Contribution of mature hepatocytes to small hepatocyte-like progenitor cells in retrorsine-exposed rats with chimeric livers

Authors

  • Ya-Hui Chen,

    1. Graduate Institute of Clinical Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Department of Pediatrics, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Mei-Hwei Chang,

    1. Graduate Institute of Clinical Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Department of Pediatrics, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    3. Hepatitis Research Center, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Chin-Sung Chien,

    1. Graduate Institute of Clinical Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Department of Pediatrics, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Shang-Hsin Wu,

    1. Graduate Institute of Clinical Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Department of Pediatrics, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Chun-Hsien Yu,

    Corresponding author
    1. Department of Pediatrics, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Department of Pediatrics, Buddhist Tzu-Chi General Hospital, Taipei Branch, and Buddhist Tzu-Chi University College of Medicine, Hualien, Taiwan
    • Department of Pediatrics, Buddhist Tzu-Chi General Hospital, Taipei Branch, No. 289, Jianguo Rd., Xindian Dist., New Taipei City 23142, Taiwan
    Search for more papers by this author
  • Hui-Ling Chen

    Corresponding author
    1. Department of Pediatrics, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Hepatitis Research Center, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    • Hepatitis Research Center, National Taiwan University Hospital. No. 7, Chung-Shan S. Road, Taipei 100, Taiwan
    Search for more papers by this author
    • fax: +886-2-23825962


  • Potential conflict of interest: Nothing to report.

  • Supported by grants from the National Science Council of the Republic of China (NSC97-2321-B-002-021-MY2; NSC99-2314-B-002-041; NSC99-2628-B-002-069-MY3; NSC100-2314-B-303-008; TCRD-TPE-101-27).

Abstract

The potential lineage relationship between hepatic oval cells, small hepatocyte-like progenitor cells (SHPCs), and hepatocytes in liver regeneration is debated. To test whether mature hepatocytes can give rise to SHPCs, rats with dipeptidyl peptidase IV (DPPIV) chimeric livers, which harbored endogenous DPPIV-deficient hepatocytes and transplanted DPPIV-positive hepatocytes, were subjected to retrorsine treatment followed by partial hepatectomy (PH). DPPIV-positive hepatocytes comprised about half of the DPPIV chimeric liver mass. Tissues from DPPIV chimeric livers after retrorsine/PH treatment showed large numbers of SHPC clusters. None of the SHPC clusters were stained positive for DPPIV in any analyzed samples. Furthermore, serial sections stained for gamma-glutamyl-transpeptidase (GGT, a marker of fetal hepatoblasts) and glucose-6-phosphatase (G6Pase, a marker of mature hepatocytes) showed inverse expression of the two enzymes and a staining pattern consistent with a lineage that begins with GGT(+)/G6Pase(−) to GGT(−)/G6Pase(+) within a single SHPC cluster. Using double immunofluorescence staining for markers specific for hepatic oval cells and hepatocytes in serial sections, oval cell proliferations with CK-19(+)/laminin(+) and OV-6(+)/C/EBP-α(−) were shown to extend from periportal areas into the SPHC clusters, differentiating into hepatic lineage by progressive loss of CK-19/laminin expression and appearance of C/EBP-α expression towards the cluster side. Cells in the epithelial cell adhesion molecule (EpCAM(+)) SHPC clusters showed membranous EpCAM(+)/HNF-4α(+) (hepatocyte nuclear factor-4α) staining and were contiguous to the surrounding cytoplasmic EpCAM(+)/HNF-4α(−) ductular oval cells. Extensive elimination of oval cell response by repeated administration of 4,4′-methylenedianiline (DAPM) to retrorsine-exposed rats impaired the emergence of SHPC clusters. Conclusion: These findings highly suggest the hepatic oval cells but not mature hepatocytes as the origin of SHPC clusters in retrorsine-exposed rats. (HEPATOLOGY 2013)

The liver has an enormous capacity to regenerate through different cellular responses depending on the nature and severity of the injury.1-3 In situations associated with acute loss of hepatic mass, mature hepatocytes can rapidly proliferate to repair liver damage or replace surgical resection.1-5 When hepatocyte proliferation is inhibited or overwhelmed, liver injury leads to the activation of hepatic progenitor cells to replace the injured hepatic parenchyma.3-5 Hepatic oval cells and small hepatocyte-like progenitor cells (SHPCs) are two of hepatic progenitor cell populations that have been described in many different experimental injury models.3-10 Hepatic oval cells are the most well-identified progenitor cell population. They originate from the Canals of Hering.8, 9 When activated to proliferate, they migrate into the hepatic lobules and differentiate into hepatocytes or cholangiocytes.3-7 SHPCs emerge and proliferate to regenerate liver tissue in several treatments in rodents, including intraperitoneal injection of D-galactosamine, dipin plus partial hepatectomy (PH), or retrorsine followed by PH.6-8, 11-13 However, the cellular origin of SHPCs is debated.

It has been shown that SHPCs expressed phenotypic characteristics of fetal hepatoblasts, hepatic oval cells, and mature hepatocytes.10 By using 3D mapping techniques, continuous streams of hepatic oval cells can be tracked to enter deep into the nodules of SHPCs in the transgenic mouse model of chronic liver injury with retrorsine treatment.13 These data suggest that SHPCs may originate from hepatic oval cells and represent a transitional cell type between hepatic oval cells and mature hepatocytes.6, 13 However, bile duct destruction by 4,4′-diaminodiphenylmethane to inhibit oval cell proliferation did not block SHPC emergence,11 suggesting that SHPCs were not derived from hepatic oval cells and may represent an independent hepatic progenitor cell population.11 SHPCs have also been suggested to derive from mature hepatocytes that escape the mito-inhibitory effect of retrorsine.14, 15 These studies employed in vivo genetic labeling of rat hepatocytes using recombinant retroviruses harboring the β-galactosidase gene.14, 15 They observed that 4% of hepatocytes were labeled, then followed the fate of labeled cells after retrorsine/PH treatment and found that 4.6% of SHPC clusters contained the transgene, thus suggesting a precursor-product relationship between hepatocytes and SHPCs.14 However, the evidence is not conclusive. The labeling technique may have labeled putative stem cells or biliary cells, thereby not fully excluding those cell types as a source of SHPCs.15-17 In addition, 4% of hepatocytes should be escaping the retrorsine effect to be genetically labeled to explain the observed results instead of the initial 4% labeled hepatocytes in each rat.14, 17 Therefore, to definitively clarify the potential lineage relationship between hepatic oval cells, SHPCs, and mature hepatocytes will require additional studies using different genetic marking models or cell transplantation approaches.12, 16

In this study we aimed to determine the cell origin of SHPCs, particularly whether mature hepatocytes can give rise to SHPCs. We took advantage of the chimeric genetic status of hepatic parenchyma in repopulated dipeptidyl peptidase IV (DPPIV)-deficient rats. The DPPIV chimeric livers display stable repopulation by two kinds of hepatocytes, endogenous DPPIV-deficient hepatocytes and transplanted DPPIV-positive hepatocytes.18 Rats with DPPIV chimeric livers were subjected to retrorsine/PH treatment. If SHPCs were derived from mature hepatocytes, we expected to observe both DPPIV-positive and DPPIV-deficient SHPC clusters.

Abbreviation

C/EBP-α, CCAAT/enhancer binding protein alpha; CK-19, cytokeratin 19; CYP, cytochrome P450; DAPM, 4,4′-methylenedianiline; DPPIV, dipeptidyl peptidase IV; EpCAM, epithelial cell adhesion molecule; GGT, gamma-glutamyl-transpeptidase; G6Pase, glucose-6-phosphatase; HNF-4α, hepatocyte nuclear factor-4α; PH, partial hepatectomy; SHPC, small hepatocyte-like progenitor cells.

Materials and Methods

Animals.

DPPIV-deficient F344 rats were kindly provided by Professor Sanjeev Gupta from the Albert Einstein College of Medicine (New York, NY). Male DPPIV-deficient rats were used as recipient animals. Normal male DPPIV-positive F344 rats (age 8-10 weeks, 200-250 g) were purchased from the National Laboratory Animal Center, Taiwan, and used as donor animals. These animals were in-house bred and maintained on standard laboratory chow and daily 12-hour light/dark cycles. All of the animals received humane care in compliance with the guidelines of the National Science Council of Taiwan (NSC, 1997). All animal experiments were approved by the Institutional Laboratory Animal Care and Use Committee of the National Taiwan University, College of Medicine and College of Public Health.

Generation of Rats With DPPIV Chimeric Livers.

The rats with DPPIV chimeric livers were generated according to our previous studies.18 Male DPPIV-deficient rats received two treatments of retrorsine (30 mg/kg, intraperitoneal; Sigma, St. Louis, MO) 2 weeks apart, at 6 and 8 weeks of age. D-galactosamine (0.7 g/kg, intraperitoneal) was used to induce acute hepatic injury 2 weeks after the second retrorsine treatment. The rats then received DPPIV-positive hepatocyte transplantation (1 × 107/mL) intraportally 24 hours after D-galactosamine treatment. The rats were left to recover and were not subjected to any other experimental procedure for the next month.

Retrorsine Administration and PH.

Proliferation of SHPCs was initiated in the wildtype F-344 rats and the rats with DPPIV chimeric livers by the protocol of Gordon et al.10 The rats with DPPIV chimeric livers received an additional two intraperitoneal injections of retrorsine (30 mg/kg) at 14 and 16 weeks of age (4 and 6 weeks after hepatocyte transplantation), and then a PH at 21 weeks of age. Liver tissue removed at PH was saved for analysis. The rats were allowed to recover and were sacrificed at the indicated timepoints. Liver tissues were fixed in 4% formaldehyde and processed for paraffin-embedded sections. In addition, tissues were snap-frozen in liquid nitrogen or embedded into optimum cutting temperature (OCT) compound and stored at −80°C.

Administration of 4,4′-Methylenedianiline (DAPM).

The biliary toxin DAPM (Sigma) was dissolved in dimethylsulfoxide at a concentration of 50 mg/dL.11 Retrorsine-exposed rats received a single injection of DAPM (50 mg/kg, intraperitoneal) 24 hours before PH or repeated administrations 24 hours before each retrorsine and before PH.

Histochemistry and Immunohistochemistry.

All histochemical and immunohistochemical stainings were performed according to previously described protocols. To identify the DPPIV-positive hepatocytes in the DPPIV chimeric liver, DPPIV expression was determined by enzyme histochemistry in liver cryosections as described.19 Gamma-glutamyl-transpeptidase (GGT) was detected by the method of Rutenberg et al.21 and glucose-6-phosphatase (G6Pase) by the method of Dabeva et al.7, 19-21 Double immunofluorescence staining for cytokeratin 19 (CK-19) (Novocastra, Newcastle, UK) and laminin (Dako, Carpinteria, CA), OV6 (R&D Systems, Minneapolis, MN) and CCAAT/enhancer binding protein alpha (C/EBP-α) (Santa Cruz Biotechnology, Santa Cruz, CA), epithelial cell adhesion molecule (EpCAM) (US Biological, Swampscott, MA) and CK-19, CK-19 and C/EBP-α, or EpCAM and hepatocyte nuclear factor-4α (HNF-4α) (Santa Cruz Biotechnology) were detected using the method described by Paku et al.22 Appropriate secondary antibodies used in various experiments included Alexa Fluor 488 donkey anti-goat IgG (Molecular Probes, Eugene, OR) and Alexa Fluor 594 donkey anti-mouse IgG (Molecular Probes). The area occupied by DPPIV-positive hepatocytes was quantitated from 3-4 liver sections stained for DPPIV activity with ImageJ software (National Cancer Institute, Bethesda, MD).18

Statistical Analysis.

Data are presented as the mean ± standard error of the mean (SEM). The significance of differences was analyzed by t test using SPSS 9.0 (Chicago, IL). P < 0.05 was considered statistically significant.

Results

Histology of DPPIV Chimeric Livers Before Retrorsine/PH Treatment.

DPPIV chimeric livers were examined at 1 month after DPPIV-positive hepatocyte transplantation. Liver histology was essentially normal. DPPIV-positive hepatocytes were completely integrated into the hepatic plates, histologically identical to the surrounding hepatocytes (Fig. 1A). The DPPIV-positive colonies were distributed in various locations of the host liver lobules, with most found in the periportal and midlobular regions. Consistent with our previous study,18, 20 DPPIV-positive hepatocytes comprised about half of the chimeric liver mass, CK-19-positive oval cells were frequently observed, and DPPIV-positive biliary ductules were not found (Fig. 1B).

Figure 1.

Histology of DPPIV chimeric livers before (A,B) and after retrorsine/PH treatment (C,D) were analyzed using histochemical staining for DPPIV (red) and dual immunofluorescence stainings for DPPIV (green) / CK-19 (red). (A) Before retrorsine/PH, liver histology is essentially normal. DPPIV-positive hepatocytes are completely integrated into the hepatic plates, histologically identical to the surrounding hepatocytes, and comprise about half of the DPPIV chimeric liver mass. (B) CK-19(+) oval cell proliferations are observed. (C) After retrorsine treatment, both DPPIV-deficient and DPPIV-positive hepatocytes become hypertrophic at the time of PH. Liver histology exhibits mild inflammatory response. DPPIV-positive hepatocytes comprise 63.6 ± 4.9% of the DPPIV chimeric liver mass. (D) Remarkable CK-19(+) oval cell proliferations were observed (original magnification: 100×). Scale bars = 100 μm.

Cellular Response in the DPPIV Chimeric Livers After Retrorsine/PH Treatment.

Retrorsine/PH treatment in the rats with DPPIV chimeric livers caused high mortality in the first 2 days after PH. Twenty rats with DPPIV chimeric livers in total were generated in three separate experiments. Four rats survived and were sacrificed at 10 days after PH.

At the time of PH the liver histology exhibited a mild inflammatory response. Both DPPIV-deficient and DPPIV-positive hepatocytes were incapable of proliferation by retrorsine treatment and became hypertrophic. The mean proportion of DPPIV-positive hepatocytes was 63.6 ± 4.9% of the chimeric liver mass in the four surviving rats (Fig. 1C) and 40.7 ± 8.1% in the nonsurviving rats (t test, P = 0.042). Remarkable CK-19(+) oval cell proliferations were observed (Fig. 1D).

At 10 days after PH the overall histology examined by hematoxylin-eosin stain was similar to that described in the literature for the retrorsine/PH model.10 Numerous SHPC clusters were found. They were of different sizes and appeared in various locations of the liver lobules. Morphologically, these SHPC clusters were highly vacuolated. They were surrounded by large hepatocytes (megalocytes).

SHPC Clusters Are Not Derived From Mature Hepatocytes.

To determine the lineage relationship between mature hepatocytes and SHPC clusters, tissues from DPPIV chimeric livers after retrorsine/PH treatment were stained for DPPIV. None of the numerous SHPC clusters was stained positive for DPPIV in any analyzed samples from the four rats (Fig. 2A). In contrast, in retrorsine/PH treated wildtype F-344 rats, SHPC clusters were DPPIV-positive (Fig. 2B). Therefore, it is highly unlikely that DPPIV-deficient SHPC clusters observed in the DPPIV chimeric livers are derived from DPPIV-positive mature hepatocytes.

Figure 2.

Mature hepatocytes do not give rise to SHPCs. (A) Whole liver section of a DPPIV chimeric liver after retrorsine/PH treatment is stained for DPPIV. Numerous SHPC clusters of different sizes are found in various locations of the liver lobules. None of the numerous SHPCs clusters are stained positive for DPPIV. (A') A representative area of the liver section circumscribed by a box is shown enlarged to illustrate the DPPIV-deficient SHPC cluster. (B) In retrorsine/PH-treated wildtype F-344 rats, SHPC clusters are DPPIV-positive (original magnification: A, 50×; A', 200×; B, 200×). Scale bars = 100 μm.

To further characterize the cell origin of SHPC clusters, we used the histochemical staining for DPPIV, GGT (a marker of fetal hepatoblasts),7 and G6Pase (a marker of mature hepatocytes)7 in serial sections. In the sections examined, SHPC clusters were uniformly negative for DPPIV staining. In contrast, SHPC clusters were highly variable for GGT or G6Pase staining. In single-level sections, some clusters were nearly all positive for GGT or G6Pase, whereas others were negative. The staining pattern of GGT was not associated with cluster sizes. In serial sections, SHPC clusters showed an inverse expression of GGT and G6Pase, that is, GGT(+)/G6Pase(−) or GGT(−)/G6Pase(+) (Fig. 3A). In addition, a staining pattern consistent with a lineage that begins with nearly all GGT(+)/G6Pase(−) progressing to inhomogeneous GGT/G6Pase staining and finally to all GGT(−)/G6Pase(+) were often present within a single SHPC nodule (Fig. 3B). These data suggest that SHPCs were a heterogeneous population of cells that were at different stages of differentiation within the same nodule. Moreover, SHPC clusters were frequently surrounded by robust GGT-positive oval cell proliferations.

Figure 3.

SHPCs are a heterogeneous population of cells at different stages of differentiation in hepatic progenitor cell and hepatocyte lineage. Shown are serial liver sections stained histochemically for DPPIV (left column), GGT (a marker of fetal hepatoblasts, middle column), and G6Pase (a marker of mature hepatocytes, right column). (A1-A3) SHPC clusters are uniformly negative for DPPIV staining, but show inverse expression of GGT and G6Pase. (B1-B6) Staining pattern consistent with a lineage that begins with nearly all GGT(+)/G6Pase(−) progressing to GGT(−)/G6Pase(+) are often present within a single SHPC nodule (original magnification: 100×). Scale bars = 100 μm.

Relationship of Oval Cell Proliferations and SHPC Clusters.

To determine the relationship between oval cell proliferations and SHPC clusters, we first used double immunofluorescence stainings for CK-19/laminin and OV-6/C/EBP-α (an early gene of hepatocyte lineage)23 in serial sections. SHPC clusters were largely negative for CK-19, OV-6, and laminin staining, but expressed C/EBP-α. Oval cell proliferations stained positive for CK-19 and OV-6, always surrounded by laminin, indicating they were formed by hepatic oval cells. By analyzing serial sections, oval cell proliferations with strong staining for CK-19/laminin and OV-6(+)/C/EBP-α(−) were shown to extend from periportal areas (Fig. 4A1,A2). They became weaker for CK-19/laminin staining and, simultaneously, were observed to express faint C/EBP-α while approaching to surround SHPC clusters (Fig. 4A3,A4). Oval cell proliferations frequently entered deep into the SHPC clusters and ductular oval cells within the cluster were CK-19(+)/laminin(−) and OV-6(+)/C/EBP-α(+) (Fig. 4A4; Supporting Fig. 1).

Figure 4.

Relationship of oval cell proliferations to SHPC clusters. (A1-A4) Double immunofluorescence stainings for CK-19 (green) / laminin (red) (A1,A3) and OV6 (green) / C/EBP-α (red) (A2,A4) in serial liver sections show that oval cell proliferations are CK-19(+), surrounded by laminin (A1, arrows), and OV-6(+)/C/EBP-α(−) (A2, arrows). They extend from periportal areas into the SHPC cluster with progressively weaker staining for CK-19 and laminin (A3, arrows), and appearance of faint C/EBP-α expression in OV-6(+) ductular cells (A4, arrows). Intermediate transitional cells (long arrows) with CK-19(+)/laminin(−) and OV-6(+)/C/EBP-α(+) can be found in the SHPC cluster (A4). (A2′ and A4′) C/EBP-α staining in A2 and A4. Single-color and merged images of A1-A4 are also shown in Supporting Fig. 1. (B1-B3) Double immunofluorescence stainings for CK-19 (green) / EpCAM (red), CK-19 (green) / C/EBP-α (red), and HNF-4α (green) / EpCAM (red) in serial sections. EpCAM(+) SHPC clusters are in close contiguity with EpCAM(+)/CK-19(+) (yellow) oval cell proliferations (B1). Membranous EpCAM(+)/HNF-4α(+) cells are observed adjacent to the surrounding cytoplasmic EpCAM(+)/HNF-4α(−) ductular oval cells. Some membranous EpCAM(+)/HNF-4α(−) cells are identified between them (B3). (B3′) An enlarged representative area of the liver section circumscribed by a box in B3. DAPI was used for nuclear staining (blue). All of the data are shown as merged images. (C) Shown were serial liver sections stained histochemically for GGT and G6Pase in the retrorsine/PH model, retrorsine/DAPM/PH model, and DAPM/retrorsine/DAPM/retrorsine/DAPM/PH model at 2 weeks after PH, respectively. SHPC clusters were observed and showed inverse expression of GGT and G6Pase in the retrorsine/PH model and retrorsine/DAPM/PH model. However, SHPC clusters were not observed in DAPM/retrorsine/DAPM/retrorsine/DAPM/PH-treated liver (original magnification: A, 200×; B1-B3, 100×; B3′, 200×; C, 100×). Scale bars = 100 μm.

Next, we used double immunofluorescence stainings for EpCAM (hepatic progenitor cell marker)24, 25 / CK-19, CK-19/C/EBP-α, and EpCAM/HNF-4α (hepatocyte-specific marker)26 in serial sections. SHPC clusters were variable for EpCAM staining. However, they were in close contiguity with EpCAM(+)/CK-19(+) oval cell proliferations. SHPC clusters were largely EpCAM(−)/CK-19(−), C/EBP-α(+), and HNF-4α(+). EpCAM(+) SHPC clusters were also strongly GGT(+) (data not shown). Within the EpCAM(+) SHPC clusters, cells showed membranous EpCAM(+)/HNF-4α(+) staining and were adjacent to the surrounding cytoplasmic EpCAM(+)/HNF-4α(−) ductular oval cells. Membranous EpCAM(+)/HNF-4α(−) cells were identified between them (Fig. 4B). These findings together suggest that a lineage relationship was present within a single oval cell proliferation and between oval cell proliferations and SHPC clusters.

To corroborate these findings, we treated retrorsine-exposed rats with DAPM, a biliary toxin that specifically destroys the biliary epithelium and is used to eliminate oval cell proliferations.27 Considering that retrorsine itself causes a moderate increase in the oval cell population and single dose of DAPM injures about 70% of the biliary ducts,28, 29 we used two protocols by single injection of DAPM 24 hours before PH (retrorsine/DAPM/PH)11 or repeated administration of DAPM 24 hours before each retrorsine and before PH (DAPM/retrorsine/DAPM/retrorsine/DAPM/PH). Because rats with DPPIV chimeric livers did not survive the retrorsine/DAPM/PH treatment, only wildtype F344 rats were examined. As previously reported,11 SHPC clusters were observed in the retrorsine/DAPM/PH-treated livers and showed similar phenotypic characteristics to those in retrorsine/PH-treated livers (Fig. 4C). In contrast, we did not find the emergence of SHPC clusters at 2 weeks after PH in the DAPM/retrorsine/DAPM/retrorsine/DAPM/PH-treated livers when biliary restoration was still under way and selective periportal hepatocyte were weakly positive for GGT. The results provide further evidence that SHPCs are closely associated with oval cells and that mature hepatocytes do not contribute to the formation of SHPCs.

Discussion

In the current study we made use of the DPPIV chimeric liver model to examine the potential contribution of mature hepatocytes to formation of SHPCs in retrorsine-exposed rats. DPPIV-positive hepatocytes comprise about half of the chimeric liver mass. However, none of the numerous SHPC clusters in the regenerated livers were DPPIV-positive. Our study clearly shows that mature hepatocytes do not give rise to SHPCs.

Our results argue against the proposed hypothesis that mature hepatocytes escaping mito-inhibition of retrorsine give rise to SHPC clusters.10, 12, 13 In the DPPIV chimeric livers, DPPIV-positive hepatocytes were the predominant cell type and were treated with two doses of retrorsine, whereas DPPIV-deficient hepatocytes were treated with four doses. If the escaping-hepatocyte theory were true, we should have observed more DPPIV-positive SHPC clusters than DPPIV-deficient SHPC clusters. However, we found no DPPIV-positive SHPC clusters.

Our results contradict the finding of Braun and Sandgren12 that small hepatocytes were derived from hepatocytes in dipin/PH-treated mice. Dipin/PH induces hepatocarcinogenesis in the mouse.8 Small hepatocyte clusters in this model appeared at 8 weeks, persisted for months (at least 6 months in this study), and developed into liver tumors.8, 12 In the retrorsine/PH model, SHPC clusters appear at 3-5 days and differentiate into hepatocytes, restoring the normal liver structure by 30 days.10 We speculate that dipin/PH treatment might have induced preneoplastic changes of transplanted oval cells and/or hepatocytes. The small hepatocyte foci in the dipin/PH model may be preneoplastic rather than regenerative.12

Although Chen et al.30 demonstrated that mature hepatocytes underwent an epithelial mesenchymal transition (EMT) and gave rise to liver progenitor cells in culture, Chu et al.,31 using Alfp-Cre × Rosa26-YFP mice, showed that hepatocytes did not undergo EMT in vivo and Malato et al.,32 using the AAV8-Ttr-Cre model, demonstrated that hepatocytes were not precursors of liver progenitor cells in vivo. Furthermore, Bhave et al.33 showed that primary hepatocytes expressed reprogramming factors in culture. Their expression was up-regulated with time in culture. These studies together suggest that cellular morphology and physiology are prone to change in vitro. Cell conversion would ideally be tested in vivo.34 Therefore, a DPPIV chimeric liver model would be a powerful tool to clarify the lineage relationship between mature hepatocytes and SHPCs.

Nevertheless, a basic concern about the DPPIV chimeric liver model must be addressed as to whether transplanted hepatocytes are the same mature hepatocytes in normal livers and might have lost the ability of cell lineage conversion. Cell proliferation is thought to facilitate cell type conversions.34 Transplanted hepatocytes have undergone several doublings to generate DPPIV chimeric livers.20, 29 Therefore, transplanted hepatocytes have been thought to be more amenable to lineage conversion than resident hepatocytes.32 Michalopoulous et al.29 have used a similar DPPIV chimeric liver model to elegantly prove that transplanted hepatocytes transdifferentiate into biliary cells after bile duct ligation and toxic biliary injury. The DPPIV chimeric liver model has additional advantages to clarify the lineage relationship between mature hepatocytes and SHPCs in retrorsine-exposed rats. First, the biliary epithelium is uniformly DPPIV-negative. DPPIV-positive biliary ductules are not observed in the DPPIV chimeric livers in either the Michalopoulous et al. study or ours.20, 29 Although the very small biliary cell contaminants could be injected with the transplanted hepatocytes,35 they should have no proliferation advantage over transplanted hepatocytes,20, 36 or need in the livers treated with retrorsine/PH or retrorsine/D-galactosamine, which induce moderate oval cell proliferations.12, 20, 28 Second, the hepatocytes with different markings were treated with different doses of retrorsine.

Retrorsine is metabolized by hepatocytes through cytochrome P450 (CYP) enzymes to exert mito-inhibition.37 Periportal hepatocytes have lower CYP enzyme expression as compared to perivenous hepatocytes, and are less susceptible to retrorsine toxicity.37, 38 Transplanted DPPIV-positive hepatocytes engraft in the periportal areas, proliferate to distribute through all zonal areas of the liver lobule in the DPPIV chimeric liver, and acquire the position-specific enzyme expression depending on their lobular location.18, 38, 39 Transplanted cells locate in the periportal areas exhibit low CYP enzyme expression.40, 41 Therefore, the predominant transplanted DPPIV-positive hepatocytes in the DPPIV chimeric liver should have theoretically more probability to escape the retrorsine effect. However, we did not observe DPPIV-positive SHPC clusters. Our results strongly indicate that SHPC clusters are derived from the cell type without expressing CYP enzymes. Hepatic progenitor cells do not express CYP enzyme and are the most probable candidate.42, 43

Furthermore, a lineage relationship within a single SHPC nodule suggested by serial sections stained for GGT and G6Pase does not support the hepatocyte origin of SHPCs. GGT is an early gene and G6Pase is a late gene expressed specifically in the differentiation of hepatic progenitor cell into hepatocytes.5, 7, 39 Proliferating hepatocytes express G6Pase, but do not express GGT.20, 39 If hepatocytes are the source of SHPC clusters in this study, they must dedifferentiate into progenitor cells and then redifferentiate into SHPCs in order to explain the expression pattern of GGT and G6Pase in the SHPC nodules. This possibility is unlikely. If dedifferentiation/redifferentiation is the potential mechanism, both DPPIV-deficient and -positive hepatocytes should have at least equal probability to develop into SHPC clusters, and we should have found both DPPIV-negative and -positive SHPC clusters. In addition, the hepatocytes toward dedifferentiation may be limited because they had been treated with retrorsine to restrict their proliferation potential.

Using two different dual staining protocols for markers specific for hepatic oval cells and hepatocytes in serial sections, we identified the potential lineage relationship between the oval cell proliferations and SHPC clusters. We observed that ductular oval cells extended from portal area into the SHPC cluster, differentiating into hepatic lineage by progressive loss of expression of surrounding laminin and gain of expression of C/EBP-α. These results are consistent with the theory that laminin always surrounds the hepatic oval cells and helps maintenance of undifferentiated hepatic oval cells.22, 23 Loss of the surrounding laminin expression allows hepatic oval cells to differentiate into hepatocytes.22, 23 This observation is further supported by the fact that cells in the EpCAM(+) SHPC clusters showed membranous EpCAM(+)/HNF-4α(+) staining and were closely contiguous to the surrounding cytoplasmic EpCAM(+)/HNF-4α(−) ductular oval cells. These findings are in accordance with the concept that EpCAM(+) hepatocytes are newly derived from stem/progenitor cells.25 It has also been shown that transplanted EpCAM(+) oval cells proliferate to repopulate the recipient liver and do not express EpCAM 2 weeks after transplantation.28 Thus, EpCAM(−) SHPC clusters may represent a more mature stage. Our results together highly suggest a hepatic oval cell origin of SHPC clusters.

High mortality in the drastic experimental conditions limited the collection of more rat numbers. However, the use of four rats do not decrease the significance of the results in the present study. DPPIV-positive hepatocytes comprised 63.6 ± 4.9% of the DPPIV chimeric liver mass. Emergence of each SHPC cluster is independent in the liver lobes. Each surviving rat with DPPIV chimeric livers is also independent. Given the probability of 0.4 for a DPPIV-negative SHPC cluster to be derived from DPPIV-negative hepatocytes, the probability for all of the numerous SPHC clusters to be DPPIV-negative in one rat is the nth power of 0.4, and for four consecutive rats the 4th power of the nth power of 0.4. The probability is extremely low and statistically significant to reject the hypothesis of a hepatocyte origin of SHPCs. Furthermore, we found that surviving rats had significantly higher repopulation percentages by DPPIV-positive hepatocytes as compared to nonsurviving rats. It is reasonable to speculate that the latter may not bias the results in this study.

We appreciate that the DPPIV chimeric liver model cannot directly exclude that potentially DPPIV-negative host hepatocyte-derived SHPCs were overlooked. However, we identified that extensive elimination of oval cell response by repeated administration of DAPM to retrorsine-exposed rats impaired the emergence of SHPC clusters. The findings corroborate that SHPCs are closely associated with oval cells and mature hepatocytes do not contribute to the formation of SHPCs.

Based on this study, we conclude that it would be almost impossible for SHPC clusters to be derived from mature hepatocytes in retrorsine-exposed rats. Instead, our results are in favor of the notion of hepatic oval cell origin of SHPCs. However, this lineage relationship of hepatic oval cells with SHPCs needs further studies using the specific genetic labeling of hepatic oval cells.

Acknowledgements

The authors thank Ms. Yi-Tian Ho for excellent technical assistance.

Ancillary

Advertisement