HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 2007-0325v1 25/12/3223    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mutskov, V. Articles by Gershengorn, M. C. Search for Related Content PubMed PubMed Citation Articles by Mutskov, V. Articles by Gershengorn, M. C.

TISSUE-SPECIFIC STEM CELLS

The Human Insulin Gene Displays Transcriptionally Active Epigenetic Marks in Islet-Derived Mesenchymal Precursor Cells in the Absence of Insulin Expression

^aLaboratory of Molecular Biology and
^bClinical Endocrinology Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland, USA

Key Words. Islet-derived precursor cells • Mesenchymal stromal cells • Epigenetic • Histone modifications • Insulin gene

Correspondence: Gary Felsenfeld, Ph.D., Laboratory of Molecular Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 20892, USA. Telephone: (301) 496-4173; Fax: (301) 496-0201; e-mail: gary.felsenfeld{at}nih.gov; or Marvin C. Gershengorn, Clinical Endocrinology Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 20892, USA. Telephone: (301) 496-4128; Fax: (301) 496-9943; e-mail: marving{at}intra.niddk.nih.gov

Received on May 1, 2007; accepted for publication on September 18, 2007.

First published online in STEM CELLS EXPRESS  September 27, 2007.

	ABSTRACT

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
Human islet-derived precursor cells (hIPCs), mesenchymal cells derived in vitro from adult pancreas, proliferate freely and do not express insulin but can be differentiated to epithelial cells that express insulin. hIPCs have been studied with the goal of obtaining large quantities of insulin-producing cells suitable for transplantation into patients suffering from type 1 diabetes. It appeared that undifferentiated hIPCs are "committed" to a pancreatic endocrine phenotype through multiple cell divisions, suggesting that epigenetic modifications at the insulin locus could be responsible. We determined patterns of histone modifications over the insulin gene in human islets and hIPCs and compared them with HeLa and human bone marrow-derived mesenchymal stem cells (hBM-MSCs), neither of which expresses insulin. The insulin gene in islets displays high levels of histone modifications (H4 hyperacetylation and dimethylation of H3 lysine 4) typical of active genes. These are not present in HeLa and hBM-MSCs, which instead have elevated levels of H3 lysine 9 dimethylation, a mark of inactive genes. hIPCs, in contrast, show significant levels of active chromatin modifications, as much as half those seen in islets, and show no measurable H3 K9 methylation. Cells expanded from a minor population of mesenchymal stromal cells found in islets exhibit the same histone modifications as established hIPCs. We conclude that hIPCs, which do not express the insulin gene, nonetheless uniquely exhibit epigenetic marks that could poise them for activation of insulin expression. This epigenetic signature may be a general mechanism whereby tissue-derived precursor cells are committed to a distinct specification.

Disclosure of potential conflicts of interest is found at the end of this article.

	INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
Stimulated by development of the Edmonton protocol [1, 2], islet transplantation has recently achieved considerable advances as a cell replacement therapy for treatment of patients with type-1 diabetes mellitus. This protocol has been successfully used by other centers around the world and leads to high initial rates of insulin independence. However, as with all transplantation therapies, the source of β cells severely limits this approach. For example, in the U.S. at least 1 million people live with type-1 diabetes mellitus, and only a few thousand donor pancreata become available each year (reviewed in [3]). To solve this problem, efforts are being made to establish in vitro systems that might generate a large number of insulin-secreting cells that can be used for cell transplantation. Adult β cells have proliferative capacity in vivo [4] and in vitro [5], but usually the cells lose differentiated function, including insulin expression and secretion, during cell expansion in vitro. However, undifferentiated precursor cells in vitro can be expanded significantly and then induced to differentiate into mature endocrine cells as a potential source for transplantable insulin-producing tissue. Several different types of precursor cells have been used, including embryonic stem cells, bone marrow-derived stem cells, and mesenchymal precursor cells derived from pancreatic islets (reviewed in [5, 6]).

In this study, we are particularly interested in the human islet-derived precursor cells (hIPCs) [7], which are mesenchymal cells that proliferate readily in vitro and exhibit no insulin expression but can then be induced to differentiate into insulin-expressing cells and establish the epithelial phenotype typical of islet cells. Our hypothesis is that hIPCs, after expansion, can undergo mesenchymal-to-epithelial transition and differentiate into insulin-expressing cells more easily than do other precursor cells [6]. One possible explanation is that these cells are "committed" to a pancreatic endocrine phenotype and retain this commitment through multiple cell divisions by epigenetic mechanisms.

The mammalian insulin gene is expressed virtually exclusively in the β cells of the pancreatic islets. Recent studies of the insulin gene promoter and the transcription factors that regulate it have expanded our understanding of the mechanisms involved in the regulation of insulin gene transcription (reviewed in [9, 10]). The role of chromatin remodeling and modifications in regulating the expression of the human insulin gene has not been well studied. However, it is known that the chromatin at active gene loci is generally marked by characteristic histone modifications, such as lysine acetylation and methylation, and it is to be expected that this is true of the insulin locus as well. We therefore determined first that such histone modifications could be detected in the neighborhood of the insulin gene in human islets, and then asked whether these modifications were present in hIPCs. We employed chromatin immunoprecipitation (ChIP) assays for this purpose. We assessed two major histone modifications associated with transcriptionally active chromatin states: histone H4 tetra-acetylation (acetylation of lysines 5, 8, 12, and 16) and dimethylation of lysine 4 on histone H3. We show that these modifications are present in the neighborhood of the insulin gene in human islets but not in HeLa cells and human bone marrow-derived mesenchymal stem cells (hBM-MSCs), which do not express insulin. In contrast, we found that hIPCs, even after multiple passages and at times in culture when no insulin expression could be detected, retained a considerable fraction of the "positive" histone modifications usually associated with an open chromatin structure. At the same time, there was no increase in another modification usually associated with silencing at this locus. The insulin gene in hIPCs thus exhibits many of the chromatin marks associated with the gene in islets.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
Human Islets
Isolated human pancreatic islets from independent cadaveric donors were obtained through the National Islet Cell Resource Center Basic Science Islet Distribution Program from the following centers: Columbia University College of Physicians and Surgeons, New York; City of Hope National Medical Center, Duarte, CA; Islet Distribution Center, University of Miami, Miami; Northwest Tissue Center, Islet Processing Laboratory, Seattle; and islets purified by Dr. Eric Liu, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD. All islets satisfied the criteria of at least 80% purity and at least 70% viability as characterized by the facilities preparing the islets. No effort to purify islets further was made. We have found no consistent differences between hIPC preparations derived from donors of different ages.

Cell Culture

hIPC Culture.   hIPCs were derived from human islets as described previously [7] and maintained in growth medium. These fibroblast-like cells were defined as "passage 0" 14 days after the islets were placed into culture. Growth medium was CMRL-1066 (Gibco, Grand Island, NY, http://www.invitrogen.com) supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and penicillin/streptomycin. Cells were passaged every 3–4 days at 80%–90% confluence. Over the initial 20–25 passages, cell number approximately doubled with each passage.

For the formation of epithelial cell clusters (ECCs), cells were monodispersed with trypsin, collected by centrifugation, and washed with differentiation medium. Differentiation medium was CMRL-1066 supplemented with 2 mM L-glutamine, 1% (wt/vol) bovine serum albumin (BSA), Fraction V Fatty Acid Free (MP Biomedicals, Irvine, CA, http://www.mpbio.com), and 1x insulin-transferrin-selenium-A, (100x; Gibco). Of note, we found that cells cluster more readily in fatty acid-free BSA than in BSA not treated to remove fatty acids. In addition, monodispersion of mesenchymal cells with trypsin is essential for differentiation since adherent cells refed with differentiation medium do not migrate, do not form clusters, and do not differentiate efficiently. After washing, cells were counted, diluted with differentiation medium, and transferred to 6-well tissue culture plates at 3.3 x 10⁵ cells per well. One day after transfer to differentiation medium and every 3 days thereafter, developing ECCs were collected by centrifugation, resuspended in fresh differentiation medium, and redistributed into their original plates.

hBM-MSC Culture.   hBM-MSCs were obtained from Cambrex (Walkersville, MD, http://www.cambrex.com) and cultured according to the vendor's protocols. All studies presented were performed using hBM-MSCs between passages 4 and 10. In an effort to differentiate hBM-MSCs into ECCs, Mesenchymal Stem Cell Basal Medium (Cambrex) was supplemented with 2 mM L-glutamine, 1% (wt/vol) fatty acid-free BSA, and 1x insulin-transferrin-selenium-A as described for differentiation of hIPCs. After monodispersion with trypsin, cells were transferred to 6-well tissue culture plates and treated like hIPCs for differentiation. After 6 days, nonadherent cell clusters were collected by centrifugation for the isolation of total RNA.

HeLa Culture.   HeLa cells were obtained from American Type Culture Collection (Manassas, VA, http://www.atcc.org) and cultured according to their protocol in minimum essential medium supplemented with 10% fetal bovine serum and penicillin/streptomycin.

Sorting of Islet Cells by Three Cell-Surface Markers
Human islets after 4 days in culture were treated with 0.05% Trypsin-EDTA solution for 5 minutes at 37°C until most cells were monodispersed. Monodispersed islet cells were washed with serum containing medium, collected by centrifugation, and washed with phosphate-buffered saline (PBS). Cells were resuspended in cold PBS containing 10% donkey serum at a concentration of 10⁷ cells per milliliter. Cells were maintained at 4°C or on ice for all subsequent procedures. After 20 minutes, primary or isotype control fluorophore-labeled antibodies were added, and incubation continued for a minimum of 40 minutes. Immediately prior to fluorescence-activated cell sorting analysis, samples were diluted 10-fold with cold PBS. Cells were sorted using a FACSVantage SE equipped with the DiVa option. Fluorophore-labeled antibodies were against human CD73 (phycoerythrin), CD90 (fluorescein isothiocyanate) (both from BD Biosciences, San Diego, http://www.bdbiosciences.com), and CD105 (allophycocyanin; eBioscience Inc., San Diego, http://www.ebioscience.com). After sorting, the CD90/73/105+ cells were cultured in our standard growth medium for hIPCs.

Quantitative TaqMan Reverse Transcription-Polymerase Chain Reaction Assay
Attached cells were treated briefly with Trypsin-EDTA at 37°C, and cells were collected by centrifugation. Total RNA isolation, DNase I treatment, and reverse transcription-polymerase chain reaction (RT-PCR) were performed as described previously [11]. The cDNA products were amplified by real-time TaqMan polymerase chain reaction (PCR) using the following Gene Expression Assays from Applied Biosystems (Foster City, CA, http://www.appliedbiosystems.com): GAPDH-Hs99999905_m1; Insulin-Hs00355773_m1; vimentin-Hs00185584_m1; PDX1-Hs00236830_m1; ACTB-Hs99999903_m1.

Formaldehyde Cross-Linking, ChIP, and Quantitative Real-Time PCR
Attached cells were monodispersed by using Trypsin-EDTA solution (0.05% [wt/vol] Trypsin-0.53 mM EDTA) followed by neutralization with 1 mg/ml Trypsin Inhibitor (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). Formaldehyde cross-linking, sonication of the chromatin, and chromatin immunoprecipitation assays were performed as described previously [11]. In the experiments with the ECCs, the nuclear isolation step was omitted, and all procedures were performed in siliconized 1.7-ml tubes because of the very small number of cells used in these experiments. Antibodies purchased from Upstate (Charlottesville, VA, http://www.upstate.com) were used for ChIP: acetylated histone H4 antibody, dimethyl histone H3 (K4) antibody, dimethyl histone H3 (K9) antibody, as well as normal rabbit IgG in the control experiments.

DNA samples from input and antibody-bound chromatin were analyzed by real-time PCR as described previously [11]. The same fractions of the purified input and antibody-bound DNA were analyzed for each individual ChIP experiment. Data quantification was performed by applying the comparative cycle threshold method as previously described [12]. Primers and TaqMan probes used (supplemental online Table 1) were selected from the human insulin gene locus using Applied Biosystems' Primer Express software. To allow comparisons in modification levels among different kinds of cells, the ChIP data are normalized to the value of enrichment for a site at the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) promoter, chosen as a point of high positive modification levels in all cells. We use this for normalization in preference to the DNA content of the immunoprecipitates, which can introduce large uncertainties when small amounts of DNA are measured. We have found in separate experiments (supplemental online Table 2) that the enrichment of GAPDH sequences for histone modifications associated with transcriptional activation is reasonably constant across the cell types used here.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
Proliferative hIPCs Do Not Express the Insulin Gene but Show Levels of Active Chromatin Modifications at the Gene As Much as Half Those Seen in the Islets Purified from a Single Human Donor
The human insulin gene is a small gene located on the short arm of chromosome 11 and consists of three exons and two introns (schematically presented in Fig. 1B). The insulin promoter, immediately upstream of the transcription start site, contains binding sites for a number of ubiquitous and islet-specific transcription factors (reviewed in [10]). We began by ascertaining the transcriptional status of the insulin gene in hIPCs as well as in fresh islets. Quantitative (TaqMan) RT-PCR was used to detect the presence of insulin transcripts. For each cell type and preparation, the level of insulin mRNA was normalized to the level of the standard housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA. We did not see a significant variation in the expression levels of the GAPDH gene among the different cell types and preparations (data not shown). Our experiments confirmed the previous observation [7] that, in hIPCs at early passages 2 and 3, insulin transcription is almost extinguished (approximately 10⁵-fold less transcript than in fresh islets) and is undetectable (typically 10⁸-fold less transcript than in fresh islets [7]) at later passages (Fig. 1A). As expected, insulin expression was also undetectable in human HeLa cells. The same pattern of gene expression as for insulin was found for the β-cell specific transcription factor PDX-1: substantial transcript of this gene was detected in islets (Fig. 1A) and at much lower levels or not at all in early passage hIPCs. In contrast, expression of vimentin, a marker of mesenchymal cells, is more abundant in hIPCs than islets ([7] and Fig. 1A).

View larger version (44K): [in this window] [in a new window]   Figure 1. Expression analyses and ChIP analysis of histone modifications in hIPCs, islets, and HeLa cells. (A): Quantitative RT-PCR was performed with fresh human islets from donor 1, hIPCs at passage 4 and passage 8, respectively, obtained from the same donor, and HeLa cells as a control. Each bar on the graph represents the relative abundance of mRNA, respectively, of human insulin, PDX-1, vimentin, or beta actin relative to human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA. (B): A map of the human insulin gene showing three exons and the promoter region, where transcription factors (ovals) are known to bind. The position of the three ChIP primer sets with TaqMan probes spanning the regions of the human insulin gene promoter (primer 1, black rectangle), exon 1 (primer 2, striped rectangle), and exon 3 (primer 3, hatched rectangle) are shown. (C): ChIP analysis of acetylated histone H4 over the human insulin gene. Isolated human islets from donor 1 and hIPCs at passage 4 obtained from the same donor and HeLa cells as a control were fixed with low concentrations of formaldehyde (up to 0.4% final concentration). The cross-linked chromatin was immunoprecipitated with antibodies against acetylated histones H4 or nonimmune IgG. Primer sets 1–3 with TaqMan probes spanning the insulin gene were used to amplify the bound and input DNA. In parallel, these DNA samples were amplified with primers specific for the human GAPDH gene as an internal control. The values from the ChIPs and the nonspecific signal derived from the normal rabbit IgG ChIP (white bars) for each individual primer set are presented. The relative abundance of insulin gene sequence was calculated as the ratio of its concentration in the immunoprecipitated fraction to that in the input. This was then normalized to similar data obtained for GAPDH, with the latter set equal to 1. (D): ChIP analysis of K4 dimethylation of histone H3 over the human insulin gene. Fresh human islets from donor 1, hIPCs at passage 4 obtained from the same donor, and HeLa cells were analyzed for the presence of this histone modification at the insulin gene locus. (E): ChIP analysis of acetylated histone H4 over the human insulin gene and the control genes and intergenic region with hIPCs at passage 8. hIPCs at passage 8 obtained from the same donor were analyzed for the presence of acetylated histone H4 at the human insulin gene and at the control genes (gray bars); epsilon globin gene promoter and exon and tyrosine hydroxylase gene exon. The intergenic region is approximately 4.4 kilobases downstream of the insulin gene (gray bar). (F): ChIP analysis of K4 dimethylation of histone H3 over the human insulin gene and the control genes and intergenic region with hIPCs at passage 8. Abbreviations: Acetyl., acetylated; ChIP, chromatin immunoprecipitation; Ex., exon; Exn., exon; hIPC, human islet-derived precursor cell; Meth., dimethylation; p., passage; Pr., promoter; RT-PCR, reverse transcription-polymerase chain reaction; TH, tyrosine hydroxylase.

We repeated the ChIP experiments with islets isolated from several different human donors (some of these data are summarized in Fig. 5A and 5B). All islet samples produced high levels of insulin transcript and showed the above positive histone modifications, but the amounts of both varied somewhat among preparations. All data were normalized to levels of modification at the GAPDH gene promoter (see Materials and Methods and supplemental online Table 2).

We asked whether the mesenchymal hIPCs, although they exhibited little or no insulin expression, might exhibit any of the epigenetic marks associated with active expression of the insulin gene. We used a hIPC line generated from islets purified from a single human donor (Fig. 1, donor 1). The culture of these cells is described in Materials and Methods. Initially, hIPCs were analyzed after passages 4 and 8, at which point they had expanded at least 16- and 256-fold, respectively; they were treated for ChIP analysis by methods identical to those used with intact islets. We assessed both histone H4 acetylation and H3 K4 dimethylation in these passage 4 and 8 hIPCs (Fig. 1C–1F). Both samples showed somewhat lower levels of these two modifications across the insulin gene (Fig. 1C–1F) compared with the fresh islet preparations (1.8- to 3.1-fold less for K4 dimethylation of H3 and 1.6- to 2-fold less for H4 acetylation). These levels of modification are nonetheless much higher than those we observe in HeLa cells, which also make no detectable insulin transcript (Fig. 1A) and in which we expect the insulin gene to be packaged in a silenced chromatin structure (Fig. 1C, 1D). Enrichment of these histone modifications in hIPCs was 6- to 10-fold higher for K4 methylation and 4.5- to 6.5-fold higher for histone H4 acetylation than in the nonexpressing HeLa cells. It was interesting to determine whether other lineage-specific genes in hIPCs also displayed the epigenetic signature that we found for the nonexpressing insulin gene. We analyzed hIPCs at passage 8 for the presence of active histone modifications at two genes, tyrosine hydroxylase, which is neuronally expressed, and epsilon globin. Neither of these genes was expressed in hIPCs (data not shown), and we confirmed that, in hIPCs, neither gene had positive modifications associated with gene expression (Fig. 1E, 1F), consistent with the idea that only genes normally expressed in islets show these modifications in hIPCs.

The Epigenetic Signature of Proliferative hIPCs Over the Insulin Gene Does Not Depend on the Source of the Human Islets
To determine the reproducibility of these results and to show that they did not depend on the origin of the islets, we used two additional hIPC populations that had been obtained from two independent human donors (donors 2 and 3; Figs. 2, 3). ChIP analyses were performed at passages 6 and 14 for donor 2 and passages 9 and 11 for donor 3. All showed relatively high levels of transcriptionally active chromatin marks, H4 acetylation, and K4 H3 dimethylation compared with HeLa cells (Figs. 2A, 2B, 3A, 3B). We continued to culture cells from donor 3 and repeated the measurements at late passages 20 and 23. In contrast to the early and middle passages (where the levels of these modification were still quite similar; Figs. 1–3), at these later stage passages hIPCs showed a decrease in the extent of the positive histone modifications (Figs. 2C, 3C). This result could be related to the potential of hIPCs to differentiate to insulin-producing cells; we showed [7] that mid-passage hIPCs generated hormone-expressing epithelial cell clusters with an induction of insulin transcript by 100-fold (or more). However, smaller inductions were observed at late hIPC passages.

View larger version (25K): [in this window] [in a new window]   Figure 2. ChIP analysis of acetylated histone H4 over the human insulin gene in independent hIPC populations. (A): hIPCs that had been obtained from independent human donor 2 were cross-linked at passage 6, passage 14, and after clustering as hormone-expressing epithelial cell clusters (ECCs) (see Materials and Methods) at passage 16. ChIP assays were performed with antibodies against acetylated histone H4, and the presence of insulin gene sequences was analyzed as described in Figure 1. (B): The same experiments were performed with another hIPC population obtained from independent human donor 3 at passage 8, passage 9, and after differentiation into ECCs at passage 11. (C): Analysis of the levels of histone H4 acetylation in cultured hIPCs from donor 3 at late passages 20 and 23. The three ChIP primer sets were from the human insulin gene promoter (black bar), exon 1 (striped bar), and exon 3 (hatched bar). Abbreviations: ChIP, chromatin immunoprecipitation; Diff., differentiated; hIPC, human islet-derived precursor cell; p., passage.

View larger version (24K): [in this window] [in a new window]   Figure 3. ChIP analysis of dimethylation of lysine 4 on histone H3 over the human insulin gene in independent hIPC populations. (A): ChIP of hIPCs obtained from human donor 2 at passage 6, passage 14, and after differentiation at passage 16 was carried out as described in Figure 1. (B): The same experiments were performed with a hIPC line obtained from human donor 3 at passage 8, passage 9, and after differentiation into epithelial cell clusters at passage 11. (C): ChIP of hIPCs from donor 3 at late passages 20 and 23. The three ChIP primer sets were from the human insulin gene promoter (black bar), exon 1 (striped bar), and exon 3 (hatched bar). Abbreviations: ChIP, chromatin immunoprecipitation; Diff., differentiated; hIPC, human islet-derived precursor cell; p., passage.

View larger version (24K): [in this window] [in a new window]   Figure 4. ChIP analysis of dimethylation of lysine 9 on histone H3 over the human insulin gene. ChIP experiments as described in Figure 1 were performed with antibodies against dimethylated histone H3 K9 with hIPCs and compared with the levels of this modification in HeLa cells. The values from the ChIPs and the nonspecific signal derived from the normal rabbit IgG ChIP (white bars) are presented. The relative abundance of insulin gene sequence was calculated as the ratio of its concentration in the immunoprecipitated fraction to that in the input. (A): ChIP of donor 2 hIPCs at passage 6, passage 14, and after differentiation at passage 16. (B): The same experiments were carried out with hIPCs obtained from human donor 3 at passage 8, passage 9, and after differentiation at passage 11. (C): ChIP of donor 3 hIPCs at late passages 20 and 23. The three ChIP primer sets were from the human insulin gene promoter (black bar), exon 1 (striped bar), and exon 3 (hatched bar). Abbreviations: ChIP, chromatin immunoprecipitation; Diff., differentiated; hIPC, human islet-derived precursor cell; p., passage.

View larger version (24K): [in this window] [in a new window]   Figure 5. Correlation between the levels of insulin gene expression and the extent of histone modifications. Comparison of the levels of insulin gene expression estimated by quantitative reverse transcription-polymerase chain reaction with the levels of histone H4 acetylation (A) and dimethylation of histone H3 K4 (B) at the gene exon 1 region (black rhombus) and at the gene exon 3 (gray square). The data from human islets donor 1 (Fig. 2B, 2C) were combined with data obtained from additional islet preparations from different donors (three more for [A] and four more for [B]) and were compared with the data obtained from hIPCs at different passages and from different donors. The x-axis indicates the levels of insulin gene expression estimated as described in Figure 1. The y-axis indicates the extent of histone modification estimated as described in Figure 2. Abbreviation: hIPC, human islet-derived precursor cell. Acet. h., acetylated histone; Meth., methylated; h, histone.

hIPCs Derived from Islet Cells and Positive for CD105/CD73/CD90 Cell-Surface Markers Are Marked by Positive Histone Modifications at the Insulin Gene
In the accompanying paper, we further characterize hIPCs as mesenchymal stromal cells (MSCs) because they adhere to plastic, express CD73, CD90, and CD105, and have the capability to differentiate in vitro into adipocytes, chondrocytes, and osteocytes [8]. We also found that both hIPCs and hBM-MSCs homogeneously express CD90/73/105, and each marker is expressed at similar levels in the two cell types [8]. We asked whether hIPCs derived from CD105/CD73/CD90-positive sorted islet cells are marked by the same histone modification as seen in our previous experiments. Trypsin-dispersed cells from human islets that had been cultured for 4 days in growth medium containing 10% fetal bovine serum (Fig. 6, donor 4) were sorted twice for CD105/CD73/CD90-positive cells. At this time, approximately 3% of the cells in this islet preparation were positive for all three markers. We cultured these sorted cells in hIPC growth medium [7] for several weeks until we had enough cells for ChIP analysis (Fig. 6A–6C) to compare the levels of histone modifications over the human insulin locus in these cells. Both the insulin promoter and gene body showed comparable levels of histone H4 acetylation and K4 dimethylation of histone H3 (Fig. 6A, 6B) in the cells expanded from CD105/CD73/CD90-positive cells with those found in established hIPC cultures. These levels of modification were much higher than those observed in control HeLa cells (Fig. 6A, 6B). Moreover, these cells did not show enrichment of histone H3 lysine 9 over the insulin gene (Fig. 6C). Thus, cells expanded from a MSC population present in day 4 islet cultures exhibit the same histone modifications as established hIPC cultures.

View larger version (25K): [in this window] [in a new window]   Figure 6. ChIP analysis of histone modifications over the human insulin gene in cells derived from CD105/CD73/CD90-positive sorted cells present in 4-day islet cultures. (A): Cells from human islets (donor 4) after 4 days in culture monodispersed and sorted twice for the CD105/CD73/CD90 cell-surface markers. After expansion of the sorted cells for several weeks, ChIP assays were performed with antibodies against acetylated histone H4, and the presence of insulin gene sequences was analyzed as described in Figure 1. (B): ChIP was performed using antibodies against dimethylated histone H3 K4. (C): ChIP was performed using antibodies against dimethylated histone H3 K9. Abbreviations: ChIP, chromatin immunoprecipitation; hIPC, human islet-derived precursor cell; p., passage, Meth, methylated.

View larger version (43K): [in this window] [in a new window]   Figure 7. Gene expression and chromatin modification in hBM-MSCs in comparison with human islets, HeLa cells, and hIPCs. (A): Q. RT-PCR was performed as described in Figure 1 with hBM-MSCs; the levels of insulin, vimentin, and beta actin expression were compared with those observed with fresh human islets from donor 1, hIPCs at passage 4, and HeLa cells (all shown in Fig. 1). (B): ChIP was performed as described in Figure 1 using antibodies against acetylated histone H4. (C): ChIP was performed using antibodies against dimethylated histone H3 K4. (D): ChIP was performed using antibodies against dimethylated histone H3 K9. The pattern of histone H3 K9 methylation across the insulin gene locus was analyzed using additionally designed primer sets upstream and downstream of the gene. (E): Q. RT-PCR of hBM-MSCs in which differentiation was attempted using the same conditions as for hIPCs (see Materials and Methods) at day 0 or after 6 days in differentiation medium. Cycle thresholds from Q. RT-PCR measurements are presented for the following genes: GAPDH, INS, GCG, NTN1, CLDN3, CLDN4, No Ct. Abbreviations: Acetyl., acetylated; ChIP, chromatin immunoprecipitation; CLDN3, claudin 3; CLDN4, claudin 4; Ex., exon; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GCG, proglucagon; hBM-MSC, human bone marrow-derived mesenchymal stem cells; hIPC, human islet-derived precursor cell; INS, proinsulin; No Ct, not detected; NTN1, netrin; p., passage; Q. RT-PCR, quantitative reverse transcription-polymerase chain reaction.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
Covalent post-translational modifications of histone tails by acetylation, phosphorylation, methylation, and ubiquitination are central elements in the epigenetic control of gene expression; the enzymes that are responsible for these modifications are important regulators of transcription (reviewed in [24]). Specific histone modifications, or certain combinations of them, may facilitate transcription by changing nucleosomal structure and interactions or by recruiting other proteins (reviewed in [13–15]). Some other histone tail modifications may have an opposite effect, promoting chromatin condensation and blocking transcriptional initiation or elongation. Since these patterns of histone modification can be maintained through cell division, they may encode heritable "epigenetic" information [25, 26]. Among these modifications, hyperacetylation of both histones H3 and H4 and methylation of lysine 4 of histone H3 generally marks active, transcriptionally competent regions, whereas methylation of lysine 9 of histone H3 and overall histone underacetylation are usually characteristic of transcriptionally silent chromatin.

In this study, we wished to characterize hIPCs, mesenchymal precursor cells derived in vitro from human islets by growth in an appropriate medium [7]. In particular, we wanted to know how histone modifications over the insulin gene locus changed during the establishment of mesenchymal cultures and how those changes related both to the accompanying extinction of insulin gene expression and to the ability of hIPCs, under appropriate conditions, to differentiate into cells which exhibit an epithelial phenotype similar to islet cells. It was of interest to determine whether this ability of hIPCs to undergo mesenchymal-to-epithelial transition was associated with a distinct chromatin structure at the insulin locus.

We tested this hypothesis by using ChIP analysis to compare histone modifications over the human insulin locus in hIPCs with those observed in islet cells, where the insulin gene is expressed at high levels, and in HeLa cells and hBM-MSCs, where it is inactive. It was previously found that the mouse insulin-1 gene promoter is hyperacetylated at histone H4 [27] and hypermethylated at histone H3 K4 [28] in insulin-producing tumor cell lines. We found that histone H4 acetylation and H3 K4 dimethylation were significantly elevated at the human insulin gene promoter and gene body in islets. hIPCs also showed considerable levels of these two modifications across the insulin gene promoter and gene body (Figs. 1C–1F, 2A, 2B, 3A, 3B). When compared with fresh islets, hIPCs obtained from a single human donor exhibited approximately 30%–50% of the levels of K4 dimethylation of H3 and approximately 50% of H4 acetylation (Fig. 1C–1F, donor 1). ChIP experiments performed with hIPCs from two additional independent donors confirmed that, in these cells, the insulin gene locus is marked by positive modifications (Figs. 2A, 2B, 3A, 3B, donor 2, donor 3), even in the absence of insulin transcripts (data not shown). In contrast, HeLa cells and hBM-MSCs showed no such positive modifications.

hIPCs thus represent an exception to the general correlation between levels of transcriptional activity and the extent of positive histone modifications. Recent genome-wide microarray (ChIP-Chip) studies of histone modifications and gene expression in yeast, Drosophila Melanogoster, and human cells have shown that there are strong positive correlations, genome-wide, among levels of hyperacetylation of H3 and H4, di- and trimethylation of H3K4, and transcript abundance [29–31]. In fact, we observe similar correlations between the extent of H4 hyperacetylation or H3K4 dimethylation over the insulin gene and insulin transcript abundance in human islets (Fig. 5A, 5B). Furthermore, most of our results are consistent with the "all-or-none" relationship between the presence of histone modifications and transcriptional activity described by Schübeler et al. [31]; modifications are always present on active genes but absent from nontranscribed genes. (The best known exception to this relationship is found in embryonic stem cells [see below].) Supporting this model, we find that positive histone modifications are essentially absent from the insulin locus in both HeLa cells and hBM-MSCs, where insulin is not expressed. However, this all-or-none pattern of histone modification appears not to apply to hIPCs: the insulin gene in these cells, although inactive, is marked by methylated H3K4 and acetylated H4 histones. Another indication of the noncanonical state of insulin chromatin conformation in hIPCs is the absence of histone H3K9 dimethylation. This modification is associated with condensed chromatin [12, 16–19]. In some cases, it has been found over the bodies (but not the promoters) of actively transcribed genes, although at lower levels than those found at mouse major satellite repeats [32]. In the case of the insulin locus, the dimethylation of histone H3K9 over the promoter is correlated with gene inactivation; it is present at the insulin locus in hBM-MSCs and HeLa cells, but not in hIPCs.

During the development of multicellular organisms, cell differentiation involves changing programs of gene expression, accompanied by corresponding changes in histone modifications. These changes are in response to transient or stage-specific signals, but in many gene systems the pattern of histone modifications and the accompanying state of activity (either "on" or "off") can be maintained through many subsequent rounds of cell division. The epigenetic propagation of histone modifications, mediated for example by the Polycomb and Trithorax families of proteins, constitutes a kind of "cellular memory" [33, 34]. No specific mechanisms have been demonstrated for propagation of these epigenetic marks, although models have been proposed [25]. In any case, we do not know what mechanisms are responsible for the persistence of active chromatin marks in the case of hIPCs. Whatever the mechanism, it appears that hIPCs derived from islets, even after many rounds of cell division accompanied by major changes in the profile of gene expression, retain over the insulin locus significant levels of epigenetic marks associated with the active state of the gene. We speculate that this epigenetic potential of hIPCs could keep the insulin gene poised for induction upon initiation of specific differentiation programs; that is, hIPCs are mesenchymal precursor cells that are committed to specification into insulin-expressing cells.

We found in our earlier kinetic studies of a transgenic cassette stably integrated in chicken erythroid cells that histone H3 and H4 deacetylation and loss of methylation at H3 lysine 4 all occurred during the same window of time as transgene silencing [12]. In fact, by the time gene transcripts were no longer detectable, the active epigenetic marks over these transgenes were completely lost. In the present study with hIPCs, we did not observe this correlation. Although the insulin gene was inactive (Fig. 1), the positive histone modifications were maintained (Figs. 1–3). Moreover, the insulin gene in hIPCs did not undergo dimethylation of H3 K9, which in our earlier study was a later event in gene silencing. In contrast, in nonexpressing HeLa cells and hBM-MSCs, the human insulin gene was marked with this inactive modification, which was enriched at the gene promoter (Figs. 4, 7). Methylation of histone H3 K9 at the promoter could target the heterochromatin protein HP1 and/or repressor complexes [19], which may play a role in maintaining the silence of the insulin gene. Our data collectively show that although proliferative mesenchymal hIPCs do not express the insulin gene, the gene is not epigenetically silenced.

As mentioned above, the best known exception to the all-or-none relationship between histone modifications and transcriptional activity has been found in ESCs [35–37]. Of note, pluripotent ESCs have been found to exhibit epigenetic marks of both transcriptional activity and repression across large, highly conserved sequences that may constitute a "chromatin-based mechanism for maintaining pluripotency" [38]. In ESCs, the presence of both active and inactive epigenetic marks leads, in general, to repression of gene expression, but it is thought that these areas are poised to undergo activation when the appropriate cue is present. Although it has been proposed that "master regulatory genes," primarily transcription factors, mediate these chromatin modifications, little is known about the mechanism of this regulation. Chromatin modifications have also been studied in hematopoietic stem cells, which are not ESCs, as they undergo lineage commitment, differentiation, and maturation, but the relationship between epigenetic marks and expression of genes of the differentiated phenotype is unclear in this cell system [39]. With regard to ESCs and insulin gene modifications, it is noteworthy that Mirmira and colleagues found H3 and H4 acetylation at the proximal promoter but not in the distal promoter of the insulin-1 gene in mouse ESCs [28]. To our knowledge, there is, however, no previous report of transcriptionally active histone modifications in any tissue-derived adult precursor cells including precursor cells from the adult pancreas. Indeed, our findings that cells expanded from a MSC population isolated from human islets (cultured for 4 days) exhibit these epigenetic marks at the insulin gene (Fig. 6) are consistent with the idea that MSC-like precursor cells may be present in islets in situ. Moreover, our observation of the presence of active chromatin modifications in the absence of insulin gene expression in hIPCs is the first description of this combination of findings in adult precursor cells. We suggest that these findings in hIPCs may be an indication of the "committed state" of hIPCs as endocrine pancreas precursor cells and that active epigenetic marks in the absence of differentiated gene expression may be present generally in committed tissue-derived adult precursor or stem cells.

	DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
The authors indicate no potential conflicts of interest.

	ACKNOWLEDGMENTS

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
We thank the Islet Cell Resource Basic Science Islet Distribution Program and Drs. David Harlan and Eric Liu, NIDDK, NIH, for providing human pancreatic islets for basic science research studies. We thank Bernice Marcus-Samuels for the culturing of hIPCs and Elizabeth Geras-Raaka for helpful technical advice concerning these cells. We also acknowledge members of the Felsenfeld laboratory for useful discussions and critical reading of this manuscript. This research was supported by the Intramural Research Program of the National Institute of Diabetes and Digestive and Kidney Diseases, NIH.

	REFERENCES

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 

1. Shapiro AM, Lakey JR, Ryan EA et al. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid free immunosuppressive regimen. N Engl J Med 2000;343:230–238.[Abstract/Free Full Text]

2. Shapiro AM, Ricordi C, Hering BJ et al. International trial of the Edmonton protocol for islet transplantation. N Engl J Med 2006;355:1318–1330.[Abstract/Free Full Text]

3. Naftanel MA, Harlan DM. Pancreatic islet transplantation. PLoS Med 2004;1:e58.[CrossRef][Medline]

4. Dor Y, Brown J, Martinez OI et al. Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature 2004;429:41–46.[CrossRef][Medline]

5. Lechner A, Habener JF. Stem/progenitor cells derived from adult tissues: Potential for the treatment of diabetes mellitus. Am J Physiol Endocrinol Metab 2003;284:E259–E266.[Abstract/Free Full Text]

6. Gershengorn MC, Geras-Raaka E, Hardikar AA. Are better islet cell precursors generated by epithelial-to-mesenchymal transition? Cell Cycle 2005;4:380–382.[Medline]

7. Gershengorn MC, Hardikar AA, Wei C et al. Epithelial-to-mesenchymal transition generates proliferative human islet precursor cells. Science 2004;306:2261–2264.[Abstract/Free Full Text]

8. Davani B, Ikonomou L, Raaka BM et al. Human islet-derived precursor cells are mesenchymal stromal cells that differentiate and mature to hormone-expressing cells in vivo. STEM CELLS 2007;25:3215–3222.[Abstract/Free Full Text]

9. Melloul D, Marshak S, Cerasi E. Regulation of insulin gene transcription. Diabetologia 2002;45:309–326.[CrossRef][Medline]

10. Ohneda K, Ee H, German M. Regulation of insulin gene transcription. Semin Cell Dev Biol 2000;11:227–233.[CrossRef][Medline]

11. Mutskov VJ, Farrell CM, Wade PA et al. The barrier function of an insulator couples high histone acetylation levels with specific protection of promoter DNA from methylation. Genes Dev 2002;16:1540–1554.[Abstract/Free Full Text]

12. Mutskov V, Felsenfeld G. Silencing of transgene transcription precedes methylation of promoter DNA and histone H3 lysine 9. EMBO J 2004;23:138–149.[CrossRef][Medline]

13. Clayton AL, Hazzalin CA, Mahadevan LC. Enhanced histone acetylation and transcription: A dynamic perspective. Mol Cell 2006;23:289–296.[CrossRef][Medline]

14. Nightingale KP, O'Neill LP, Turner BM. Histone modifications: Signalling receptors and potential elements of a heritable epigenetic code. Curr Opin Genet Dev 2006;16:125–136.[CrossRef][Medline]

15. Sims RJ 3rd, Reinberg D. Histone H3 Lys 4 methylation: Caught in a bind? Genes Dev 2006;20:2779–2786.[Free Full Text]

16. Litt MD, Simpson M, Gaszner M et al. Correlation between histone lysine methylation and developmental changes at the chicken beta-globin locus. Science 2001;293:2453–2455.[Abstract/Free Full Text]

17. Kondo Y, Shen L, Issa JP. Critical role of histone methylation in tumor suppressor gene silencing in colorectal cancer. Mol Cell Biol 2003;23:206–215.[Abstract/Free Full Text]

18. Nguyen CT, Weisenberger DJ, Velicescu M et al. Histone H3-lysine 9 methylation is associated with aberrant gene silencing in cancer cells and is rapidly reversed by 5-Aza-2'-deoxycytidine. Cancer Res 2002;62:6456–6461.[Abstract/Free Full Text]

19. Nielsen SJ, Schneider R, Bauer UM et al. Rb targets histone H3 methylation and HP1 to promoters. Nature 2001;412:561–565.[CrossRef][Medline]

20. Chen LB, Jiang XB, Yang L. Differentiation of rat marrow mesenchymal stem cells into pancreatic islet beta-cells. World J Gastroenterol 2004;10:3016–3020.[Medline]

21. Choi KS, Shin JS, Lee JJ et al. In vitro trans-differentiation of rat mesenchymal cells into insulin-producing cells by rat pancreatic extract. Biochem Biophys Res Commun 2005;330:1299–1305.[CrossRef][Medline]

22. Moriscot C, de Fraipont F, Richard MJ et al. Human bone marrow mesenchymal stem cells can express insulin and key transcription factors of the endocrine pancreas developmental pathway upon genetic and/or microenvironmental manipulation in vitro. STEM CELLS 2005;23:594–603.[Abstract/Free Full Text]

23. Tang DQ, Cao LZ, Burkhardt BR et al. In vivo and in vitro characterization of insulin-producing cells obtained from murine bone marrow. Diabetes 2004;53:1721–1732.[Abstract/Free Full Text]

24. Berger SL. Histone modifications in transcriptional regulation. Curr Opin Genet Dev 2002;12:142–148.[CrossRef][Medline]

25. Felsenfeld G, Groudine M. Controlling the double helix. Nature 2003;421:448–453.[CrossRef][Medline]

26. Turner BM. Cellular memory and the histone code. Cell 2002;111:285–291.[CrossRef][Medline]

27. Mosley AL, Ozcan S. Glucose regulates insulin gene transcription by hyperacetylation of histone H4. J Biol Chem 2003;278:19660–19666.[Abstract/Free Full Text]

28. Chakrabarti SK, Francis J, Ziesmann SM et al. Covalent histone modifications underlie the developmental regulation of insulin gene transcription in pancreatic beta cells. J Biol Chem 2003;278:23617–23623.[Abstract/Free Full Text]

29. Bernstein BE, Kamal M, Lindblad-Toh K et al. Genomic maps and comparative analysis of histone modifications in human and mouse. Cell 2005;120:169–181.[CrossRef][Medline]

30. Pokholok DK, Harbison CT, Levine S et al. Genome wide map of nucleosome acetylation and methylation in yeast. Cell 2005;122:517–527.[CrossRef][Medline]

31. Schubeler D, MacAlpine DM, Scalzo D et al. The histone modification pattern of active genes revealed through genome-wide chromatin analysis of a higher eukaryote. Genes Dev 2004;18:1263–1271.[Abstract/Free Full Text]

32. Vakoc CR, Mandat SA, Olenchock BA et al. Histone H3 lysine 9 methylation and HP1gamma are associated with transcription elongation through mammalian chromatin. Mol Cell 2005;19:381–391.[CrossRef][Medline]

33. Hake SB, Allis CD. Histone H3 variants and their potential role in indexing mammalian genomes: The "H3 barcode hypothesis". Proc Natl Acad Sci U S A 2006;103:6428–6435.[Abstract/Free Full Text]

34. Ringrose L, Paro R. Epigenetic regulation of cellular memory by the Polycomb and Trithorax group proteins. Annu Rev Genet 2004;38:413–443.[CrossRef][Medline]

35. Gan Q, Yoshida T, McDonald OG et al. Epigenetic mechanisms contribute to pluripotency and cell lineage determination of embryonic stem cells. STEM CELLS 2007;25:2–9.[Abstract/Free Full Text]

36. Roloff TC, Nuber UA. Chromatin, epigenetics and stem cells. Eur J Cell Biol 2005;84:123–135.[CrossRef][Medline]

37. Szutorisz H, Dillon N. The epigenetic basis for embryonic stem cell pluripotency. Bioessays 2005;27:1286–1293.[CrossRef][Medline]

38. Bernstein BE, Mikkelsen TS, Xie X et al. A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 2006;125:315–326.[CrossRef][Medline]

39. Higgs DR, Vernimmen D, De Gobbi M et al. How transcriptional and epigenetic programmes are played out on an individual mammalian gene cluster during lineage commitment and differentiation. Biochem Soc Symp 2006;73:11–22.[Medline]

This article has been cited by other articles:

				C. Evans-Molina, R. D. Robbins, T. Kono, S. A. Tersey, G. L. Vestermark, C. S. Nunemaker, J. C. Garmey, T. G. Deering, S. R. Keller, B. Maier, et al. Peroxisome Proliferator-Activated Receptor {gamma} Activation Restores Islet Function in Diabetic Mice through Reduction of Endoplasmic Reticulum Stress and Maintenance of Euchromatin Structure Mol. Cell. Biol., April 15, 2009; 29(8): 2053 - 2067. [Abstract] [Full Text] [PDF]

				M. V Joglekar, V. M Joglekar, S. V Joglekar, and A. A Hardikar Human fetal pancreatic insulin-producing cells proliferate in vitro J. Endocrinol., April 1, 2009; 201(1): 27 - 36. [Abstract] [Full Text] [PDF]

				S. A. Litherland Immunopathogenic Interaction of Environmental Triggers and Genetic Susceptibility in Diabetes: Is Epigenetics the Missing Link? Diabetes, December 1, 2008; 57(12): 3184 - 3186. [Full Text] [PDF]

				B. Davani, L. Ikonomou, B. M. Raaka, E. Geras-Raaka, R. A. Morton, B. Marcus-Samuels, and M. C. Gershengorn Human Islet-Derived Precursor Cells Are Mesenchymal Stromal Cells That Differentiate and Mature to Hormone-Expressing Cells In Vivo Stem Cells, December 1, 2007; 25(12): 3215 - 3222. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 2007-0325v1 25/12/3223    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mutskov, V. Articles by Gershengorn, M. C. Search for Related Content PubMed PubMed Citation Articles by Mutskov, V. Articles by Gershengorn, M. C.