Physoxia Influences Global and Gene-Specific Methylation in Pluripotent Stem Cells

Pluripotent stem cells (PSC) possess unlimited proliferation, self-renewal, and a differentiation capacity spanning all germ layers. Appropriate culture conditions are important for the maintenance of self-renewal, pluripotency, proliferation, differentiation, and epigenetic states. Oxygen concentrations vary across different human tissues depending on precise cell location and proximity to vascularisation. The bulk of PSC culture-based research is performed in a physiologically hyperoxic, air oxygen (21% O2) environment, with numerous reports now detailing the impact of a physiologic normoxia (physoxia), low oxygen culture in the maintenance of stemness, survival, morphology, proliferation, differentiation potential, and epigenetic profiles. Epigenetic mechanisms affect multiple cellular characteristics including gene expression during development and cell-fate determination in differentiated cells. We hypothesized that epigenetic marks are responsive to a reduced oxygen microenvironment in PSCs and their differentiation progeny. Here, we evaluated the role of physoxia in PSC culture, the regulation of DNA methylation (5mC (5-methylcytosine) and 5hmC (5-hydroxymethylcytosine)), and the expression of regulatory enzyme DNMTs and TETs. Physoxia enhanced the functional profile of PSC including proliferation, metabolic activity, and stemness attributes. PSCs cultured in physoxia revealed the significant downregulation of DNMT3B, DNMT3L, TET1, and TET3 vs. air oxygen, accompanied by significantly reduced 5mC and 5hmC levels. The downregulation of DNMT3B was associated with an increase in its promoter methylation. Coupled with the above, we also noted decreased HIF1A but increased HIF2A expression in physoxia-cultured PSCs versus air oxygen. In conclusion, PSCs display oxygen-sensitive methylation patterns that correlate with the transcriptional and translational regulation of the de novo methylase DNMT3B.


Introduction
Epigenetics play a key role in pluripotency determination in human pluripotent stem cells (hPSCs) [1]. A powerful laboratory tool, hPSC, is exploited as an in vitro model system to explore differentiation and epigenetic changes during development processes [2,3]. Human-induced pluripotent stem cells (hiPSCs) are an effective complementary aid for hESCs (embryonic stem cells) studies due to their matched properties, including selfrenewal and differentiation into multiple cell lineages [4,5]. Further, both hESCs and hiPSCs have a key role to play as cell sources for regenerative medicine and drug screening research [6,7]. hPSCs display unique epigenetic patterning when compared to differentiated and somatic cells that may contribute to unravelling developmental processes reliant on appropriate gene expression [8,9].
Epigenetic mechanisms consist of DNA methylation, histone modifications, and noncoding RNAs. These can modify chromatin and genomic structures and affect gene expression without changing DNA sequence [10,11]. DNA methyltransferase enzymes (DNMTs) include DNMT1, DNMT3A, DNMT3B, and accessory protein DNMT3L, and establish and maintain DNA methylation patterns in mammals [12]. The structure, mechanism, and function of DNMTs are well defined. DNMT1 is predominantly involved in the maintenance of DNA methylation during cell division while DNMT3A and DNMT3B, which work in coordination with DNMT1, are responsible for de novo methylation, typically during early development in embryonic stem cells [13,14]. DNMT3A and DNMT3B expression is high in undifferentiated ESCs, down-regulated post-differentiation, and remains low in adult somatic tissues [15,16]. DNMT3A and DNMT3B knockout cells can differentiate into the three existing germ layers (ectoderm, endoderm, mesoderm) [16]. Further, early passage DNMT3A and DNMT3B knockout-mouse ESCs can initiate and proceed to terminally differentiated cardiomyocytes and hematopoietic cells where DNMT1 is essential for the maintenance of their differentiation potential [17,18]. The knockout of DNMT3A and DNMT3B in ESCs does not impact their replication potential but does so in differentiated progeny [18].
Oxygen is fundamental for cellular processes. hESCs are derived from preimplantation blastocysts that undergo routine physiological exposure to a 2-5% oxygen environment [19] but paradoxically, air oxygen remains the predominant culture condition. Physiological oxygen tensions promote the maintenance of key properties in embryonic and mesenchymal stem cells, are an essential component of the stem cell microenvironment, and act as signalling molecules in the regulation of stem cell development [20]. Physoxia promotes hESC expansion, clonogenicity, transcriptional (mRNA and miRNA) changes, translational changes, genomic stability, altered metabolic activity, migration, and differentiation features [21][22][23][24][25].
The methylation of genomic CpG islands is correlated to transcriptional silencing and is required for normal development and the differentiation of hESCs [26]. Key transcription factors (Nanog and Oct4/Pou5f1), which maintain stem cell features in ESCs, are unmethylated in undifferentiated stem cells, and become methylated as differentiation progresses [27]. Physiological oxygen (1%) promotes a significant reduction in global DNA methylation levels in human colorectal, melanoma, and neuroblastoma cancer cell lines [28,29]. Moreover, in neuroblastoma, this is accompanied by increased TET1 transcription and elevated global 5hmC levels with the accumulation of 5hmC density at the hypoxia response genes [29]. Similarly, hMSC (human mesenchymal stem cell) isolation and continuous exposure to 2% O 2 resulted in a significant decrease in global 5mC and 5hmC levels, accompanied by a reduced expression of DNMT3B and HIF1A but increased expression of HIF2A [30]. TET1 and TET2 expression is reported as playing a key role in cell fate and differentiation potential in undifferentiated ESCs [31]. TET1 activity, for example, is regulated by physiological O 2 in embryogenesis and is specifically inhibited by low (1%) O 2 [32]. The downregulation of TET1 is associated with decreased global 5hmC, both in undifferentiated and 3-days differentiated mESC in 1% O 2 ; in addition, silencing TET1 increases the expression of pluripotency markers and inhibits the differentiation potential od mESCs [32]. Our previous observations detailed reductions in global 5mC, 5hmC levels and TET1 expression in hMSCs cultured in physoxia [30]. DNMT1 and DNMT3B promoters both possess HIF1A binding sites, suggesting that oxygen levels can play a role in the regulation of DNMT1 and DNMT3B expression via HIF1A transcription factor [33,34]. A prolonged exposure to low oxygen culture is reported to be associated with decreased HIF1A expression in hESCs [19]. Consistent with this report, decreased DNMT3B and HIF1A expression was observed after long term culture in 2% O 2 in hMSCs [30]. The effects of physoxia on DNA methylation in hPSCs and differentiated progeny have received little attention so far.
In this study, we report a significant decrease in global methylation (5mC and 5hmC) and DNMT3B/TET1 expression, which is consistent with increased promoter-specific methylation patterns linked to physoxia culture in PSCs. hPSCs have a great potential to provide new research tools for clinical applications such as drug screening and cell replacement therapies. Understanding the methylation patterning associated with hPSCs differentiation in physiological settings is essential for the development of functionally relevant models.

Cell Culture
SHEF1 and SHEF2 were obtained from the laboratory stock (Guy Hilton Research Centre) and used under the approval of the UK Stem Cell Bank (UKSCB) [35]. The humaninduced pluripotent stem cell line (iPSC) (ZK2012L) derived from human dermal fibroblasts was kindly provided by Professor Susan Kimber and Dr Zoher Kapacee, Faculty of Biology Medicine and Health, University of Manchester. PSCs were routinely cultured in an E8 medium (Life Technologies) with Essential 8 Supplement (1×) in culture vessels coated with 5 mg/mL vitronectin (Recombinant Human Protein; Life Technologies, London, UK). The spontaneous differentiation medium was Knockout DMEM, 10% FBS, 1% NEAA, 1% L-glutamine and β-mercaptoethanol. hPSCs were grown in vitronectin-coated culture plates for 48 h in an E8 medium before switching into a spontaneous differentiation medium. Confluent cells were passaged enzymatically using the 0.5 mM EDTA (Fisher Scientific, Loughborough, UK) and incubated for 3-5 min at 37 • C until cells began to detach. Then, cells were collected in pre-warmed media and centrifuged at 1000 rpm for 3 min. Next, the supernatant was removed, the pellet was re-suspended in fresh media, and seeded at a minimum split ratio of 1:2. Cells were maintained in three different oxygen settings; 21% air oxygen (21% AO), a fully defined 2% O 2 environment (workstation) (2% WKS), and a standard 2% O 2 incubator (2% PG), where samples were handled in a standard class II biological safety laminar flow cabinet. Media utilised in a 2% O 2 setting was deoxygenated to a 2% O 2 level prior to use, using the defined Hypoxycool (Baker Ruskinn, Bridgend, UK) cycle settings.

Metabolic Activity of hPSCs
PSCs were seeded into 96-well plates at a density of 1 × 10 4 cells/cm 2 , incubated for 24 h, an MTT stock solution (15 µL) was added to 150 µL of culture media and they were further incubated for 4 h at 37 • C. Following incubation, a 125 µL aliquot was removed and 50 µL of Dimethyl sulfoxide (DMSO) was added, then they were again incubated for 45 min at 37 • C. For the Alamar blue assay (Fisher Scientific, Loughborough, UK), 1 × 10 4 cells/cm 2 cells were cultured in flat-bottom 96-well plates for 48 h. The medium was then removed and replaced with a 100 µL fresh E8 medium supplemented with 10 µL Alamar blue reagent and incubated for 4 h at 37 • C. After the final incubation steps, both assays were measured with a plate reader (BioTek, Synergy 2, Agilent, Cheshire, UK) at a 570 nm wavelength. Data are presented as a mean ± standard deviation (SD) of the three triplicate experiments.

Flow Cytometry
Cells were characterised using the hPSC imaging kit (Human Pluripotent Stem Cell Marker Antibody Panel kit, SC008, R&D system, Abingdon, UK) with the analysis performed on a flow cytometer (Beckman Coulter Cytomics FC 500). hPSCs were labelled with pluripotency markers (SSEA-4, TRA-1-60 and SSEA-1) and 5 × 10 5 cells were used for each experiment. After harvest, pellets were re-suspended in the flow cytometry buffer (PBS with 0.5% (w/v) BSA and 2 mM EDTA, (Fisher Scientific). The percentage of positive events was determined via the gate exclusion of 99% of the control events. Flowing Software was used for data analysis.

DNA Methylation Analysis
The DNeasy Blood and Tissue kit (Qiagen, Manchester, UK) was used to extract genomic DNA isolation from hPSCs. The total 5mC levels within genomic DNA were measured with the MethylFlashTM Methylated DNA Quantification Kit (Epigentek, Farmingdale, NY, USA) using 100 ng of input DNA. Total 5hmC quantification was performed with the MethylFlashTM Hydroxymethylated DNA Quantification Kit (Epigentek, Farmingdale, NY, USA) using 200 ng of input DNA. Methylation levels for both assays were established by reading absorbance at A450 nm via a microplate reader (BioTek, Synergy 2, Agilent, Cheshire, UK) using program Gen5 1.10 for quantification via the provided standards.

Pyrosequencing
Bisulphite conversion of 500 ng input genomic DNA per sample was performed using EZ DNA Methylation-Gold™ Kit (Zymo Research, Orange, CA, USA). Primers specific for DNMT1, DNMT3A, DNMT3B, DNMT3L, TET1, TET2, and TET3 were designed via the PyroMark Q24 Software 2.0 (Qiagen, Manchester, UK). Primer sequences and product sizes are listed in Table S2 and were supplied by Biomers (Ulm, Germany). Converted DNA (2-4 µL) was used as a template in the PCR reactions and amplifications were performed with the GoTaq ® G2 Flexi DNA Polymerase kit (Promega, Southampton, UK). The cycling parameters were one cycle of 95 • C for 5 min for initial denaturation followed by touch-down cycling for the first 14 cycles, where the temperature was reduced by 0.5 • C in each successive cycle. This was followed by 35 cycles of 95 • C for 45 s, annealing at 55-63 • C for 45 s, elongation at 72 • C for 30 s, and a final elongation step at 72 • C for 5 min. PCR amplification product quality was confirmed via 2% agarose gel electrophoresis. Biotin-labelled PCR amplicons were captured by mixing PCR products with streptavidinsepharose beads (GE Healthcare). The Vacuum Prep Tool with filter probes was used to capture sepharose beads with biotin-labelled PCR amplicons. Filter probes were washed with a 70% ethanol solution, then with a denaturation solution (NaOH) to denature PCR product and a washing buffer to purify the final biotinylated strands. The filter probes were carefully placed into the Q24 pyrosequencing plate wells and the beads were released into an annealing mix. Then, the pyroMark Gold Q24 Reagents, including the four nucleotides, the substrate and enzyme mix were loaded into a pyrosequencing dispensation cartridge. Pyrosequencing was started after the cartridge and Q24 pyrosequencing plate were inserted into the pyrosequencing instrument. Data were analysed using the PyroMark Q24 Software 2.0 (Qiagen, Manchester, UK).

Gene Expression
RNA was isolated from hPSCs with the RNeasy ® Mini Kit (Qiagen, Manchester, UK) and the concentration was quantified using a NanoDropTM 2000/2000c Spectrophotometer (Thermo Scientific, Waltham, MA, USA ). RT-PCR was performed using the QuantiFast SYBR Green OneStep RT-PCR kit (Qiagen, Manchester, UK). PCR primers were obtained from Invitrogen Ltd. and ThermoFisher-Scientific (London, UK). The primer sequences and product sizes are listed in Table S1. PCR amplification was performed on 25 ng of isolated RNA and the relative quantification of gene expression was measured using the 2 −∆∆CT method. All primers had a 55 • C annealing temperature except DNMT3A at 56 • C.

DNMT3B Activity
The EpiQuik™ DNA Methyltransferase 3B Activity/Inhibitors Assay Core Kit (Insight Biotechnology, Wembley, UK) was used to analyse DNMT3B activity.
Blank control wells were labelled, then 3 µL (10 µg) samples were added to the appropriate wells. After samples were prepared, 27 µL of the DNMT Assay Buffer and 3 µL of diluted Adomet were added, 8 mM was added to each well, and they were incubated at 37 • C for 90 min. After incubation, the wells were washed three times with 150 µL of 1× Wash Buffer. Diluted Capture Antibody solution (50 µL) was added to each well, and the plate was incubated at room temperature for 60 min on an orbital shaker (50-100 rpm).
Then, the wells were aspirated and washed three times with 150 µL of the 1× Wash Buffer. The detection antibody (50 µL) was added to each strip well and incubated at room temperature for 30 min. Then, the wells were washed with 150 µL of the 1× Wash Buffer four times. The enhancer solution (50 µL) was added to each strip well and they were incubated at room temperature for 30 min. Again, the wells were washed with 150 µL of the 1× Wash Buffer four times. The developing solution (100 µL) was added to each well and they were incubated at room temperature for 2-10 min in the dark. The colour change was monitored in the wells where a blue colour developed in the wells with DNMT3B enzyme activity. The stop solution (50 µL) was added to stop the reaction and then absorbance was read on a microplate reader at 450 nm within 5-15 min.

Statistical Analysis
The data were analysed using statistical software SPSS (IBM SPSS Statistics 21, Hampshire, UK). A one-way analysis of variance (ANOVA) was performed to assess the comparison among the three groups. The threshold for statistical significance was accepted as p < 0.05. GraphPad Prism 5 (GraphPad Software, La Jolla, San Diego, CA, USA) was used to graphically display the data. Experimental data are represented as mean ± standard deviation (SD) and describe a minimum of 3 independent replicates.

Reduced Oxygen Increases Proliferation and Metabolic Activity in hPSC
We determined the proliferation of hPSCs cultured in air oxygen (21% AO) and reduced oxygen conditions (2% PG and 2% WKS) over a 6-day period. A significant increase in proliferation between days 2 and 6 in physoxia settings was noted in comparison to those cultured in air oxygen (AO, p < 0.05). Their metabolic activity was explored via MTT and Alamar blue assays. A significant increase was again noted for hPSCs cultured in physoxia when compared to AO (p < 0.05). SHEF1 and ZK2012L displayed a significant increase in MTT after day 3 while SHEF2 cells displayed a significant increase in MTT after day 5. Alamar blue indicated increased metabolic activity and the proliferation of cells cultured under both reduced oxygen conditions from day 4 ( Figure S1).

Physoxia Decreases Global DNA Methylation in hPSCs
DNA from undifferentiated and differentiated hPSCs was isolated to explore the global DNA methylation levels in 21% AO, 2% PG, and 2% WKS. Decreased global DNA methylation in hPSCs cultured in physoxia (2% PG and 2% WKS) versus 21% AO was noted. Further, global DNA methylation was reduced in a time-dependent manner as differentiation progressed in PSCs.

Discussion
Epigenetics marks including DNA methylation (5mC and 5hmC) impact cellular characteristics and fate, and are critical for various cellular processes [36]. The role of the methylation of Cytosine residues (5mC) at CpG islands is well studied across a range of biological processes including gene expression, cell differentiation, and development [37]; however, 5hmC is still being explored in cell differentiation and transcriptional regulation in mammalian cells [38]. Oxygen variations can mediate stem cell fate, where, for example, low, physiologic, oxygen conditions increase clonal recovery, maintain a higher number of colony-initiating cells, alter metabolism and morphology, and enhance the proliferation and stemness of PSCs [22,23,25,39]. Here, we focused on the role of physoxia on PSC characteristics in conjunction with global DNA methylation, the regulation of methylation enzymes, and the methylation of their promoters in human PSCs. We determined, in agreement with others, that reduced oxygen conditions enhanced the proliferation and metabolic activity of PSCs when compared to air-oxygen cultured cells [19,40]. We observed that global methylation is oxygen-sensitive and associated with the regulation of DNMT3B transcription and translation.
Immunocytochemistry and flow cytometry demonstrated that undifferentiated PSCs expressed the anticipated pluripotency markers, namely OCT-4, NANOG, ALP, SSEA-4, TRA-1-60, and were negative for SSEA-1. Previous reports suggested that reduced-oxygen culture conditions have no detrimental effect on the expression of stem cell markers in PSCs [19,23,41,42]. We noted elevated OCT-4 and SOX-2, although not NANOG, expression in 2% WKS condition with flow cytometry determining increased TRA-1-60 expression in reduced oxygen conditions. Therefore, we believe that a controlled oxygen environment of 2% WKS provides a more consistent environment for PSCs than 2% PG. Consistent with our observations, hESCs exposed to AO displayed the decreased proliferation and reduced expression of the SOX2, NANOG, and OCT4 gene and protein expression when compared with those cultured at 5% O 2 tri-gas incubators [19,43]. In contrast to the above, other reports detail no significant differences in SOX2, NANOG, or OCT4 expression in hESCs cultured at 2-4% O 2 (again with tri-gas incubators) compared to AO [44,45].
DNA methylation and hydroxymethylation have essential roles in PSC functioning and in the characteristic determination of physiological oxygen conditions [29,46]. Oxygen can alter the global epigenetic landscape, for instance, global histone methylation was decreased at 5% oxygen tension in hPSCs [47]. These observations are consistent with ours, in which we noted decreased global methylation (vs. AO) in 5mC and 5hmC levels in undifferentiated hPSCs exposed to physiological oxygen conditions (2% PG and 2% WKS). Recent reports detail that both 5mC and 5hmC levels are high in ESCs and that these reduce after differentiation [17,[48][49][50][51]. Furthermore, essential transcription factors involved in the ESCs maintenance of pluripotency such as OCT-4 and NANOG are unmethylated in undifferentiated cells and become methylated as differentiation progresses [27,52,53]. We also noted progressively decreased global 5mC and 5hmC levels during the differentiation of PSCs and reduced global methylation in physiological oxygen conditions (2% PG and 2% WKS) in differentiated stem cells compared to AO.
Notably, global methylation levels correlated with decreased DNMT3B and TET1 expression in reduced oxygen conditions. ESCs exhibit robust de novo methyltransferase enzyme (DNMT3A and DNMT3B) expression when undifferentiated, with levels of DNMT3B decreasing with progressive differentiation states [16,[54][55][56]. A previous report described that the expression of DNMT3B decreased at 5% oxygen in undifferentiated hPSCs [47]. Consistent with this, we noted reduced DNMT3B expression in physiological oxygen conditions in undifferentiated hPSCs and subsequently differentiated cells. Methylation levels in the DNMT3B promoter were significantly higher under reduced oxygen settings when compared to AO. CpG island methylation can impair transcription factor binding, recruit repressive methyl-binding proteins, and stably silence gene expression [26]. Global methylation levels and DNMT3B enzyme activity decreased by almost two times after 20 days differentiation in three oxygen conditions with a gradual decrease in both global methylation and DNMT3B enzyme activity observed across this time course. Decreased global methylation levels in physoxic culture were accompanied by reduced DNMT3B, TET1 and TET3 gene expression. Taken together, we evidenced the reduced transcriptional expression of DNMT3B in association with increased methylation of its promoter. In contrast, there was a significant decrease in the TET1 mRNA level but no significant changes occurred for its promoter CpG content in physoxia versus AO.
Long term exposure of hESC to low oxygen decreased the expression of HIF1A mRNA in hESCs [19]. We also noted decreased HIF1A gene expression in physoxia. In contrast, the long-term physoxic culture resulted in upregulated HIF2A at transcript levels. Consistent with this observation, previous reports detailed that HIF2A was upregulated following the long-term culture of hESC in 5% O 2 when compared with air-cultured counterparts [19]. Further, and indicative of a conserved behaviour, HIF2A was upregulated at transcriptional and translational levels in human MSCs in a 5% O 2 culture setting [57]. The repression of HIF2A via siRNA resulted in the decreased proliferation and reduced expression of OCT-4, NANOG and SOX-2 in hESCs cultured under 5% O 2 [19]. In addition, DNMT1 and DNMT3B promoters each displayed hypoxia response elements to HIF binding [33,46]. Therefore, we hypothesise that increased HIF2A and reduced HIF1A under physoxia may associate with, or even drive, the observed decrease in DNMT3B and overall increase its promoter methylation.
In summary, we demonstrated that the physiological oxygen tension increased pluripotency marker expression, proliferation, and metabolic activity with a concurrent decrease in DNMT3B, TET1 gene expression, and global DNA methylation in hPSCs. Further, DNMT3B expression was methylation-regulated in an oxygen-sensitive manner. We show that physoxia conditions modify levels of global methylation linked to DNMT3B expression and that hPSCs display oxygen-sensitive methylation patterns. Importantly, air oxygen is almost never normoxic in mammalian cell culture and increased global methylation can be considered as an artefact of air oxygen culture. Physoxia is an essential component for regenerative medicine to safely realise cell potential and facilitate clean characterisation. Our previous data showed no impact of low oxygen conditions at the histology level in PSCs, but we observed transcriptional changes, which led us to explore molecular alterations in greater detail. Here, we used spontaneous differentiation as a model but there are limitations to this approach as an experimental tool. Future studies may, for example, explore differentiation markers and methylation changes in induced or directed differentiated cells in different oxygen settings. In addition, in vivo differentiation models can be utilised to explore other essential epigenetic markers including histone modifications that may contribute to an enhanced understanding of epigenetics in embryonic development in physoxia.