SEARCH

SEARCH BY CITATION

Keywords:

  • Bone marrow cells;
  • Migration;
  • Hepatocyte differentiation;
  • Quantitative polymerase chain reaction;
  • Tissue regeneration

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

In vitro and in vivo studies have shown that bone marrow (BM) stem cells can differentiate into hepatocytes. However, it is not known whether such a differentiation event occurs during normal liver regeneration process. We investigated the role of endogenous BM cells in liver regeneration following acute injury and phenotypically characterized them. We showed that LinSca-1+ cells proliferate in the BM and subsequently mobilize in the peripheral blood in response to liver injury by CCl4 or an injury simulating condition. In vitro studies confirmed that the damaged liver tissue was capable of inducing migration of a distinct population of BM cells, phenotypically characterized as LinCXCR4+OSMRβ+, which can differentiate into albumin and cytoketarin-18 expressing cells. In order to study the migration of BM cells to the regenerating liver, the hematopoietic system was reconstituted with green fluorescent protein (GFP)+ BM cells by intra-bone marrow transplantation prior to liver damage. The BM-derived cells were found to express hepatocyte-specific genes and proteins in the regenerating liver. Quantitative polymerase chain reaction analysis for a recipient specific gene (sry) in sorted GFP+Alb+ donor cells suggested that fusion was a rare event in this experimental model. In conclusion, we first demonstrated the potential phenotype of BM cells involved in regeneration of liver from acute injury, primarily by the process of direct differentiation.

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


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

During mid to late stages of gestation, fetal liver acts as the major site of hematopoiesis in mice [1]. It has been shown that both hematopoietic and hepatic systems interact with each other in the fetal and neonatal stages of development [2]. Various reports suggest that even in the adult life, the resident hematopoietic stem cells of liver or migrated bone marrow (BM) cells are involved in immunological and pathological functions of the liver [3, [4], [5]6]. Earlier studies described the potential of BM hematopoietic stem cells to cross germ layer boundaries by differentiating into other tissues [7, 8]. Liver epithelium tissues have been the most extensively studied systems in this connection [9, [10], [11], [12], [13]14]. Petersen et al. [9] presented the first account of transdifferentiation of marrow-derived cells into hepatocytes, after which many researchers presented evidence in support of plasticity of adult hematopoietic stem cells with respect to hepatocytes [10, [11], [12]13]. Subsequent studies proposed the role of cell fusion in perceived differentiation of hematopoietic cells into hepatocytes [14, 15]. Although in some investigations [16, 17] it was shown that liver epithelium can be generated from BM cells without fusion, these studies precluded phenotypic characterization of the cells and their involvement in the liver regeneration model.

In this study, we have explored (a) the function of endogenous BM cells in liver regeneration following acute injury, (b) the potential BM subpopulation responsible for such regeneration, and (c) whether BM cells directly differentiate into hepatocytes in the regenerating liver. We adopted the CCl4-induced acute liver injury model described by others [18, 19]. Our preliminary in vitro studies show that a distinct subset of Lin BM cells, coexpressing CXCR4 and oncostatin M receptor β (OSMRβ), with/without stem cell antigen-1 (Sca-1), respond to the liver injury simulating condition. In order to comprehend the migration of cells from BM to the damaged liver, we transplanted green fluorescent protein (GFP)+ cells directly in the BM before damaging the liver. We demonstrated that endogenous BM cells migrate specifically to the damaged liver, where they differentiate into albumin and cytokeratin (CK)-18 expressing hepatocytes. Furthermore, by real-time quantitative polymerase chain reaction (PCR) for the recipient-specific gene sry, we confirmed that the conversion of BM-derived cells into hepatocytes was primarily independent of fusion.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Animals

Six to ten-week-old C57BL6/J, GFP transgenic mice (FVB.Cg-Tg[GFPU] 5NAGY/J) and FVB/NJ mice were used in this investigation. Mice were obtained from the Jackson Laboratory (Bar Harbor, ME, http://www.jax.org) and maintained in the institute's experimental animal facility. During the experiments, mice were kept in an isolator and fed with autoclaved acidified water and irradiated food ad libitum. All experiments using mice were conducted as per procedures approved by the Institutional Animal Ethics Committee.

Liver Damage and Collection of Sera

Mice were given one injection of either 200 μl of mineral oil (control) or 10% solution of CCl4 via intraperitoneal route [20]. Mice were euthanized at different time intervals to collect BM and peripheral blood (PB) cells. To collect sera, two groups of mice (n = 20) were taken; the first group was given one injection of mineral oil (200 μl) and the other group received CCl4, as above. Serum was collected 24 hours after injection, and the pooled serum was termed as normal serum (NS) (mineral oil injected mice) or damaged liver serum (DLS) (CCl4 injected mice).

Cell Migration Assay

In vitro cell migration assay was conducted using a Falcon cell culture insert (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) fitted on a 24-well tissue culture plate. The insert had a 0.3-cm2 membrane with 5-μm pores at the bottom. Twenty-four hours after the liver damage with CCl4, a small piece of the same liver tissue was placed on one side of the lower companion well containing 1 ml of Iscove's modified Dulbecco's medium (IMDM) supplemented with 10% DLS. GFP+ whole BM or PKH-26 labeled [21] Lin BM cells were used as test cells for migration studies. One hundred thousand cells in 200 μl of IMDM were loaded on the insert, and the cells were allowed to migrate for 3 hours across the membrane. The migrated cells were analyzed by surface staining for lineage markers Sca-1, CXCR4, and OSMRβ by flow cytometry and immunocytochemistry. Lineage antibody cocktail was purchased from Miltenyi Biotec (Bergisch Gladbach, Germany, http://www.miltenyibiotec.com), anti-Sca-1 and anti-CXCR4 antibodies from BD Pharmingen (San Diego, http://www.bdbiosciences.com/index_us.shtml), and anti-OSMRβ from R&D Systems Inc. (Minneapolis, http://www.rndsystems.com). To study hepatic differentiation potential, the migrated cells as a whole and their OSMRβ+ fraction and LinCXCR4+OSMRβ+ cells from BM were cultured in the presence of 10% DLS for 3 days. Prior to culture, plates were coated with a mixture of gelatin (10 mg/ml), laminin (50 μg/ml), and hyaluronic acid (100 μg/ml). The cultured cells were examined for the expression of albumin (Nordic Immunological Laboratories, Tilburg, The Netherlands, http://www.nordiclabs.nl) and cytokeratin-18 (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com) by immunocytochemical analysis. Before staining of intracellular proteins, cells were fixed using 4% paraformaldehyde and then permeabilized by treating with 0.1% saponin solution for 20 minutes at room temperature. LinCXCR4+OSMRβ+ cells were separated from BM by following a three-step magnetic activated cell sorting protocol using specific antibodies and magnetic beads.

Intra-Bone Marrow Transplantation and Immunohistochemistry

FVB/NJ male mice (n = 30) were irradiated at 900 cGy (137Cs source) 2 hours prior to transplantation. Intra-bone marrow transplantation (IBMT) of GFP+ BM cells was performed as described previously [22] with some modifications. Ten million GFP+ cells were directly injected in the femur of anesthetized FVB/NJ mice (supplemental online Methods). After 15 days of marrow repopulation, mice were injected with either mineral oil or CCl4. To analyze liver tissue, mice were sacrificed at 3 days, 1 month, and 1 year after damaging the liver.

Formalin-fixed tissues were embedded in paraffin blocks, and 5-μm sections were cut. Immunostaining was performed using biotinylated anti-albumin antibody, rat anti-GFP, and mouse anti-CK-18 antibodies. The secondary antibodies used were either conjugated with Alexafluor488 or Alexafluor594 (Molecular Probes, Carlsbad, CA, http://probes.invitrogen.com). Nuclei were stained with 4,6-diamidino-2-phenylindole. Serial sections of each liver sample were examined in a fluorescence microscope (Olympus, Tokyo, http://www.olympus-global.com).

Flow Cytometry

Cells (BM, PB, and liver) were stained by incubating with primary antibodies at 4°C for 45 minutes followed by detection using fluorochrome-labeled streptavidin. Cells were analyzed by flow-cytometry (BD LSR; BD Biosciences, San Diego, http://www.bdbiosciences.com). The antibodies and the conjugates used in this study were Sca-1/fluorescein isothiocyanate (FITC), streptavidin-FITC/phycoerythrin/Cy5.5 (BD Pharmingen), biotinylated anti-albumin, and lineage antibody cocktail. Isotype control antibodies were procured from BD Pharmingen.

Isolation of Total RNA and Reverse Transcription-PCR Analysis

The expression of different genes was analyzed by reverse transcription (RT)-PCR. In brief, total RNA was recovered using TRI Reagent (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) according to the manufacturer's instructions. cDNA was synthesized from 0.5 μg of total RNA using ProtoScript First Strand cDNA Synthesis Kit (NEB, Beverly, MA, http://www.neb.com) and was subjected to PCR amplification with primers selective for various target genes. The primers and amplification conditions for PCR reactions were as follows: albumin: GTGCAAGAACTATGCTGAGG (forward), ACTCACTGGGTCTTCTCAT (reverse) (amplicon—465 base pairs [bp], 58.1°C); tyrosine aminotransferase (TAT): GAGGAGTGTACAAATAAGGC (forward), AGAGGACACTCCTGTGTCAG (reverse) (amplicon—422 bp, 51.7°C); tryptophan dioxygenase (TDO): CATGGCTGGAAAGAACACCT (forward), TCGAGGCTCTTCCCTGTAAA (reverse) (amplicon—289 bp, 59°C); CK-18: CACCACCTTCTCCACCT (forward), GCCTCGATTTCTGTCTCCAG (reverse) (amplicon—573 bp, 50°C); stroma-derived factor-1α (SDF-1α): GTCCTCTTGCTGTCCAGCTC (forward), GGGGGTCTACTGGAAAGTCC (reverse) (amplicon—356 bp, 58.9°C); α-fetoprotein (AFP): CGCTCTCTACCAGACCTTAGGA (forward), CTCCTCTGTCAGTTCAGGCTTT (reverse) (amplicon—451 bp, 59.2°C); γ-glutamyl transferase (GGT): CACAGACAGTGGCTCAGACTTGG (forward), AGTGTGTGGTCCTCCAGGATGG (reverse) (amplicon—306 bp, 59.8°C); CK-19: CACCACCTTCTCCACCAACT (forward), GCCTCGATTTCTGTCTCCAG (reverse) (amplicon—573 bp, 57.3°C); CD45: TCTCCCAGGAGTATGAGTCCAT (forward), GGCCAATACTGATCACACTTCA (reverse) (amplicon—339, bp, 51.1°C); glyceraldehyde-3-phosphate dehydrogenase: GAATACGGCTACAGCAACAG (forward), CTAGGCCCCTCCTGTTATTA (reverse) (amplicon—209 bp, 57.9°C). The reactions were performed with hot start at 95°C for denaturation, annealing at specific temperatures, and finally amplification at 72°C for 25–35 cycles (depending on the target gene). The amplified products were resolved in 1.5%–2.0% agarose gel and visualized by ethidium bromide staining, and images were recorded in UVP BioImaging System (UVP, Upland, CA, http://www.uvp.com).

Real-Time Quantitative PCR

Real-time quantitative (q)PCRs were performed by means of SYBR Green technology and ABI PRISM 7000 apparatus (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com). For each sample, ABI PRISM 7000 software plotted an amplification curve by relating the fluorescence signal intensity (ΔRn) to the cycle number. The ΔRn value corresponded to the variation in reporter fluorescence intensity during each PCR cycle, normalized to the fluorescence of an internal passive reference. A specific crossing point (Ct) was determined for each PCR. The Ct was defined as the cycle number at which a significant increase in the fluorescence signal was first detected (the higher the starting copy number, the lower the Ct). The PCR reaction parameters were as follows: (a) reaction mix: 100 ng of DNA mixed with 12.5 μl of 2X Master Mix (Applied Biosystems) and 200 nM each primer in a final volume of 25 μl; (b) PCR cycles: 10 minutes at 95°C, 40 amplification cycles (95°C for 30 seconds, 60°C for 30 seconds, 95°C for 30 seconds), and 4°C for 2 minutes. Dissociation protocol was added.

To evaluate the validity and the sensitivity of real-time qPCR sry gene analysis, standard amplification curves were plotted for recipient (male)- and donor (female)-specific DNA samples isolated from 10 serial halved dilutions of recipient cells in donor cells. The cell mixtures were made as follows: male cells were maintained 100%, 50%, 25%, 12.5%, 6.25%, 3.2%, 1.6%, 0.8%, 0.4%, 0.2%, and 0% and the corresponding female cells were 0%, 50%, 75%, 87.5%, 93.75%, 96.8%, 98.4%, 99.2%, 99.6%, 99.8%, and 100%. The genomic DNA was isolated from these standard cell mixtures using a Qiagen genomic DNA isolation kit (Hilden, Germany, http://www1.qiagen.com). In PCR, constant amounts (100 ng) of these DNA mixtures were used. To determine loading control in each standard mix, β-actin qPCR was performed. The Ct values of unknown samples were determined by qPCR, and the corresponding percentage of Y chromosome bearing cells was directly determined from the standard plot.

Statistical Values

Results of multiple experiments were reported as the mean ± SEM. One-way analysis of variance was followed to calculate the significance between two means.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Liver Damage Induces Proliferation and Mobilization of Hematopoietic Progenitor Cells from BM to PB

Initially, we studied the proliferation of hematopoietic progenitor cells (LinSca-1+) in BM and their mobilization in the peripheral blood in CCl4-injected mice. Mice were sacrificed at different times, and the number of progenitor stem cells in BM and PB was analyzed by flow cytometry. The analysis of LinSca-1+ cells indicated a wave of increasing stem cell population in BM on day 1, which passes on to the PB by the next day (Fig. 1A). In BM, LinSca-1+ cells reached a maximum level by day 1 post-liver damage, which declined to the basal level by day 3 (Fig. 1A). The increase in LinSca-1+ cells at day 1 was 2.8-fold higher (p < .01) as compared with the control mice at day 0. The circulating LinSca-1+ cells at day 0 were 0.46% ± 0.03% (Fig. 1A), which increased to 1.95% ± 0.17% (p < .01) by day 2 of liver injury. We examined the cell cycle status of BM cells by quantitating DNA content per cell (supplemental online Figure 1, panel IB). It was found that liver injury induced proliferation of BM cells, as S+G2/M phase cell population was significantly (p < .05) higher than the control (Fig. 1B).

thumbnail image

Figure Figure 1.. Effect of liver damage on LinSca-1+ cell population in BM and PB. Mice were sacrificed after 0, 1, 2, and 3 days of CCl4 injection, and the mononuclear cells were analyzed (supplemental online Fig. 1, panel IA). (A): Change in LinSca-1+ population in BM and PB cells at different time intervals. (B): Percentage of S+G2/M phase cells in BM at different time intervals. Mice were injected with either NS or DLS, and the mononuclear cells of BM and PB were analyzed. (C): Change in LinSca-1+ population in BM cells at different time intervals. (D): Change in LinSca-1+ population in PB cells at different time intervals. Mean ± SEM values from three to four independent experiments are shown. ∗, p < .01; ∗∗, p < .05 in comparison with the respective control at day 0. Abbreviations: BM, bone marrow; DLS, damaged liver serum; NS, normal serum; PB, peripheral blood.

Download figure to PowerPoint

In order to examine whether the above change in BM-progenitor cells was a consequence of liver damage or due to direct effects of CCl4, instead of injecting CCl4 we injected NS and DLS in two different groups of mice. Although NS did not cause any change in the LinSca-1+ population in BM and PB cells, DLS induced mobilization of these cells to PB (Fig. 1C, 1D). These results suggest that the serum of liver-damaged mice contains some soluble factors, which can induce the proliferation of LinSca-1+ cells.

CXCR4/SDF-1α Mediated Migration of BM Cells Takes Place in Response to Liver Injury

We analyzed the progenitor stem cell population in BM and PB cells for the expression of chemokine receptor CXCR4. The expression of CXCR4 on LinSca-1+ cells was increased to a maximum value at day 1 and was twofold (p < .05) higher as compared with the 0-day control (Fig. 2A). The LinSca-1+CXCR4+ BM cells were mobilized to the PB, as a significant (p < .05) increase in their number was observed in PB by day 2 of liver damage as compared with the control (Fig. 2A). These findings prompted us to hypothesize that mobilization of CXCR4+ progenitor cells in PB may be a direct consequence of liver damage. Since these progenitor cells can potentially home to the damaged liver, we envisaged a plausible role of chemotactic signals in this process. SDF-1α is a chemokine that has been shown to be involved in stress-induced recruitment of transplanted CD34+ cells in liver [23]. We therefore analyzed SDF-1α gene expression in the CCl4-damaged liver tissue. The expression of the SDF-1α gene was found to increase substantially over time with a maximum level at day 2 of liver damage (Fig. 2B). Together, these results suggested the possible involvement of CXCR4/SDF-1α in migration of BM cells to the site of CCl4-induced damaged liver.

thumbnail image

Figure Figure 2.. Effect of liver damage on migration of BM cells in vivo and in vitro. The effect of liver damage on expression of chemokine receptor CXCR4 and gene of cognate ligand SDF-1α was analyzed in BM and liver, respectively. (A): Proportion of LinSca-1+ cells in BM and PB expressing CXCR4 in response to CCl4 treatment. (B): Expression of SDF-1α gene in damaged liver tissue. Damaged liver tissue induces migration of Lin BM cells in vitro. The GFP+ BM cells were incubated in the upper chamber of culture insert in contact with damaged liver tissue placed in the lower chamber. Photomicrographs showing (Ca) GFP+ cells in the upper chamber, (Cb) migrated GFP+ cells in the lower chamber, (Cc) migration in absence of damaged liver tissue, (Cd) GFP+ cells near the damaged liver tissue, and (Ce–Cj) migration patterns of GFP+ cells. The GFP+ cells were crowded near the damaged liver tissue, and the numbers of migrated cells reduced with increase in distance between liver tissue and the plane of migration. (D): Analysis of migrated GFP+ cells for the expression of lineage markers by flow cytometry. Mean ± SEM values from three independent experiments are shown. Representative results of one experiment are shown in (C) and (D). ∗∗, p < .05 in comparison with the respective control at day 0. Abbreviations: BM, bone marrow; d, day(s); GFP, green fluorescent protein; PB, peripheral blood; SDF-1α, stroma-derived factor-1α.

Download figure to PowerPoint

In order to ensure that BM cells migrate to the damaged liver, we performed in vitro migration studies. GFP+ mononuclear BM cells were allowed to migrate, as mentioned in Materials and Methods. Within 1 hour, GFP+ cells started migration, and significant numbers of cells were detected in the lower chamber within 3 hours (Fig. 2Cb). In contrast, there was no migration of GFP+ cells in the absence of damaged liver tissue (Fig. 2Cc). Most interestingly, the number of GFP+ cells that migrated near the liver tissue was more (Fig. 2Cd), which gradually reduced with increase in distance between cell migrating front and the liver tissue (Fig. 2Cd–2Cj). Based on these preliminary studies and RT-PCR results (Fig. 2B), we conclude that the concentration gradient of the chemoattractants, secreted by the damaged liver tissue, directed the migration of BM cells. We analyzed the migrated cells in terms of the expression of lineage markers and revealed that the migrated cells in the lower chamber were free from any lineage-committed cells (Fig. 2D).

A Distinct Population of Cells Expressing CXCR4 and OSMRβ Migrate Under the Influence of Damaged Liver Tissue

In vivo results showed that LinSca-1+CXCR4+ cells are mobilized in the PB in response to the liver damage. Furthermore, in vitro migration experiments confirmed that only Lin cells were capable of migration. In order to phenotypically characterize the migrating cells, PKH-26 labeled Lin cells were used for similar in vitro migration experiments for 3 hours. The migrated cells were subjected to both immunocytochemical and flow-cytometric analyses for the cell surface markers. The results further confirmed that the migratory cells were Lin fraction (<98%) of the BM cells (Fig. 3A, upper panel; Fig. 3Ba). Further analyses of these migrated cells revealed that a significant fraction of them expressed CXCR4 (89.9%), OSMRβ (67.7%), and Sca-1 (51.2%) as shown in Figure 3A (lower three panels) and Figure 3Ba. Besides CXCR4 and Sca-1, we analyzed the surface expression of OSMRβ, as its ligand OSM is known to play an important role in hepatic differentiation of precursor cells in fetal liver [24]. These types of analyses of surface proteins on individual cells indicated that a significant fraction of Lin cells expressed CXCR4, OSMRβ, and Sca-1. However, it is inadequate to conclude on the phenotype of the cells migrated in response to liver injury. So, we did a four-color analysis of the cells migrated under identical conditions as above (Fig. 3Bb). Since Lin cells were labeled with PKH-26 dye, we selected PKH-26+ fraction (R1). These Lin PKH-26+ fraction cells were analyzed with respect to the expression of CXCR4, OSMRβ, and Sca-1. It was evident that approximately 90% of migrated cells expressed CXCR4, out of which only 62.4% of cells coexpressed with OSMRβ (Fig. 3Bb). These results showed that a significant fraction of migrated CXCR4+ cells also expressed receptor for OSM, perhaps suggesting that not all migrating cells were potential for hepatocytic differentiation. We further analyzed LinCXCR4+OSMRβ+ cells for the expression of Sca-1, a hematopoietic stem cell marker, which is also expressed in the hepatic stem cells [25]. It was revealed that approximately 67% of LinCXCR4+OSMRβ+ cells express Sca-1 antigen (Fig. 3Bb). These results suggest that there are two types of OSMRβ+ cells in migrated population, and they differ from each other by the expression of Sca-1.

thumbnail image

Figure Figure 3.. Phenotypic characterization of Lin bone marrow (BM) cells migrated in response to damaged liver. In vitro migration experiments were performed using PKH-26 labeled Lin BM cells, same as above. The migrated cells were stained with respective antibodies. Photomicrographs showing Lin cells ([A], top panel), CXCR4+ cells ([A], second top panel), OSMRβ+ cells ([A], second bottom panel), and Sca-1+ cells ([A], bottom panel). (B): Four-color flow-cytometric analysis of migrated cells. Histograms (Ba) showing the cells expressed individual markers; contour plot ([Bb], middle) showing expression of OSMRβ and CXCR4 on the migrated LinPKH-26+ cells (R1); cells staining positively for both these markers were further gated to show the expression of Sca-1 antigen ([Bb], right). Abbreviations: DAPI, 4,6-diamidino-2-phenylindole; FSC, forward scatter; OSMRβ, oncostatin M receptor β; Sca-1, stem cell antigen-1.

Download figure to PowerPoint

Before initiating further analysis on above cells, we ensured the differentiation potential by culturing them in the presence of 10% DLS. Although approximately 60% of migrated cells expressed albumin and CK-18 (Fig. 4A), the results demonstrated that a considerable fraction of them do not participate in the differentiation program. In order to reanalyze migrated cells, as well as the cells present in BM for differentiation potential, they were subjected to sorting for the phenotype expressing LinCXCR4+OSMRβ+, as CXCR4 and OSMRβ were found important for migration and hepatic differentiation, respectively. The sorted cells from these two sources were again cultured and revealed that 100% of them expressed albumin and CK-18 (Fig. 4B, 4C). This implies that hepatic progenitor cells present in BM express OSMRβ and are denoted by LinCXCR4+OSMRβ+Sca-1+/−.

thumbnail image

Figure Figure 4.. In vitro hepatic differentiation potential of cells. Different fractions of cells were cultured for 3 days in the presence of damaged liver serum. Prior to immunostaining with anti-albumin and CK-18 antibodies, harvested cells were spread on glass slide with the help of a cytospin apparatus. Photomicrographs showing expression of albumin and CK-18 on (A) migrated cells as in Figure 3. (B): Migrated cells as in Figure 3, followed by positive selection for oncostatin M receptor β (OSMβ+). (C): Three-step purified fresh bone marrow cells designated as LinCXCR4+OSMRβ+ fraction. Abbreviations: CK, cytokeratin; DAPI, 4,6-diamidino-2-phenylindole.

Download figure to PowerPoint

BM Cells Migrate to the Damaged Liver and Differentiate into Hepatocytes

Increase in progenitor stem cell population in the BM, its mobilization to the PB, increasing expression of chemokine receptor-ligand pair, and in vitro migration of a selected population of BM cells toward damaged liver tissue predict the involvement of these cells in liver regeneration from CCl4-induced acute injury. For further assessment of these cells, we conducted liver regeneration experiments. Bone marrow of FBV/J male mice was repopulated with GFP+ BM cells of congenic female mice by IBMT. This method of transplantation circumvents trapping of BM cells in the liver, seen in cases of injections through the intravenous route, and allows us to directly deliver cells in the bone marrow. After repopulating BM with the donor cells (Fig. 5A), the liver was damaged by injecting CCl4 (control mice received mineral oil). The presence of GFP+ cells in liver tissue was examined by immunostaining with anti-GFP antibody. The control mice showed no GFP+ cells in any of the liver tissue sections examined (Fig. 5Be, 5Bi), whereas considerable numbers of GFP+ cells could be visualized in CCl4 induced damaged liver sections (Fig. 5Bd). This migration of GFP+ cells was specific to the damaged liver (Fig. 5B), as we did not observe any such cells in the pancreas (Fig. 5Ba), heart (Fig. 5Bb), and kidney (Fig. 5Bc) tissue sections of CCl4 treated mice. Immunohistochemical analysis of serial liver sections showed that as early as 3 days after injury, a distinct number of GFP+ cells expressed albumin (Fig. 5Bf). Although some of the BM-derived cells expressed albumin at an early time point, they did not express CK-18, an epithelial marker (Fig. 5Bj). To ensure that the BM-derived cells were retained in the damaged liver tissue, we extended the study up to 1 year. At later times, GFP+ cells not only expressed albumin, but a considerable number of them also expressed CK-18 (Fig. 5Bk). The long-term (1-year) study confirmed that BM-derived GFP+ cells persisted in the liver lobules, and nearly 80% of them continued expression of liver-specific markers, such as albumin (Fig. 5Bh) and CK-18 (Fig. 5Bl). The BM-derived hepatocytes were found to be morphologically comparable to the normal hepatocytes. Summarizing the immunohistochemical analysis, the expression of albumin but not CK-18 could be seen in the donor-derived GFP+ cells within 3 days of liver damage, although both of them were consistently expressed at later times.

thumbnail image

Figure Figure 5.. Assessment of bone marrow (BM) engraftment with GFP+ cells, its migration to the damaged liver, and immunohistochemical analysis of liver sections. Hematopoietic system was reconstituted with GFP+ cells as mentioned in Materials and Methods. Prior to liver damage with CCl4, BM reconstitution by donor cells was studied by analyzing GFP+ cells in the peripheral blood. Mice showing GFP chimerism above 40% were inducted for in vivo migration experiments. (A): Analysis of GFP+ cells in 15-day postengrafted BM followed by mineral oil injection (3-day, left) and CCl4 injection (3-day, middle and 30-day, right). (B): Immunohistochemical analysis of liver sections. Mice were sacrificed at 3 days, 1 month, and 1 year after the damage; 5-μm serial liver sections were analyzed with different antibodies. Top panel: Micrographs of different tissues ([Ba]: pancreas, [Bb]: heart, [Bc]: kidney, and [Bd]: liver) stained for GFP and nuclei (×100). Middle panel showing GFP (green), albumin (red), and nuclear (blue) staining (×400): micrographs. (Be): Three-day control after mineral oil injection; (Bf) 3 days, (Bg) 1 month, and (Bh) 1 year after CCl4 injection. Merged photographs (Bf–Bh) showing yellow color GFP+Alb+ cells and nuclei stained blue with DAPI. Lower panel showing GFP (green), CK-18 (red), and nuclear (blue) staining (×400): micrographs. (Bi): Three-day control after mineral oil injection; (Bj) 3 days, (Bk) 1 month, and (Bl) 1 year after CCl4 injection. Merged photographs (Bj–Bl) showing yellow color GFP+CK-18+ cells and nuclei stained blue with DAPI. Three-day and 1-month experiments were conducted in four mice each, and 1-year experiments were conducted in two mice. Representative results of one experiment are shown. Abbreviations: Alb, albumin; CK, cytokeratin; DAPI, 4,6-diamidino-2-phenylindole; eGFP, enhanced green fluorescent protein; GFP, green fluorescent protein.

Download figure to PowerPoint

Furthermore, de novo expression of hepatic genes was examined in sorted BM-derived liver cells (supplemental online Fig. 1, panel II) by RT-PCR. This analysis revealed that albumin and other hepatocyte-specific genes, such as TAT, TDO, and CK-18, were significantly expressed in the donor-derived cells. The expression of hepatocarcinogenic markers AFP and GGT was negligible in the 3-day samples but was completely absent at the later time points (Fig. 6). CK-19 expression, which is indicative of differentiation into bile ductular cells, was completely absent in BM-derived cells (Fig. 6). Hepatocyte-specific gene expression in GFP+ cells confirmed the presence of an inductive microenvironment in damaged liver that mediates reprogramming of progenitor stem cells in favor of hepatocytes. This was further evident by the downregulation of the expression of CD45 gene in the sorted BM-derived cells. These results suggest that the BM-derived cells were comparable to normal adult hepatocytes in terms of gene and protein expression and also in morphological features.

thumbnail image

Figure Figure 6.. Liver-specific gene expression in sorted cells recovered from the liver. Experiments were performed as described in Figure 5. GFP+ (3-day and 1-month) and GFP+Alb+ (1-year) sorted cells were analyzed for the target genes by reverse transcription-polymerase chain reaction. BM cells were taken as negative control for the target genes (BM cells were supplemented with 2% liver cells, as the purity of GFP+/GFP+Alb+ sorted cells was 98%–99%). Adult liver cells were used as positive control for albumin, TAT, TDO, CK-18, and CK-19, and fetal liver cells were used as positive control for AFP and GGT. Three-day and 1-month experiments were conducted in four mice each, and 1-year experiments were conducted in two mice. Representative results of one experiment are shown. Abbreviations: AFP, α-fetoprotein; BM, bone marrow; CK, cytokeratin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GGT, γ-glutamyl transferase; TAT, tyrosine aminotransferase; TDO, tryptophan dioxygenase.

Download figure to PowerPoint

BM-Derived Cells Differentiate into Hepatocytes in the Regenerating Liver Without Fusion

The observation that liver injury induced migration of BM cells to the liver and reprogrammed into hepatocytic cells suggests their physiological role in liver regeneration. With this understanding, we examined the possibility of fusion of the BM-derived cells with the host hepatocytes. We conducted real-time qPCR analysis for the Y chromosome-specific gene sry in sorted GFPAlb+ (recipient hepatocytes) and GFP+Alb+ (donor-derived hepatocytes) cells from the same mice. The premise behind this study was that, only if BM-derived cells (female) fused with host cells (male), they would possess sry gene; otherwise, GFP+Alb+ cells will not show the amplification of sry gene. To test the presence of sry gene in the sorted GFP+Alb+ cells, initially artificial chimeric genomic DNA samples were prepared from 10 serial halved dilutions (0.2%–100%) of recipient (male) cells in donor (female) cells. The specific PCR amplification plots obtained from these chimeric DNA mixtures were overlaid and are shown in Figure 7B. The amplification plot shifted to the right to higher threshold cycles as the input target sry gene quantity was reduced. By contrast, no amplification was visualized for the negative control DNA (100% female + 0% male cells DNA) up to 40 cycles. Figure 7C represents the Ct values plotted versus the log values of percentage male cells in artificial chimeric DNA samples. Standard curve slope ranged from −3.17 to −3.41, and the corresponding PCR efficiency was found to be 0.97–0.99. In 100 ng of total DNA, the quantification of sry gene was shown to be linear over three logs of DNA concentrations of Y-chr+ cells, and the assay could measure as low as 0.2% cells containing target gene.

thumbnail image

Figure Figure 7.. Real-time polymerase chain reaction (PCR) analysis for recipient-specific marker for sry gene. Experiments were performed as described in Figure 5 for 1-year postdamage of liver. (A): Sorting of GFPAlb+ and GFP+Alb+ cells using BD FACSAria. Dot plots of presorted and postsorted cells are shown to examine the purity of respective fractions of cells. (B): Amplification curve by relating the fluorescence signal intensity (▵Rn) to the cycle number for different concentrations of chimeric DNA mixture, as mentioned in Materials and Methods. (C): Crossing point (Ct) curve plotted against each concentration of male chromosome. Values of unknown samples (Y-chr+ cells) of two mice (marked as 1 and 2) were directly quantified from the Ct plot. (D): Percentage of Y-chr+ cells in the recipient (GFPAlb+) and donor (GFP+Alb+) cell fraction. PCR experiments were conducted for both the mice in duplicate samples. Mean ± SEM values from 2 × 2 experiments are shown. Abbreviations: Alb, albumin; E, efficiency; GFP, green fluorescent protein; S, slope.

Download figure to PowerPoint

To determine the presence of Y-chr+ cells, we sorted GFP+Alb+ and GFPAlb+ liver cells of the experimental mice. The purity of these fractions of cells was greater than 98.5% (Fig. 7A). The purpose of using GFPAlb+ cells in qPCR analysis was to show that the recipient mice were indeed male, if Ct value almost matches with that obtained with 100% male standard cells used to establish amplification plots (Fig. 7B). The results of unknown samples (GFP+Alb+ cells) of two different experimental mice are shown in Figure 7C. The calculated values of Y-chr+ cells in GFPAlb+ and GFP+Alb+ sorted cells were 99.26% ± 0.45% and 5.37% ± 0.83%, respectively (Fig. 7D). These results confirmed that (a) recipient mice were male and (b) fusion between the recipient and donor-derived cells was not the cause of hepatocyte-specific gene expression in BM-derived cells, indicating that direct differentiation was the principal mechanism of conversion of BM cells into hepatocytes. The presence of only 5% Y-chr+ DNA in GFP+Alb+ sorted fraction could be either due to contamination with the recipient hepatocytes or due to the formation of few fused cells between BM-derived cells and recipient hepatocytes. These results indicate that cell fusion was not the predominating mechanism involved in differentiation of BM-derived cells into hepatocytes in the present experimental model.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Primary hepatocytes remain the choice for transplantation to treat patients with severe acute end-stage chronic liver and metabolic liver diseases [26]. Due to worldwide shortage of donor hepatocytes and perceived inherent risks of infection and rejection in xenogenic cells, there is a need to work on alternate sources of hepatocytes. In this connection, embryonic stem cells hold much promise in the future. The circulating hematopoietic stem cells contribute to repair solid organs, like liver, and could be promising for the future [27]. In this investigation, we provide direct evidence to show that endogenous BM cells are involved for regeneration of liver after acute injury, which appears to be an important mechanism for autonomous regeneration of liver. Using in vitro migration assay, we have characterized the potential population of BM cells, which participate in the regeneration process.

As a first step toward understanding the role of BM cells in liver regeneration, we examined the effect of liver injury on cell cycle activation and proliferation of progenitor stem (LinSca-1+) cells [28, 29] in the BM. It was observed that the increase in LinSca-1+ cells and mitotic activity in the BM was the consequence of physiological response to liver damage. The response in BM was identical irrespective of whether the mice received CCl4 or DLS, implicating that the observed cellular activities in BM were the effect of the soluble putative growth factors present in the serum. Studies in humans have shown that the PB-hematopoietic stem cell population is increased after extensive liver resection [30] and in patients suffering from alcoholic hepatitis [31].

In the CCl4-treated mice, the major fraction of mobilized progenitor stem cells expressed chemokine receptor CXCR4. At the same time, the mRNA level of its ligand (SDF-1α) in the damaged liver tissue was increased. These results provided us a clue that possibly CXCR4/SDF-1α interaction was responsible for the mobilization of progenitor stem cells from BM to the damaged liver for its regeneration, but it was not ascertained by these studies. The above findings were strengthened by the results of in vitro migration studies, in which only Lin cells were attracted toward damaged liver tissue, out of which 90% of cells expressed CXCR4. SDF-1α dependent stress-induced recruitment of cord blood cells (CD34+CXCR4+) in injured mouse liver has been shown earlier when transplanted through intravenous route [23]. Based on our in vivo and in vitro studies, we concluded that stress-induced liver regeneration by endogenous progenitor stem cells from BM also occurs following similar mechanism. After ascertaining migration of LinCXCR4+ cells toward damaged liver tissue, we observed that a significant fraction of them was also expressed OSMRβ. Based on earlier findings [23, 32] and present results, we believe that the expression of CXCR4 in mouse progenitor stem cells may be sufficient for its recruitment in the damaged liver; however, for hepatic differentiation, the expression of OSMRβ is imperative. Thus, we conclude that LinCXCR4+OSMRβ+ BM cells are candidate hepatocyte progenitors in BM, which are involved in regeneration of damage liver.

Earlier investigators followed intravenous route for transplantation of BM cells either in irradiated or in CCl4-injected mice [9, [10], [11], [12], [13], [14], [15], [16]17]. This route of transplantation involves trapping of a significant number of cells in the liver. In studies pertaining to the migration of BM cells into damaged liver, we did not consider intravenous route as ideal for transplantation of hematopoietic cells in BM. To avoid cell trapping in any organ, the donor cells were directly transplanted in the BM. The results showed that GFP+ cells initially remain confined to the BM, later they appear in circulation, and they migrate to the liver tissue only after damage. The damaged liver tissue only showed the presence of GFP+ cells among other tested organs, indicating that the migration of these cells was specific and might be chemoattraction dependent. We not only showed that BM cells migrate to liver in response to damage, but also a significant number of them attained hepatic fate. We obtained a higher level of BM-derived albumin expressing cells in the damaged liver than other investigators [13, 33]. Moreover, the engrafted cells expressed TAT, TDO, and CK-18 genes. In this study we did not inject BM cells for the sake of liver regeneration. They spontaneously migrated to the damaged site in response to the physiological needs, which itself explains why the outcome of this study differs from others [13, 33]. We observed BM-derived albumin and CK-18 expressing cells even after 1 year of liver damage, which confirmed that these cells had been stably integrated in the recipient tissue. The results of RT-PCR experiments also indicate that reprogramming of BM cells into hepatocytes did not cause carcinogenic transformation, as hepatocarcinogenic markers AFP and GGT were not expressed in sorted GFP+Alb+ cells. Most interestingly, qPCR data showed that more than 95% of GFP+Alb+ cells were Y-chr. These results strongly favored direct differentiation of BM cells into hepatocytic cells, and if fusion is the other mechanism for such conversion it was significantly low in this model. Our results deviated from the earlier reports [14, 15, 34, 35], showing that cell fusion was the primary mechanism for attaining hepatic phenotype in BM-derived cells. However, one of the reports [14] described the possibility for the formation of normal donor male nonfusion karyotypes (XY, XXYY) in serially transplanted female recipients. The authors [14] interpreted the results by reductive division of tetraploid hybrids followed by normal polyploidization. Our qPCR data certainly demonstrated that donor-derived hepatocytes are predominantly of female origin. We believe that the extent and the nature of damage (acute vs. chronic) and the agent used to damage the liver have an important role on deciding whether BM cells will be fused with host hepatocytes or will be directly differentiated into hepatocytes. For example, BM differentiation program may not be activated unless liver is in regenerating mode (produces specific growth factors). In chronic liver injury model, hepatocytes and liver progenitor cells were shown to be in nonproliferating stage due to lack of critical growth factors produced by the normal liver mesenchyme [36, 37]. This probably explains why earlier investigators [34, 35] did not observe direct differentiation. The inductive microenvironment developed (manuscript submitted) in the CCl4-damaged model used by us and others [16] induced differentiation of BM cells into hepatocytes.

Overall, we conclude that liver regeneration after acute injury may not be completely autonomous, and the BM-derived cells might play an important role in such a regenerative process. A distinct characteristic of cells present in the BM are involved in liver regeneration. In response to acute liver damage, progenitor stem cells proliferate and mobilize to the PB and, finally, competent hepatic progenitor cells migrate to the damaged liver for participation in regeneration by direct differentiation process. The results of this investigation offer a conceptual basis for the repair of vital organs prone to damage in everyday functions, the liver being one such organ.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Authors are indebted to National Center for Biological Sciences, Bangalore, India, and the Jackson Laboratories for providing the GFP-expressing mice for this investigation. A part of the research was supported by grants from the Department of Biotechnology, Government of India.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
Supplementary_Fig1_1.JPG25KSupplemental Figure

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.