SEARCH

SEARCH BY CITATION

Keywords:

  • Induced pluripotent stem cells;
  • Donor memory;
  • Hepatocyte lineage cells;
  • Hepatic differentiation

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Recent studies suggested that induced pluripotent stem cells (iPSCs) retain a residual donor cell gene expression, which may impact their capacity to differentiate into cell of origin. Here, we addressed a contribution of a lineage stage-specific donor cell memory in modulating the functional properties of iPSCs. iPSCs were generated from hepatic lineage cells at an early (hepatoblast-derived, HB-iPSCs) and end stage (adult hepatocyte, AH-iPSCs) of hepatocyte differentiation as well as from mouse embryonic fibroblasts (MEFs-iPSCs) using a lentiviral vector encoding four pluripotency-inducing factors Oct4, Sox2, Klf4, and c-Myc. All resulting iPSC lines acquired iPSCs phenotype as judged by the accepted criteria including morphology, expression of pluripotency markers, silencing of transducing factors, capacity of multilineage differentiation in teratoma assay, and normal diploid karyotype. However, HB-iPSCs were more efficient in directed differentiation toward hepatocytic lineage as compared to AH-iPSCs, MEF-iPSCs, or mouse embryonic stem cells (mESCs). Extensive comparative transcriptome analyses of the early passage iPSCs, donor cells, and mESCs revealed that despite global similarities in gene expression patterns between generated iPSCs and mESCs, HB-iPSCs retained a transcriptional memory (seven upregulated and 17 downregulated genes) typical of the original cells. Continuous passaging of HB-iPSCs erased most of these differences including a superior capacity for hepatic redifferentiation. These results suggest that retention of lineage stage-specific donor memory in iPSCs may facilitate differentiation into donor cell type. The identified gene set may help to improve hepatic differentiation for therapeutic applications and contribute to the better understanding of liver development. STEM CELLS 2012;30:997–1007


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Induced pluripotent stem cells (iPSCs) are derived by introducing a combination of four transcription factors (KLF4, Oct4, Sox2, and Myc) into somatic cells. iPSCs have been shown to be similar to embryonic stem cells (ESCs) in many aspects, including morphology, self-renewal, pluripotency, and the capacity to generate any lineage-specific cell types including hepatocytes [1–4]. The generation of iPSCs has significantly changed our perspective on regenerative medicine including cell replacement therapy and disease modeling in vitro by generating patient-specific iPSC [5–8]. Although iPSCs share most characteristics of ESCs, recent evidence reveals important differences between iPSCs and ESCs at the level of DNA methylation, global gene expression [9–11], genome stability [12, 13], and differentiation potential [14, 15]. The unique properties of iPSCs are thought to be defined by a donor memory reflecting the epigenetic and transcriptional characteristics inherited from the cell of origin during the reprogramming process [16–18]. Although the molecular mechanisms of transcriptional donor memory retention in iPSCs are not fully understood, evidence is growing that it is due to the incomplete DNA methylation during reprogramming of parental cells into iPSC. Among the identified transcriptional memory genes, C9orf64 is implicated in regulating the efficiency of iPSCs generation from human foreskin fibroblasts [18].

Functionally, the retention of donor memory in iPSCs has been shown to facilitate the redifferentiation capacity of iPSCs into the donor cells rather than into other cell types [16, 17, 19]. Recent work demonstrated that a unique pattern of methylation and/or gene expression characteristic for low-passage iPSCs is not stable but gradually reverses to the ESC-type gene expression profile upon extended iPSC culture [9–11, 16–18]. The latter suggests that donor memory can be one of the potential factors to selectively regulate the redifferentiation potency of iPSCs. However, it remains unknown whether a lineage stage-specific donor memory can modulate iPSCs functional properties.

In this study, to address the lineage stage-specific donor memory, we generated iPSC from the early and late stages of hepatocytic differentiation, including E16.5 hepatoblasts (HB-iPSC) and adult hepatocytes (AH-iPSC). We chose the hepatocyte lineage as a reliable experimental system with well-described structural and molecular changes along the hepatocytic differentiation and availability of cell isolation methods [20–22]. Furthermore, in the past, hepatocytes have been successfully generated from both iPSC and ESC [23–25] and are considered to be an attractive tool for treating end-stage liver diseases hampered by a shortage of donor organs for transplantation and the difficulties in cryopreservation and long-term culture of mature hepatocytes [26, 27]. Using a comparative analysis of transcriptional profiles of the resulting iPSC and mouse ESCs (mESCs), we identified the HB lineage stage-specific donor memory genes that may be functionally relevant during induced redifferentiation toward hepatocytes.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Isolation of HBs, Hepatocytes, and Mouse Embryonic Fibroblasts

HBs, AH, and mouse embryonic fibroblasts (MEFs) were isolated from C57BL/6 mice (Jackson Laboratories, Rochester, New York). HBs isolated from a pool of fetal (E16.5) livers were purified using magnetic-activated cell sorting (MACS) system and HB-specific E-cadherin antibody (Miltenyi, Auburn, CA), as described previously [22]. In brief, cells were first blocked with 5% normal goat serum, incubated with a rat anti-mouse ECCD-2 antibody (Clontech, Mountain View, CA) for 15 minutes at 4°C in phosphate buffered saline (PBS) with 1% bovine serum albumin (BSA) followed by washing and incubation with goat anti-rat IgG microbeads (Miltenyi, Auburn, CA). HBs were plated on collagen I-coated plates (BD) in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS). After 4 hours attachment, a medium was replaced with a fresh DMEM containing 50% DMEM conditioned by fetal liver cells cultured without the separation of hepatocytes and nonparenchymal cells (1:1). AHs were isolated from 3-month-old male mice by a two-step collagenase perfusion method as reported previously [28] and seeded at 1.5 × 103 cells per square centimeter in hepatocyte growth medium supplemented with 10% FBS. MEFs were isolated by a standard protocol as reported [29].

KOSM Lentiviral Generation and Establishment of iPSC Lines from Mouse Hepatic Cells and MEF

The 293T cells were cultured in 100-mm dish at 75% confluence and transfected with a mixture of DNA containing 4 μg of pLentG-KOSM, 3.5 μg of pCMV-VSVG, and 2 μg of psPAX2 (Addgene, Cambridge, MA) by Fugene HD (Roche, Indianapolis, IN), according to the manufacturer's instruction. Twenty-four hours after transfection, the supernatant of transfected cells was collected and filtered through a 0.45-μm pore-size filter. The filtered lentiviral particles were concentrated by ultracentrifugation. For virus infection, MEF and mouse hepatic lineage cells were seeded in a six-well plate at 1 × 105 cells per well 1 day before transduction. The medium was replaced with virus-containing supernatant, incubated for 2–3 hours and then cultured up to 7 days with fresh media. For iPSC induction, the infected cells were trypsinized at day 7, plated in 1:5 ratio into six-well plates containing feeder cells, and incubated until appearance of ES-like cells. The mESC medium was changed every day.

Culture of mESCs and iPSCs on Feeder Cells

mESCs (ASE-9005, Applied StemCell, Inc., Menlo Park, CA) and iPSCs were maintained as undifferentiated cells on inactivated MEF feeder cells as described previously [25].

Immunofluorescence and Alkaline Phosphatase Staining

For immunofluorescence staining, cells were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) for 30 minutes at room temperature. After washing with PBS, cells were incubated with PBS containing 1% BSA and 0.1% Triton X-100 for 1 hour at room temperature, followed by incubation with primary antibodies, including OCT3/4 (1:50, Santa Cruz, CA), Sox2 (1:100, Cell Signaling, Danvers, MA), Nanog (1:200, Cosmo Bio, Carlsbad, CA), SSEA1 (1:50, Santa Cruz, CA), HNF4α (1:50, Santa Cruz, CA), alpha-fetoprotein (AFP) (1:500, Dako, Carpinteria, CA), and albumin (1:100, Bethyl Laboratories, Montgomery, TX). Texas Red-conjugated goat anti-mouse IgG (1:100, Invitrogen, Carlsbad, CA), fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG (1:100, Invitrogen), or Alexa Fluor 647 donkey anti-goat IgG (1:500, Invitrogen) were used as secondary antibodies. Alkaline phosphatase staining was performed using the alkaline phosphatase detection kit (Millipore, Billerica, MA) according to the manufacturer's instruction.

RNA Preparation and Real-Time Quantitative Polymerase Chain Reaction

Total RNA was prepared from cells as indicated using RNase mini kit (Qiagen, Valencia, CA). A total of 1 μg of RNA was reverse transcribed using SuperScript First-Strand cDNA synthesis kit (Invitrogen). Quantitative polymerase chain reaction (qPCR) was performed using the SYBR Green PCR Core Reagents kit (Applied Biosystems, Foster City, CA). All reactions were performed in triplicate. mRNA expression levels were normalized to endogenous glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and expressed relative to control cells. Primers and optimal annealing temperatures are listed in Supporting Information Table S1.

Karyotyping of iPSC Lines

Karyotyping of HB-iPSC (passage 4), AH-iPSC (passage 3), and MEF-iPSC (passage 4) was performed by Sandra S. Burkett, Cytogenetic Core Facility, MCGP, Center for Cancer Research, National Cancer Institute at Frederick, Frederick, Maryland, U.S.

Teratoma Formation

The iPSCs were collected by trypsinization, resuspended in PBS, and subcutaneously inoculated into nude mice (3 × 106). Tumors were dissected 4–5 weeks after transplantation, embedded in paraffin, and stained with hematoxylin and eosin (H&E) for routine histology.

Microarray Analysis

A total of 200 ng RNA from three to four independent biological replicates of MACS-sorted mESCs and iPSCs were linearly amplified according to manufacture's specification (Ambion, Austin, TX). For in vitro transcription, reactions were incubated for 16 hours at 37°C. The efficiency of the single-round amplification was measured by NanoDrop (ND1000, Thermo Scientific, Waltham, MA). Hybridization, washing, detection (Cy3-streptavidin, Amersham Biosciences, GE Healthcare, Piscataway, NJ), and scanning were performed on an illumina iScan system (Illumina, San Diego, CA) using reagents and following protocols supplied by the manufacturer. The biotinylated cRNA (750 ng per sample) was hybridized on Sentrix beadchips human Ref-8v3 for 18 hours at 58°C while rocking (5 rpm). The beadchip covers ∼24,000 RefSeq transcripts. Image analysis and data extraction were performed automatically using illumina GenomeScan Software. Gene expression values were adjusted by subtracting a background noise in each spot by GenomeStudio (illumina, San Diego, CA) and normalized by quantile normalization method across all samples. Signal intensity with a detection p > .05 was treated as a missing value, and only genes with sufficient representation across the samples were included in further data analysis (presence in ≥2 replicates per group). Differentially expressed genes were identified by the Bootstrap analysis of variance (ANOVA) and Contrast test with 10,000 repetitions. Genes with a Bootstrap p value ≤ .05 were considered significantly different. Ingenuity Pathway Analysis tool (Ingenuity Systems Inc., Redwood City, CA) was used to explore the functional relationships among the differentially expressed genes. The significance of each network, function, and pathway was determined by a scoring system provided by Ingenuity Pathway Analysis tool. All microarray data were submitted to Gene Expression Omnibus database with the accession number of GSE33110.

In Vitro Differentiation of mESC and iPSC into Hepatocytes

The protocol for differentiation of mES and iPS cells into hepatocytes was as described by Morrison et al. [30] and Si-Tayeb et al. [24] with minor modifications. Briefly, 1 × 104 cells were seeded on collagen I-coated six-well plates and cultured in media containing 6-bromoindirubin-3-oxime (BIO), glycogen synthase kinase 3 (GSK-3)-specific inhibitor, as a feeder cell-free culture [31] for 4 days to maintain undifferentiated state. Before starting differentiation, cells were cultured in media without leukemia inhibitory factor (LIF) and Bio for 1 day to remove the effect of BIO. The RPMI 1640 differentiation media contained B27 (2%, Gibco), glutamine (1%, Invitrogen), and penicillin/streptomycin (1%, Gibco) supplemented with activin A and BMP4 for 2 days. After that, cells were cultured in RPMI 1640 media containing insulin-transferrin-selenium (ITS) (2%, Gibco), glutamine (1%, Invitrogen), and penicillin/streptomycin (1%, Gibco) supplemented with activin A and basic fibroblast growth factor (FGF) for 4 days for induction of endoderm differentiation. To induce hepatocytic differentiation, mES and iPSCs were continuously cultured in hepatocyte differentiation media [25] containing sequentially bone morphogenetic protein4 (BMP4) and acidic FGF for 5 days, HGF for 5 days and OSM/Dex for 6–9 days. The supernatants were collected at day 22 or 25 for the determination of albumin secretion using mouse albumin enzyme-linked immunosorbent assay (ELISA) Kit (Bethyl Laboratories).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Generation of Lineage Stage-Specific iPSCs

Our experimental approach was as follows: isolation of hepatic lineage cells at different stages of differentiation, followed by generation of iPSC and microarray analysis (Fig. 1A). The HBs were isolated from E16.5 fetal livers and purified by MACS sorting using anti-E-cadherin, a cell-specific surface marker of immature hepatocytes [22]. Fully differentiated AHs were isolated from 3-month-old mice by a two-step collagenase perfusion technique followed by Percoll purification [28]. The vast majority (>99%) of the isolated HB and AH expressed E-cadherin and albumin, another specific marker of hepatic lineage cells, respectively (Fig. 1B, and not shown). Reflecting a high purity, only isolated HB expressed AFP [32], an early marker of hepatic differentiation, as measured by a highly sensitive real-time qPCR (RT-qPCR) analysis, whereas AHs exhibited the highest levels of liver-specific transcription factors HNF4α [33], essential for hepatocytic differentiation (Fig. 1C). To generate iPSC, HBs, AHs, and MEFs were transduced with lentiviral vector pLentG-KOSM, a single polycistronic vector containing a fusion of four genes (Oct4, Sox2, Klf4, and c-Myc) and green fluorescent protein (GFP) reporter gene for easy monitoring the expression of the reprogramming factors [34]. One of the critical characteristics of complete reprogramming into iPSC is the acquisition of exogenous factor independence required for maintaining iPSC pluripotency during reprogramming [35]. In agreement with these reports, we found the ES-like colonies that lost GFP expression from passage 1 (Supporting Information Fig. S1). We further characterized the established iPSC clones by alkaline phosphatase staining, expression of SSEA1, a surface marker of mES, and karyotyping (Fig. 2A). Only clones with typical morphology of mESC (a round shape, large nucleoli, and scant cytoplasm), normal karyotype, and uniform expression of the major pluripotent markers, such as Oct4, Sox2, and Nanog, as judged by immunofluorescence and RT-qPCR, were selected for further analysis (Fig. 2A, 2B). The pluripotency of the selected iPSC was confirmed by embryoid body (EB) formation in vitro and development of teratoma upon subcutaneous transplantation into immunodeficient mice (Fig. 2C).

thumbnail image

Figure 1. Schematic representation of the experimental design. (A): HBs (E16.5), AHs, and MEFs (E16.5) were isolated from 3-month-old C57/B6 mice. iPSCs were generated by a transfection with a pLentG-KOSM vector containing a fusion of four genes, including Klf4, Oct3/4, Sox2, and c-Myc, and green fluorescent protein, followed by global gene expression analysis. (B): Immunostaining with anti-E-cadherin before and after MACS sorting demonstrates a high purity of HBs used for HB-iPSC generation. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Scale bar = 20 μm. (C): Quantitative reverse transcription polymerase chain reaction analysis with primers specific to hepatic lineage cells, AFP and HNF4α, and GAPDH as a control. Only HBs expressed AFP, an early marker of hepatic lineage cells, whereas AHs expressed higher levels of HNF4α, a marker of differentiated hepatocytes. The data are presented as mean expression levels ± SD relative to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). All experiments were performed in duplicate using three independent cell isolations. Abbreviations: AFP, alpha-fetoprotein; AH-iPSC, adult hepatocyte-derived induced pluripotent stem cells; HB-iPSC, hepatoblast-derived iPSC; HNF, HNF4 alpha; MACS, magnetic-activated cell sorting; MEF-iPSC, mouse embryonic fibroblast-derived iPSC.

Download figure to PowerPoint

thumbnail image

Figure 2. Characterization of iPSCs generated from hepatic lineage cells at different stages of differentiation. (A): The morphology, karyotyping, AP staining, and immunostaining with antibodies against the pluripotency markers, including SSEA1, Sox2, Oct4, and Nanog in HB-iPSC, AH-iPSC, and MEF-iPSC. Nuclei were counterstained with DAPI. Scale bar = 20 μm. (B): Quantitative reverse transcription polymerase chain reaction analysis with primers specific to endogenous Nanog, Sox2, and Oct4. The data are presented as mean ± SD relative to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) of duplicate experiments. (C): Evidence of multilineage differentiation. Formation of embryoid bodies (EB) in vitro (left images) and teratomas in vivo (three images on the right) by HB-iPSC, AH-iPSC, and MEF-iPSC. Hematoxylin and eosin (H&E) staining of teratoma sections shows spontaneous differentiation into all three germ layers, including ectoderm, entoderm, and mesoderm. Black arrows point to pigmented epithelia. EB images, ×100 magnifications, H&E images, ×200 magnifications. Abbreviations: AH-iPSC, adult hepatocyte-derived induced pluripotent stem cell; AP, alkaline phosphatase; EB, embryoid body; HB-iPSC, hepatoblast-derived iPSC; MEF-iPSC, mouse embryonic fibroblast-derived iPSC; mESC, mouse embryonic stem cell.

Download figure to PowerPoint

HB-iPSCs Exhibit Higher Capacity for Hepatic Differentiation than AH-iPSCs

The iPSCs derived from different cell types can differentiate more effectively into the parental rather than other cell types [16–19]. Therefore, we asked whether a lineage stage-specific differentiation state of the donor cell may be a contributing factor to the redifferentiation potency of iPSC. To address this question, we first modified the differentiation protocols developed for the ESCs [24, 30] to exclude the requirement for feeder cells and EB formation (Supporting Information Fig. S2A). We confirmed that the feeder-free monolayer culture of ES did not affect either the pluripotency or capacity to differentiate toward hepatocytes (Supporting Information Fig. S2B--S2D). In fact, hepatic differentiation in monolayer was more efficient than that driven by EB (Supporting Information Fig. S2E). Using this system and a sequential application of various combinations of growth factors and cytokines reported to be essential for targeted hepatocyte differentiation, we have shown that ESCs grown as feeder-free monolayer cultures for four passages gave rise to hepatocyte-like cells as judged by morphology and expression of hepatocytic-specific markers, including AFP, albumin (ALB), HNF4α, and G6P (Supporting Information Fig. S2).

All reprogrammed cell lines were then subjected to hepatocyte differentiation protocol in feeder- and serum-free conditions, and secretion of albumin into culture medium was determined as a quantitative measure of hepatocytic-differentiation efficiency. The media was collected at 25 days, the peak time of hepatic differentiation as estimated by morphological criteria. As expected, iPSCs generated from early (HB-iPSC) and terminal (AH-iPSC) stages of hepatic differentiation produced more albumin than either mESCs or MEFs (Fig. 3A). The albumin secretion reached maximum values in HB-iPSC, which were more than fivefold higher as compared to the similarly treated AH-iPSC cultures. Consistent with the mouse albumin ELISA data, the HB-iPSCs also showed a considerably stronger induction of HNF4α and G6P, other known markers of hepatic differentiation as shown by RT-PCR analysis (Fig. 3B) and immunoflourescence staining (Fig. 3C). Thus, the HB-iPSCs derived from the early progenitor-type cells displayed higher hepatocyte-forming potential than iPSCs derived from AHs indicating that the differentiation potency of iPSCs may depend on the lineage stage-specific differentiation state of original cells.

thumbnail image

Figure 3. HB-iPSCs have a greater potential to differentiate into hepatocyte-like cells in vitro. mESC, HB-iPSC, AH-iPSC, and MEF-iPSC were subjected to a hepatocyte differentiation protocol in serum- and feeder layer-free conditions as described in Materials and Methods. (A): Albumin secretion at day 25. Albumin protein levels detected in culture medium and normalized to total protein are shown in the box and whisker plots of duplicate measurements from three independent experiments for each clone (n = 6 in mESC and MEF-iPSC from one clone, n = 18 in AH-iPSC and HB-iPSC from three clones). Boxes represent upper and lower quartiles, lines within boxes represent median, and the error bars comprise the whiskers that extend to the maximum and minimum value datasets. **, p < .01; ***, p < .001; Bootstrap t test with 10,000 random replications as compared to mESC. (B): Quantitative reverse transcription polymerase chain reaction analysis with primers specific to HNF4α, G6P, and CK18 at day 25. mRNA expression values relative to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) are shown in the box and whisker plots of triplicate measurements for each clone (n = 3 in mESC and MEF-iPSC from one clone, n = 9 in AD-iPSC and HB-iPSC from three clones). Boxes represent upper and lower quartiles, lines within boxes represent median, and the error bars comprise the whiskers that extend to the maximum and minimum value datasets. *, p < .05; **, p < .01; ***, p < .001; Bootstrap t test with 10,000 random replications as compared to mESC. (C): Triple immunofluorescence staining with antibodies against E-cadherin, AFP, and albumin (a) and immunofluorescence staining with anti-HNF4 alpha (b). Representative images of HB-iPSC cultures at day 25 are shown. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Scale bar = 20 μm. Abbreviations: AFP, alpha-fetoprotein; AH-iPSC, adult hepatocyte-derived induced pluripotent stem cell; HB-iPSC, hepatoblast-derived iPSC; HNF, hepatocyte nuclear factor; MEF-iPSC, mouse embryonic fibroblast-derived iPSC; mESC, mouse embryonic stem cell.

Download figure to PowerPoint

Transcriptional Profiling of mES, iPS, and Parental Cells

To gain a better understanding of the genomic traits of the HB-iPSC that may support the enhanced capacity for hepatocytic lineage differentiation, we performed a global gene expression analysis of parental cells (AH, HB, and MEF) and the corresponding iPSCs as well as mESCs. Only low-passage iPSCs (passage 4) were used, which are most likely to retain the donor memory [16–18]. To achieve a high purity, mESCs and various iPSCs were separated from the feeder cells using SSEA1-based MACS sorting (Supporting Information Fig. S3). The transcriptome profiling was performed in four replicate experiments using illumina microarrays (Fig. 4). An unsupervised hierarchical clustering analysis based on the similarity in the expression pattern of all genes in all samples showed a clear separation between the donor and stem (mES and iPS) cells. iPSCs clustered together with mESCs, indicating that reprogrammed iPSCs were closer to mES than to the parental cells (Fig. 4). However, each individual group of iPSCs (HB-iPSC, AH-iPSC, and MEF-iPSC) clustered tightly together indicating that they had a unique gene expression signature (Fig. 4).

thumbnail image

Figure 4. Transcriptional profiling of mESC, iPSC, and donor cells. mESCs and early passage iPSCs (p4) were separated from the feeder cells using SSEA1-based magnetic-activated cell sorting (MACS) sorting to achieve a high purity of cells used for microarray analysis. Average linkage unsupervised hierarchical clustering based on correlation coefficient of mESC, iPSCs, and donor cells, including AHs, HBs, and MEFs. The dendrogram demonstrates a clear separation of stem cells (red) and donor cells (blue) into two clusters. The tight cluster of stem cells (red) confirms that the donor cells have been successfully reprogrammed. Each cell in the matrix represents the expression level of a gene feature in an individual sample. Expression values are adjusted to the median value across all samples and genes (red and green indicate high and low expression relative to median expression). Colored bars between dendrogram and heat-map represent samples as indicated. Abbreviations: AH-iPSC, adult hepatocyte-derived induced pluripotent stem cell; HB-iPSC, hepatoblast-derived iPSC; MEF-iPSC, mouse embryonic fibroblast-derived iPSC; mESC, mouse embryonic stem cell.

Download figure to PowerPoint

To identify genes related to the superior capacity of HB-iPSC for hepatic redifferentiation, we next analyzed genes that were differentially expressed in HB-iPSC as compared to AH-iPSC and MEF-iPSC using Bootstrap ANOVA and contrast tests with 10,000 replications. In total, we identified 109 upregulated and 115 downregulated genes using 1.5-fold expression changes and p ≤ .05 as selection criteria (Fig. 5A; Supporting Information Table S2). We also investigated functional networks that might be associated with hepatic differentiation. Ingenuity Pathway Analysis of HB-iPSC gene expression signature revealed that the 109 upregulated genes were significantly related to hepatic system development and function while 115 downregulated genes were linked to embryonic, organism, and tissue development (Fig. 5B, and data not shown). Among the top upregulated canonical pathways with high scores were oncostatin M and BMP signaling, known to be involved in hepatocyte maturation during liver development [36, 37] (Fig. 5C). The HB-iPSC-specific genes seem to be capable of facilitating rather than driving hepatic differentiation since early passage HB-iPSC maintained the pluripotency state and self-renewal under unstimulated conditions (Fig. 2). This notion is supported by the observation that mice deficient in oncostatin M or BMP signaling are characterized by normal or delayed liver development [38, 39]. Furthermore, a comparison of the differentially expressed genes between mESCs and HB-iPSCs showed that the upregulated HB-iPSC-specific genes were expressed at low levels in mESCs.

thumbnail image

Figure 5. Gene network analysis of HB-iPSC gene signature. (A): Unsupervised hierarchical clustering of HB-iPSC, AH-iPSC, and MEF-iPSC: the gene expression signature of 224 differentially expressed genes (109 up and 115 down) from HB-iPSC identified by Bootstrap ANOVA and contrast test with 10,000 random replications (n = 4, p ≤ .05 and fold changes ≥1.5). (B): Ingenuity Pathway Analysis of 109 upregulated genes in HB-iPSC. The top five gene networks include hepatic system development and function. *, The score represents the likelihood that the focus genes within the network are found therein by random chance. Upregulated genes are marked with red arrow. (C): Canonical pathways associated with 109 upregulated genes in HB-iPSC as defined by Ingenuity Pathway Analysis tools (p ≤ .05, Fisher's exact test). Abbreviations: AH-iPSC, adult hepatocyte-derived induced pluripotent stem cell; AMPK, 5′ AMP-activated protein kinase; HB-iPSC, hepatoblast-derived iPSC; JAK, Janus kinase; Maml1, mastermind-like 1; MEF-iPSC, mouse embryonic fibroblast-derived iPSC.

Download figure to PowerPoint

HB-iPSC Retain the HB-Specific Donor Memory Genes

Since the gene expression signature of iPSC may represent a transcriptional donor memory [18, 40, 41], we searched for the donor memory genes in each iPSC group. For this, we have adopted a strategy described by Ghosh et al. [41], which is based on the sequential comparisons of the commonly differentially expressed genes (either upregulated or downregulated, p ≤ .05, at least twofold expression changes) between parental (donor) cells and mESCs, defined as gene set A, and between iPSCs and mESCs, defined as gene set B. This strategy allowed for identifying transcriptional donor memory genes in each category of iPSCs (Fig. 6A). Noteworthy, among these genes, we did not find those encoding liver-enriched transcription factors (e.g., HNF4α) or hepatocyte-specific proteins (e.g., AFP and albumin) supporting a proposed suppression of master transcriptional regulators during cellular reprogramming [18]. Using Venn diagram comparisons of donor memory genes in HB-iPSC, AH-iPSC, and MEF-iPSC, we then identified 204 genes (62 upregulated and 142 downregulated genes) as donor memory genes unique for HB-iPSC (Supporting Information Fig. S4; Table S3). Since donor memory genes are thought to facilitate the efficiency of redifferentiation to the original cells [16, 17, 19], we sought to compare the commonly dysregulated genes in the HB-iPSC gene signature (Supporting Information Table S2) and the unique donor memory genes in HB-iPSC (Supporting Information Table S3). This strategy permitted identification of the donor memory genes in HB-iPSC retained in HB-iPSC gene signature (herein referred to as HB-iPSC-specific donor memory genes), which may support a greater hepatocytic-differentiation potential of HB-iPSCs as compared to AH-iPSCs or MEF-iPSCs. In total, 24 donor memory genes (including seven upregulated genes and 17 downregulated genes) were identified in HB-iPSC (Fig. 6B, 6C). Although some of these genes were also expressed at similar levels in other parental cells, such as AH and MEF (Supporting Information Fig. S6A, S6C), only HB-iPSC retained their stable expression pattern consistent with a cell-specific donor memory preserved in the HB- but not AH or MEF-derived iPSCs (Supporting Information Fig. S6B, S6D). This result is an agreement with a previous report in human iPSC [18]. It has been suggested that incomplete DNA methylation underlies a transcriptional memory of somatic cells in human iPSCs and may be explained by the differences in DNA methylation and/or chromatin modifications during cellular reprogramming.

thumbnail image

Figure 6. HB-iPSC-specific donor memory. (A): The number of donor memory genes in AH-iPSC, HB-iPSC, and MEF-iPSC. The donor memory genes, as described by Ghosh et al. [41], are identified based on the sequential comparisons of the commonly differentially expressed genes (p ≤ .05, at least twofold expression changes) between parental (donor) and mES cells and between iPS and mES cells (B, C) seven upregulated and 17 downregulated HB-iPSC-specific donor memory genes. The data are shown as means of log 2 fold changes relative to mESCs ± SEM (n = 4). Abbreviations: AH-iPSC, adult hepatocyte-derived induced pluripotent stem cell; HB-iPSC, hepatoblast-derived iPSC; Maml1, mastermind-like 1; MEF-iPSC, mouse embryonic fibroblast-derived iPSC; mESC, mouse embryonic stem cell.

Download figure to PowerPoint

We next assessed whether this small number of donor memory genes is functionally relevant. To do so, we correlated the persistence of the donor cell gene expression with the capacity of HB-iPSCs to undergo directed differentiation into hepatocytes upon extended passaging. RT-qPCR confirmed upregulation of the HB-iPSC-specific donor memory in early passage HB-iPSC (p4) as compared to the mESCs (Fig. 7A). However, in late passage HB-iPSCs (p20 and p30), the expression levels of four out of six annotated upregulated genes, including Smarca2 and mastermind-like 1 (Maml1) that belong to the top functional networks involved in hepatic system development and function as well as regulation of cell cycle and gene expression, were gradually decreasing albeit with a different kinetics (Fig. 7A). The transcript levels of Lmna2l did not change, whereas Rnf135 levels continued to increase. Furthermore, HB-iPSC at passages 20 and 30 lost the advantage of the early passage HB-iPSC for hepatocytic differentiation (Fig. 7B, 7C). Together, these findings suggest that the HB-iPSC-specific donor memory could contribute to facilitating hepatic differentiation.

thumbnail image

Figure 7. The reduced capacity of HB-iPSCs to undergo directed differentiation into hepatocytes upon extended passaging. (A): Quantitative reverse transcription polymerase chain reaction analysis with primers specific to the HB-iPSC-specific donor memory, gene including Irf1, Smarca2, Maml1, Rbm15, Lman2l, and Rnf135. The mRNA expression values were normalized to GAPDH and shown as means ± SEM of triplicate measurements expressed relative to mESC. (n = 3, Bootstrap t test with 10,000 random replications. *, p < .05; **, p < .01; and ***, p < .001 as compared to HB-iPSC at passage 4. (B): Albumin secretion in culture medium at day 25 during in vitro hepatic differentiation of mESC and HB-iPSC at indicated passages. Albumin protein was detected in culture medium using enzyme-linked immunosorbent assay and normalized to total protein. The data are shown as means ± SEM of duplicate measurements from two independent experiments (n = 4, Bootstrap t test with 10,000 random replications, *, p < .05; **, p < .01; and ***, p < .001, as compared to HB-iPSC at passage 6). (C): Quantitative reverse transcription polymerase chain reaction analysis with primers specific to hepatic-specific markers albumin and alpha-fetoprotein (AFP) at day 25. The data are shown as means ± SEM of triplicate measurements normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and expressed relative to the mESC levels. (n = 3, Bootstrap t test with 10,000 random replications, *, p < .05; **, p < .01, as compared to HB-iPSC at passage 6. Abbreviations: HB-iPSC, hepatoblast-derived induced pluripotent stem cell; Maml1, mastermind-like 1; mESC, mouse embryonic stem cell; P, passage.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

This research suggests that cellular origin may affect in vitro differentiation capacity of iPSCs [16, 17, 19]. Here, we have extended these studies to demonstrate that differentiation potential of iPSCs may depend on the lineage stage-specific differentiation state of donor cells. Using hepatic lineage as a well-defined experimental model, we have generated iPSC clones from cells at early (HB-iPSC) and end (AH-iPSC) stages of hepatocyte differentiation through lentiviral expression of Oct4, Sox2, Nanog, and c-Myc. All reprogrammed cells expressed known pluripotency markers, formed EB in vitro, gave rise to the cells from three embryonic germ layers in vivo, and maintained a typical ESC-like morphology for 30 passages demonstrating their complete reprogramming (Fig. 2, Supporting Information Fig. S1). Nevertheless, despite the global similarities with ESC transcriptome (Fig. 4), the iPSCs generated from early hepatic progeny cells (HB-iPSCs) retained a transient transcriptional memory characteristic of donor cells as previously reported [9–11, 16–18, 40–42] and displayed a greater hepatocyte-forming potential in vitro as compared to either mESCs or iPSCs obtained from terminally differentiated AHs and MEF (Fig. 3).

Consistent with being more proficient in generating albumin-producing hepatocyte-like cells, HB-iPSCs also exhibited a unique gene expression signature, which clearly differentiated them from both AH-iPSC and MEF-iPSC (Fig. 5). The stem cells used for microarray analysis were separated from the contaminating feeder cells by MACS sorting using a surface marker of pluripotent cells, SSEA1 (Supporting Information Fig. S3). This approach yielded on average 93%–97% enrichment of stem cells and allowed a high reproducibility of gene expression data (Fig. 4). Although no differences in the expression of lineage-specific transcription factors were found, analysis of the gene expression networks in iPSCs reprogrammed from HBs revealed a significant association of HB-iPSC gene expression signature with hepatic system development and function (Fig. 5). Among the top upregulated canonical pathways were oncostatin M and BMP signaling controlling hepatocyte maturation [36] and induction of liver budding [37] during liver development. These results suggest that the gene expression differences between HB-iPSC and AH-iPSC may be responsible for the variability in their differentiation potency as compared to both ESCs and other iPSCs.

This notion was further supported by identification of donor memory genes (Supporting Information Fig. S4) postulated to predispose iPSC to differentiate more effectively into original cells [16, 17, 19, 42]. Using a strategy first described by Ghosh et al. [41], we compared the gene expression profiles of the generated iPSCs and donor cells with respect to mESC. These analyses allowed categorizing 24 genes as the HB-iPSC-specific donor memory including seven upregulated genes (Fig. 6B, 6C and Supporting Information Fig. S6A--S6D). Among the genes with annotated functions was Smarca2, SWItch/Sucrose NonFermentable (SWI/SNF) related matrix-associated regulator of chromatin also known as BRM [43]. The Smarca2 plays an important role in enhancing expression of albumin [44] and hepatic differentiation [45] during liver development as well as in germ-layer formation in mESC in vivo [46]. Noteworthy, two of the upregulated genes (i.e., Maml1 and Rbml5) are implicated in cell fate decisions via modulation of Notch signaling. Thus, Maml1, transcriptional coactivator of Notch signaling [47], has been shown to prevent the development of erythroid/megakaryocytic cells [48, 49] as well as pancreatic acinar cells [50], whereas enforced expression of Rbm15 has been found to inhibit myeloid differentiation [51]. This gene set also contained Irf1, a transcriptional regulator of type 1 interferons and interferon-inducible genes required for DNA damage-induced growth arrest and apoptosis [52]; Rnf135, a widely expressed gene containing a RING finger domain at the N terminus and a B30.2/SPRY domain at the C terminus implicated in diverse functions [53] as well as Lman2l and LOC100044190 with little known or unidentified functions.

The expression of 4/6 upregulated genes, including Smarca2, Maml2l, Irf1, and Rbm15, was progressively decreasing to the levels similar to the ESC state upon extended passaging of HB-iPSCs (p30) (Fig. 7) in agreement with evidence that genes reflecting the cell of origin do not represent a permanent feature of iPSC signature [9–11, 16–18, 35].This was paralleled by a remarkable reduction in hepatocyte-forming potential of the late passage HB-iPSCs, which did not exceed that of ESCs or iPSCs obtained from terminally differentiated AHs and MEF (Figs. 3, 7). Although the functional relevance of the HB-iPSC-specific donor memory awaits further investigation, these data suggest that they may enhance the efficiency of hepatic differentiation via direct (e.g., Smarca2) or indirect (e.g., Maml1 and Rbm15) effects.

The transcriptional differences between iPS and ESCs have been explained by epigenetic memory inherited from parental cells, which facilitates differentiation into the donor cells [16]. In support of importance of DNA methylation during reprogramming, DNA methyltransferase (DNMT) 1 was strongly and uniformly upregulated [1, 54] across all generated iPSC lines to the levels of ESCs as compared to the parental cells. Furthermore, all reprogrammed cells showed a significantly higher albeit variable expression of DNMT3b, the de novo methyltransferase induced upon cellular reprogramming [9, 55], than found in ESCs and original cells (Supporting Information Fig. S5).

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Our findings demonstrate that iPSCs generated from progenitor-like HBs retained a transient transcriptional memory characteristic of donor cells and were more efficient to differentiate into hepatocytes in vitro as compared to either mESCs or iPSCs obtained from terminally differentiated AHs and MEF. This is the first example of lineage stage-specific donor memory genes to be associated with facilitating targeted differentiation of iPSCs. The identification of donor memory genes that are functionally relevant for induced differentiation of iPSCs to hepatocytic lineage may help to advance therapeutic application and contribute to a better understanding of liver development. It remains to be explored whether these findings are relevant for iPSC generated from other cell lineages.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

We thank Susan Garfiled and Langston Lim for the assistance with confocal imaging and Barbara Taylor for the help with fluorescence-activated cell sorting analysis. This research was supported by the Intramural Research Program of the US National Institutes of Health, National Cancer Institute, Center for Cancer Research.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
STEM_1074_sm_SuppFig1.pdf272KFig. S1. Loss of GFP expression after reprogramming. Hepatic lineage cells (AH and HB) and MEF were transduced with pLentG-KOSM vector containing fusion of 4 genes, including KLF, Oct4, Sox2, Myc and GFP. The infected cells were cultured on feeder cells until appearance of ES-like cells which expressed GFP (P0). At passage 1 (P1), the reprogrammed cells lost GFP expression. Images were taken with original magnification, ×100. iPS, induced pluripotent stem cells; HB-iPSC, hepatoblast-derived iPSC; AH-iPSC, adult hepatocyte-derived iPSC; MEF-iPSC, mouse embryonic fibroblast-derived iPSC.
STEM_1074_sm_SuppFig2.pdf790KFig. S2. Protocol development for induced differentiation of mES cells towards hepatocytes in monolayer without feeder cells. (A) Schematic representation of the key factors required for differentiation of mES cells towards hepatic lineage cells. (B) The morphology and immunofluorescence staining of mESC with antibodies against pluripotency markers at passage 4. The cells were cultured on collagen-coated plates in mESC media containing a GSK3 inhibitor BIO (1μM). Sale bar, 20 μM. (C) Phase-contrast micrograph (x200). By day 25, the cells acquired a hepatocyte-like cells morphology and sporadic binuclearity (arrow). (D) Double immunufluorescence staining with AFP and Albumin at day 25. Scale bar, 20 μM. (E) Quantitative reverse transcription polymerase chain reaction analysis with primers specific to AFP, Albumin, HNF4α, and G6P. The data are shown as means expression values ± SEM relative to GAPDH of triplicate measurements. EB, embryoid body, Mono, monolayer culture.
STEM_1074_sm_SuppFig3.tif2454KFig. S3. Purification of mES and iPS cells from feeder cells for microarray analysis. (A) Ten million of the bulk mES or iPS and feeder cells were isolated and stained with PE-SSEA1, a cell surface-specific marker of mESC. SSEA1-positive cells were separated from feeder cells using MACS PE-conjugated MicroBeads followed by FACS analysis of purity. (A) Representative FACS plots of mES cells stained with isotope control (top) and PE-SSEA1 before after MACS separation demonstrate a high purity of MACS-isolated mES cells. (B) Fractions of cells expressing SSEA1 (red histograms) after MACS separation as shown by FACS analysis. Black histograms, isotype control. iPS, induced pluripotent stem cells; HB-iPSC, hepatoblastderived iPSC; AH-iPSC, adult hepatocyte-derived iPSC; MEF-iPSC, mouse embryonic fibroblast-derived iPSC.
STEM_1074_sm_SuppFig4.tif1190KFig. S4. Venn diagram analsysis of donor memory genes in AH-iPSC, HB-iPSC and MEFiPSC. The numbers in the Venn diagrams indicate both unique and commonly dysregulated donor memory genes in AH-iPSC, HB-iPSC, and MEF-iPSC. The unique donor memory genes in HB-iPSC comprise of 62 up-regulated genes (red) and 142 down-regulated genes (blue). iPSC, induced pluripotent stem cells; HB-iPSC, hepatoblast-derived iPSC; AH-iPSC, adult hepatocytederived iPSC; MEF-iPSC, mouse embryonic fibroblast-derived iPSC.
STEM_1074_sm_SuppFig5.tif935KFig. S5. Upregulation of DNMT1 and DNMT3b in iPS cells generated in hepatic lineage cells. (A) The expression levels of DNMT1 and (B) DNMT3b based on the analysis of microarray. The data represent the mean expression levels ± SEM (n=4). iPS, induced pluripotent stem cells; HB-iPSC, hepatoblast-derived iPSC; AH-iPSC, adult hepatocyte-derived iPSC; MEF-iPSC, mouse embryonic fibroblast-derived iPSC; mESC, mouse embryonic stem cells.
STEM_1074_sm_SuppFig6.tif2183KFig. S6. HB-iPSC specific donor memory genes. (A, B) Expression pattern of up-regulated HB-iPSC specific donor memory genes in the parental (A) and corresponding iPSC cells (B). The data are shown as means of log2 fold changes relative to mES cells ± SEM (n=4). (C, D) Expression pattern of down-regulated HB-iPSC specific donor memory genes in parental (C) and corresponding iPSC (D) cells. The data are shown as means of log2 fold changes relative to mES cells ± SEM (n=4). iPS, induced pluripotent stem cells; HB, hepatoblast; AH, adult hepatocyte; MEF, mouse embryonic fibroblast; HB-iPSC, hepatoblast-derived iPSC; AH-iPSC, adult hepatocyte-derived iPSC; MEF-iPSC, mouse embryonic fibroblast-derived iPSC. Red line indicates a two-fold expression differences as compared to mESC.
STEM_1074_sm_SuppTab1.pdf37KSupplementary Table 1
STEM_1074_sm_SuppTab2.pdf93KSupplementary Table 2
STEM_1074_sm_SuppTab3.pdf92KSupplementary Table 3
STEM_1074_sm_SuppInfo.pdf8KSupplementary Data

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.