferred to 100 mm dishes, covered by 2% agarose diluted 1:1 with differentiation medium, allowed to form 3D structures and subjected to XRD and FTIR after 3 months.

Real Time PCR analysis
Total RNA was extracted from undifferentiated and differentiated P5 DPSC, positive and negative control cells, using TRIzol (Gibco-Invitrogen, Carlsbad, CA). Quantitative PCR was conducted using SYBR green (Roche, Mannheim, Germany) on a 96well-plate using Lightcycler480 (Roche). The total volume (20 ml) of each PCR reaction contained SYBR Green PCR Master Mix (Roche), 10 ng cDNA, and 0.4 mM of each of the forward and reverse primers. Real Time PCR (n = 4) was performed on undifferentiated and differentiated P5 DPSC, positive and negative controls and primer sequences, PCR product sizes, annealing temperatures and gene bank accession numbers were Oct4A

rt-PCR analysis
Total RNA was extracted from about 2610 6 undifferentiated and differentiated P5 DPSC, positive and negative control cells, using TRIzol (Gibco). RNA samples were quantified by spectrophotometer. After DNase treatment (Ambion), first strand cDNA synthesis was performed with 2 mg of total RNA using 100 ng random hexa-nucleotides and 10 U/ml M-MLV reverse transcriptase (Gibco). PCR amplification was carried out in a final volume of 50 ml; using 0.1 mg cDNA; 0.2 mM dNTPs; 50 pmol of each primer (MWG operon, Ebersberg, GE) and 2 U/ml Taq I polymerase (Gibco). Tissue specific primer sequences, PCR product sizes, annealing temperatures and gene bank accession numbers are reported in Table 1. Optimal conditions and number of cycles were chosen to allow sample amplifications within PCR linear phase. Reaction products were visualized on 1-2% ethidium bromide stained gels. Semi-quantitative analysis was performed on triplicate experiments, using b-actin as normalizing gene and Quantity-One software (Bio-Rad, Hercules, CA). . Unconjugated antibodies for SSEA4 and SSEA1 (0.1 mg/10 6 cells) were revealed using PE conjugated (0.02 mg/10 6 cells) anti-mouse IgG (Chemicon, Cat#AQ326H) and Cy5 anti-mouse IgM (Jackson, Cat#715-176-020), respectively. Conjugated isotype-matching antibodies were used as negative controls. Data (20.000 events) were collected from three independent experiments using a FACS-Calibur (BD) and were expressed as mean 6 standard deviations (SD).

X-ray diffraction (XRD) and Fourier Transform Infra Red spectroscopy (FTIR)
Three month osteo-differentiated DPSC aggregates were powder reduced and subjected to XRD and FTIR. The powder X-ray diffraction patterns were recorded using a X'Celerator diffractometer (PANalytical, Almelo, NL) with Cu Ka radiation (l = 1.5418Å ) and a Ni filter in a 2 h range between 10u and 50u using a resolution of 0.05u. Prior the collection of the diffraction data, samples have been ground in a mortar and put onto a low background silica holder. In the FTIR analysis each powdered sample (approximately 0.1 mg) was mixed with about 10 mg of anhydrous KBr. The mixtures were pressed into 7 mm diameter discs. Pure KBr discs were used as background. The analysis was performed at 4 cm 21 resolution using a Nicolet 380 FTIR spectrometer.

Immunoprecipitation (IP)
Total cell extracts (n = 3) of 40610 6 undifferentiated DPSC, Ntera2 and MCF7 cells were subjected to immunoprecipitation (IP). After a pre-clearing step with Protein A/G agarose beads (SantaCruz, Cat#sc2003) for 3 hours at 4uC, lysates were incubated with protein A/G conjugated anti-Oct4A antibody (SantaCruz, Cat#sc5279-AC) o/n at 4uC. The immunocomplexes were pelleted at 4uC, subjected to 12% SDS-PAGE and were reacted with p-serine (Sigma, Cat#P3430) 1:3000, p-tyrosine (Santa Cruz, Cat#sc7020) 1/3000, incubated o/n at 4uC to identify phosphorylated residues. Primary antibodies were reacted with HRP conjugated anti-mouse IgG and anti-rabbit IgG, both 1:7000, for 1 hour at RT; antibody/antigen complexes were detected using ECL advance (GE Healthcare, Milano, IT). Whole undifferentiated DPSC cell extracts were used as input and were reacted with Oct4A antibody as previously described.

Protein identification and phospho-analysis
Immunoprecipitated Oct4A samples, obtained from undifferentiated DPSC, were subjected to 12% SDS-PAGE and stained by MS compatible FireSilver Staining Kit (Proteome factory, Berlin, GE). The bands corresponding to those obtained in WB were excised, destained and washed for 1 h in dH2O. Proteins were allowed to diffuse out of the crushed gel overnight at 30uC, by incubation in 0.5 ml of elution buffer (50 mM Tris-HCl, 150 mM NaCl, and 0.1 mM EDTA; pH 7.5), and supernatants were recovered after centrifugation at 10.0006 g for 10 minutes [36]. Then samples were reduced, alkylated with iodoacetamide, and digested with trypsin (Sigma) at 20 ng/ml and 37uC for 90 min. The resulting peptides were concentrated on a ZipTip C18 (Millipore) micropurification column, and eluted onto an anchorchip target for analysis on a Bruker Autoflex III MALDI TOF/TOF instrument (Bruker Daltonik, Bremen, GE). The peptide mixture was analyzed in positive reflector mode for accurate peptide mass determination. MALDI MS/MS was performed approximately 12 peptides for partial peptide sequencing and the MS and MS/MS spectra were combined and used for database searching using the MASCOT software. Separately, Oct4A higher band derived from undifferentiated DPSC was electrophoresed into 12% SDS-PAGE gel, excised and processed as previously reported. Then Oct4A samples and phospho protein standard were digested with Glu-C at 10 ng/ml (Sigma), diluted in 50 mM Tris buffer for 2 hours at 37uC. Samples were also purified on TiO 2 POROS R3 micro-column (Glygen Corp, Columbia, MD), the amount purified was ,2 pmol of the standard and ,100 pmol of the samples. Finally, elutes were used for detection by mass spectrometry and the MS and MS/MS spectra were combined and used for database searching using the MASCOT software.

Statistical analysis
Data from experiments are expressed as mean 6 SD of at least three independent experiments. Statistical significance was determined by unpaired Student's t test; P,0.05 was considered significant.
Cardiomyocyte differentiation was shown after 4 weeks in differentiation, during this period DPSC changed morphology to become elongated and irregular (Fig. 2G). Differentiated DPSC showed organized filaments for a-sarcomeric actin (Fig. 2I) and gap-junctions presence was demonstrated by connexin-43 (Cx-43) in proximity of contact sites (Fig. 2J). Serca 2 ATPase pump (Fig. 2K) was also identified in differentiated cells. a-smooth muscle actin (SMA) was highly expressed during differentiation showing a specific filamentous structure as demonstrated by IF (Fig. 2L). It has been reported that SMA is present in embryonic or fetal but not in adult cardiomyocytes, suggesting that these phenotype may represent an early differentiation stage [40,41].
Morphology of neural-differentiated DPSC closely resembled mature neurons: they had a large number of neurites, increased 3 to 4 weeks after differentiation, with significant branching (Fig. 3A) [42]. Semiquantitative rt-PCR showed that DPSC increased m-RNA expression for ß3-Tubulin (ß3-Tubulin) (2.2360.01) fold,  (Fig. 3B). ß3-Tubulin was largely expressed in DPSC, showing a typical filamentous expression pattern, after differentiation as shown by immunofluorescence (Fig. 3C) and at the same time neuro-differentiated DPSC expressed tyrosine hydroxylase along neurites (Fig. 3D). Structural neurofilaments NF-M were expressed only during differentiation, the positivity was about 50% (Fig. 3E) [32]. Consistently with the evidence [43], which proves that GFAP and ß3-Tubulin are co-expressed during early differentiation phase, GFAP (Fig. 3F), an astrocyte marker, was always expressed in DPSC but afterward differentiation was less expressed and organized.

ES and adult stem cell markers expression during differentiation
To establish whether differentiation of P5 DPSC was associated with changes in the expression of cluster differentiation (CDs) markers, which were previously demonstrated to be expressed on human mesenchymal stem cells [16][17][18][19][20]31,32,[45][46][47][48], FACS analysis were performed. Results evidenced that differentiation process was associated with a significant reduction (p,0.05) in the expression of CD10 (from 9265% to 0.160.09%), CD29 (from 9862% to 0.760.5) and CD117 (from 1562% to 0.760.05%), while only small differences were detected for the other tested antigens which showed a general decreased expression after differentiation had taken place (Fig. 4A, 4B and Table 2). In addition we also analyzed whether P5 DPSC differentiation was associated with changes in the expression of stemness-related genes, a Real Time PCR was performed quantifying the transcripts of Oct4A, Oct4AB, Oct4B, Nanog, Klf4 and c-Myc with respect to the housekeeping gene RNA polymerase type II (RP II) and expressing the results as a ratio over undifferentiated cells (Fig. 5A). Interestingly, differentiated DPSC expressed, with respect to the undifferentiated cells, significantly reduced level of Oct4A (0.2460.072; p,0.05) fold, Oct4AB (0.3960.094; p,0.05) fold, Nanog (0.4060.094; p,0.05) fold and Klf4 (0.4860.048; p,0.05) fold. Conversely, the differentiation process was associated with a significant increase in Oct4B transcripts (2.9160.04; p,0.05) fold and c-Myc transcripts (1.4060.019203; p,0.05) fold.
Characterization was further improved by detecting Nanog and Oct4A proteins by means of immunofluorescence. Both undifferentiated DPSC (Fig. 7A, 7B) and Ntera2 (Fig. 7C, 7D) cells showed high nuclear Nanog protein presence. By contrast, differentiated    DPSC showed reduced cytoplasmic and nuclear Nanog expression (Fig. 7E, 7F). Evaluating the expression of the Oct4A protein, it was apparent that in undifferentiated DPSC Oct4A was localized not only in the nucleus but also in the cytoplasm (Fig. 7G, 7H).
Conversely, Ntera2 showed a predominantly nuclear positivity and a weak cytoplasmic signal, probably because of the high nuclear to cytoplasmic ratio of these cells (Fig. 7I, 7J). In differentiated DPSC, the localization of the pluripotent-state specific transcription factor Oct4A was mainly cytoplasmic (Fig. 7K, 7L).

Oct4A protein analysis
In the attempt to explain the high Oct4A expression rate and acquire new evidences about its sub-cellular localization afterward differentiation; we performed WB, extending electrophoresis running time, loading both whole cell extracts, cytoplasmic and nuclear extracts of undifferentiated, differentiated DPSC, Ntera2 and MCF7 cells. Results indicated that Oct4A protein can be resolved in two distinct bands for whole extracts (Fig. 8A), as well as for nuclear and cytoplasmic extracts (Fig. 8B) of undifferentiated, differentiated DPSC and Ntera2 cells. Nuclear and cytoplasmic bands evidenced in WB were then quantified by densitometry and the resulting data (Fig. 8C), reported as differentiated over undifferentiated DPSC expression ratio, were: higher Oct4A nuclear band (1.0260.027) fold, lower Oct4A nuclear band (0.5360.017) fold, higher Oct4A cytoplasmic band (1.3260.037) fold, lower Oct4A cytoplasmic band (1.2560.019) fold (P,0.05).
To confirm Oct4A identity, whole cell extracts of undifferentiated DPSC and Ntera2 were subjected to immunoprecipitation using anti Oct4A antibody, then the immunocomplexes were subjected to WB. Separated proteins were evidenced by MS compatible silver staining and the bands corresponding to those previously obtained, which we supposed to be Oct4A, were excised and subjected to protein identification by MS. Protein identification confirmed that both proteins correspond to Oct4A, as evidenced by relative m/z graphs and by Oct4A sequence matching (bold underlined peptides), respectively for higher (Fig. 8D) and lower (Fig. 8E) band.
Based on studies describing post-translational modifications of Oct4 [2], we hypothesized that the higher Oct4A band might represent one or more post-translational modifications (PTM) and specifically phosphorylations. For this reason Oct4A protein sequence was analyzed using PhosphoMotif Finder, www.hprd. org, and the results revealed 35 serine and 10 tyrosine kinase/ phosphatase motifs. Therefore, taking into consideration these results, undifferentiated DPSC samples were immunoprecipitated, using anti Oct4A antibody, and were reacted with anti-phospho serine and anti-phospho tyrosine antibodies, to evidence the presence of one or more phosphorylation sites. Results showed that Oct4A immunoprecipitated samples possessed one or more phosphorylated serine (Fig 8F), instead we were not able to detect tyrosine phosphorylated residues (Fig. 8G).
To confirm Oct4A post-translational modification (PTM) Oct4A higher band, derived from undifferentiated DPSC, was extracted from the gel and was subjected to phospho-analysis using MS. MS spectrum of TiO 2 purified sample and MS/MS annotated spectra demonstrated the presence of one possible phosphorylation site (Fig. 8H), which corresponds to Oct4A serine residue S 105 or S 107 (Fig. 8I). Unfortunately from the data it was not possible to establish whether the phosphorylation was exactly placed, however our previous phospho motif analysis evidenced that S 105 is a putative Casein Kinase II (CK-II) phosphorylation site (Fig. 8I), and phospho-motif analysis suggested that serine 107 has to be phosphorylated already for the enzyme to recognize the motif and phosphorylate serine 105.

Discussion
There are many investigations describing the role of human DPSC for tissue regeneration, advancing their therapeutic relevance as a valuable stem cell source [14][15][16]19]. DPSC are commonly cultured in FBS, which poses risk of transferring infections and induction of immune reactions upon transplantation. HS has been considered to be a safer alternative excluding the transfer of animal derived infections and related immunogenic reactions and our data evidence that low HS percentages support the isolation of DPSC, showing a well-defined stemness related phenotype and multilineage differentiation properties.
The aim of this study was to establish the role played by the ES core transcription factor, which includes c-Myc, Klf4, Nanog, and Oct4, as well as that of the stemness related CDs in adult stem cells. With this intent we investigated the expression pattern of these markers both in undifferentiated and fully differentiated DPSC.
It is well established that stemness in ES cells is under the control of Oct4 [1][2][3][4]49,52,62], in particular Niwa et al. [3] reported a direct correlation between Oct4 expression and stemness maintenance, evidencing that the Oct4 expression level drops below 650% of normal levels when cells lose their pluripotency.
Subcellular compartimentalization of transcriptions factors is an important mechanism to regulate their activity. Many transcriptions factors localize to the cytoplasm in their basal, unstimulated state, needing to be activated and imported into nucleus to initiate the expression of their target genes, others, as Oct4A in stem cells, are located mainly in the nucleus where initiates the expression of their target genes; however both types are exported from the nucleus for their recycling and/or regulation [63][64][65][66]. It has been also demonstrated in different cell types, as well as in ES cells, the identification of a phosphorylation mediated mechanism, which permits nuclear/cytoplasmic translocation degradation and consequent cell differentiation [64][65][66]. In this research we provide evidence that nuclear Oct4A is down-regulated to about 50% during DPSC differentiation, as well as that a post-translationally modified Oct4A form increases in cytoplasm afterward the acquisition of the differentiated phenotype. Moreover, MS permitted us to suppose a CK-II-dependent serine phosphoryla-tion site at residue 105 or 107 in Oct4A protein, and phosphomotif analysis suggested that serine 107 has to be phosphorylated already for the enzyme to recognize the motif and phosphorylate serine 105.
CK-II is a ubiquitous Ser/Thr protein kinase that plays a central role in the regulation of a variety of cellular processes, and has been found to facilitate phosphorylation of nuclear proteins [65] and mediate their nuclear/cytoplasmic shuttling [66] (i.e. via CRM1-dependent mechanism [67]). Oct4 m-RNA and protein disappear relatively quickly following differentiation [1][2][3][4], but at the same time have been also noted their re-appearance [5,6] suggesting that a critical amount of Oct4 is required to maintain stemness [3]. Taking into consideration these observations, this study confirms Oct4A importance in stemness maintenance even in adult stem cells, and suggests that a CK-II-dependent phosphorylation mechanism could be involved in its nucleo/ cytoplasmic shuttling, by which could balance stemness versus differentiation in DPSC. Considering the Oct4A importance, we think that this could not be the sole mechanism involved in the regulation of its nuclear amount and we also suppose the presence of a redundant control mechanism, both for Oct4A as well as for the other core transcription factors, as already reported for Klf4 [64]. In conclusion understanding and controlling these mechanisms may be of great importance for stemness maintenance and stem cells clinical use, as well as for cancer research.