Hepatocyte transplantation activates hepatic stellate cells with beneficial modulation of cell engraftment in the rat


  • Daniel Benten,

    1. Departments of Medicine and Pathology, Marion Bessin Liver Research Center, Comprehensive Cancer Research Center and General Clinical Research Center, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Vinay Kumaran,

    1. Departments of Medicine and Pathology, Marion Bessin Liver Research Center, Comprehensive Cancer Research Center and General Clinical Research Center, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Brigid Joseph,

    1. Departments of Medicine and Pathology, Marion Bessin Liver Research Center, Comprehensive Cancer Research Center and General Clinical Research Center, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Jörn Schattenberg,

    1. Departments of Medicine and Pathology, Marion Bessin Liver Research Center, Comprehensive Cancer Research Center and General Clinical Research Center, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Yury Popov,

    1. Department of Medicine I, University of Erlangen-Nuernberg, Germany
    Search for more papers by this author
  • Detlef Schuppan,

    1. Department of Medicine I, University of Erlangen-Nuernberg, Germany
    2. Division of Gastroenterology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA
    Search for more papers by this author
  • Sanjeev Gupta

    Corresponding author
    1. Departments of Medicine and Pathology, Marion Bessin Liver Research Center, Comprehensive Cancer Research Center and General Clinical Research Center, Albert Einstein College of Medicine, Bronx, NY
    • Albert Einstein College of Medicine, Ullmann 625, 1300 Morris Park Avenue, Bronx, NY 10461
    Search for more papers by this author
    • fax: 718-430-8975

  • Potential conflict of interest: Nothing to report.


We investigated whether transplanted hepatocytes interact with hepatic stellate cells, as cell–cell interactions could modulate their engraftment in the liver. We transplanted Fischer 344 rat hepatocytes into syngeneic dipeptidyl peptidase IV–deficient rats. Activation of hepatic stellate cells was analyzed by changes in gene expression, including desmin and α-smooth muscle actin, matrix proteases and their inhibitors, growth factors, and other stellate cell-associated genes with histological methods or polymerase chain reaction. Furthermore, the potential role of hepatic ischemia, Kupffer cells, and cytokine release in hepatic stellate cell activation was investigated. Hepatocyte transplantation activated desmin-positive hepatic stellate cells, as well as Kupffer cells, including in proximity with transplanted cells. Inhibition of Kupffer cells by gadolinium chloride, blockade of tumor necrosis factor alpha (TNF-α) activity with etanercept or attenuation of liver ischemia with nitroglycerin did not decrease this hepatic stellate cell perturbation. After cell transplantation, soluble signals capable of activating hepatic stellate cells were rapidly induced, along with early upregulated expression of matrix metalloproteinases-2, -3, -9, -13, -14, and their inhibitors. Moreover, prior depletion of activated hepatic stellate cells with gliotoxin decreased transplanted cell engraftment. In conclusion, cell transplantation activated hepatic stellate cells, which, in turn, contributed to transplanted cell engraftment in the liver. Manipulation of hepatic stellate cells might provide new strategies to improve liver repopulation after enhanced transplanted cell engraftment. Supplementary material for this article can be found on the HEPATOLOGYwebsite (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). (HEPATOLOGY 2005;42:1072–1081.)

Transplanted hepatocytes engraft, proliferate, and function in the liver of healthy or diseased animals. Liver repopulation with cells has potential in treating liver failure and genetic disorders. However, most transplanted cells (70%–80%) are cleared from liver sinusoids within 24 to 48 hours.1 In contrast, superior cell engraftment accelerates liver repopulation. Therefore, analysis of mechanisms promoting transplanted cell engraftment remains appropriate.

Cell engraftment in the liver involves sinusoidal endothelial disruption, entry of transplanted cells in liver plates, reconstitution of plasma membrane structures with restoration of cell polarity, and integration of cells in the liver parenchyma.1 Deleterious events interfering with cell engraftment have been identified, although these can be modulated; for example, hepatic vasodilators ameliorate hepatic ischemia, and the Kupffer cell response can be blunted.2–4 Similarly, pharmacological manipulation of cell–cell interactions, such as disruption of the sinusoidal endothelium, can improve transplanted cell engraftment.5

Insights into the role of additional liver cell types could be helpful in cell engraftment. The hepatic stellate cell (HSC) represents an attractive candidate for this consideration.6–8 HSC are located in the space of Disse, which transplanted cells must circumvent during translocation from liver sinusoids to the parenchyma.1 HSC regulate multiple physiological processes in the liver, including extracellular matrix (ECM) deposition during the onset, as well as regression, of hepatic fibrosis.7 Paracrine signals from Kupffer cells, liver sinusoidal endothelial cells (LSEC), and hepatocytes can activate HSC.6 Activation of HSC causes multiple early events, including gene regulation and release of soluble factors.6–8 Other changes include HSC proliferation and increased desmin or α-smooth muscle actin (SMA) expression, when HSC undergo myofibroblast-like change.

Previously, we established the efficacy of morphological methods for identifying transplanted hepatocytes in dipeptidyl peptidase IV deficient (DPPIV−) rats, which offer convenient assays of cell engraftment and proliferation.1–5 Using DPPIV− rats, we demonstrate here that hepatocyte transplantation partially activated HSC, including expression of growth factors and ECM-modifying matrix metalloproteinases (MMPs), which are largely produced in HSC.7, 8 Moreover, depletion of HSC interfered with cell engraftment.


HSC, hepatic stellate cells; ECM, extracellular matrix; LSEC, liver sinusoidal endothelial cells; SMA, smooth muscle actin; DPPIV, dipeptidyl peptidase IV; MMP, matrix metalloproteinases; ATPase, adenosine triphosphatase; GGT, gamma glutamyltransferase; RT-PCR, reverse transcription polymerase chain reaction; TNF-α, tumor necrosis factor alpha; NTG, nitroglycerin; HGF, hepatocyte growth factor; bFGF, basic fibroblast growth factor; VEGF, vascular endothelial growth factor; TGF-β, transforming growth factor beta; uPA, urokinase-type plasminogen activator; TIMP, tissue inhibitor of metalloproteinase.

Materials and Methods


The Special Animal Core of the Marion Bessin Liver Research Center provided 6- to 10-week old DPPIV− rats weighing 160 to 250 g. Normal syngeneic F344 rats were from National Cancer Institute (Bethesda, MD). The institutional Animal Care and Use Committee approved animal use in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals (United States Public Health Service publication, revised 1996).

Cell Isolation and Transplantation.

The liver of donor F344 rats was digested by collagenase (Worthington Biochem. Co., Lakewood, NJ) perfusion.9 Cell viability was determined by trypan blue dye exclusion and cell attachment to culture dishes. Cells were used only if they had viability >80%. For transplantation, freshly isolated hepatocytes were washed thoroughly to remove collagenase and suspended without culture in serum-free RPMI 1640 medium. Rats anesthetized with ether received a left subcostal incision to isolate the spleen and 2 × 107 or fewer hepatocytes, as indicated, in 0.5 mL were injected into the splenic pulp over a few seconds. Rats with liver injury received 1 × 107 cells. Control rats received plain RPMI medium alone, without cells.

Additional Treatment of Animals.

The experimental design is shown in Fig. 1 and described further in Results. Various drugs and their uses are provided in supplementary material (Available at the HEPATOLOGY website (http://www.interscience.wiley.com/jpages/0270-9139/suppmat/index.html).

Figure 1.

Experimental strategy. See Materials and Methods and Results for interpretation. All studies were repeated at least twice to assure reproducibility.

Histochemistry and Immunohistochemistry.

Hepatic DPPIV, adenosine triphosphatase (ATPase) and gamma-glutamyl transpeptidase (GGT) activities were visualized according to published protocols.1, 2 Immunohistochemical methods and assessment of Kupffer cell activity are described in the supplementary material. For morphometric analysis, the number of desmin-positive HSC was determined in 100 microscopic fields centered on portal areas (400× magnification), per time point in multiple animals. Only HSC with a visible nucleus and at least 1 cytoplasmic process were included.

Analysis by Western Blot.

Blots were probed with antibodies against α-SMA (clone 1A4, Sigma) and desmin (clone D33, DAKO), diluted 1:1000, or protein disulfide isomerase, diluted 1:6000. See Supplementary Material for further detail.

Reverse Transcription Polymerase Chain Reaction and Real-Time Polymerase Chain Reaction.

These procedures are described in Supplementary Material.

Statistical Analysis.

Data are presented as means ± SD. The significances were analyzed by t tests, Mann-Whitney rank sum tests, chi square tests, or ANOVA using the Tukey test for pairwise comparisons of mean responses with SigmaStat software (Jandel Scientific, San Raphael, CA). P values <.05 were considered significant.


Cell Transplantation Induced a Partially Activated Phenotype in HSC.

Initial studies used the plan in Fig. 1A to determine HSC activation after cell transplantation (n = 4 rats per time-point). Transplanted cells were in portal vein radicles and sinusoids after 6 hours (Fig. 2A), whereas 24 to 48 hours after transplantation, most transplanted cells were cleared from portal vein radicles. Transplanted cells were in the liver parenchyma after 3 to 7 days (Fig. 2B).

Figure 2.

Cell transplantation and HSC activation. (A) DPPIV-positive transplanted cells (red color, arrows) within vascular spaces in the liver 6 hours after cell transplantation. (B) Transplanted cells in the liver 3 days after cell transplantation. In this situation, transplanted cells were cleared from vascular spaces and are present in only the liver parenchyma. (C) Desmin immunostaining showing only occasional HSC with desmin expression (arrows) 6 hours after cell transplantation. Pa = portal area. (D) Much more widespread appearance of desmin-positive HSC 3 days after cell transplantation. Inset shows an enlarged view of a desmin-positive HSC. (E) Cumulative analysis of the kinetics and extent of HSC perturbation (n = 4 rats each). This indicated that HSC perturbation peaked 3 days after cell transplantation. (F) Appearance of desmin-positive HSC (arrows) in the immediate vicinity of transplanted DPPIV-positive cells (red color, arrowhead) 3 days after cell transplantation. (G) Real-time RT-PCR data showing that α-SMA mRNA expression increased in the liver, especially between 12 and 24 hours after cell transplantation. (H) Western blot showing that α-SMA protein expression was not detectable in the liver by this method after cell transplantation (α-SMA, top panel; bottom panel, protein disulfide isomerase [PDI]). (I) Immunostaining of the rat liver for α-SMA 3 days after cell transplantation showed staining of blood vessel walls and only an occasional HSC (arrowhead). (J) α-SMA immunostaining in the liver of a “positive control” rat treated with CCl4 showed extensive HSC activation (Cv, central vein). Original magnification, A and B, ×200; C, D, and F, ×400; insets, ×600. DPPIV, dipeptidyl peptidase IV; RT-PCR, reverse transcription polymerase chain reaction; α-SMA, alpha-smooth muscle actin; HSC, hepatic stellate cell.

Desmin and α-SMA were expressed prominently in blood vessel walls (Fig. 2C–D). However, desmin-positive HSC were only rarely observed 6 hours after cell transplantation (Fig. 2C), similar to sham-treated controls. The number of desmin-positive HSC increased 24 hours after cell transplantation and was most prominent in zone 1 (periportal areas) of the liver lobule after 3 days (Fig. 2D). HSC exhibited characteristic morphology with small central nuclei and desmin-positive cytoplasmic processes. Morphometric analysis showed few desmin-positive HSC in either sham-treated controls or in rats 6 hours after hepatocyte transplantation (2.0 ± 1.8 vs. 2.4 ± 2.1 desmin-positive HSC per microscopic field, zone 1, respectively, P = NS, t test) (Fig. 2E). In contrast, desmin-positive HSC were more frequent 24 hours, 3 days, and 7 days after cell transplantation (4.0 ± 3.7, 14.3 ± 5.9, and 9.3 ± 4.1 cells per field, respectively, P < .001, ANOVA with the Tukey test showing P < .05 in all pairwise comparisons vs. controls). This 7.2-fold increase in number of desmin-positive HSC 3 days after cell transplantation, followed by a decline after 7 days, suggested transient perturbation of HSC.

Co-localization of desmin-positive HSC and DPPIV-positive transplanted cells in recipients of 3 different cell preparations showed that, 3 days after transplantation, 64% ± 7% of transplanted cells were in the immediate proximity of 1 or more desmin-positive HSC (Fig. 2F), suggesting that these cells could interact with one another. Despite transplantation of fewer hepatocytes in animals (1, 5, or 10 × 106 cells), desmin-expressing HSC were observed in the vicinity of transplanted cells, which was similar to Fig. 2F. However, in these animals, the number of activated HSC with desmin increased by only 1.2-fold, 1.5-fold, and 1.4-fold, respectively, compared with sham-treated controls, indicating the requirement of a threshold effect for extensive HSC response, such as after transplantation of 2 × 107 cells.

Real-time reverse transcription polymerase chain reaction (RT-PCR) verified the onset of α-SMA expression 24 hours after cell transplantation, with a return to normal values thereafter (Fig. 2G). However, Western blotting or immunostaining did not show increased α-SMA expression in cell recipients, although this was apparent in control animals with CCl4-induced liver injury (Fig. 2H–J). Immunoblotting for desmin in either cell recipients or CCl4-treated control rats was not successful.

Relationship Between Hepatic Ischemia and HSC Activation.

To determine the role of ischemia–reperfusion in HSC activation, after sinusoidal occlusion by transplanted cells, we used GGT as a reporter of hepatic ischemia.2 Bile duct cells expressed GGT in the normal adult rat liver, whereas after cell transplantation, GGT was extensively expressed in native hepatocytes (Fig. 3A–B). In contrast with sham-treated control rats, we observed multiple desmin-positive HSC within ischemic areas with GGT expression, 3 days after cell transplantation (Fig. 3C–D). To further examine whether occlusion of sinusoids played a role in HSC activation, we injected 2 × 107 MAA particles, which serve as effective biodegradable cell surrogates to reproduce physical events.10 We observed the appearance of multiple desmin-expressing HSC (Supplementary Fig. 1; available at the HEPATOLOGY website (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html)), indicating that mechanical factors, including the onset of hepatic ischemia,3 were involved in this process. However, additional studies with transplantation of small numbers of cells and other studies described below showed that mechanical factors alone were not responsible for inducing desmin expression in HSC.

Figure 3.

Liver ischemia and HSC perturbation. (A) Histochemical staining for GGT in the normal liver showing its expression restricted to the bile ducts (arrows). (B) Extensive GGT expression in periportal hepatocytes (asterisk) 6 hours after cell transplantation indicated the onset of hepatic ischemia. (C) GGT plus desmin staining of the normal liver with GGT-positive bile ducts (arrow) and only occasional desmin-positive HSC (arrowheads). (D) GGT plus desmin staining of the liver 3 days after cell transplantation showing desmin-positive HSC (arrowheads) in ischemic areas with GGT expression additional to bile ducts (arrows). (E) Liver showing carbon-containing Kupffer cells (thin arrows), DPPIV-positive transplanted hepatocytes (red color, arrows) and desmin-positive HSC (arrowheads) next to one another 6 hours after cell transplantation. (F) Carbon-containing Kupffer cells (thin arrows), DPPIV-positive transplanted hepatocytes (red color, arrows) and desmin-positive HSC (arrowheads) 3 days after cell transplantation. Original magnification, ×400; B, ×100. GGT, gamma-glutamyl transferase; HSC, hepatic stellate cells; DPPIV, dipeptidyl peptidase IV.

Role of Kupffer Cells in HSC Activation.

Because Kupffer cells are activated by ischemia–reperfusion after cell transplantation,4 and serve roles in HSC acitvation,6, 7 we examined relationships between Kupffer cells, HSC, and transplanted cells. Activated Kupffer cells containing carbon particles were adjacent to transplanted cells within 6 hours after cell transplantation (Fig. 3E). We found that, 3 days after cell transplantation, DPPIV-positive transplanted hepatocytes, carbon-containing Kupffer cells, and desmin-positive HSC were frequently located next to one another (Fig. 3F). Analysis of 75 liver lobules (n = 2 rats each) showed significant mean concordances between colocalization of desmin-positive HSC and carbon-containing Kupffer cells (69% and 75%, P = .003 and <.001, respectively, chi square tests). In further studies of Kupffer cells in HSC activation, we treated rats with gadolinium chloride (GdCl3), a potent Kupffer cell inhibitor,4, 11 before cell transplantation (Fig. 1B). However, despite appropriate interference with Kupffer cell function, as verified by less carbon uptake in animals (not shown), the number of desmin-positive HSC in GdCl3-treated and control rats 3 days after cell transplantation was unchanged (11.6 ± 4.4 vs. 11.2 ± 3.8 desmin-positive HSC per microscopic field, n = 3 each, P = NS). Similarly, because tumor necrosis factor alpha (TNF-α) is a major effector of the Kupffer cell response, and TNF-α messenger RNA (mRNA) expression increased 2.1-fold, 12 hours after cell transplantation, we transplanted cells in animals pretreated with etanercept to block TNF-α. The activity of etanercept used here, which is also used clinically, is validated by the manufacturer. However, etanercept treatment did not alter the numbers of desmin-positive HSC after cell transplantation (17.4 ± 5.4 vs. 15.0 ± 5.4 in etanercept-treated, and control rats, respectively, n = 3 each, P = NS).

Ischemia Alone Was Not Responsible for HSC Activation.

To prevent ischemic liver injury after cell transplantation, we infused nitroglycerin (NTG) intrasplenically during cell transplantation, which attenuated hepatic GGT expression, as described previously.2 However, 3 days after cell transplantation in NTG-treated animals, more desmin-positive HSC were observed in all three zones of the liver lobule, whereas animals subjected to cell transplantation alone showed desmin-positive HSC primarily in zone 1, as observed earlier (Fig. 4 and Supplementary Fig. 2). Morphometry showed a mean 2.7-fold increase in desmin-positive HSC in NTG-treated rats compared with controls treated with only cells (78 ± 13 vs. 29 ± 9 desmin-positive cells throughout the liver lobules, n = 3 each, P < .001, t test). Administration of NTG or normal saline alone, without cells, did not induce desmin expression in HSC. In NTG-treated rats, more transplanted cells survived (6.5 ± 0.7 × 104 vs. 1.8 ± 0.5 × 104 transplanted cells per cm3 liver, P < .05), which was in agreement with studies showing that sinusoidal vasodilatation improves cell engraftment.3 Moreover, analysis of conjoint hybrid bile canaliculi containing ATPase and DPPIV−positive components from native and transplanted cells, respectively, verified that transplanted cells integrated more rapidly in NTG-treated recipients (not shown).

Figure 4.

Effect of nitroglycerin on HSC perturbation. Panels A-C show liver from control animals without NTG treatment with presence of desmin-positive HSC in zone 1 (A), zone 2 (B), but not zone 3 (C). Panels D-F are from animals treated with NTG, where more widespread desmin expression, with desmin-positive HSC distributed in all 3 zones of the liver lobule (arrows), was observed. (G) Morphometric analysis, which indicated a 2.7-fold increase of desmin-positive HSC per liver lobule in NTG-treated rats. NTG, nitroglycerin; HSC, hepatic stellate cells.

Cell Transplantation Altered Expression of HSC-Related Genes.

RT-PCR analysis of HSC-related genes showed increased expression of hepatocyte growth factor (HGF), basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), and transforming growth factor (TGF)-β1 within 1 to 6 hours after cell transplantation (Fig. 5A). Real-time RT-PCR verified 2.1- to 2.6-fold upregulation of TGF-β, platelet-derived growth factor β, and TNF-α expression between 6 and 12 hours after cell transplantation (Fig. 5B) and a return to near normal values thereafter.

Figure 5.

RT-PCR analysis of HSC-related cytokine expression. (A) RT-PCR from untreated controls (lanes 1, 2) and after cell transplantation (lanes 3–14) (n = 2 rats each), PCR mix alone as negative control (lane 15), and liver from a rat 24 hours after a single CCl4 dose (lane 16), with increased gene expression. VEGF165 and VEGF121 transcripts were amplified. (B) Real-time RT-PCR with depiction of mean fold-change in gene expression from these samples. RT-PCR, reverse transcription polymerase chain reaction; HSC, hepatic stellate cells; VEGF, vascular endothelial growth factor.

Expression of MMP and Tissue Inhibitor of Metalloproteinase mRNA Was Upregulated After Cell Transplantation.

Cell transplantation rapidly and markedly upregulated MMP-13 expression, which encodes the major interstitial collagenase in the rat liver, for up to 12 hours after cell transplantation (Fig. 6A). Real-time RT-PCR readily detected MMP-13 mRNA 6 and 12 hours after cell transplantation (mean Ct = 26.26 and 27.32 cycles, respectively, vs. 19.75 cycles for β2-microglobulin), whereas it was undetectable in controls or at other time-points after cell transplantation. MMP-3 mRNA level was increased more than 20-fold 12 hours after cell transplantation (Fig. 6B), and expression of the gelatinase A/basement membrane collagenase (MMP-2), as well as its activator protease, membrane-type MMP-14, increased moderately after 1 to 12 hours and 1 to 24 hours, respectively (Fig. 6C). MMP-9 mRNA levels rose 6 hours after cell transplantation and persisted at a higher level subsequently for 72 hours (Fig. 6A). The expression of urokinase-type plasminogen activator (uPA) showed similar kinetics, whereas its inhibitor PAI-1 was upregulated 4-fold between 1 and 6 hours (Fig. 6D). Greater levels of tissue inhibitor of metalloproteinase (TIMP-1) mRNA were observed within 1 hour after cell transplantation, and TIMP-1 expression peaked with a 17-fold increase after 12 hours (Fig. 6E). Notably, whereas MMP expression rapidly returned to almost baseline, TIMP-1 mRNA remained increased at later times (7-fold at 48 hours, 4-fold at 72 hours). In contrast, TIMP-2 mRNA levels were only minimally altered after cell transplantation. Therefore, MMPs, as well as their activators and inhibitors, were regulated in a highly ordered temporal manner after cell transplantation.

Figure 6.

RT-PCR analysis of ECM-regulating genes. (A) RT-PCR from samples in Fig. 5 with expression of specific MMPs and TIMPs as shown. (B-E) The gene expression profiles were verified by real-time RT-PCR of the samples. RT-PCR, reverse transcription polymerase chain reaction; ECM, extracellular matrix; MMP, matrix metalloproteinases; TIMP, tissue inhibitor of metalloproteinase.

Prior Depletion of HSC Impaired Transplanted Cell Engraftment.

To demonstrate whether HSC activation directly affected cell engraftment, we performed “loss-of-function” studies by depleting activated HSC with gliotoxin.12–14 In contrast to results in animals with liver injury,12 we observed by oil-red-O or desmin stainings that gliotoxin did not deplete HSC in the healthy rat liver. No other method is available to deplete quiescent HSC. Therefore, we reasoned that prior activation of HSC would facilitate gliotoxin-induced cell depletion in an effort to study the role of HSC on cell engraftment. For this purpose, we administered animals 4 doses of CCl4, separated by intervals of 3 to 4 days each (n = 6 rats). Three days after the final dose of CCl4, animals were divided into groups of 3 rats each for administration of either 3 mg/kg gliotoxin to deplete activated HSC or vehicle only.12 This was followed 2 days later by cell transplantation, as depicted in Fig. 1C. In additional animals treated in this fashion, except for cell transplantation, we performed further studies to elicit the effect of gliotoxin on hepatocytes, Kupffer cells, and HSC (also see Supplementary Material).

To demonstrate that gliotoxin did not produce concomitant hepatocytic injury, we measured serum alanine aminotransferase after completing CCl4 treatment and immediately before, as well as after, 3 mg/kg gliotoxin. This excluded a hepatotoxic effect of gliotoxin with alanine aminotransferase before gliotoxin, 76 ± 28 U/L (n = 6), 6 hours after gliotoxin, 91 ± 34 U/L (n = 3), and 24 hours after gliotoxin, 89 ± 35 U/L (n = 3), P = NS, ANOVA). Light microscopy of liver samples showed hepatocytes with similar morphology in all conditions. In control animals treated with CCl4 and vehicle alone, desmin staining and α-SMA immunoblotting showed significant activation of HSC (Supplementary Material including Supplementary Fig. 3). Gliotoxin treatment produced depletion of such activated HSC. Conversely, Kupffer cell activity was not affected by gliotoxin.

Morphometric analysis of cell engraftment 3 days after gliotoxin or vehicle treatment showed transplanted cells in equivalent numbers of portal vein branches (37% ± 8% gliotoxin-treated vs. 31% ± 3% in vehicle-treated controls, P = NS, t test), indicating that CCl4 and gliotoxin did not perturb the intrahepatic distribution of transplanted cells. However, fewer transplanted cells engrafted in gliotoxin-treated rats compared with rats treated with CCl4 alone (Fig. 7A-B), leading to a mean 36% decline in the number of transplanted cells in the liver (gliotoxin-treated rats, 5.6 ± 0.8 × 105 transplanted cells vs. 8.8 ± 1.3 × 105 transplanted cells per cm3 liver in vehicle-treated rats, P < .001, t test) (Fig. 7C). In a repeat study using additional rats, gliotoxin treatment resulted in a mean 57% decline in cell engraftment (2.6 ± 0.6 × 105 transplanted cells vs. 6.0 ± 0.3 × 105 transplanted cells per cm3, n = 2 rats each). Through damage to native hepatocytes, CCl4 treatment increased transplanted cell engraftment in the animals (6.0 ± 0.3 × 105 transplanted cells vs. 2.5 ± 1.4 × 104 transplanted cells per cm3 in vehicle-treated controls, n = 2). To exclude whether gliotoxin alone, without prior CCl4, would alter transplanted cell engraftment, we administered either gliotoxin or dimethylsulfoxide to animals (n = 4). Morphometric analysis showed that cell engraftment actually increased in gliotoxin-treated rats compared with controls (2.5 ± 1.4 × 104 transplanted cells vs. 6.3 ± 2.9 × 104 transplanted cells per cm3, P < .05).

Figure 7.

HSC depletion with gliotoxin and cell engraftment. (A) Rat treated with 4 doses of CCl4 showing DPPIV-positive transplanted cells (arrows) 3 days after transplantation. (B) Liver from rats treated with CCl4 plus gliotoxin before cell transplantation showed fewer transplanted cells after 3 days. (C) Morphometric analysis of the difference in transplanted cell numbers in these two conditions. Asterisk represents P < .05. Original magnification, A, B, ×400. HSC, hepatic stellate cells; DPPIV, DPPIV, dipeptidyl peptidase IV.


These data established that hepatocyte transplantation rapidly initiated HSC activation. The changes in HSC phenotype and gene expression profiles are characteristic of the initiation or preinflammatory stage of HSC activation, as well as additional responses, such as cell proliferation or production of cytokines and matrix-regulating enzymes, which characterize the perpetuation phase.15 Cell transplantation is associated with perturbations of additional liver cell compartments, such as those associated with ischemia–reperfusion injury, that is, Kupffer cell activation, and endothelial disruption, which increases the potential for cell–cell interactions.1–5 Activation of HSC is associated early with increased desmin expression and cell cycle entry, which is in agreement with intense and widespread desmin expression in HSC within 3 days after cell transplantation. However, these changes were transient, because the number of desmin-positive HSC declined within 7 days after cell transplantation. Overall, this kinetics resembled responses after acute CCl4-induced liver injury, where HSC activation rapidly subsides after peaking in 2 to 4 days.16, 17 Conversely, cell transplantation induced “partial activation” phenotype of HSC with only transient α-SMA expression at the mRNA level.

Transplantation of 2 × 107 cells represented approximately 2% of the liver cell mass, which will be clinically appropriate, and is expected to produce an element of hepatic ischemia. Reversible activation of HSC after ischemic liver injury has been shown.18, 19 Similarly, the proximity of activated HSC with Kupffer cells could be relevant because Kupffer cells and leukocytes, which have major effects on HSC,6 are activated after cell transplantation.1, 4 However, use of GdCl3 to inactivate Kupffer cells or of etanercept to block TNF-α, did not prevent HSC perturbation after cell transplantation, although GdCl3 might not abolish all of the Kupffer cell responses.4, 11 Whereas TNF-α is involved in HSC activation, it may inhibit proliferation of HSC.20, 21 Also, quiescent HSC may not respond to TNF-α.21 Multiple other Kupffer cell–derived effectors regulate HSC activation, and these substances may mediate interactions between these cell types after cell transplantation.22

Other potential interactions during the initiation of HSC activation could include those with LSEC, which are disrupted after cell transplantation.1, 5 Endothelial injury stimulates conversion of latent TGF-β to its active form via plasmin,15 and release of fibronectin isoforms may activate HSC.23 Among HSC-related genes expressed after cell transplantation, VEGF activates HSC and induces HSC proliferation.24, 25 Our observation of VEGF mRNA expression here was in agreement with that of previous immunostaining studies showing VEGF expression within a few hours after cell transplantation.1 TGF-β, which was upregulated early after cell transplantation, plays roles in activating HSC during liver fibrosis. Quiescent HSC are more responsive to TGF-β than fully activated ones,25 although the role of TGF-β in regulating HSC proliferation is less clear.6, 26, 27 Furthermore, platelet-derived growth factor and bFGF can induce proliferation or mediate the mitogenic effects of TGF-β in HSC.28, 29 Such paracrine signaling by neighboring liver cells should be effective in initiating HSC activation.

Our studies with NTG were in further agreement with the role of soluble factors in activating HSC and excluded an exclusive role of ischemia–reperfusion in HSC activation. Previous studies showed that NTG restores hepatic microcirculatory disruptions after cell transplantation.3 In control rats subjected to cell transplantation alone, similar to healthy rats, pericentral HSC expressed desmin less well.30 In contrast, after cell transplantation in NTG-treated rats, desmin-expressing HSC were widely distributed in the liver lobule. Consistent with this finding, nitric oxide donors reproduced the effects of hypoxia in stellate cell lines, for example, with increased VEGF production,31 while inhibiting proliferation and procollagen expression in cultured HSC.32, 33 Because the latter studies were carried out in fully activated cell lines, comparing all mechanisms of NTG in vitro with its effects in vivo may not be possible.

Activation of HSC could play multiple roles in cell engraftment. For instance, transplanted cells require adhesion factors for binding to LSEC. Breaching of the endothelial barrier during entry of transplanted cells into the space of Disse is facilitated by VEGF-induced endothelial permeabilization.1 Similarly, release of hepatotrophic factors, such as HGF and bFGF, could facilitate the subsequent cell engraftment process. HGF is expressed well in early cultures of activated primary HSC, although transition of HSC to α-SMA–expressing myofibroblast-like phenotype decreases HGF expression.34

The need for physical disruption of the sinusoidal barrier should be commensurate with the induction of matrix degrading enzymes. During the initial phase of ischemic liver injury, Kupffer cell activation, phagocytic infiltration, and stimulation of sinusoidal endothelial cells, multiple soluble factors and cytokines with HSC-activating properties are released.6–8 Initiation of HSC activation should promote an early matrix-degrading phenotype. The uPA-plasmin system induces pro-MMPs during early remodeling in liver regeneration,7, 35 and expression of plasmin-activating uPA, as well as its inhibitor, PAI-1, is upregulated immediately after cell transplantation. More importantly, MMP-13, which was maximally induced 12 hours after cell transplantation, is the major interstitial collagenase in the rodent liver, and cleaves collagen type I in a manner that destabilizes collagen crosslinks and facilitates its degradation by gelatinases.36 Multiple cytokines upregulate MMP-13 expression.37, 38 Moreover, MMP-13 activation is regulated by MMP-14, which is poorly inhibited by TIMP-1, and was coordinately induced shortly after cell transplantation.7 MMP-3, which was markedly upregulated after cell transplantation, degrades a variety of ECM components and may be activated by neutrophils infiltrating nonviable transplanted cells.2 MMP-3 expression is primarily triggered by TNF-α and interleukin-1 released from Kupffer cells/macrophages and T lymphocytes.

Subsequently, as transplanted cells integrate in the liver parenchyma, inhibition of excessive matrix degradation and ECM reconstitution would likely be needed. It was noteworthy that following cell transplantation TIMP-1 expression was sustained longer than that of MMPs. TGF-β, and possibly TNF-α, stimulate TIMP-1 expression,39, 40 which inhibits virtually all MMPs, except for MMP-9. TIMP expression after acute liver injury follows peak MMP expression and may subsequently limit matrix degradation.41, 42 TIMP-1 expression after cell transplantation coincided with the time required (3–7 days) for resumption of cell polarity and plasma membrane reconstitution.1 Therefore, these cumulative effects of MMPs and TIMPs could result in regulated disruption of basement membranes anchoring native hepatocytes, as well as degradation of denatured interstitial collagens and proteoglycans to facilitate the entry and integration of transplanted cells in the liver plate. Although we studied MMP expression at only the mRNA level, MMP activity is posttranscriptionally regulated, and TIMP availability is subject to its local pools. However, good correlations have been established between the mRNA expression and proteolytic activity of MMPs.42 Although hepatocytes and Kupffer cells also may express MMPs and TIMPs, HSC are considered the major source of these molecules in the liver.7, 41 For instance, MMP-2 and TIMP-1 were expressed in HSC, and not hepatocytes, after turpentine-induced acute liver injury.43

Finally, our studies with gliotoxin, which induces apoptosis in activated HSC,12–14 showed impaired transplanted cell engraftment. These findings cannot be accounted for by gliotoxin toxicity in transplanted cells, because gliotoxin did not cause hepatotoxicity in the doses used here and in previous studies,12 and did not impair cell engraftment when used by itself. Therefore, taken together, these findings indicate that examining the potential of strategies to modify HSC before cell transplantation should be appropriate.


Ms. Chaoying Zhang provided technical assistance. Drs. M. Bauer and M. Koda, University of Erlangen-Nuernberg, designed some of the TaqMan primers.