Disruption of RING and PHD Domains of TRIM28 Evokes Differentiation in Human iPSCs

TRIM28, a multi-domain protein, is crucial in the development of mouse embryos and the maintenance of embryonic stem cells’ (ESC) self-renewal potential. As the epigenetic factor modulating chromatin structure, TRIM28 regulates the expression of numerous genes and is associated with progression and poor prognosis in many types of cancer. Because of many similarities between highly dedifferentiated cancer cells and normal pluripotent stem cells, we applied human induced pluripotent stem cells (hiPSC) as a model for stemness studies. For the first time in hiPSC, we analyzed the function of individual TRIM28 domains. Here we demonstrate the essential role of a really interesting new gene (RING) domain and plant homeodomain (PHD) in regulating pluripotency maintenance and self-renewal capacity of hiPSC. Our data indicate that mutation within the RING or PHD domain leads to the loss of stem cell phenotypes and downregulation of the FGF signaling. Moreover, impairment of RING or PHD domain results in decreased proliferation and impedes embryoid body formation. In opposition to previous data indicating the impact of phosphorylation on TRIM28 function, our data suggest that TRIM28 phosphorylation does not significantly affect the pluripotency and self-renewal maintenance of hiPSC. Of note, iPSC with disrupted RING and PHD functions display downregulation of genes associated with tumor metastasis, which are considered important targets in cancer treatment. Our data suggest the potential use of RING and PHD domains of TRIM28 as targets in cancer therapy.


Introduction
Due to self-renewal and pluripotency maintenance properties, induced pluripotent stem cells (iPSC) exhibit several features that are also characteristic for cancer cells, e.g., the similar expression profile of many genes or activity of signaling pathways regulating self-renewal [1][2][3]. Epigenetic and transcriptional dysregulations in tumor cells disturb many signaling pathways also responsible for maintaining the phenotype of normal stem cells, leading to progressive dedifferentiation and acquisition of stemness features [2,3]. The stemness score is the lowest in somatic cells, increased in primary tumors, and reaches the highest level in tumor metastases. This indicates that tumor progression usually involves the process of oncogenic dedifferentiation [3]. Because of many similarities, iPSC can Nuaillé, France) for 3 min at 37 • C. After incubation, the solution was removed and residues inactivated with a culture medium. The colonies were gently detached from the plate using a soft silicone scraper, carefully broken up in suspension into smaller aggregates by pipetting, and transferred to new Matrigel-coated culture vessel.
All plasmid vectors were transformed into E. coli JM109 (#P9751, Promega, Madison, WI, USA Research) and isolated with JETSTAR 2.0 Plasmid Maxiprep Kit (#220100, Gentaur, Kampenhout, Belgium). The volumes given are sufficient for one 100 mm plate. The transfection mixture was incubated for 5 min at RT and added to the cells. Cells were incubated for 6 h under standard culture conditions. Then, the medium was changed to 6 mL of fresh medium. The supernatant containing the virus particles was collected 48 h after transfection and centrifuged at 3000 rpm, 5 min, RT.

Lentiviral Vectors Production, Purification, and Titration
Virus-containing supernatant was purified and concentrated by ultracentrifugation in Ultra-Cone Polyallomer Centrifuge Tubes (Seton Scientific, Petaluma, CA, USA), in Sorvall Discovery 100S Ultracentrifuge (Kendro Laboratory Products, Asheville, NC, USA). Supernatants (20 mL) were carefully added dropwise on a 4 mL layer of 20% saccharose (#107651, Merck KGaA, Darmstadt, Germany) in distilled H 2 O. Tubes were centrifuged at 26,000 rpm, for 1.5 h at 4 • C. The supernatant was decanted, and the pellet (barely visible/invisible) was suspended in 2% BSA (#A9418, Sigma-Aldrich, St. Louis, MO, USA) in DPBS (#L0615, Biowest, Nuaillé, France) and incubated for 20 min at RT. The precipitate dissolved in the buffer was intensively pipetted, aliquoted, and stored at −80 • C. Viral titers were determined by real-time PCR. HEK 293T cells were seeded at 20,000 cells/well of a 6-well plate in a complete culture medium. After 24 h, cells were transduced with each lentivirus produced, in volumes of 2, 5, and 10 µL, in duplicates. The flow cytometry-titered pWPXL lentiviral vector expressing EGFP was used as the reference. The medium was replaced 24 h after transduction with a fresh one, and the cells were cultured for another 5 days with the medium changed every other day. Cells were harvested by trypsinization (#25-053-CI, Corning, Corning, NY, USA), and DNA was isolated with Quick-DNA Miniprep (#D3025, Zymo Research, Irvine, CA, USA). All procedures were performed according to the manufacturer's protocol. We used primers amplifying WPRE (present in integrating vector fragment) and GAPDH sequence. Primer sequences are listed in Table S2. Amplicons were detected with TaqMan hydrolysis probes (#04683633001, Roche, Basel, Switzerland). The number of infectious particles was calculated as described in the protocol [32].

Reprogramming of Human Fibroblasts towards iPSC
PHDF cells, in an early (2-3) passage, were seeded 10,000 cells/well of a 6-well plate in a complete culture medium. After 24 h and 48 h, cells were transduced with Stemcca-TetO lentiviral vector (50 IU/well). On day 7 after transduction, cells were passaged into 6-well plates, coated with BD Matrigel™ Matrix, Basement Membrane, GFR (#354230, BD Biosciences, San Jose, CA, USA), and MEF cells (#PMEF-CF, Millipore Merck KGaA, Darmstadt, Germany), approx. 4000/well. Cells were cultured in an iPSC medium as described in Cell culture, p. 3. The medium was changed every other day. Twenty-one days after transduction, the clusters of cells with stem cell-like morphology were manually transferred to freshly prepared wells of 6-well plates coated with Matrigel and MEF cells at 100% confluency. From this point, the iPSC medium was changed daily. After 2 passages, iPSC were transduced with LVE-HK lentiviral vector (hygromycin resistance), 10 IU/well, to silence the expression of exogenous reprogramming agents. IPSC were selected for 7 days with 50 µg/mL Hygromycin B (#H3274, Sigma-Aldrich, St. Louis, MO, USA).

Generation of TRIM28-Depleted or Mutated iPSC Populations
Silencing of TRIM28 with siRNA in iPSC (obtained by reprogramming) was performed in 5 iPSC lines generated from different fibroblast donors, in duplicate. The results collected in this paper are representative results from the iPSC 26.6. iPSC were treated with a mix of two equimolar (50 nM) siRNA particles (Table S2) and Lipofectamine RNAiMAX Transfection Reagent (#13778-075 The next day, the cells were washed 3 × 5 min with DPBS and incubated 1 h in the dark, RT, with 200 µL fluorescently labeled secondary antibodies in DPBS with 1% BSA, in a ratio of 1:1000. After washing the cells three times with DPBS, the cell nuclei were stained with a DAPI (#32670, Sigma-Aldrich, St. Louis, MO, USA) in distilled water (1:10,000) for 5 min in the dark at RT. After washing the cells three times with DPBS, the cells were analyzed with Leica DMI3000B fluorescence microscope (Leica Microsystems, Wetzlar, Germany) and Leica Application Suite (RRID:SCR_016555, Leica Microsystems, Wetzlar, Germany). Primary and secondary antibodies are listed in Table S3.

cDNA Samples Preparation
Total RNA was isolated from 2 biological replicates (for siRNA-treated iPSC) or 3 biological replicates (for LV-treated ND41658 iPSC) using TRI Reagent (#T9424, Sigma-Aldrich, St. Louis, MO, USA). Reverse transcription was performed with an EvoScript Universal cDNA Master (#07912439001, Roche, Basel, Switzerland), according to the manufacturer's protocol. One µg of total cellular RNA was used for each reaction. Obtained cDNA was diluted 10-fold in sterile DEPC water and used as an RT-PCR and real-time PCR template.

Real-Time PCR Quantification
Samples were amplified on a LightCycler ® 480 instrument (Roche, Basel, Switzerland). The gene expression level was quantified with 2 −∆∆CT method, relative to the control sample. Primer sequences are listed in Table S2.
siRNA-treated iPSC were analyzed every 7 days after the first siRNA transfection for 3 weeks.    Table S2.

Proliferation Assay
The proliferation ratio was determined in 3 biological replicates, with Cell Proliferation ELISA, BrdU, and colorimetric (#11 647 229 001, Roche, Basel, Switzerland), according to the manufacturer's protocol. Briefly, 5000 cells were seeded in a 96-well plate in 100 µL/well and incubated at hypoxic conditions at 37 • C for 48 h. BrdU labeling was performed for 2 h. Cells were incubated with the anti-BrdU solution for 90 min and then with a substrate solution for 10 min. The absorbance was measured at 370 nm and calculated by subtracting the blank control absorbance value. Statistical results were calculated in GraphPad Prism6 (RRID:SCR_002798, GraphPad Software). One-way ANOVA variance analysis with a post-hoc Dunnett's test was conducted.

Spontaneous In Vitro Differentiation Potential Assessment (Embryoid Bodies Formation)
Analysis was performed on 3 biological replicates of the iPSC 6 passages after lentiviral transduction. All variants of iPSC were harvested and washed twice with DPBS to remove Matrigel residues. Cells were then counted and seeded at 5000 cells/well on a 96 well, non-adherent, U-shaped plate in Essential 6™ Medium (#A1516401, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) in standard culture conditions. Images and videos were taken with the Incucyte SX1 Live-Cell Analysis System (#4788, Sartorius, Göttingen, Germany), and the spheres' area was calculated with ImageJ (RRID:SCR_003070). Statistical results were calculated in GraphPad Prism6 (RRID:SCR_002798, GraphPad Software) with a Kruskal-Wallis test followed by a Dunns' test. . Membrane fragments were incubated with primary antibodies in 5 mL 5% milk in TBST buffer, at 4 • C, overnight. Membranes were washed 3 times with 10 mL TBST buffer, for 10 min, and incubated with secondary HRP-conjugated antibodies in 5 mL 5% milk in TBST buffer, followed by triple washing with 10 mL TBST buffer. Antibodies were visualized with a WesternBright Quantum HRP substrate (#K-12042, Advansta, San Jose, CA, USA). Antibodies are listed in Table S3.

Teratoma Formation
iPSC were harvested with 0.1% collagenase IV (#17104019, Gibco, Thermo Fisher Scientific, Waltham, MA, USA), and 2 × 10 6 of cells were resuspended in 50 µL iPSC medium. Before injection, 50 µL of BD Matrigel™ Matrix Basement Membrane GFR (#354230, BD Biosciences, San Jose, CA, USA) was added to the cell suspension at 4 • C, and the mixture was injected subcutaneously into the lower flank of immunodeficient NOD SCID mice NOD.CB17-Prkdc scid /NCrCrl (Charles River Laboratories, Wilmington, MA, USA). All animal experiments were performed following institutional guidelines. After 7-9 weeks, tumors were resected, measured, and subjected to RNA isolation and immunohistochemical staining. Paraffin sections of formalin-fixed teratomas were stained with hematoxylin and eosin (H + E) and antibodies specific for markers of three germ layers: endoderm cytokeratins, ectoderm GFAP, and mesoderm desmin. Analysis was performed in the Department of Tumor Pathology, Greater Poland Cancer Centre in Poznań.

Karyotyping of Generated iPSC Lines
Karyotype analyses were performed by the Cytogenetic Laboratory, Cancer Centre-Maria Sklodowska-Curie Institute in Warsaw, according to the standard protocol for G-banding.

Proteomic Profiling-Reverse Phase Protein Array (RPPA)
Analysis was performed on 1 biological replicate. Pellets of PHDF cells and derived iPSC were washed with DPBS (#L0615, Biowest, Nuaillé, France) and lysed in 100 µL of RIPA buffer (#J63306, Alfa Aesar, Ward Hill, MA, USA). Cell lysates were centrifuged at 13,000 rpm, at 4 • C for 30 min, and supernatant was collected. Protein concentration was determined by BCA reaction with a Pierce™ BCA Protein Assay Kit (#23225, Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer's protocol. Each sample (40 µL) was mixed with 4X SDS sample buffer: 40% Glycerol (#A1123, AppliChem, PanReac, Darmstadt, Germany), 8% SDS (#L4509, Sigma-Aldrich, St. Louis, MO, USA), 0.25 M Tris-HCl (#10812846001, Roche, Basel, Switzerland), pH 6.8, and 10% β-mercaptoethanol (#ES-007-E, Merck KGaA, Darmstadt, Germany), and boiled for 5 min. Samples were stored in −80 • C. Samples were analyzed in the RPPA Core Facility, The University of Texas, MD Anderson Cancer Center (Houston, TX, USA). Each sample was diluted in five 2-fold serial dilutions in 1% SDS lysis buffer. Serially diluted lysates were arrayed on nitrocellulose-coated ONCYTE ® Film slides (Grace Bio-Labs, Bend, OR, USA) by Aushon 2470 Arrayer (Aushon BioSystems, Billerica, MA, USA) in 11 × 11 format. Each slide was probed with a validated primary antibody plus a biotin-conjugated secondary antibody. A total of 305 unique antibodies were used. The signal obtained was amplified using a Dako Cytomation-catalyzed system (Agilent Technologies, Santa Clara, CA, USA) and visualized by DAB colorimetric reaction. The slides were scanned, analyzed, and quantified using MicroVigene (RRID:SCR_002820, VigeneTech Inc., Carlisle, MA, USA) software to generate spot intensity. Each dilution curve was fitted with a logistic model SuperCurve Fitting (Department of Bioinformatics and Computational Biology in MD Anderson Cancer Center, Houston, TX, USA) utilizing the R environment (RRID:SCR_001905, CRAN family, http://www.r-project.org/, accessed on 27 June 2021). The fitted curve was plotted with the signal intensities on the y-axis and the log2-concentration of proteins on the x-axis. The protein concentrations of each set of slides were then normalized by median polish, corrected across samples by the linear expression values. A correction was performed using the median expression level of all antibody experiments to calculate a loading correction factor for each sample.

Principal Component Analysis
Principal components were calculated using the ClustVis web tool [34] (RRID:SCR_017133, University of Tartu, Tartu, Estonia). The imputation and Singular Value Decomposition (SVD) were performed iteratively until estimates of missing values converge was performed. As an input, linear normalized RPPA data were used for all tested samples. The first two principal components (PC1, PC2) were plotted.

Differential Expression Analysis
Proteins from RPPA data were filtered based on the adjusted p-value < 0.05 and presented in volcano plots. All significantly differentially expressed proteins were clustered with the Morpheus tool [35] (RRID:SCR_014975, Dresden University of Technology, Dresden, Germany) and visualized as heatmaps. The distance was calculated with one minus Pearson's correlation coefficient metric.

Gene Set Enrichment Analysis (GSEA)
GSEA (http://www.broad.mit.edu/gsea/index.html access date: 26 March 2020) was used to detect the coordinated expression of a priori defined groups of genes within the tested samples. Gene sets are available from the Molecular Signatures Database (MSigDB, RRID:SCR_016863, Broad Institute, Cambridge, MA, USA, http://software.broadinstitute. org/gsea/msigdb/index.jsp access date: 26 March 2020). Briefly, GSEA generated an enrichment score (ES) reflective of the degree to which a gene set is overrepresented at the extremes (top or bottom) of the entire list of RPPA data. Genes are ranked according to expression difference (signal/noise ratio) between the tested group of samples: PHDF-WT/PHDF-CTRL and iPSC-WT/iPSC-CTRL, where CTRL cells are treated with a control siRNA. The ES calculation and estimation of the p-value, together with the normalized enrichment score (NES) and FDR calculations, have been previously described in detail [36]. A total of 305 markers (previously ranked based on their log2FC between analyzed groups) were imported for GSEA. The GSEA was run according to the default parameters: each probe set was collapsed into a single gene vector (identified by its HUGO gene symbol), permutation number = 1000; permutation type = "gene-sets." The FDR was used to correct for multiple comparisons and gene set sizes.

Over-Representation Enrichment Analysis (ORA)
The ORA [37] was performed with a WEB-based Gene SeT AnaLysis Toolkit (We-bGestalt; http://www.webgestalt.org/ access date: 28 March 2020) with the "pathway" database. Protein names were transferred into gene symbols, and the reference gene list was set at "genome protein-coding". Upregulated and down-regulated markers were considered separately.

Generating Human iPSC with Doxycycline-Inducible System Results in Repression of Transgene Expression in Established Clones
IPSC were generated by reprogramming primary human dermal fibroblasts (PHDFs) with Yamanaka factors: OCT3/4, SOX2, KLF4, and c-MYC (OSKM) [38], delivered in Stemcca-TetO lentivirus, under the control of the doxycycline-inducible system ( Figure S2A). PHDF cell lines were established from the healthy skin margins surrounding breast cancers excised during mastectomy. Application of the inducible system allowed to switch off OSKM expression in obtained iPSC colonies by transducing them on day 21 with LV-HK lentivirus. LV-HK vector carried the expression of the KRAB tet-repression domain, which, in the absence of doxycycline, inhibits EF-1α promoter ( Figure S2A). Re-induction of transgene expression would require the presence of doxycycline in culture media.
Obtained iPSC displayed typical [38] round colony morphology with a clear, regular peripheral outline ( Figure S2B). Colonies expressed intra-and extra-cellular pluripotency markers (positive for OCT3/4, NANOG, SSEA-4, TRA-1-60, and TRA-1-81, and negative to differentiation marker SSEA-1), comparable to the expression profile of hESC ( Figure S2C). The CpG methylation status in the promoter regions of OCT3/4 and NANOG was evaluated by bisulfite sequencing. The analyzed promoters were unmethylated in iPSC compared to parental PHDF cells, indicating the activity of these promoters in iPSC ( Figure S2D). Moreover, we confirmed a similar expression profile of pluripotency markers (REX1, NODAL, DNMT3B, OCT3/4, GABRB3, NANOG) in hESC and obtained iPSC ( Figure S2E). Chromosomal G-band analysis showed normal karyotypes with no chromosomal aberrations in generated lines ( Figure S2F). Immunohistochemical and H + E staining of iPSC-derived teratoma sections proved iPSC potential to differentiate into ecto-, endo-and mesoderm ( Figure S2G). We also confirmed no transgene expression from the integrated Stemcca-TetO vector in teratomas ( Figure S2H).

Silencing of Endogenous TRIM28 Induces Downregulation of Pluripotency Markers and Differentiation of Human iPSC
To evaluate the role of TRIM28 in stemness, we silenced endogenous TRIM28 expression with small interfering RNA (siRNA) in two generated iPSC lines. IPSC treated with siRNA with no target sequence served as a control (siCTRL) of the experiment. Upon TRIM28 silencing, the stemness and differentiation status was examined every 7 days for 3 weeks ( Figure 1A). The silencing efficiency was confirmed on transcriptional and protein level by qRT-PCR and immunofluorescence staining ( Figure 1B). The significant downregulation of TRIM28 in the CTRL sample at day 21 might result from that control siRNA, and transfection reagents can influence mRNA and protein levels [39]. Upon TRIM28 silencing, we observed decreased expression of extracellular pluripotency markers, SSEA-4, TRA-1-60, TRA-1-81 ( Figure 1C), and progressive loss of intracellular pluripotency markers OCT3/4 and NANOG ( Figure 1D). Finally, differentiation-associated markers (CDX2, MSX1, FSP1, PAX6, SOX1) were upregulated in siTRIM28 cells on the transcript level ( Figure 1E). Our data indicate that TRIM28 knock-down facilitates differentiation of iPSC.

iPSC with Silenced TRIM28 Display Metabolic Changes, and Their Proteomic Profile Differs from the Control iPSC
Two weeks after silencing TRIM28, we evaluated metabolic changes by immunofluorescence staining (Figure 2A). During reprogramming of somatic cells into iPSC, the metabolic profile shifts from oxidative phosphorylation (OXPHOS) to glycolysis [40][41][42]. Upon cellular differentiation, the metabolic profile shifts back to OXPHOS, and mitochondrial activity is restored. Silencing of TRIM28 resulted in high upregulation of mito-chondrial ATP synthase subunit D (ATP5PD), which indicates OXPHOS metabolism in differentiated cells. Evident downregulation of glycolysis-related markers, including phosphoglycerate kinase (PGK), pyruvate kinase (PKM2), and hexokinase 1 (HK1), compared to wild type (WT) cells, also confirmed inhibition of glycolysis processes as a result of differentiation. The expression of HK2 did not change significantly upon TRIM28 silencing.  iPSC with silenced TRIM28 were subjected to RPPA analysis, and their proteomic profile was compared with WT and siCTRL iPSC. Raw data are openly available in GEO at www.ncbi.nlm.nih.gov/geo access date: 3 July 2020, reference number: GSE153726. Principal Component Analysis (PCA) showed that the samples from each group clustered together, and the groups were clearly segregated ( Figure 2B). We determined 185 differentially expressed proteins, of which 64 were significantly downregulated and 121 upregulated between iPSC siTRIM28 and reference iPSC (WT and siCTRL) ( Figure 2C). Differentially expressed proteins showed clustering of markers from iPSC_WT, iPSC_siCTRL, and iPSC_siTRIM28 samples ( Figure 2D). Pathway enrichment analysis using Gene Ontology datasets showed significant upregulation of pathways related to apoptosis, differentiation, cellular response to DNA damage stimulus, and cell cycle regulation in iPSC-siTRIM28, relative to reference iPSC ( Figure 2E). In contrast, reference iPSC demonstrated enrichment of the processes involved in the cellular response to organonitrogen and nitrogen compounds and processes related to the regulation of phosphorylation and cell proliferation ( Figure 2F). The list of markers assigned to individual processes is presented in Figure S3. Among the markers upregulated upon TRIM28 silencing, we found some tumor suppressor genes, e.g., MSH2, CHEK2, ANXA, or CAV1 ( Figure S3A). One of the downregulated markers in iPSC-siTRIM28 (upregulated in reference iPSC) was TRIM28, and a few protooncogenes, e.g., EIF4E, BRAF, ARAF ( Figure S3B). These results indicate the role of TRIM28 in the regulation of several signaling pathways implicated in maintaining self-renewal and stemness of iPSC, and probably of highly dedifferentiated metastatic tumor cells as well.

Selection of Mutation Sites Impairing the Function of TRIM28 Protein Domains
To determine the TRIM28-dependent mechanisms responsible for the self-renewal and pluripotency maintenance, we selected eight different mutation sites within TRIM28. Mutations impaired the functions of its particular domains by their effect on phosphorylation of the domain structure ( Figure 3A,B). All mutations introduced into the TRIM28 sequence and predicted effects on protein activity are summarized in Table 1.
Five of the TRIM28 mutation sites were chosen based on the literature reports related to the key amino acids of the TRIM28 protein undergoing phosphorylation that affect particular domains' activity. Phosphorylation on Ser474, which is located directly before HP1BD, lowers the ability to bind HP1 protein and inhibits TRIM28 transcription repressor activity [31,43]. TRIM28 phosphorylation on three tyrosines, Y449F/Y458F/Y517F (3YF), flanking HP1BD, also reduces the HP1 binding ability, preventing silencing of gene expression [19]. Phosphorylation on Ser824 affects the activity of BROMO by decreasing TRIM28 sumoylation and is associated with relaxed chromatin. It blocks the differentiation of mouse pluripotent cells and induces the expression of SOX2 and NANOG [18,44]. PDB-annotated in 3D structure browser. FI score-functional impact combined score. VC score-variant conservation score. VS score-variant specificity score. Gaps in MSA-portion of gaps in variant position in MSA, MSA height-number of diverse sequences in multiple sequence alignment. (C) Lentiviral system generated to silence endogenous TRIM28 by shRNA (shTRIM28) and express exogenous, shRNA-resistant (shTRIM-res), and FLAG-tagged TRIM28 sequences. (D) Silencing of endogenous TRIM28 and shRNA specificity to endogenous TRIM28 sequence evaluated by immunofluorescence staining (green-TRIM28, blue-DAPI). Scale bar: 100 µm. (E) Silencing of endogenous TRIM28 and expression of exogenous FLAG-tagged proteins confirmed by Western Blot. (F) Efficient transgene expression assessed by immunofluorescence staining against FLAG-tag attached to N-terminus of exogenous TRIM28 protein (red-FLAG, blue-DAPI). Scale bar: 100 µm.  [50]. The algorithm analyzes protein family multiple sequence alignments (MSA) of homologous sequences, exploits sequence homologs 3D structures, and generates conservation scores to predict functional specificity. The functional impact score of mutation is calculated based on the evolutionary conservation in a protein family and, separately, in every subfamily [50].
The C91A mutation in the RING domain results in the abolition of E3 ubiquitin ligase function and inhibition of binding to transcription factors containing the KRAB domain, leading to the loss of the transcription repressor function [45][46][47][48]. Structural C628R mutation in PHD inhibits its endogenous E3 SUMO ligase function and hindrance of BROMO sumoylation [28]. It impedes the interaction with the NuRD complex and SETDB1 methyltransferase and restricts the function of TRIM28 as a transcription repressor [30,49]. The N773G structural mutation, located in BROMO, was selected as a structural control mutation that has a low impact on the impairment of protein function.

Mutations in RING and PHD Domains Are Classified as Mutations with a High Impact on TRIM28 Function
Predictive algorithms indicated C91A and C628R to be mutations having a high impact on the function of the TRIM28 protein ( Figure 3B). Many previous reports demonstrated common TRIM28 phosphorylation sites (S473, 3YF, S824) as crucial for TRIM28 function. Experiments were performed on mouse ESC [18], HEK-293 cells [31,43,44], breast cancer cell line MCF-7 [44], or cervical cancer cell line HeLa [19,31,43]. In contrast, our predictive analysis indicated that phosphorylation site mutations have a low or neutral functional impact.

Constructed Lentiviral System Enables Efficient Transgene Expression and Silencing of Endogenous TRIM28
To obtain more homogenous populations and reduce the factors that might influence the subtle phenotype changes, we decided to perform further experiments on commercial feeder-free human iPSC line ND41658*H (NINDS Human Genetics DNA and Cell Line Repository, Coriell, Camden, NJ, USA). We also wanted to evaluate whether previously observed differentiation due to TRIM28 silencing will be observed on other iPSC lines.
First, we evaluated the shRNA effect on endo-and exogenous TRIM28 expression by immunofluorescence staining (Figure 3D). At the initial stages (2nd passage), we observed some colonies expressing TRIM28 in cells transduced with the shTRIM28 vector, probably due to ongoing and not yet complete puromycin selection. After the 6th passage, we reassessed TRIM28 silencing, and this time TRIM28 expression was decreased in all studied colonies. The control RESCUE population expressed TRIM28 among all the colonies, confirming resistance of exogenous sequence to shRNA.
Analysis of anti-FLAG antibodies proved the functionality of applied lentiviral vectors. We detected transgene expression in all populations of iPSC with exogenous variants of TRIM28, but the intensity varied within transduced populations ( Figure 3E) and cells in colonies ( Figure 3F). Heterogenous transgene expression might result from transducing iPSC as small aggregates, not as a single cell suspension, as well as from non-clonal selection.

Mutations of RING and PHD Domains Influence Human iPSC Morphology and Pluripotency Markers Expression
First, morphological changes appeared right after the 1st passage in the variant with the silenced endogenous TRIM28 transcript (shTRIM28). After the 2nd passage, similar changes were observed in iPSC with mutations within RING (C91A) and within PHD (C628R) ( Figure 4A). These populations demonstrated the loss of morphology typical for pluripotent cells [38]. Non-modified cell colonies (WT) and cells transduced with control vectors (CTRL and RESCUE) showed no changes in the characteristic uniform morphology. Among other modified variants of iPSC, no morphology changes indicating differentiation were noted. Occasionally, differentiating single cells were left among the undifferentiated colonies to not interfere with the differences arising between individual variant populations.
Silencing of endogenous TRIM28 (shTRIM28) resulted in downregulation of OCT3/4 expression compared to other analyzed populations ( Figure 4B). On the other hand, a downregulated level of NANOG was demonstrated in the shTRIM28 population and, to a smaller extent, in populations with mutated RING (C91A) or PHD domains (C628R). In differentiating populations (shTRIM28, C19A, and C628R), the SOX2 level was also reduced. Also, in some cells of the S824D population, SOX2 was decreased, but no other signs of cell differentiation were noted, and no changes in colony morphology after the 10th passage were found. The remaining phospho-mutants and phospho-mimetics did not affect the expression of the analyzed pluripotency markers.

Dysfunction of RING and PHD Domains Results in Decreased Proliferation and Inhibition of Embryoid Bodies Formation
We also evaluated the influence of TRIM28 mutations on proliferation and in vitro spontaneous differentiation potential ( Figure 5). As expected, differentiating populations (shTRIM28, C19A, and C628R) displayed a decreased proliferation compared to WT or CTRL/RESCUE iPSC ( Figure 5A). However, analyzed phospho-mutants and phosphomimetics did not significantly affect proliferation. Differentiation potential was analyzed by EBs formation. EBs were formed by forced aggregation in a non-adherent 96-well plate and monitored with an Incucyte SX1 Live-Cell Analysis System (#4788, Sartorius, Göttingen, Germany). We observed smaller or no spheres derived from shTRIM28, C91A, and C628R compared to the rest of the populations ( Figure 5B,C, Videos S1-S12). We assumed this effect was caused by differentiation induction after impairment of the RING and PHD function. Therefore we evaluated differentiation markers expression in the populations used to obtain EBs ( Figure 5D). Unfortunately, due to late or no amplification in many samples, SD is very high and disabled drawing conclusions. However, we observed a clear trend in the expression of MAP2 (ectoderm) in shTRIM28 and SMA (mesoderm) in shTRIM28 and C628R cells. MAP2 is a neuronal marker, yet its expression can be found in differentiating EBs, as soon as 16 days upon EBs formation, which might explain its presence in differentiating shTRIM28 population [51]. We also captured the amplification of SOX17 (endoderm), but only in one out of three biological replicates. Due to high deviations, the graph for SOX17 expression represents only one replicate. The result is not statistically significant, but it is interesting enough to mention it in this study.
Until this point, our results suggested that TRIM28 phosphorylation does not significantly affect the mechanisms contributing to pluripotency and self-renewal maintenance in human iPSC. However, in the 3YF population, we noticed that the expression level of some differentiation markers ( Figure 5D) was lower than in WT cells. This may support findings suggesting the inhibition of triple tyrosine phosphorylation impact on HP1BD interaction with HP1, which enables transcription repression by TRIM28 [19,52,53].
Still, our results imply that mutations of Ser473 or Ser824 do not affect TRIM28 function in human iPSC. These data stand in opposition to previous reports indicating the impact of Ser473 and Ser824 phosphorylation on TRIM28 function [18,31,43,44]. What is important, mentioned reports did not include the research on human iPSC. Nevertheless, the phosphorylation-dependent regulation of genes responsible for sustaining the undifferentiated state cannot be completely ruled out. Obtained data, however, indicates a much stronger influence of structural mutations on the mechanisms supporting pluripotency. Thus, only the populations with mutants in RING (C91A) or PHD (C628R) and shTRIM28 were subjected to further experiments.

Impairment of RING and PHD Functions Results in Dysregulation of Stem Cell-Associated Signaling Pathways
The differentiating shTRIM28, C19A, and C628R populations exhibited downregulation of extracellular pluripotency markers TRA-1-60 and TRA-1-81 ( Figure 6A). The decreased expression of surface markers is more evident in cells with silenced TRIM28 (shTRIM28) than in cells with a structural defect of RING or PHD domain. The SSEA-4 level was only slightly reduced in C19A and C628R mutants. However, the literature data indicate that the most rapid changes occur in the expression of TRA-1-81 and TRA-1-60 antigens, and the expression of SSEA-4 decreases much slower during differentiation of hESC [54]. QRT-PCR analysis of differentiating populations also showed significant downregulation of pluripotency markers NANOG and LIN28A ( Figure 6B).
We next analyzed the expression of various markers implicated in the maintenance of self-renewal and stemness ( Figure 6C-H). These genes are important for the functioning of both pluripotent and cancer cells. First, we investigated the expression of the genes engaged in chromatin modification ( Figure 6C). Analyzed populations displayed downregulation of DNA methyltransferases (DNMT) 3A and 3B, which contributes to pluripotency maintenance and inhibition of differentiation [55,56]. Many reports indicate rapid repression of TERT within a few days upon differentiation [57,58]. However, according to our observations, TRIM28 RING and PHD domain dysfunctions do not seem to affect TERT expression level. EZH2, a set-domain containing histone methyltransferase specific to H3K27 [59], was upregulated only in the PHD-mutated population. However, this shift cannot be compared to the expression level in control or silenced TRIM28 populations, and the results remain inconclusive. Self-renewal regulation also depends on several signaling pathways, such as Wingless (Wnt) [60], Hedgehog (Hh) [61], phosphoinositide 3-kinase (Pi3K)/Akt kinase [5], or mitogen-activated protein kinases (MAPK) [62]. Therefore, we analyzed the expression of critical genes involved in these pathways. We observed that the downregulation of FGF receptor (FGFR), PI3K, and BMP7 mRNA levels correlated with the impairment of RING and PHD function and the silencing of TRIM28 ( Figure 6D). The expression of key genes involved in the Hh pathway, receptor patched 1 (PTCH1) and smoothened (SMO), was also remarkably reduced in iPSC populations with silenced TRIM28, and in cells with RING and PHD mutants, indicating the inhibition of Hh signal transduction ( Figure 6E). Structural mutations of TRIM28 resulted in a significant decrease in the expression of EPCAM, pluripotency, and proliferation marker in mouse and human stem cells [63][64][65] ( Figure 6F). Populations with silenced TRIM28 and the population with PHD mutant showed a slight increase in WNT expression and downstream WNT signaling inhibitor DKK1 [66]. Silencing of TRIM28 and defects within the PHD contributed to a double increase in NOTCH1 ( Figure 6G). The expression of MAML1, which acts as a transcriptional coactivator for NOTCH signaling [67,68], did not alter under the influence of TRIM28 dysfunction. However, due to high deviations, the results for WNT, DKK1, NOTCH1, and MAML1 expression were not considered statistically significant.
Finally, we examined the expression of genes that are known to be associated with migration and metastasis of cancer cells, and we observed a significant reduction in E-cadherin (CDH1), an inhibitor of differentiation 1 (ID1) and TWIST2 ( Figure 6H). We assumed that CDH1 shift is directly related to the loss of typical compact colony morphology, regularly maintained in an undifferentiated state by cell-cell contact via E-cadherin [69,70]. Altogether, our results demonstrated that TRIM28 affects the expression of the genes implicated in the pathways common for stem cells and cancer cells.

Discussion
The TRIM28 protein was shown to have a crucial impact on self-renewal ability and maintenance of pluripotency in mouse and human ESC [8,[16][17][18]71,72], as well as in human cancer cells [9,19,43]. In this study, we confirmed that TRIM28 preserves stemness and self-renewal in human iPSC. In addition, for the first time in human iPSC, we analyzed the function of individual TRIM28 domains. We identified RING and PHD as the principal domains responsible for these TRIM28 properties.
Impairment of RING or PHD activity, as well as TRIM28 silencing, had a very rapid influence on iPSC populations. IPSC dissociated into individual cells, lost a typical compact colony structure, and showed a decrease in NANOG and SOX2 protein level. In contrast to previous studies suggesting the impact of Ser473 and Ser824 phosphorylation on TRIM28 function, examined phospho-mutants and phospho-mimetics did not affect the self-renewal in human iPSC [18,31,43,44]. Nevertheless, the inhibition of triple tyrosine phosphorylation (Y449F/Y458F/Y517F) slightly reduced the occasional differentiation of the 3YF population. This observation may support reports indicating 3YF mutation enables HP1BD interaction with HP1, resulting in transcription repression by TRIM28 [19,52,53]. Therefore, TRIM28 phosphorylation-dependent regulation of genes responsible for sustaining the undifferentiated state cannot be excluded. However, our data indicate a much stronger influence of structural mutations on the mechanisms supporting pluripotency.
Both mutations with observed effects on stemness properties (C91A, C628R) were predicted to have a high functional impact in our initial analysis with the Mutation Assessor algorithms. Indeed, our results underline the importance of both substituted amino acids for early developmental processes associated with the maintenance of self-renewal and pluripotency. Therefore, only the iPSC populations exhibiting phenotype changes were selected for further study.
Silencing TRIM28 and impairment of RING or PHD domains activity led to downregulation of intra-and extra-cellular pluripotency markers. It also resulted in downregulation of methyltransferases DNMT3A and DNMT3B, associated with the ongoing switch in gene expression profile, progressive cell differentiation, and loss of the parental phenotype [55,56]. Our observations regarding RING domain function in stemness maintenance support the results indicating the engagement of KRAB-ZFPs in pluripotency maintenance. Silencing particular KRAB-ZFPs was shown to induce differentiation of pluripotent stem cells by epigenetic repression of crucial differentiation genes [8]. This may conclude that the RING domain, which is responsible for interaction with the KRAB domain of ZFPs, maintains self-renewal and stemness in human pluripotent stem cells through mediating KRAB-ZFPs repression function. On the other hand, we report that stemness maintenance in human iPSC is also determined by the activity of PHD, which is responsible for sumoylation, and therefore activation of the BROMO domain. Cooperation of PHD and BROMO domains results in chromatin remodeling, histone deacetylation, enhanced methylation of H3K9, and finally binding HP1 to tri-methylated H3K9 and silencing gene expression [30,73]. Our data indicate that TRIM28, to support pluripotency and self-renewal mechanisms, requires functional RBCC and PHD domains that are responsible for the activity of TRIM28 as a transcriptional co-repressor. Gene repression mediated by TRIM28 is also compromised by phosphorylation of S824 TRIM28 residue. In opposition to previous data indicating the important role of this post-translational modification in maintaining pluripotency in murine cells [18], S824A phospho-mutant was shown to be insufficient to initiate differentiation in human iPSC in our study.
In mouse ESC, Trim28, along with Cnot3 co-repressor, were identified as factors necessary to maintain self-renewal capacity [71]. They were shown to bind to numerous gene promoters, creating a unique module distinct from the main module of the core pluripotency network, formed by OCT3/4, SOX2, and NANOG. The sequences regulated by TRIM28 included genes involved in the cell cycle, cell death, and tumorigenesis [71]. Here, we show that during differentiation of iPSC due to TRIM28 depletion, there is a significant shift in cellular signaling. We demonstrate that silencing TRIM28 leads to upregulation of processes related to the regulation of apoptosis, differentiation into a multicellular organism, positive regulation of the developmental process, and regulation of the cell cycle, which is also confirmed by earlier evidence [9,74,75].
As previously reported, TRIM28 overexpression correlates with poor prognosis in many cancer types [20][21][22][23][24][25]. Furthermore, there is growing evidence that cancer stemness is associated with the expression of OCT3/4, SOX2, c-MYC, and other genes involved in the self-renewal regulation of malignant cancer cells, as well as normal stem cells [3]. Numerous studies have been performed to characterize tumor cells in terms of their stemness score and similarity to normal stem cells. Genes upregulated in ESC, for instance genes regulated by NANOG, OCT3/4, SOX2, and c-MYC, are often overexpressed in undifferentiated tumors compared to differentiated ones [76][77][78]. Collectively, with many reports presenting TRIM28 contribution to cancer, these findings suggest that regulatory networks controlling self-renewal in stem cells may also be active in some types of cancer and may constitute new cancer cell therapy.
Our results also indicate that the dysfunction of RING or PHD domains decreases proliferation and EBs formation of hiPSC, which might be very important considering TRIM28 correlation with some tumors. For this reason, we also analyzed the expression of various markers involved in the maintenance of self-renewal, stemness, and the functioning of pluripotent stem cells, as well as of highly dedifferentiated cancer cells.
In human ESC, the pivotal factor in maintaining the state of pluripotency is FGF, which activates PI3K/Akt and MAPK/Erk signaling cascades [62]. Our results reveal a significant reduction in FGFR2 and PI3K expression in all differentiating iPSC populations. PI3K/Akt signaling maintains pluripotency by regulating OCT4, SOX2, and NANOG [79,80]. Zhou et al. demonstrated that in hESC inhibition of mTOR, which is a PI3K effector, this results in a decrease in OCT4, SOX2, and NANOG expression [79]. On the contrary, our results indicate that the most significant changes in the expression of core pluripotency transcription factors occurring within the timescale of our experiment are limited to NANOG. Thus, our data may suggest that the differentiation effect was caused not by the primary inhibition of FGF/PI3K signaling but by the lower NANOG expression caused by disrupted TRIM28 activity.
Recently Do et al. determined that Trim28 prevents the degradation of Oct4, and Trim28 overexpression stabilizes Oct4 in mouse ESC [81]. Furthermore, they found the Trim28 CC domain to be responsible for interaction with Oct4. This supports our findings, indicating that only silencing of TRIM28 resulted in downregulation of OCT4 expression, and none of the introduced mutations were able to induce such an effect. Data presented by Do et al. may also support reports regarding TRIM28 overexpression correlation with poor prognosis in certain cancer types [20][21][22][23][24][25].
We also demonstrate that dysfunction of the RING or PHD domains and TRIM28 silencing leads to the downregulation of the Hedgehog pathway and EPCAM, implicated in stemness maintenance in iPSC. These factors are also therapeutic targets for cancer. In adults, the mutation or deregulation of the Hedgehog pathway plays a key role in both proliferation and differentiation, leading to tumorigenesis or accelerated tumor growth in many different tissues [4,82,83]. EPCAM is frequently overexpressed in tumor cells [84], while its suppression is considered a new approach for the treatment of colon cancer [85]. Furthermore, we show that TRIM28 significantly impacts the expression of metastasisrelated genes, TWIST2 and ID-1. Many reports indicate the involvement of TRIM28 in the induction of EMT through regulation of TWIST [9,86,87]. ID-1 contributes to many cellular processes, including cell growth, aging, differentiation, apoptosis, angiogenesis, and neoplastic transformation [88][89][90]. Recent studies suggest that ID-1 knock-down in endothelial cells derived from angioma inhibits proliferation and induces apoptosis by inhibiting PI3K/Akt/mTOR signaling [90]. Our data also indicate a significant reduction in ID-1 expression, as well as a decrease in PI3K activity in cells with impaired TRIM28 functions, which supports previous findings.
Although both RING and PHD domains contribute to the self-renewal and stemness of human iPSC, they may use different interactions and mechanisms to repress gene expression. We found that WNT1, DKK1, and NOTCH1 showed a trend of increased expression in the population with mutated PHD compared to RESCUE and RING mutants. The expression level of WNT1, DKK1, and NOTCH1 in the PHD mutant population is similar to the expression in cells with silenced TRIM28. In stem cells, the activation of Wnt signaling can induce the expression of Notch pathway components [91]. This may suggest that in contrast to the RING domain, PHD takes part in Wnt/Notch signaling.

Conclusions
In conclusion, our study provides new insights into the role of TRIM28 protein domains in the regulation of pluripotency processes and self-renewal mechanisms in human induced pluripotent stem cells. Of the numerous biological functions of TRIM28 protein, our results indicate the activity of RING and PHD domains in the transcriptional repression to be one of the main molecular mechanisms responsible for maintaining self-renewal and pluripotency. In addition, we demonstrate that in iPSC, TRIM28 influences the expression of the genes involved in the cell cycle, self-renewal, cell death, mobility, and other gene characteristic for cancer cells. Therefore, regulatory networks dependent on TRIM28 signaling that control self-renewal in stem cells may also be active in some types of cancer cells. As such, TRIM28 RING and PHD domains may be considered new targets for cancer therapy, e.g., by designing a drug inhibitor complementary to selected domains. The application of an oncolytic virus expressing a short peptide, binding to crucial TRIM28 amino acids and blocking domain interactions, could also prove useful.