Unlocking cellular plasticity: enhancing human iPSC reprogramming through bromodomain inhibition and extracellular matrix gene expression regulation

Abstract

The transformation from a fibroblast mesenchymal cell state to an epithelial-like state is critical for induced pluripotent stem cell (iPSC) reprogramming. In this report, we describe studies with PFI-3, a small-molecule inhibitor that specifically targets the bromodomains of SMARCA2/4 and PBRM1 subunits of SWI/SNF complex, as an enhancer of iPSC reprogramming efficiency. Our findings reveal that PFI-3 induces cellular plasticity in multiple human dermal fibroblasts, leading to a mesenchymal-epithelial transition during iPSC formation. This transition is characterized by the upregulation of E-cadherin expression, a key protein involved in epithelial cell adhesion. Additionally, we identified COL11A1 as a reprogramming barrier and demonstrated COL11A1 knockdown increased reprogramming efficiency. Notably, we found that PFI-3 significantly reduced the expression of numerous extracellular matrix (ECM) genes, particularly those involved in collagen assembly. Our research provides key insights into the early stages of iPSC reprogramming, highlighting the crucial role of ECM changes and cellular plasticity in this process.

Graphical Abstract

Graphical Abstract.

Significance Statement.

In this study, we found the bromodomain inhibitor PFI-3 increases reprogramming efficiency in human dermal fibroblasts. Our findings provide valuable insights into the mechanisms underlying cellular plasticity and the contribution of the ECM in this process. Understanding these processes could potentially lead to improved iPSC reprogramming techniques and advancement in therapeutic intervention in regenerative medicine.

Introduction

Cellular plasticity is the ability of cells to transform from one fate to another.¹ It is a crucial process in various cellular functions, including cell development, wound repair, and cancer progression.² Epithelial to mesenchymal transition (EMT) exemplifies cellular plasticity in cancer cells. During iPSC reprogramming, fibroblasts, of mesenchymal origin, are initially transformed into an epithelial-like state through a process known as mesenchymal-to-epithelial transition (MET).³ MET is an essential step in cell reprogramming since iPSCs exhibit characteristics similar to epithelial cells, such as tight junctions and polarity.⁴ The plasticity of fibroblasts, as demonstrated by MET during iPSC reprogramming, is closely associated with the efficiency of reprogramming.⁵ This cell type switch involves extensive changes in gene expression, including the activation and inhibition of genes in both cell types.⁶

Cellular plasticity is regulated by various factors, including epigenetics, chromatin modifiers,⁵ and extracellular matrix (ECM).¹ Bromodomains (BRD) are protein domains that recognize and bind acetylated lysine on histones.⁷ These BRD proteins orchestrate gene expression via mechanisms such as chromatin remodeling and acetylated histone recognition.^8,9 PFI-3, a bromodomain inhibitor, specifically targets SMARCA2/4 (BRG1/BRM) and PBRM1 (BAF180) subunits of SWI/SNF chromatin remodeling complex.^10,11 Recent studies have shown that PFI-3 inhibits the binding of BRG1 to chromatin and displaces BRG1, BRM, and BAF180 from chromatin, depending on cellular context.¹² However, the specific role of these SWI/SNF chromatin remodeling subunits in cell plasticity is not fully understood. Previous research has shown that BRG1 regulates gene expression related to extracellular matrix (ECM) and adhesion in melanoma cells,¹³ as well as genes involved in cell polarity and cell adhesion in retinal development.¹⁴ Similarly, BAF180 has been found to regulate expression of genes related to cell adhesion in renal carcinoma.¹⁵ These studies indicate that SWI/SNF chromatin remodeling subunits collaborate with ECM to regulate gene expression and cell plasticity.¹⁶

The ECM is a dynamic protein network¹⁷ that regulates biochemical and mechanical signals essential for determining cell fate, including differentiation, adhesion, and migration.^1,18 The ECM constantly undergoes remodeling process such as assembly and degradation.¹⁹ Remarkably, changes in collagen deposition and collagen cross-linking alter the stiffness of ECM,^20,21 which in turn promotes cell migration, invasion, and EMT through mechanotransduction.²² The process of EMT in cancer metastasis has been extensively studied,^23,24 while the reverse process, MET, and its underlying mechanisms are not well understood. Interestingly, both carcinogenesis and iPSC reprogramming processes have similarities.²⁵ During iPSC reprogramming, MET is an essential step before establishing pluripotency.²⁶ However, the induction of MET during reprogramming has yet to be fully elucidated.

In this study, we found that bromodomain inhibitor PFI-3 increased reprogramming efficiency in human dermal fibroblasts. Further experiments revealed that PFI-3 altered the expression of ECM-related genes and promoted MET during iPSC reprogramming. These findings provide valuable insights into the mechanisms underlying cellular plasticity and the regulation of gene expression by the ECM. Understanding these processes could potentially lead to improved iPSC reprogramming techniques and advancement in regenerative medicine.

Materials and methods

Cell lines and cell culture

Protocols for derivation of primary dermal fibroblasts are described in detail in our previous study.²⁷ Briefly, fibroblast lines were generated from skin biopsies acquired from NIEHS Clinical Research Unit under institutional review board approved protocol human subjects 10-E-0063. Skin punch biopsies, measuring 4 mm in diameter, were cut into small pieces and air-dried on a 60-mm Petri dish for 5 minutes before being immersed in DMEM (Gibco 11965-092) with 10% FBS (Atlanta Biologicals) and 1% Pen/Strep (Sigma). Media was changed daily until fibroblasts emerged. The cell lines included AF31 (African American Female, 28 years), EF49 (European American Female, 20 years), AM50 (African American Male, 24 years), and EM87 (European American Male, 21 years). For the reprogramming process, cells at passage 4 were used.

Reprogramming assay

Induced PSC generation was performed as previously described.²⁷ Briefly, fibroblasts were reprogrammed using lentiviral vectors that carried 6 transcription factors (ADDGENE/PSIN4-EF2-N2L, ADDGENE/PSIN4-EF2-O2S, and ADDGENE/PSIN4-CMV-K2M). Fibroblast cells were plated in 12-well plates at 5 × 10⁴ cells per well (day 0). On day 1, these cells were transduced with viral supernatant (MOI = 6) and 8 µg/mL of polybrene, along with DMSO or PFI-3. Virus supernatant was removed after 24 hours and replaced with fresh DMEM complete media for 48 hours. Cells were then transferred to 6-well plate coated with Matrigel (Corning Cat# 354234 diluted in DMEM/F-12, Gibco). Forty-eight hours later, cells were fed with a 1:1 mixture of complete DMEM and E8 media (TeSR-E8 STEMCELL Technologies). Cells were subsequently maintained in E8 daily for 21 days. PFI-3 was purchased from Sigma (SML0939) and was applied during different timeframes with DMSO as the control. The iPSC colonies were picked using a 20-µL pipet and grew up individually in E8 media. ReLeSR (STEMCELL Technologies) was used for harvesting and splitting.

Fibroblast was also reprogrammed using a nonintegrating reprogramming system, which used Sendai virus (SeV) vectors (CytoTune-iPS 2.0 reprogramming kit, ThermoFisher). Fibroblast cells were transduced with MOI of 5:5:3 (KOS MOI = 5, hc-Myc MOI = 5, and hKlf4 MOI = 3) and 4 µg/mL of polybrene on day 1, with the reprogramming process following similar protocols described above.

The resulting iPSC colonies were quantified by TRA-1-60 staining at day 21 and visualized with BioTek Cytation 5 (Agilent). Colony counts were determined using the Gen5 software (Agilent) and averaged from triplicate wells. Additionally, reprogramming efficiency was evaluated using alkaline phosphatase staining (Stemgent AP Staining Kit II). Triplicate reprogrammed dishes (10 cm) containing colonies were AP stained and scanned to images. The software program ImageJ (National Institutes of Health, USA) was used for counting colonies from color threshold-adjusted and binary-converted images of each dish. Triplicate plates were averaged and reported as colony counts or percent reprogramming efficiency ([number of colonies/150 000] × 100).

Fibroblast-derived iPSC characterization

Embryoid body formation and differentiation

Induced PSC colonies were detached with ReLeSR and aggregated in a 6-well low attachment plate containing E8 media for 48 hours to form embryoid bodies (EBs). Subsequently, the loose cells were removed from the EBs using a reversible strainer (STEMCELL Technologies Cat# 27259). The EBs were then transferred to a new ultralow attachment plate with EB media (STEMCELL Technologies Cat# 05893). The media was replaced halfway daily. On day 7, the EBs were transferred to Matrigel-coated 6-well plate and cultured in EB medium for 12 days, allowing them to differentiate into the 3 germ layers.

NanoString nCounter stem cell characterization

Induced PSCs and day 7 EBs were collected as described above. Total RNA was isolated using Qiagen RNeasy kit and quantified using Nanodrop. The levels of mRNA transcript of pluripotency and 3 germ layer markers were quantified using the Human nCounter Stem Cell Panel from NanoString Technologies (Seattle, WA, USA) at the NIEHS Molecular Genomics Core.

The data obtained from NanoString were analyzed in R version 4.2.3 and normalized using housekeeping control genes. ComplexHeatmap version 2.15.4²⁸ was used to create heatmaps displaying gene expression patterns specific to pluripotency, endoderm, mesoderm, and ectoderm lineages, with values normalized to the mean expression in H9 embryonic stem cells.

Genomic integrity analysis

G-band karyotyping and digital PCR were performed and analyzed by KaryoLogic, Inc., RTP, NC.

Quantitative RT-PCR

Cells were washed twice with cold PBS and then scraped with PBS. After that, the cells were spun down at 4 °C for 5 minutes. Total RNA was extracted using RNeasy Mini Kits (Qiagen) with on-column treatment of DNase I (Invitrogen). Two micrograms of RNA was reverse transcribed using the SuperScript First Strand kits (Invitrogen). The qRT-PCR analysis was carried out using the Brilliant III Ultra-Fast SYBR Green QPCR Master Mix (Agilent Technologies) with 40 cycles of 20 seconds at 95 °C and 20 seconds at 60 °C. The housekeeping gene, TBP, was used to calculate relative gene expression. RT-PCR Primer sets for 23 genes are listed in Supplementary Table S2.

Western blot analysis

Cells were harvested as described above. Whole cell lysate was isolated using Buffer X (100 mM Tris-HCL pH 8.5, 250 mM NaCl, 1% (v/v) NONIDET P40, 1 mM EDTA) combined with protease inhibitors, phosphatase inhibitor, and PMSF (Sigma). Protein concentrations were determined using the colorimetric Bradford Protein Analysis method with Bio-Rad reagents. Western blotting was performed on precast 18-well 7.5% Tris-HCL gels (Criterion cat# 345-0006) loaded with 50 µg of whole lysate. The gel was then transferred to nitrocellulose membranes using Tris/Glycine buffer containing 20% methanol. Membrane blocking was carried out in 5% nonfat milk for an hour, followed by overnight incubations of primary antibodies. The antibodies used were as follows: COL11A1 (Sigma, HPA058335), E-cadherin (Cell Signaling, 24E10, #3195), FN1 (Sigma, F3648), LIN28 (PTG, 11724-1-AP), NANOG (Bethyl, A300-398A), OCT4 (Santa Cruz, sc-5279), SOX2 (Santa Cruz, sc-20088), β-Actin AC-15 (Sigma cat# A1978), and GAPDH FL-335 (Santa Cruz sc-25778). After incubation with appropriate secondary antibody, blots were quantified using digital fluorescence with LI-COR imaging reagents and the Odyssey CLX Imaging System.

Immunostaining

Cells were washed with cold PBS twice, then fixed with 4% paraformaldehyde for 20 minutes at room temperature, then permeabilized with 0.5% Triton-X100 for 10 minutes. This was followed by blocking with 0.1% Triton-X100, and 5% normal donkey serum in PBS for 1 hour at room temperature. After washing 3 times with PBS, cells were incubated overnight with the primary antibody at 4 °C. The antibodies used were as follows: alpha-fetoprotein (AFP, Millipore 2004189), alpha-smooth muscle actin (α-SMA, Millipore CBL171), E-cadherin (Cell Signaling, 24E10, #3195), LIN28 (PTG, 11724-1-AP), NANOG (Chemicon AB5731), Nestin (NES, Millipore ABD69), OCT4 (Santa Cruz, sc-5279), SOX2 (Chemicon AB5603), TRA-1-60 (Millipore MAB4360), TRA-1-81 (Millipore MAB4381), donkey anti-rabbit IgG (H + L) secondary antibody, Alexa Fluor Plus 594 (ThermoFisher, A32754), and donkey anti-mouse IgG (H + L) secondary antibody, Alexa Fluor 594 (ThermoFisher, A21203).

Flow cytometry

Fibroblasts undergoing reprogramming at days 7 and 14 were cultured on Matrigel in TeR-E8 Medium. Cells were rinsed with cold PBS twice then treated with 0.25% Trypsin-EDTA for 3 minutes and collected using room temperature DMEM media. Subsequently, the cells were centrifuged and resuspended in room-temperature PBS, followed by another round of centrifugation. Next, the cells were resuspended at room temperature in PBS to maintain a cell concentration between 1 and 2E6 cells/mL before counting. Following this, cells were washed with PBS and stained with fixable viability dye. Subsequently, they were washed with FACS Buffer (PBS, 2% BSA, 1mM EDTA, and 0.1% sodium azide), followed by incubation in FACS buffer with fluor-conjugated surface antibodies for 2 hours at room temperature. Cells were then washed with FACS Buffer and incubated in fixation at 4 °C overnight. Fixed cells were washed with permeabilization wash buffer followed by incubation in permeabilization wash buffer with intracellular target antibodies. After further washing with FACS Buffer, cells were analyzed using a Becton Dickinson LSRFortessa flow cytometer. Flow cytometry data were processed using FlowJo v10. Debris, cell clusters, and permeability dye-positive cells were excluded from downstream analysis. Gating was determined by known positive and negative populations in control cells and antibody control samples. The antibodies used were as follows: EPCAM (BD 566841) and SSEA-3 (BD 562706).

LOX activity assay

Extracellular LOX enzymatic activity was quantified using LOX activity assay kit (Abcam, ab112139), which measures the hydrogen peroxide generated by LOX through a proprietary red fluorescence substrate for HRP-coupled reactions. A mixture of 50 µL of cell culture supernatant and 50 µL of reaction buffer was incubated at 37 °C for 10 minutes, protected from light. Assays were performed using black 96-well plates (Porvair, Cat# 205003). Fluorescence was then detected at Ex/Em = 540/590 nm using a fluorescence microplate reader. As the assay is semiquantitative and does not contain a LOX standard, LOX activity is expressed as absolute fluorescence value.

Knockdown of gene expression by shRNA

Col11a1 shRNA plasmid (GIPZ Lentiviral shRNA, ThermoFisher) was obtained from NIEHS Genomic core. Lentiviral packaging was done by NIEHS Viral Vector Core, following a previously established protocol.²⁹

shRNA KD and reprogramming: Fibroblast cells were plated in 12-well plates at 5 × 10⁴ cells per well on day 0. On day 1, the cells were transduced with shRNAs (MOI = 5) and OKSMLN (MOI = 6) simultaneously. The cells were then reprogrammed as described above.

Gene expression profiling and analysis

Total RNA was isolated using Qiagen RNeasy kit and quantified using Agilent Nanodrop. RNA quality was assessed using a 2100 Bioanalyzer instrument and an Agilent 6000 RNA Pico Kit (Agilent Technologies). Microarray analysis was performed by the Molecular Genomics Core at NIEHS. Affymetrix Clariom D human transcriptome array was processed using oligo package³⁰ and analyzed using limma³¹ R packages. Probe-level statistical results were combined to genes by Fisher’s method, and P values were corrected by Benjamini Hochberg.³² Significant genes had an adjusted P value of .05 and a fold change of 1.25. Analyses were performed independently within batch, then statistical results were combined for downstream visualization.

Gene ontology and pathway analysis

Pathway analyses used clusterProfiler³³ with significant genes across days 3, 5, and 7, versus Hallmark and canonical pathways from MSigDB.³⁴ Significant pathways had an adjusted P value of .05 and at least 4 genes.

Pathway enrichment analysis

Ingenuity pathway analysis (IPA; Qiagen), was run using default settings for genes with Spearman correlations with P ≤ .01 (0.388 for n = 36 and 0.274 for n = 72) for the total cohort combined and ancestries separately. Enriched pathways with an FDR-adjusted P value <.01 were used to generate a pathway-gene heatmap, which was clustered using Euclidean distance. Concept network (Cnet) plots were created using exemplar pathways from each cluster.

Enrichment map network

Enrichment Map networks were created using pathways with an adjusted P value of .05, where pathways were colored by the direction of gene regulation, and were connected between pathways with a Jaccard gene overlap at least of 0.2.

Results

Bromodomain inhibitor PFI-3 enhances the reprogramming efficiency of human dermal fibroblasts

Small-molecule PFI-3 specifically targets SMARCA2/4 (BRG1/BRM) and PBRM1 (BAF180) subunit of SWI/SNF complex.^10,11 Given the important roles of bromodomain proteins in transcriptional regulation and developmental processes, we investigated whether PFI-3 could enhance iPSC generation. Our initial reprogramming dosage test showed that treating fibroblasts with PFI-3 at 1 and 10 µM resulted in increased efficiency, with 10 µM being more effective (Supplementary Figure S1A).

To determine the effect of PFI-3, human dermal fibroblasts (AF31, African American Female) were reprogrammed using pluripotency factors OCT4, SOX2, KLF4, c-MYC, LIN28, and NANOG (OSKMLN) and treated with PFI-3 for varying time periods (Figure 1A). Treatment of PFI-3 for the entire 21-day reprogramming process led to a 3-fold increase in iPSC colony formation (Figure 1B). However, pretreatment of fibroblasts with PFI-3 for 10 days prior to reprogramming had no effect on reprogramming efficiency (Supplementary Figure S1B). To validate this finding, multiple fibroblast cell lines from different ancestry and sex were tested, and it was found that PFI-3 consistently improved reprogramming efficiency in human dermal fibroblasts (Supplementary Figure S1C). Additionally, we observed a similar effect of PFI-3 in a nonintegrating reprogramming system using Sendai virus vectors (Supplementary Figure S1D).

Figure 1.

Bromodomain inhibitor PFI-3 enhances the reprogramming efficiency of human dermal fibroblasts. (A) Schematic protocol for PFI-3 treatment intervals during iPSC reprogramming. (B) PFI-3 promoted human iPSC reprogramming. Human dermal fibroblast cells were transduced with OSKMLN and were treated with 10 µM PFI-3 for the entire 21 days. TRA-1-60 staining at day 21 shows a high reprogramming activity of PFI-3. n = 3. P values were determined by a 2-tailed Student’s t test; **P value < .01, n = 3. P value was .00165. (C) Time course of PFI-3 reprogramming activity. TRA-1-60 staining at day 21 shows that PFI-3 has an effect on the first 7 days. ***P value < .001, n = 3. P values were .00097, .43, and .15.

Further investigations were conducted to determine the optimal timing for PFI-3 treatment to enhance iPSC generation. Human dermal fibroblasts were transduced with OSKMLN and treated with PFI-3 at different time intervals (days 1-7, 8-14, and 15-21). The results revealed that PFI-3 inhibition was most effective within the first 7 days after OSKMLN transduction. Treatment during later time periods had no significant impact on reprogramming efficiency (Figure 1C). Based on these findings, we concluded that PFI-3 bromodomain inhibition acts early in the reprogramming process to enhance iPSC generation.

PFI-3 iPSCs are pluripotent

We next investigated whether the iPSCs generated using PFI-3 bromodomain inhibitor were pluripotent. PFI-3 iPSC clones were stained to test the expression of cell surface markers TRA-1-60 and TRA-1-81 as well as the transcription factors OCT4, SOX2, and NANOG. The presence of these markers indicates pluripotency. As expected, all the PFI-3 iPSC clones exhibited positive expression of these pluripotency markers (Figure 2A), suggesting they had successfully reprogrammed to a pluripotent state. To further confirm the pluripotency of the PFI-3 iPSCs, quantitative RT-PCR and Western blot analysis were conducted to analyze the RNA and protein expression levels of key pluripotency markers. The levels of OCT4, LIN28, SOX2, and NANOG in the PFI-3 iPSCs were found to be comparable to those of a human embryonic stem cell (ESC) line H9 and control iPSCs (Supplementary Figure S2A, S2B). These findings suggest that PFI-3 iPSCs possess similar pluripotency characteristics to human ESCs.

Figure 2.

PFI-3 iPSCs are pluripotent. (A) Immunofluorescent staining of PFI-3iPSCs shows expression of pluripotency makers OCT4, SOX2, and NANOG as well as the surface markers TRA-1-60, and TRA-1-81. (B) Immunofluorescent staining of PFI-3iPSC-EBs shows expression of the markers of 3 germ layers. (C) Heatmap of gene expression by NanoString stem cell characterization panel. Markers of pluripotency and the 3 germ layers (endoderm, mesoderm, and ectoderm) are differentially expressed between iPSCs and EBs using H9-centered expression. (D) G-band karyotype. Cytogenetic analysis was performed on 20 G-banded metaphase spreads of PFI3iPSC. Every spread displayed an apparently normal female karyotype.

To assess the developmental potential of PFI-3 iPSCs, their ability to differentiate into the 3 germ layers was evaluated through the formation of embryoid bodies (EBs). The PFI-3 iPSCs were aggregated in a low attachment plate for 48 hours to form EBs. These EBs were then cultured in EB medium on Matrigel-coated plates for 14 days. The resulting cells were stained positive for markers specific to endodermal, ectodermal, and mesodermal lineages, namely AFP, NES, and SMA, respectively (Figure 2B).

We also used the NanoString Stem Cell Characterization panel to validate the pluripotency and trilineage differentiation potential of human iPSCs. The significant increases in expression of ectoderm, mesoderm, and endoderm gene markers at day 7 in iPSC-EBs further confirm their pluripotent state (Figure 2C). Additionally, G-band chromosome analysis and digital PCR revealed normal karyotypes and copy numbers in iPSCs (Figure 2D and Supplementary Figure S2C, S2D),

Gene expression analysis of the effects of PFI-3 on iPSC reprogramming

To further examine the molecular mechanism of PFI-3 on reprogramming, we compared gene expression profiles between PFI-3 treated and control samples. To capture the dynamic changes in gene expression during this process, we performed time course analysis on days 3, 5, and 7 of reprogramming. A total of 428 differentially expressed genes (DEGs) were identified, with 106 genes upregulated and 322 genes downregulated in the PFI-3 treated samples compared to the control samples (Figure 3A).

Figure 3.

Gene expression profiling analysis. (A) Venn diagram showing differentially common expressed genes across time points. (B) Heat map of 428 DEGs (FC > 1.5 and FDR < 0.05) in the comparison of PFI-3 and control across time points of reprogramming. The degree of change in gene expression is denoted by the intensity of signal reflected on the legend. Color scale represents the log2-fold change in centered expression compared to the control at each time point. Genes are divided into 5 clusters according to their dynamic expression profiles. (C) Gene Ontology (GO) analysis of PFI-3 effects on reprogramming. Gene ontology [− log10 (P value)] biological processes of DEGs (log2FC > ±0.5 and P < .05) at first 7 days of reprogramming. (D) Cnet plot of enriched pathways analysis over time, showing representative genes in each node.

Our results showed that the gene expression profiles of PFI-3-treated sample were distinguishable from the control at each time point, indicating that PFI-3 has a significant impact on gene expression during reprogramming (Figure 3B). The 428 DEGs were clustered into 5 groups based on their expression patterns. The genes are listed in an additional file (Supplementary Table S1). Cluster A consisted of immune response-related genes, which were highly enhanced by PFI-3 at day 7. This suggests that PFI-3 may modulate the immune response during iPSC reprogramming. Cluster B included genes involved in the cell cycle, mitosis, cytokinesis, and keratinization. These genes were upregulated by PFI-3 in the early phase of reprogramming, but their expression levels decreased over time. Keratinization plays a key role in promoting epithelialization and safeguarding against stress-induced damage.^35,36 Thus, it indicates that PFI-3 may promote the cell cycle progression and keratinization, ultimately facilitating the reprogramming process. Cluster C consisted of ECM glycoproteins and proteoglycans-related genes. We observed that PFI-3 repressed the expression of these genes at very early time point and further downregulated them over time. ECM glycoproteins are essential components of the extracellular matrix, providing structural support and facilitating cell adhesion and migration.³⁷ On the other hand, proteoglycans play a vital role in cell signaling processes by regulating the activity of growth factors and cytokines, as well as controlling cell adhesion, migration, and proliferation.³⁸ The downregulation of these genes suggests a disruption in signal pathways essential for reprogramming.³⁹ Cluster D was annotated as ECM organization, focal adhesion, and collagen assembly-related genes. These genes regulate cell proliferation, differentiation, and adhesion,¹⁸ PFI-3 slightly downregulated their expression at days 3 and 5, but their expression was highly repressed at day 7. This implies that PFI-3 may impair cell-matrix interactions and remodeling processes, potentially impacting the morphological changes during reprogramming. Lastly, cluster E consisted of genes involved in receptor kinase signaling, which were highly repressed upon PFI-3 treatment at day 7. This finding suggests that PFI-3 may affect receptor-mediated pathways, further highlighting its potential role in modulating cellular response during reprogramming.

We next examined the expression of the 428 DEGs from days 0 to 21 of reprogramming using a global-centered expression analysis. The results further support the significant impact of PFI-3 on gene expression during reprogramming, as distinct gene expression patterns were observed at each time point when comparing PFI-3 treated samples to the control (Supplementary Figure S3A). The complexity of different cell types within day 21 cells, including unreprogrammed fibroblasts, partially reprogrammed cells, and iPS cells, poses a challenge in understanding the impact of PFI-3 on gene expression in iPSCs. To address this issue, we compared the expression profiles of PFI-3 downregulated genes between human fibroblasts and human iPSCs. This comparison used RNA-seq data obtained from 72 sets of matched human dermal fibroblasts (DF) and their corresponding iPSCs examined from our previous study.⁴⁰ Our findings show that 176 out of 311 genes downregulated by PFI-3 exhibit moderate to high levels of expression in dermal fibroblasts but are significantly repressed in iPSCs (Supplementary Figure S3B). These results suggest that PFI-3 plays a crucial role in regulating gene expression patterns that facilitate the transition from a differentiated fibroblast state to a pluripotent iPSC state. To gain insight into the functional roles of DEGs at each time points, we conducted Gene Ontology (GO) enrichment analysis. Our results revealed that ECM organization, collagen assembly, cell adhesion, cell cycle and cell division, and immune response were the major pathways affected (Figure 3C). These findings were further validated through Cnet plot analysis, which illustrated the linkage between the genes of interest represented in the 4 enriched pathways over time (Figure 3D).

Taken together, the gene expression analysis highlights the impact of PFI-3 on gene regulation and functional pathways involved in the iPSC reprogramming process. The extensive transcriptional changes observed during the reprogramming process with PFI-3 treatment, particularly the massive gene repression on day 7, signify a critical stage in the cell transformation process. PFI-3 treatment alters the expression of genes associated with ECM organization, collagen assembly, cell-matrix interaction, and ECM remodeling, indicating a shift in cell identity and remodeling of the extracellular environment.

PFI-3 promotes MET

Gene expression profiling indicated that PFI-3 significantly influences the gene expression associated with ECM. The ECM is a complex protein network that affects cell growth, differentiation, migration, and invasion.¹⁸ MET is a process in which migratory cells dedifferentiate into polarized epithelial cells. This process is essential in iPSC reprogramming.²⁶ Interestingly, EMT, the opposite process of MET, is associated with alteration in the ECM during cancer progression. ECM has been known to drive EMT during tumorigenesis. Hence, we can infer that PFI-3 might promote MET by altering the ECM composition and cell-matrix interaction, consequently changing cell morphology and plasticity.

To investigate this hypothesis, we examined the expression of the MET marker gene E-cadherin (CDH1) during the reprogramming process. We observed that E-cadherin RNA expression was upregulated in PFI-3 treated cells at day 7 (Figure 4A), while E-cadherin protein expression was detected by Western blot at day 10, significantly upregulated by PFI-3 (Figure 4B). Moreover, the immunostaining demonstrated higher E-cadherin expression in PFI-3-treated cells compared to the control (Figure 4C). We also examined the expression of another MET marker, EPCAM, at the single-cell level, and found a higher percentage of EPCAM-positive cells in PFI-3-treated cells on reprogramming days 7 and 14 (Supplementary Figure S4A). Furthermore, density plots of SSEA-3 and EPCAM showed that PFI-3 treatment leads to a higher proportion of EPCAM-positive cells expressing pluripotency markers (Supplementary Figure S4B, S4C). Additionally, the expression of E-cadherin correlates with the induction of pluripotent markers (Supplementary Figure S5A, S5B). Therefore, the presence of these markers in PFI-3-reprogrammed cells suggests that PFI-3 promotes MET during iPSC reprogramming.

Figure 4.

PFI-3 promotes mesenchymal-to-epithelial transition (MET). (A) Relative expression level of E-cadherin (Cdh1) on days 0, 7, and 14 of reprogramming. Transcript level was normalized to TBP expression. **P value < .01, n = 3. P values were .0011 and .0014. (B) E-cadherin protein expression on days 10, 12, and 14 of reprogramming. (C) Immunofluorescent staining of PFI-3-treated cells shows enhanced expression of MET maker E-cadherin on day 14 of reprogramming. Immunostaining of E-cadherin positive epithelial colonies, co-stained with DAPI. The presence of untransformed fibroblasts contributes to the background signal in the image. An inset is an enlarged area showing epithelial colonies and unprogrammed fibroblasts. Cell counts and signal intensities are quantified using BioTek Cytation 5 Gen5 software.

PFI-3 inhibits ECM gene expression specifically involved in collagen assembly and cross-linking

Gene expression profile revealed that many ECM genes were modified by PFI-3 (Figure 3B). Next, we investigated how ECM changes affected MET and reprogramming. Enrichment analysis of ECM-associated gene networks revealed that genes involved in collagen assembly and focal adhesion were among the most repressed (Figure 5A). This suggested that PFI-3 may affect the structural integrity and mechanical properties of the ECM. Additionally, biological network analysis demonstrated that PFI-3 had significant effects on cell cycle and cell division, cell morphology, and mechanical stimulus (Figure 5B). This further supports the notion that PFI-3 plays a role in ECM remodeling and cellular processes related to ECM dynamics.

Figure 5.

Prediction of PFI-3 molecular mechanism model. (A) Pathway enrichment network map. Enrichment networks were created using pathways with an adjusted P value of .05, where pathways were colored by direction of gene regulation and were connected between pathways with a Jaccard gene overlap of at least 0.2. (B) Biological process network map.

Further gene expression analysis showed the downregulation of a subset of 16 ECM-related genes (Figure 6A; Supplementary Figure S6A), including those crucial for collagen assembly and cross-linking, at both the RNA and protein level (Figure 6B, 6C; Supplementary Figure S6B). For instance, PFI-3 significantly decreased COL11A1, which is involved in collagen assembly.⁴¹ Likewise, Fibronectin, a gene that interacts with collagen, cell receptors, and other ECM components to regulate collagen synthesis, deposition, and remodeling,^42,43 was also decreased by PFI-3. In addition, we evaluated the activity of the collagen cross-linking enzyme LOX and found a 15% decrease in LOX activity in PFI-3 treated cells on day 3 of reprogramming (Figure 6D). This suggests that PFI-3 may also affect the cross-linking and stabilization of collagen fibers.

Figure 6.

PFI-3 inhibits the expression of genes associated with collagen assembly and cross-linking. (A) Heatmap of 16 ECM genes which were downregulated by PFI-3. Color scale represents the log2-fold change compared to the control. (B) Quantitative PCR analysis for collagen assembly and collagen cross-linking genes. *P value < .05, n = 3. P values were .012, .023, .007 (Lox), .005, .048, .003 (Loxl4), .4, .06, .004 (Fn1), .002, .03, and .001 (Col11a1). (C) Protein expression for collagen assembly and collagen cross-linking genes. (D) LOX enzyme activity assay. *P value ≤ .05, n = 3. P value was .05. (E) Quantitative PCR analysis for Col11a1knockdown efficiency. Human dermal fibroblasts were infected with lentiviruses containing Col11a1 shRNA. Cells were harvested after 72 hours for qPCR analysis. ***P value < .001, n = 3. P value was .0002. (F) TRA-1-60 staining of iPSC colonies generated with knockdown of COL11A1. The quantity of TRA-1-60 positive colonies was greater in shCol11a1 KD cells compared to non-targeting control cells. **P value < .01, n = 3. P value was .005.

Given the important role of COL11A1 in EMT regulation in cancers,^44,45 we explored the functionality of COL11A1 in reprogramming. Knocking down efficiency of COL11A1 in dermal fibroblasts was evaluated first (Figure 6E), followed by simultaneous reprogramming of fibroblasts with COL11A1 knockdown (Figure 6F). We found the number of TRA-1-60-stained iPSC colonies was 2-fold increase in the COL11A1 KD cells compared to the non-targeting shRNA control (NTC) (Figure 6F). The result shows that knocking down COL11A1 facilitates iPSC reprogramming and that COL11A1 is not only an EMT regulator but also a reprogramming barrier. In conclusion, our findings indicate that PFI-3 enhances iPSC reprogramming by promoting MET through repression of genes involved in ECM collagen assembly and collagen cross-linking.

Discussion

In this study, we evaluated the effect of bromodomain inhibitor PFI-3 on human iPSC reprogramming. Our results indicate that PFI-3 bromodomain inhibition enhances reprogramming efficiency by targeting the gene expression of ECM and promoting the process of MET.

Previous research has shown that ECM acts as a barrier to reprogramming,⁴⁶ with genes associated with ECM organization being repressed during the early stages of the reprogramming process.⁴⁷ Our data show that PFI-3 can further repress the expression of ECM-related genes (Figure 3). Interestingly, while other bromodomain inhibitors have been shown to downregulate fibroblast gene signatures to alter cell identity,^48,49 our study did not observe a significant downregulation of fibroblast markers, such as Vimentin and Thy1. Instead, we found that PFI-3 inhibited EMT master regulator COL11A1 and upregulated MET marker E-cadherin (Figure 4). The genes affected by PFI-3, such as Col11a1, Lox, Loxl4, Fn1, Postn, Igfbp5, Lum, Eln, and Dcn, are expressed in fibroblasts, but they are not fibroblast-specific as they are also expressed in other cell types and tissues.^50-55 These genes are involved in the production and remodeling of ECM, which plays a crucial role in regulating cellular plasticity.¹ It has been demonstrated that cellular plasticity begins with the expression of the Yamanaka factors, leading to the acquisition of pluripotency. It was demonstrated that the initial step in this process involves making the fibroblasts more amenable to reprogramming.⁵⁶ The results from our study demonstrate that PFI-3 acts early and alters the expression of ECM-related genes, indicating bromodomain inhibition by PFI-3 may synergize with reprogramming factors for enhanced cellular plasticity.

Human dermal fibroblasts have an extensive ECM which is rich in collagen.⁵⁷ Fibroblasts are firmly attached to the ECM through the binding of integrins to collagen,⁵⁸ and integrins link intercellular cytoskeleton to regulate cell adhesion and cell migration.⁵⁹ Single-cell imaging reveals that only small and fast-dividing cells can be reprogrammed into iPSCs. Because fibroblasts display a stretched, elongated morphology, they have to reduce cell size to achieve a faster proliferation rate at the early phase of reprogramming.⁶⁰ The effect of PFI-3 on ECM gene expression suggests that fibroblasts must overcome the cell-matrix adhesion first to reduce cell size. The downregulation of collagen assembly-related genes by PFI-3 may reduce this cell-matrix adhesion and facilitate cell morphology changes.

While the role of ECM in driving EMT induction through mechanotransduction has been extensively studied,^61-63 the induction of MET in iPSC reprogramming, particularly from the ECM perspective, remains relatively unexplored. In cancers, collagen cross-linking produces stiffer ECM^20,21; integrins then sense this mechanical force and drive the rearrangement of actin cytoskeleton to create tension, which triggers a signaling cascade to the nucleus to drive EMT.^64,65 Our results indicate that PFI-3 modifies the expression of ECM-related genes, specifically inhibiting genes involved in collagen assembly and cross-linking, which could lead to a softer ECM. Soft ECM reduces integrin clustering and focal adhesion,^66,67 promoting the expression of E-cadherin and driving polarity establishment and apical-basal junction formation.⁶⁸ Studies on the impact of biomaterial stiffness have shown that softer hydrogels lead to increased MET and higher reprogramming efficiency in mouse embryonic fibroblasts.⁶⁹ Therefore, mechanical cues from the microenvironment not only drive EMT but also induce MET. In addition, the ECM plays a crucial role in modulating immune responses. Changes in the ECM composition, structure, or remodeling can significantly impact immune responses.⁷⁰ we found that PFI-3 upregulated genes involved in immune responses, which have been shown to enhance reprogramming efficiency in human cells.⁷¹ Future studies investigating ECM stiffness, collagen cross-linking, and the effects of PFI-3 on EMT in metastatic cancer cells and protein complexes will further elucidate the mechanisms involved.

Finally, PFI-3 has been demonstrated to inhibit the bromodomain activity of SWI/SNF chromatin remodeling subunits.^10,11 Further research has shown that reducing the activity of SWI/SNF, specifically by knocking down subunits such as SMARCA2 (BRM) and SMARCC2 (BAF170), can significantly improve the reprogramming efficiency in mouse embryonic fibroblasts.⁷² This indicates that both PFI-3 inhibition and subunit knockdown share a common mechanism that disrupts BRM function, resulting in enhanced reprogramming outcomes. These findings illuminate the complex roles of SWI/SNF subunits in the cellular reprogramming process, underscoring their importance in driving cellular plasticity.

In conclusion, our study demonstrates the potential of epigenetic regulators such as the bromodomain inhibitor PFI-3 as tools for enhancing iPSC reprogramming efficiency and highlights the crucial role of ECM changes and cell plasticity in this process. These findings offer compelling insights into the mechanisms behind successful human iPSC reprogramming and open new possibilities for future research in this important therapeutic area.

Supplementary material

Supplementary material is available at Stem Cells online.

Author contributions

Conceptualization: Jun Yang, Trevor K. Archer; Methodology: Jun Yang, H. Karimi Kinyamu, Trevor K. Archer; Formal analysis: Jun Yang, H. Karimi Kinyamu; Investigation: Jun Yang, H. Karimi Kinyamu; Data curation: Jun Yang, Erica Scappini, Ginger Muse; Writing—original draft: Jun Yang, Erica Scappini; Writing—reviewing final draft: Jun Yang, H. Karimi Kinyamu, Trevor K. Archer; Visualization: Jun Yang, James M. Ward; Supervision: Trevor K. Archer; Funding and resources: Trevor K. Archer.

Funding

This research was supported by the Intramural Research Program of the National Institute of Environmental Health Sciences Z01 ES071006.

Conflicts of interest

The authors declare no competing interests.

Data availability

Data are deposited in GEO with the accession number GSE241435.