The nuclear receptor THRB facilitates differentiation of human PSCs into more mature hepatocytes

Abstract

To understand the mechanisms regulating the in vitro maturation of hPSC-derived hepatocytes, we developed a 3D differentiation system and compared gene regulatory elements in human primary hepatocytes with those in hPSC-hepatocytes that were differentiated in 2D or 3D conditions by RNA-seq, ATAC-seq, and H3K27Ac ChIP-seq. Regulome comparisons showed a reduced enrichment of thyroid receptor THRB motifs in accessible chromatin and active enhancers without a reduced transcription of THRB. The addition of thyroid hormone T3 increased the binding of THRB to the CYP3A4 proximal enhancer, restored the super-enhancer status and gene expression of NFIC, and reduced the expression of AFP. The resultant hPSC-hepatocytes showed gene expression, epigenetic status, and super-enhancer landscape closer to primary hepatocytes and activated regulatory regions including non-coding SNPs associated with liver-related diseases. Transplanting the hPSC-hepatocytes resulted in the engraftment of human hepatocytes into the mouse liver without disrupting normal liver histology. This work implicates the environmental factor-nuclear receptor axis in regulating the maturation of hPSC-hepatocytes.

Keywords: 3D culture; epigenetics; hepatocytes differentiation and maturation; human pluripotent stem cells; nuclear receptors; pBAF; transcriptional regulation.

Figure. 1.. Regulome analysis identified differences in THRB motif enrichment in accessible DNA and active enhancers between primary hepatocytes and PSC-differentiated hepatocytes.

(A) Development of a spheroid-based hepatocytes differentiation system from hPSC. (B) A cartoon diagram of experimental design. Uncultured primary hepatocytes, 2D PSC-hepatocytes, and 3D PSC-hepatocytes were subjected to RNA-seq, ATAC-seq, and H3K27Ac ChIP-seq. Two biological duplicates samples were sequenced for RNA-seq and H3K27Ac ChIP-seq, and one set of samples was sequenced for ATAC-seq. (C) Principal component analysis (PCA) of 2D hPSC-hepatocytes, 3D hPSC-hepatocytes, uncultured primary hepatocytes, and undifferentiated PSCs. RNA-seq results from fetal hepatocytes (Xie et al., 2019). (D) Genome browser gene tracks of representative RNA-seq results of the ALB/AFP and CYP3A4 loci from 2D PSC-hepatocytes (2D), 3D PSC-hepatocytes (3D), and primary hepatocytes (Primary). (E) ATAC-seq tracks of ONECUT1 and AFP loci from 2D PSC-hepatocytes (2D), 3D PSC-hepatocytes (3D), and primary hepatocytes (3D). (F) Enrichment of THRB motifs (-logP) in the ATAC-seq peaks and H3K27Ac ChIP-seq peaks in 3D-PSC-hepatocytes (blue) and primary hepatocytes (green). (G) Differential analysis of H3K27Ac ChIP-seq between primary hepatocytes and PSC-hepatocytes. (H) Representative H3K27Ac ChIP-seq tracks of ONECUT1 and NFIC loci from 2D PSC-hepatocytes (2D), 3D PSC-hepatocytes (3D), and primary hepatocytes (3D). (I) Expression of THRB in 3D PSC-hepatocytes and primary hepatocytes. Data showing mean ± standard error, n=2 experiments. The scale bar in (A): 100 μm.

Figure. 2.. Correlation between gene expression and H3K27Ac ChIP-seq signals.

(A) Overlapping between super-enhancer genes (green) and top 10000 expressed genes from RNA-seq from primary hepatocytes. (B) A metagene plot of H3K27Ac ChIP-seq signals along the 5’ to 3’ of non-super-enhancer genes (orange), super-enhancer genes (green) and all genes (black) expressed in primary hepatocytes. (C) Expression of non-super-enhancer genes (orange), super-enhancer genes (green) and all genes (black) in primary hepatocytes. Solid bars denote median expression levels, and dashed lines denote quantile expression levels. (D) Super-enhancer genes list was used to query KEGG database, and proportion of super-enhancer genes in the pathways were plotted (green). A list of highly expressed genes that are not super-enhancer genes were subjected to the same analysis and proportion of highly expressed non-super-enhancer genes in the pathways were plotted (orange). Randomly selected genes were included as a control (grey).

Figure. 3.. Thyroid hormone depended upregulation of CYP3A4 transcription is mediated by binding of THRB to the proximal enhancer of CYP3A4.

(A) Effects of T3 on CYP3A4 and ALB expression in PSC-hepatocytes differentiated in 3D spheroids cultures. Plotted data are relative gene expression compared to primary hepatocytes (mean ± s.e.m.), n=3 experiments. * denotes P<0.05, two-sided Student’s t test. (B) Immunofluorescence with an anti-CYP3A4 antibody (red) on undifferentiated PSCs, control 3D PSC-hepatocytes, T3 treated 3D PSC-hepatocytes, and human liver tissue. (C) CYP3A4 activities (measured with P450-Glo luciferase assay) of control 3D PSC-hepatocytes and T3 treated 3D PSC-hepatocytes. Plotted data are mean mean ± s.e.m., n=3 experiments. * denotes P<0.05, two-sided Student’s t test. (D) ATAC-seq track of the CYP3A4 locus was overlayed with anti-FLAG ChIP-seq data (GEO: GSM2534017) from THRB-FLAG HepG2 cells (HepG2 cells with FLAG tagging to the endogenous THRB gene). The promoter and two 5’ enhancer elements of CYP3A4 were denoted as P, PE (proximal enhancer), and DE (distal enhancer). (E) Overlapping of THRB ChIP-seq and ATAC-seq peaks from 3D PSC-hepatocytes (top panel) and primary hepatocytes (bottom panel). The percentage of ATAC-seq peaks that overlapped with THRB ChIP-seq peaks were labelled. (F) THRB-FLAB HepG2 cells were cultured with T3 or not (CTL), followed by cut-and-run experiment with an anti-FLAG antibody. The isolated DNA from MNase-proteinA treated cells were purified and subjected to q-PCR analysis with primers spanning CYP3A4 proximal enhancer (PE in c). Plotted data are mean ± s.e.m., n=4 experiments. * denotes P<0.05, two-sided Student’s t test. (G) DOX-inducible CRISPR activation H1 human PSCs were transduced with control sg lentivirus (CTL) or lentivirus encoding sg RNA targeting CYP3A4 proximal enhancer (PE). DOX was added to differentiated hepatocyte-like cells for about 2 days, and RNA samples from cells were subjected to q-RT-PCR analysis with CYP3A4 primers (closed circles) and ALB primers (open squares). Plotted data are mean ± s.e.m., n=4 experiments. * denotes P<0.05, two-sided Student’s t test. The scale bar in B: 10 μm.

Figure. 4.. Generation of PSC-hepatocytes with advanced maturity by modulating thyroid hormone signaling.

(A) Representative immunofluorescence staining of anti-HNF4A, human Albumin (ALB), CYP3A4, and AFP antibodies in control PSC-hepatocytes (CTL) and PSC-hepatocytes treated with T3. (B) Left panel: a representative flow cytometry analysis of T3 treated PSC-hepatocytes after intracellular staining with anti-hALB and anti-CYP3A4 antibody. Right panel: quantification of proportion of hALB and CYP3A4 double positive cells. Plotted data are mean ± s.e.m., n=4 experiments. ** denotes P<0.01, two-sided Student’s t test. (C) KeyGenes analysis of gene expression comparison of 3D control PSC-hepatocytes, 3D PSC-hepatocytes treated with T3, and primary hepatocytes. (D-E) Genome browser gene tracks representing RNA-seq results of the CYP3A4 (D) and CYP2C9 (E) loci from primary hepatocytes (Primary), control 3D PSC-hepatocytes (3D-CTL), and 3D PSC-hepatocytes treated with T3 (3D-T3). (F) Expression of UGT1A1 and GBE1 in 2D-PSC-hepatocytes (red), control 3D-PSC-hepatocytes (light blue), T3 treated 3D-PSC-hepatocytes (dark blue), and primary hepatocytes (green). Plotted data are mean ± standard error, n=2 experiments. (G) Expression of representative genes underlying liver functions in PSC, 2D-PSC-hepatocytes, 3D-PSC-hepatocytes, T3 treated 2D-PSC-hepatocytes, and primary hepatocytes. (H) Comparison of enrichment of GWAS signal in open chromatin in 2D- PSC-hepatocytes, 3D-PSC-hepatocytes, and primary hepatocytes versus open chromatin across all ENCODE and Roadmap Epigenomic tissues for a set of liver-relevant diseases and traits. Dotted line indicates P <0.05 and colors denote the hepatocyte cell type tested. (I) ATAC-seq tracks (top panel) and H3K27Ac ChIP-seq tracks (bottom panel) of the region surrounding a common non-encoding SNP rs12740374 that regulates SORT1 expression in hepatocytes and plasma LDL-C levels. (J) Relative chromatin accessibility of AFP 3’ chromatin accessible peak, CYP3A4 proximal enhancer peak, rs12740374 peak from ATAC-seq libraries were measured by normalization to an ACTB chromatin accessible peak with using quantitative-PCR. The ratio of PSC-hepatocytes treated with T3 comparted to control PSC-hepatocytes were plotted as mean ± s.e.m. (n=3 experiments). * denotes P<0.05, two-sided Student’s t test (K) Upregulation of SORT1 expression by T3 treatment. Data shown are mean ± standard error, n=2 experiments. (L) Gene tracks representing H3K27K27 ChIP-seq data of the ONECUT1 and NFIC loci of control PSC-hepatocytes (3D-CTL), T3 treated PSC-hepatocytes (3D-T3) and primary hepatocytes (Primary). (M) PCA analysis of super-enhancers from H1 PSC, fetal liver tissue, 2D PSC-hepatocytes, 3D PSC-hepatocytes treated with T3 (T3), control 3D PSC-hepatocytes (CTL), and primary hepatocytes (AQL and SMC). Scale bars in (A): 10 μm.

Figure 5.. Interactions between THRB and pBAF components in THRB-FLAG HepG2 cells.

(A) Experimental design to identify THRB-binding proteins by immunoprecipitation with anti-FLAG antibody in THRB-FLAG HepG2 cells followed by mass-spectrometry and IP-Western blotting analysis. (B) Identification of high confidence nuclear proteins co-immunoprecipitated with THRB-FLAB (P<0.01). © Control THRB-FLAG HepG2 cells or cells treated with 3 nM T3 for 1 day were lysed and subjected to anti-FLAG immunoprecipitation and Western blotting with pBAF components PBRM1 and ARID2. (D) Q-RT-PCR analysis of expression PBRM1 in control esiRNA transfected and PBRM1 esiRNA transfected HepG2-THRB cells. Data were normalized to control esiRNA expression levels, and were shown as mean ± S.E.M., n=3 experiments. ** denotes P<0.01, two-sided Student’s t test. (E) Q-RT-PCR analysis of expression CYP3A4 in control esiRNA transfected and PBRM1 esiRNA transfected HepG2-THRB cells. Data were normalized to control esiRNA expression levels, and were shown as mean ± S.E.M., n=4 experiments. * denotes P<0.05, two-sided Student’s t test. (F) Effects of T3 treatment on CYP3A4 activity in HepG2-THRB cells transfected with control siRNA or PBRM1 esiRNA. Data were shown as mean ± S.E.M., n=3 experiments. * denotes P<0.05, two-sided Student’s t test.

Figure 6.. In vitro proliferation of PSC-hepatocytes with signaling molecules associated with liver regeneration.

(A) Representative bright-field images of PSC-hepatocytes from 2D and 3D differentiation at 1 day and 6 days after plating in medium with EGF, HGF, Wnt agonist, and YAP agonist. (B) Quantification of live cells after 4 passages in the expansion medium. Data were shown as mean ± S.E.M., n=3 experiments. (C) Formation of organoids when 3D PSC-hepatocytes were dissociated and embedded in Matrigel and cultured in expansion medium. (D) PSC-hepatocytes expanded in 2D culture were dissociated to single cells followed by reaggregation to form spheroids in Aggrewell. Representative images of 1 day and 3–4 days post aggregation were shown. (E) Q-RT-PCR analysis showed increased hepatocytes marker genes AFP, ALB, and CYP3A4 upon spheroids formation and culture in S5 medium. Data were shown as ration of undifferentiation PSC (mean ± S.E.M.), n=3 experiments, * denotes P<0.05, and ** denotes P<0.01, two-sided Student’s t test. (F) Survival rate of frozen and thawing of PSC-hepatocytes from 2D and 3D differentiation.Scale bars in (A), (C), and (D): 100 μm.

Figure 7.. Engraftment of 3D PSC-hepatocytes into undamaged mouse liver.

(A) Left panel: transplantation of 2D PSC-hepatocytes (n=3 mice) resulted in liver tumor formation whereas transplantation of 3D PSC-hepatocytes did not (n=5 mice). Right panel: a representative fluorescent micrograph of the engraftment of 3D PSC-hepatocytes differentiated from tdT-expressing H1 cells (Ma et al., 2018). Arrows denote tdT expressing cells. (B) Left two panels: H&E staining of liver from control mice (top panels) or from mice with splenic transplantation of dissociated human 3D PSC-hepatocytes (bottom panels) at 1 month post transplantation (left panel) or 6 months post transplantation (right panels). Right two panels: Anti-human albumin (hALB) immunohistochemistry of liver from control mice (top panels) or from transplanted mice (bottom panels) at 6 months post transplantation. (C) Representative immunofluorescence micrographs with an anti-human albumin (hAlbumin) antibody (green) and an anti-CYP3A4 antibody (red) on control mouse liver (top panel), liver from transplanted mice (middle panel) 1 month post transplantation, or human liver (bottom panel). (D) Representative immunofluorescence micrographs with an anti-human albumin antibody staining (green) of control mouse liver (top panel), or liver from transplanted mice (bottom panel) 6 months post transplantation. (E) Quantification of human mitochondrial DNA from DNA samples isolated from slices from control mice or transplanted mice. Plotted data are mean ± s.e.m., n=4 mice for control, n=5 for transplanted mice. * denotes P<0.05, two-sided Student’s t test. (F) Quantification of human albumin in plasma samples isolated from control mice or transplanted mice.Plotted data are mean ± s.e.m., n=3 mice for the control group, n=3 mice for the transplanted group. ** denotes P<0.01, two-sided Student’s t test. Scale bars in (A) and (B): 100 μm. Scale bars in (C) and (D): 10 μm.