SEARCH

SEARCH BY CITATION

Keywords:

  • Human;
  • Hepatocyte;
  • Stem cell;
  • Biliary epithelial cell;
  • Liver failure

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The liver in subacute hepatic failure may become enriched for hepatic progenitor cells. Liver tissue from such a patient was collagenase digested and, from the nonparenchymal cell fraction, epithelioid colonies were developed. Albumin and alpha-1-antitrypsin (AAT) were secreted for greater than 120 days from these colonies. Reverse transcription-polymerase chain reaction showed expression of markers of both hepatocyte and biliary epithelial phenotypes (cytokeratins 7, 18, and 19, albumin and AAT, hepatocyte growth factor receptor, transforming growth factor beta receptor type II, gamma-glutamyl transpeptidase, biliary glycoprotein). The cell cycle regulator p21 was also expressed. The POU domain transcription factor octamer-binding protein 4 was present in these cells, but not in RNA or cDNA prepared from adult human liver. These markers were maintained even after 165 days culture. Proliferating epithelial-like cells with combined hepatocyte- and biliary-epithelial-specific functional markers and a stem cell marker can be isolated from the nonparenchymal fraction of liver cells in subacute hepatic failure.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Concepts of the regenerative potential of adult liver recently have been strikingly revised. While proliferation of normally quiescent mature adult hepatocytes occurs readily in response to moderate inflammatory insults or after surgical resection of part of the liver, many lines of evidence demonstrate additionally the potential for progenitor cells to proliferate and differentiate into mature hepatocytes in the adult [14]. The canals of Hering contain bipotential liver cell progenitors that can proliferate in response to severe liver injury along hepatocyte or biliary epithelial differentiation pathways, particularly if the normal proliferative potential of hepatocytes is blocked prior to a regenerative stimulus [5, 6]. After bone marrow transplantation, hepatocytes and biliary cells of donor marrow origin are detectable within the liver in animals and man [79]. After liver transplantation between non-gender-identical humans, hepatocytes that must have derived from extrahepatic sources are recognized [6, 10].

Identifying and characterizing liver cell progenitors offer major therapeutic opportunities in fields as diverse as gene therapy, initiation of liver repair, and biological liver support systems. We hypothesized that, if acute liver damage with loss of functioning hepatocyte mass occurs but liver regeneration is delayed, within such livers, there will be an enrichment of hepatocyte precursors. Those precursors might derive from cells resident within the liver or from cells recruited from other sites, particularly the bone marrow. Such livers should, therefore, be a source for culturing cells with the capacity to proliferate and express hepatocyte function. In this study, we characterize a proliferating cell line, expressing hepatocytic and biliary characteristics and a stem cell-associated marker, from the liver of a patient who underwent orthotopic liver transplantation after 6 weeks of subacute liver failure.

Observations were made on the liver explanted at orthotopic transplantation, performed 7 days after a 52-year-old Caucasian woman was admitted with a 7-week history of malaise, jaundice, and progressive coagulopathy. The patient gave specific consent for the use of her explanted liver for research purposes under a protocol approved by the local research ethics committee.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Patient Details

On admission, the patient's serum bilirubin was 495 μmol/l (normal is <17), alkaline phosphatase was 172 iU/l (normal is <130), and aspartate aminotransferase was 95 iU/l. Her prothrombin time was 100 seconds (normal is <16 seconds). Hepatic imaging showed a small liver with hypodense areas, suggestive of patchy necrosis, and ascites. Transjugular hepatic biopsy demonstrated features of acute hepatic necrosis without diagnostic features. After virological testing, a diagnosis of non-A-E subfulminant hepatic failure was made. Conventional supportive therapy for acute liver failure was commenced and, to encourage liver regeneration and reduce sepsis, 10 μg/kg of G-CSF were administered on day -2 and day -1 prior to transplantation, leading to a peak leukocyte count of 22 × 109/l (normal is <10) on day -1.

Cell Isolation

The explanted liver was collagenase digested. Approximately 150 g of liver tissue were cannulated via the portal vein in three places. Blood was removed with Hank's balanced salt solution (HBSS)-10 mM HEPES, pH 7.2, followed by a flush with HBSS-10 mM HEPES-0.5 mM EGTA, pH 7.2. The tissue was warmed to 37°C by perfusion for 10 minutes in HBSS-HEPES at a flow rate achieving a pressure of not more than 40 cm H2O. Collagenase P (Roche Diagnostics Ltd.; Lewes, UK; http://www.rocheuk.com) was then used at 0.05% (1.25 U/ml) for 15 minutes at 37°C. Thereafter, the capsule was gently disrupted and cells were released from the perfused areas. Centrifugation at 50 g for 2 minutes removed most of the hepatocytes. The supernatant was centrifuged at 400 g for 10 minutes. The pellet, containing mostly nonparenchymal cells, was resuspended in 20 ml of HBSS-HEPES containing 100 units of DNAse and filtered though a 15-μm nylon mesh to remove mature hepatocytes. This suspension was layered onto Lymphoprep (Nycomed; Oslo, Norway; http://www.nycomed.no) and centrifuged (400 g, 25 minutes, room temperature) to remove red cells, cell debris, and any remaining hepatocytes. Cells at the interface were collected and washed at 400 g for 10 minutes, followed by a further wash at 200 g for 10 minutes. The pellet was resuspended in minimal essential medium alpha (MEM-α) and plated on plastic at 250,000 cells per well in 500 μl of medium in a 24-well plate in a 37°C humidified incubator with 5% CO2, 95% air.

Cell Culture

MEM-α supplemented with antibiotics, 10% fetal bovine serum, hepatocyte growth factor (HGF) (20 ng/ml), epidermal growth factor (EGF) (10 ng/ml), glucose (25 mM), thyrotropin-releasing hormone (1 μm), hydrocortisone (1 μm), insulin (10 μg/ml), linoleic acid albumin (0.05 mg/ml), selenite (0.1 μm), ferrous sulphate (0.5 mg/l), zinc sulphate (0.75 mg/l), and nicotinamide (10 mM) was used. The medium was changed twice weekly. Conditioned medium was collected, centrifuged, and stored at −20°C for enzyme-linked immunosorbent assay (ELISA) for albumin and alpha-1-antritrypsin (AAT).

ELISA

Sandwich ELISAs containing one primary antibody and one secondary horseradish-peroxidase-conjugated antibody to human albumin or human AAT (10 mg/l; DAKO Ltd.; Carpenteria, CA; http://us.dakocytomation.com) were used to determine secreted protein levels in conditioned medium; the detection limit for both assays was 25 ng/ml. These assays were set up in the laboratory and are used routinely [11, 12].

Reverse Transcription-Polymerase Chain Reaction (RT-PCR)

cDNA was prepared directly from cells using the cells-to-cDNA protocol supplied by Ambion USA (Austin, TX; http://www.ambion.com). In brief, cells were lysed, DNAse treated, and subjected to RT prior to conventional PCR amplification using a hot-start protocol. The following primer pairs were used, each crossing exons to avoid false positives resulting from amplification of genomic DNA: β-actin forward 5′ TGGCACCACACCTTCTACAATGAGC and reverse 5′ GCACAGCTTCTCCTTAATGTCACGC; albumin forward 5′ CCTTTGGCACAATGAAGTGGGTA ACC and reverse 5′ CAGCAGTCAGCCATTTCACCAT AGG; AAT forward 5′ AGACCCTTTGAAGTCAAGGAC ACCG and reverse 5′ CCATTGCTGAAGACCTTAGTGAT GC; HGF receptor forward 5′ AGAAATTCATCAGGCTG TGAAGCGCG and reverse 5′ TTCCTCCGATCGCACAC ATTTGTCG; transforming growth factor beta (TGF-β) receptor II forward 5′ TCCAGCTCATCTAGATGAGGA GCTC and reverse 5′ GTCCCATGGCCTAAATGCCTCT CAG; cytokeratin 8 (CK-8) forward 5′ AAGGGCTGAC CGACGAGATC and reverse 5′ GCTTCCTGTAGGTGGCG ATC; CK-18 forward 5′ TACAAGCCCAGATTGCCAGC and reverse 5′ AGTCCTCGCCATCTTCCAGC; CK-7 forward 5′ GCAGGATGTGGTGGAGGACTTC and reverse 5′ TGGCACGCTGGTTCTTGATG; CK-19 forward 5′ GGAC CTGCGGGACAAGATTC and reverse 5′ CCTCGGACCTG CTCATCTGG; Oct-4 forward 5′ CGRGAAGCTGGAGAA GGAGAAGCTG and reverse 5′ CAAGGGCCGCAGCTT ACACATGTTC; Rex-1 forward 5′ GCGTACGCAAATTA AAGTCCAGA and reverse 5′ CAGCATCCTAAACAGCT CGCAGAAT; p21 forward 5′ CTAGTTCTACCTCAGGC AGCTCAAG and reverse 5′ TCCAGGCCAGTATGTTAC AGGAGCT; gamma glutamyl transpeptidase (GGT) forward 5′ GACGACTTCAGCTCTCCCAG and reverse 5′ CTTGT CCCTGGATTGCTTGT; biliary glycoprotein forward 5′ ATGGAACATTCCAGCAAAGC and reverse 5′ GGAGTG GTCCTGAGTGTGGT; cytochrome P450 1B1 (CYP 1B1) forward 5′ GAGAACGTACCGGCCACTATCACT-3′ and reverse 5′ GTTAGGCCACTTCAGTGGGTCATGAT-3′.

PCR conditions for β-actin, albumin, AAT, and TGF-β receptor were: 95°C for 15 minutes, 94°C for 30 seconds, 68°C for 30 seconds, 72°C for 1 minute, 40 cycles, 72°C for 10 minutes. PCR conditions for c-met (HGF receptor) and Oct-4 were as above but annealing temperature was 58°C. PCR conditions for cytokeratins were as above except annealing temperature was 50°C. PCR conditions for p21 were 95°C for 15 minutes, 95°C for 30 seconds, 62°C for 45 seconds, 72°C for 45 seconds, 40 cycles, 72°C for 10 minutes, 15-μl reaction. PCR conditions for Rex-1 were 95°C for 15 minutes, 94°C for 30 seconds, 56°C for 30 seconds, 72°C for 1 minute, 72°C for 10 minutes, 30 cycles, 15-μl volume [13]. PCR conditions for GGT and biliary glycoprotein were 95°C for 15 minutes, 95°C for 30 seconds, 62°C for 30 seconds, 72°C for 2 minutes, 72°C for 10 minutes, 40 cycles, 20-μl reaction. Hot-start master mix (Qiagen Ltd.; West Sussex, UK; http://www.qiagen.com) was used. Products were separated on 1% agarose gels with ethidium bromide.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

At macroscopic examination, the explanted liver was severely shrunken with a wrinkled capsular surface and weighed 514 g. The cut surface showed small yellow nodular areas alternating with dark tan areas. Histology showed extensive multiacinar and panacinar liver cell loss with parenchymal collapse and periportal ductular proliferation (Fig. 1). Areas of ductular proliferation also showed an intervening lymphoid infiltrate composed of many small and scanty large lymphoid cells. The yellow nodular areas described macroscopically corresponded to nodules of morphologically viable hepatocytes.

thumbnail image

Figure Figure 1.. A) Histology of explanted liver shows multiacinar and panacinar liver cell loss and parenchymal collapse (arrowhead) with periportal ductular proliferation (arrow).Hematoxylin & eosin (H&E) staining, 100× magnification. B) Higher magnification of A. An area of ductular reaction (arrow) with an intervening lymphoid infiltrate composed of many small and some large lymphoid cells. H&E staining, 200× magnification.

Download figure to PowerPoint

The liver after collagenase perfusion yielded a mixed population of cells, but no normal viable hepatocytes were evident by light microscopy and trypan blue exclusion or by phase contrast microscopy. Most cells from the semipurified fraction were less than 10 μm in diameter. Contaminating red cells did not attach and were removed at the first medium change. There was no evidence of mature hepatocytes in culture by phase contrast microscopy. Fibroblast colonies, as expected, emerged during the first 5 days of culture. By 16 days in culture, colonies with an epithelioid appearance were evident. These had a cuboidal morphology (Fig. 2A and 2B, days 16 and 17) with a granular cytoplasm and were proliferating. Removal of half the cells led to proliferation to fill the space. Over time, cells initially increased in size (Fig. 2C, day 32) and exhibited striking cytoskeletal structures (Fig. 2D). As disrupted cell monolayers became confluent, cell size decreased and granularity increased (Fig. 2D and 2E). By day 52, some cells had a large singular, rather agranular, appearance with prominent nucleoli (Fig. 2F, day 52). In contrast, the bulk of the cells remained small and cuboidal (Fig. 2E, day 52).

thumbnail image

Figure Figure 2.. Epithelioid colonies of cells.A) day 16, 10× magnification; B) day 17, 20× magnification; C) day 32, 20× magnification; D) day 40, 40× magnification; note the greater granularity in the smaller cells and cytoskeletal structures in large cells; E) day 50, 20× magnification, small granular; F) day 52, 20× magnification, some highly enlarged cells. Magnifications are original microscope magnification.

Download figure to PowerPoint

Conditioned medium (3- and 4-day collections) collected over 123 days in culture contained the liver-specific secreted proteins, albumin and AAT (Fig. 3). In wells with only fibroblastoid colonies, no such protein secretion was demonstrable in conditioned medium.

thumbnail image

Figure Figure 3.. A) Albumin secretion into culture medium over 86 days in culture.Results are expressed as mean ± standard deviation (SD). B) AAT secretion into culture medium up to day 123 in culture. Results are expressed as mean ± SD.

Download figure to PowerPoint

PCR amplification of cDNA prepared from isolated RNA from cells at day 89 showed markers of hepatocyte and biliary epithelial phenotypes with respect to cytokeratin expression. CK-7, CK-18, and CK-19 were expressed (Fig. 4) at the expected fragment size. CK-8 was not apparent, although it was detected in control human liver. The presence of the hepatocyte-specific genes AAT and albumin was confirmed by PCR (Fig. 4). These cells also expressed β-actin, the HGF receptor c-met, and TGF-β receptor type II. Expression of two genes associated specifically with biliary epithelial cells was also noted, GGT and biliary glycoprotein (Fig. 4). The POU transcription factor octamer-binding protein 4 (Oct-4) was also present in these cells, but not in RNA or cDNA prepared from adult human liver. The cell cycle regulator p21 was also expressed (Fig. 4). These markers were still expressed after cells had been cultured for 165 days.

thumbnail image

Figure Figure 4.. mRNA expression by RT-PCR of cuboidal epithelial colony harvested for RT-PCR on two occasions, day 89 and day 165.A) Cells express albumin (suggestive of a hepatocyte phenotype), β-actin, TGF-β receptor, and HGF receptor. B) Left: cells expressed the stem cell marker Oct-4 at both days 89 and 165; center: cells expressed biliary markers, e.g., biliprotein and GGT, and the cell cycle regulator p21; right: cells expressed CYP 1B1 and AAT, the latter a hepatocyte marker. C) Cells expressed cytokeratins suggestive of hepatocytes (CK-18) and biliary cells (CK-7 and CK-19). M = hyperladder 4 markers, bp sizes as marked.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Isolating and characterizing hepatic progenitor cells are major aims of current liver cell biology. Studies have been undertaken mostly in rodents [7, 9, 1417], but in humans, bone marrow [8, 1820], amniotic tissue [21], cord blood [22], and adult and fetal liver have been studied.

The adult liver has been shown to be the site of apparent stem cells that have differentiated in culture to express a biliary epithelial cell (BEC) phenotype. Crosby et al. described two populations of cells isolated from normal and diseased liver, which were either CD34+ (suggestive of hemopoietic origin) or c-kit+ cells, both of which developed in culture to the BEC phenotype [23]. Malhi et al. isolated hepatoblasts from human fetal liver in the second trimester that expressed biliary (CK-19) and hepatocyte (glycogen) markers at initial isolation, which were maintained in 5%-10% cultured cells over several population doublings; moreover, some other biliary markers (GGT and dipeptidylpeptidase IV) and hepatocyte markers (albumin) were expressed in the longer term, albeit at lower levels than observed initially [24]. Human embryonic stem (ES) cells have also been shown to differentiate into hepatocytes in vitro [25, 26], as have mouse ES cells [14, 25, 27, 28].

Both animal and human studies indicate that progenitor cells in the liver become more prominent when the parenchymal cells of the liver are unable to proliferate to repair damage and/or provide functional mass. The livers of patients with subfulminant hepatic failure—clinically defined by acute liver failure with delayed repair and once referred to as ‘regeneration-deficient’ liver failure—should be enriched for such progenitors. We isolated colonies of cells expressing hepatocyte, biliary, and stem cell markers from such a patient. At a functional level, secreted albumin and AAT were readily measurable (by ELISA) at levels of the same order of magnitude as those produced by primary cultures of human hepatocytes [29]. Moreover, colonies proliferated in culture, unlike primary hepatocytes, and also expressed Oct-4, suggestive of stem cells. Although not phenotypically like primary human hepatocytes, in culture they are cuboidal with obvious cytoskeletal structure and a dense cytoplasm with prominent nuclei. Clearly, under the circumstances of this observation, made on explanted liver in a patient listed for urgent transplantation, we cannot identify whether these cells entered the liver from a source such as the bone marrow or developed from an intrahepatic population such as that identified in the canal of Hering. Evidence of bone-marrow-derived hepatocytes in vivo is still controversial, with the recent evidence from Wang et al. [30] and Vassilopoulos et al. [31] that cell fusion between transplanted bone marrow cells and resident liver cells gives rise to hepatocytes thought initially to be differentiated directly from the transplanted bone marrow population.

Verfaillie's group described human multipotent adult progenitor cells (hMAPCs) isolated from healthy bone marrow [32] that can express hepatocyte markers when grown, for example, on matrigel to provide an extracellular matrix [20]. The cells described in this study appear to differ from those. Specifically, hMAPCs, when differentiated, undergo cell cycle arrest and express only one cell phenotype. The cells in our study, in contrast, expressed hepatocyte markers at the mRNA and protein levels while maintaining proliferation. These cells expressed GGT and biliary glycoprotein at the mRNA level. As this was identified using RT-PCR, data for coexpression at the cellular level are not available. Colonies of these cells in our study were visible only after more than 10–12 days in culture, suggesting that they were the result of clonal expansion of cells.

The cells in this study also differ from those described by Crosby et al., which differentiated only into cells of the biliary epithelial or endothelial phenotype, as evidenced by morphology and CK-19 or CD31 positivity, but were negative for albumin [23]. The cells described by Malhi et al., derived from human fetal liver, although maintaining significant proliferative potential, were not apparently responsive to HGF [24]. These characteristics differ from our isolated cells that expressed the HGF receptor c-met and continued to express secreted liver-specific proteins (e.g., AAT, albumin) at similar levels throughout the culture period.

The late appearance of colonies of cells reported here is in agreement with other descriptions of putative progenitor cells. Studies identifying potential stem cells with hepatic markers derived from bone marrow described by Fiegel et al. showed albumin and CK-19 expression by RT-PCR only after culture in a collagen matrix for 35 days [18], and Jiang et al. also noted the differentiated phenotype after many weeks in culture [32].

The stem cell transcription factor Oct-4 in the mouse is restricted to the embryonic carcinoma and ES cells and becomes downregulated on subsequent differentiation [33]. We used this as a putative stem cell marker and noted its expression both early and late in the culture period and demonstrated its coexpression with markers of fully differentiated cells. Assady et al. showed, in human ES cells differentiated to produce insulin, that the high Oct-4 expression seen in the undifferentiated cells diminished over 3 weeks after the stimulus to differentiation was given but was still identifiable at the time of insulin production [34]. Henderson et al., comparing mouse and human ES cell characteristics, found that Oct-4 expression was less tightly regulated in primates than in the mouse [13], which supports our observation that this stem cell transcription factor is expressed while the cells are also expressing markers of fully differentiated cells. Interestingly, Oct-4 expression was not apparently associated with Rex-1. The Rex-1 gene encodes a developmentally regulated acidic zinc finger protein, regulating transcription [3537]. The Rex-1 promoter contains an octamer motif (ATTTCGAT at −220) that is the binding site for transcription factors of the POU domain family, including Oct-4, which can activate or repress Rex-1 depending upon the level of expression. The lack of Rex-1 expression in these cells may indicate that the levels of Oct-4 were sufficient to repress rather than activate Rex-1 [33].

While bipotential hepatic precursor cells (clonogenic hepatoblasts) have been clearly recognized in the liver using fetal tissue prepared from rodents [38] and man [24] and transplants of bone marrow cells have been shown to give rise to both biliary epithelial cells and hepatocytes [5, 6, 10], although, as discussed, these latter observations may reflect cell fusion [5, 6], we are not aware of any other studies using adult human liver that have isolated proliferating colonies expressing markers of both biliary epithelial and hepatocytes in vitro. Whether these cells derived from a progenitor resident in the liver or entered the liver after the onset of liver failure, possibly stimulated by the administration in this patient of G-CSF, remains unclear. It is also not possible to know whether there were particular circumstances (e.g., the presence of an unknown hepatotrophic virus) that led to the emergence of this cell population in this patient. However, this paper clearly indicates the value of explanted liver tissue in subacute liver failure as a source of hepatocyte progenitors for study and exploitation. That these cells, present in the patient during a prolonged episode of subacute liver failure, were not sufficient to ‘rescue’ the patient before a transplant was required would indicate that they are not competent to provide significant levels of function in the setting of the toxic milieu of acute liver failure. However, out of the toxic environment, it may be possible to expand such a population for therapeutic purposes.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We are grateful to J. Draper and P. Andrews of the Department of Biomedical Science, University of Sheffield, for the kind gift of PCR primers to human Rex-1 (accession number AF450454). This work was funded by a grant from The Liver Group Charity.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References