- Email: [email protected]
Human liver slices as an in vitro model to study toxicity-induced hepatic stellate cell activation in a multicellular milieu
Chemico-Biological Interactions 162 (2006) 62–69
Human liver slices as an in vitro model to study toxicity-induced hepatic stellate cell activation in a multicellular milieu M. van de Bovenkamp a,∗ , G.M.M. Groothuis a , D.K.F. Meijer a , M.J.H. Slooff b , P. Olinga a a
Department of Pharmacokinetics and Drug Delivery, University Center for Pharmacy, Groningen, The Netherlands b Department of Surgery, Division of Hepatobiliary Surgery and Liver Transplantation, University Hospital Groningen, The Netherlands Received 18 April 2006; accepted 11 May 2006 Available online 17 May 2006
Abstract Introduction: Hepatic stellate cell (HSC) activation is a key event in wound healing as well as in fibrosis development in the liver. Previously we developed a technique to induce HSC activation in slices from rat liver. Although this model provides a physiologic, multicellular milieu that is not present in current in vitro models it might still be of limited predictive value for the human situation due to species-differences. Therefore, we now aimed to evaluate the applicability of human liver slices for the study of HSC activation. Method: Liver slices (8 mm diameter, 250 m thickness) were generated from human liver tissue and incubated for 3 or 16 h with 0–15 l of carbon tetrachloride (CCl4 ) after which ATP-content and expression levels of HSC (activation) markers was determined. Results: Human liver slices remained viable during incubation as shown by constant ATP levels. Incubation with CCl4 caused a dose-dependent decrease in viability and an increase in mRNA expression of the early HSC activation markers HSP47 and ␣Bcrystallin, but not the late markers for HSC activation, ␣SMA and pro-collagen 1a1. Synaptophysin mRNA expression remained constant during incubation with or without CCl4 , indicating a constant number of HSC in the liver slices. Conclusion: We developed a technique to induce early toxicity-induced HSC activation in human liver slices. This in vitro model provides a multicellular, physiologic milieu to study mechanisms underlying toxicity-induced HSC activation in human liver tissue. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Hepatic stellate cells; Carbon tetrachloride; Precision-cut human liver slices
1. Introduction Liver fibrosis due to viral or metabolic liver injury is one of the leading causes of death worldwide [1]. ∗
Correspondence to: University of Groningen, Department of Pharmacokinetics and Drug Delivery, A. Deusinglaan 1, 9713 AV Groningen, The Netherlands. Tel.: +31 50 3637565; fax: +31 50 3633247. E-mail address: [email protected] (M. van de Bovenkamp).
It is well known that the injury-induced activation of hepatic stellate cells (HSC) is a key event in the natural process of scar tissue formation as well as excessive fibrogenesis in the liver [1,2]. However, to date no curative treatment for liver fibrosis is available and patients are dependent on liver transplantations. In order to develop effective anti-fibrotic therapies it is of importance to elucidate the mechanisms underlying initial injury-induced HSC activation and subsequent fibrosis development in human liver. Current in vivo and
0009-2797/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.cbi.2006.05.006
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
in vitro models contribute significantly to the understanding of HSC biology. However, experimental animal models for liver fibrosis are difficult to extrapolate to human disease and in vitro cell culture models cannot incorporate cell–cell and cell–extracellular matrix interactions that play an important role in the development of fibrosis. Thus, there persists a need for an in vitro model that can mimic the in vivo situation more closely. Previously we showed that precision-cut rat liver slices could be used to study early activation of HSC in a physiological milieu [3]. The main advantage of this system over current available in vitro models for the study of HSC activation is that in the liver slice all liver cell types are present in a physiological extracellular matrix. This feature of liver slices provides the opportunity to study HSC activation while preserving cell–cell and cell–extracellular matrix interactions [4]. However, when using rodent liver tissue for this purpose, the model may still be of limited predictive value for the human situation. Therefore, the aim of our present study was to evaluate the applicability of precision-cut liver slices derived from human liver tissue for the study of HSC activation. The design of the study was two-fold. Firstly, we evaluated the viability of human liver slices during incubation and determined whether responses of the liver slices were dependent on the origin of the liver tissue used. Secondly, we determined if HSC activation could be induced in the liver slices. For this purpose, we incubated the slices with carbon tetrachloride (CCl4 ), a known fibrogenic compound whose toxicity and fibrogenic activity requires conversion into a free radical by hepatocytes. Its fibrogenic activity, therefore, has been linked to oxidative damage to hepatocytes, yielding lipid peroxides and other mediators that activate HSC [5,6]. To assess HSC activation in the liver slices heat shock protein 47 (HSP47) and ␣B-crystallin expression were used as early marker for HSC activation, and mRNA expression of ␣ smooth muscle actin (␣SMA) and pro-collagen 1a1 as late marker for HSC activation and fibrogenesis. Synaptophysin mRNA, which is expressed in quiescent as well as activated HSC [7] was used as a marker for the amount of HSC present. The development of this in vitro model enables the study of human HSC activation in a physiological milieu and may contribute to the understanding of processes underlying early HSC activation in human liver. In addition, the model may contribute substantially to the reduction, refinement, and potential replacement of animal experiments.
63
2. Materials and methods 2.1. Human liver tissue Human liver tissue was obtained from multi-organ donors (Tx-livers) or from patients after partial hepatectomy because of metastasis of colorectal carcinoma (PH-livers). Consent from the legal authorities and from the families concerned was obtained for the use of Txlivers for transplantation-related research. The Tx-liver was perfused with cold University of Wisconsin organ storage solution (UW, DuPont Critical Care, Waukegan, IL, USA) in situ before explantation and stored in cold UW until the liver was reduced in order to perform reduced-size or split liver transplantation. During reduction the liver was immersed in UW cooled with ice slush. The liver tissue remaining after bipartition was stored in cold UW solution until the start of the slicing procedure. In case of PH-livers, consent to use liver tissue for research purposes was obtained from the patients concerned. The technique of partial hepatectomy was performed as described earlier [8] after which a wedge of non-cancerous liver tissue was cut from the resected liver lobe at distance from the metastases. Immediately after excision the biopsy wedge was perfused with cold UW and transported to the laboratory were the slicing procedure was started within 30 min. The medical ethical committee of the University of Groningen approved the research protocols. 2.2. Preparation of human liver slices Precision-cut liver slices (8 mm diameter, 250 m thickness) were prepared in ice-cold Krebs–Henseleit buffer saturated with carbogen (95% O2 /5% CO2 ) and containing 25 mM glucose (Merck, Darmstadt, Germany), 25 mM NaHCO3 (Merck), and 10 mM HEPES (ICN Biomedicals Inc. Aurora, OH, USA) using the Krumdieck tissue slicer [9,10]. Slices were stored on melting ice in UW until the start of the experiments. 2.3. Incubation of human liver slices Slices were preincubated for 1 h in 3.2 ml Williams Medium E with glutamax-I (Gibco, Paisly, Scotland) supplemented with 25 mM d-glucose and 50 g/ml gentamycin (Gibco) (WEGG) under carbogen atmosphere at 37 ◦ C in six well culture plates (Greiner bio-one, Frickenhausen, Germany) while gently shaken. Preincubation enables the slices to restore their ATP-levels [4]. After preincubation slices were transferred to 25 ml Erlenmeyer flasks containing 5 ml carbogen-saturated WEGG
64
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
and incubated at 37 ◦ C for 3 or 16 h, while gently shaken. At the start of the incubation period 0, 5, 10 or 15 l carbon tetrachloride (CCl4 , Fluka Chemie, Steinheim, Switzerland) was added in the 20 ml headspace of the flask to a paper attached to the stopper. During incubation CCl4 evaporates and equilibrium is reached between the gas phase and the medium. 2.4. Viability Each slice was transferred to a sonication solution containing 70% ethanol and 2 mM EDTA, snap frozen in liquid nitrogen, and stored at −80 ◦ C until ATP determination. ATP was determined in the supernatant that was obtained after sonication the samples 15 s and centrifuging the homogenate 2 at 13,000 rpm using the ATP Bioluminescence assay kit CLSII (Roche diagnostics, Mannheim, Germany). 2.5. Real-time PCR RNA was isolated from three snap-frozen slices using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). cDNA was synthesized from 2 g total RNA, using the Promega Reverse Transcription System (Promega, Madison, WI, USA). 0.63 l cDNA was used in real-time PCR reactions using Taqman reaction mixture (Applied Biosystems, Warrington, UK), and the appropriate primers and probes (Assays-On-Demand, Applied Biosystems, Foster City, CA, USA). The comparative threshold cycle (CT ) method was used for relative quantification. CT is inversely related to the abundance of mRNA transcripts in the initial sample. Mean CT of duplicate measurements was used to calculate the difference in CT for target and reference GAPDH gene (CT ), which was compared to the corresponding CT of the control (CT ). Data are expressed as fold-induction or repression of the gene of interest according to the formula 2(−CT ) . 2.6. Western blot analysis Total protein was isolated from three snap-frozen liver slices by homogenizing in RIPA (50 mM Tris (Roche), 1% Igepal, 0.5% Na-deoxycholate, 0.1% SDS (all from Sigma–Aldrich, Steinheim, Germany), and 150 mM NaCl) containing protease inhibitors. Liver slice homogenates (20 g protein) were electrophoresed in a 10% acrylamide gel and transferred electrophoretically to a polyvinylidene difluoride membrane (Roche). The membrane was blocked with 4% bovine serum albumin (ICN Biomedicals Inc.) and subsequently incubated overnight with the primary antibody against HSP47
(StressGen Biotechnology, Victoria, BC, Canada, 1:500 dilution) or ␣B-crystallin (StressGen, 1:2000 dilution). Afterwards the membrane was washed with PBSTween (Sigma–Aldrich) and incubated for 3 h with peroxidase-conjugated secondary antibody (rabbit–antimouse IgG, Dako, Denmark, 1:100 dilution). After washing, 0.4 mg/ml Luminol (Sigma–Aldrich) and 1 mg/ml enhancer (4-iodophenol, Acros Organics, Fair Lawn, NJ, USA) were added and signals were visualized on a Kodak Biomax Light film. Optical density × mm2 of the bands was measured using Quantity One (Bio-Rad Laboratories). 2.7. Statistics Experiments were performed with at least six livers using slices in triplicate from each liver. Data were compared using a two-tailed unpaired Students-t-test. Results obtained with real-time PCR were compared using mean CT values, and are presented as mean fold induction 2(−CT ) . A P-value < 0.05 indicated significance. Data are presented as mean ± standard error of the mean (S.E.M.). 3. Results 3.1. Characteristics of the human liver tissue used The characteristics of the human liver tissue used for the experiments with respect to age of the donor or patient, and the ATP-content of the liver slices are indicated in Table 1. No significant differences were observed in these parameters when comparing the different liver origins, although the age of the patients donating liver tissue after partial hepatectomy (PH) tended to be higher than the age of multi-organ donors (Tx). Importantly, neither the marker expression in the liver slices in response to incubation with CCl4 nor the toxic effect of Table 1 Characteristics of the human liver tissue used for the experiments Origin
Age mean ± S.E.M. [range]
ATP-content mean nmol/slice ± S.E.M.
Tx (N = 17) PH (N = 11) P-value PH + Tx
44 ± 3 [22–59] (N = 15) 51 ± 3 [33–57] (N = 8) 0.08 47 ± 2 [22–59]
6.0 ± 0.8 (N = 17) 5.2 ± 1.4 (N = 10) 0.59 5.7 ± 0.7
Origin of the liver tissue, age of the donors or patients as well as ATPcontent of the generated liver slices after 3 h of incubation are indicated. PH: partial hepatectomy, Tx: transplantation liver. In some cases, the age of the donor was not known and/or the ATP-content of the slices was not measured after 3 h of incubation (see N-values).
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
65
Fig. 1. (A) ATP-content of human liver slices after different incubation intervals relative to slices directly after preincubation (0). (B and C) ATPcontent of human liver slices after 3 h (B) and 16 h (C) of incubation with increasing doses of CCl4 relative to slices incubated without CCl4 (0). The mean of at least nine independent experiments ± S.E.M. is shown. * P < 0.05 compared to the control.
CCl4 showed a correlation with the origin of the liver, the age of the donor or patient, or the ATP-content of the liver slices after 3 h of incubation (data not shown). Therefore, in the present study all experiments were analyzed together. 3.2. Viability of human liver slices during incubation Human liver slices remained viable in the incubation system used during the incubation time of the experiment, as shown by constant ATP levels (Fig. 1A). Addition of 5–15 l CCl4 during 3 or 16 h of incubation caused a dose-dependent decrease in ATP-content of the slices (Fig. 1B and C). 3.3. Induction of HSC activation in human liver slices by CCl4 To assess whether the CCl4 -induced toxicity results in activation of HSC in the human liver slices, mRNA expression of markers for HSC-activation and fibrosis was measured in the liver slices after 3 and 16 h of incubation with 0–15 l CCl4 . Incubation of human liver slices in the presence of CCl4 resulted in a dosedependent increase in mRNA expression of the early markers HSP47 and ␣B-crystallin (Fig. 2A and B). This increase was observed as early as 3 h after addition of CCl4 and maintained after 16 h of incubation. A small, but non-significant increase in protein expression of these markers was observed (Fig. 2C). In contrast, mRNA expression of the late markers for HSC activation, ␣SMA and pro-collagen 1a1, was not increased (Fig. 3A and B). mRNA expression of synaptophysin in the liver slices remained constant during control incubation and during incubation with CCl4 , indicating a constant number of HSC in the liver slices (Fig. 4A and B).
4. Discussion Previously we have developed a technique to induce activation of HSC in rat liver slices, which provides the opportunity to study early toxicity-induced HSC activation in a physiologic, multicellular milieu [3]. However, due to species-differences this model may still be of limited predictive value for human disease. In the present study, we therefore evaluated the applicability of liver slices derived from human liver tissue for the study of HSC activation. 4.1. Viability of human liver slices during culture Human liver slices remained viable during the incubation period of 16 h as determined by ATP-content of the liver slices, a parameter generally used as indicator of viability [4]. The sustained mRNA expression of the HSC-specific marker synaptophysin during incubation (Fig. 4) indicates that HSC in the liver slices remain viable. In addition, as will be discussed below, the increased expression of HSP47 and ␣B-crystallin in response to incubation with CCl4 (Fig. 2) indicates activation of HSC, implying that these cells remain functionally active during preparation and subsequent culture of human liver slices. Previously we showed that hepatocytes, Kupffer cells, and endothelial cells remain viable and functional in human liver slices during culture up to 24 h [9,11]. Since HSC activation in vivo is a multicellular process the presence and functionality of all liver cell types could offer a major advantage over current in vitro models for human HSC activation, which are mainly single-cell cultures. 4.2. Toxicity-induced HSC-activation To evaluate the use of human liver slices for the study of toxicity-induced HSC activation, the slices were incu-
66
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
Fig. 2. (A and B) mRNA expression of HSP47, and ␣B-crystallin, respectively, in human liver slices after 3 and 16 h of incubation with 0–15 l CCl4 ; (C) protein expression of ␣B-crystallin and HSP47, in human liver slices after 16 h of incubation with 0 or 10 l CCl4 . The mean of at least four independent experiments ± S.E.M. is shown. Data are expressed relative to slices incubated without CCl4 (0). * P < 0.05 compared to slices incubated without CCl4 .
bated for 3 or 16 h with the fibrogenic compound CCl4 after which ATP-content of the liver slices and mRNA expression of markers for HSC activation and fibrosis were determined and compared to those in unstimulated slices. CCl4 was added in the headspace of the culture flasks to a paper attached to the stopper. During incu-
bation CCl4 evaporates and dissolves into the medium until equilibrium is reached between the gas phase and the medium. This way of administration yields relative constant concentrations of CCl4 in the culture medium and the liver slice and is more reproducible than administering CCl4 directly into the medium [12].
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
67
Fig. 3. mRNA expression of ␣SMA (A), and pro-collagen 1a1 (B) in human liver slices after 3 and 16 h of incubation with 0–15 l CCl4 . The mean of six independent experiments ± S.E.M. is shown. Data are expressed relative to slices incubated without CCl4 (0).
Incubation of human liver slices in the presence of CCl4 resulted in a dose-dependent decrease in ATPcontent, indicating toxicity. In vivo liver toxicity of CCl4 is caused by cytochrome P450 mediated conversion of CCl4 into free radicals in the hepatocytes, which yields lipid peroxides and other mediators that can activate HSC [5,6]. In accordance, CCl4 -induced toxicity in human liver slices resulted in a dose-dependent increase of HSC activation in the liver slices, as reflected by the increased mRNA expression of HSP47 and ␣Bcrystallin, two early markers for HSC activation and fibrogenesis [13,14]. We were unable to measure significant changes in protein expression of these markers in human liver slices after incubation with CCl4 . This could be ascribed to differences in sensitivity between real-time PCR and Western blot analysis or to the relatively short incubation times used. The constant expression of synaptophysin and ␣SMA in human liver slices incubated with CCl4 indicates that the observed changes in HSP47 and ␣B-crystallin expression are due to HSC activation, rather than to an increase in the amount of (activated) HSC present in the
liver slices. Increased mRNA expression of HSP47, and ␣B-crystallin is also observed within the first 24 h of spontaneous HSC activation in in vitro cell culture models [15,16]. In contrast, pro-collagen 1a1 and ␣SMA are expressed later during activation of HSC [17,18]. In line with these studies, no differences in mRNA expression of these late markers for HSC-activation were observed in human liver slices after 16 h of incubation with CCl4 . Given the mechanism of action of CCl4 -induced fibrosis, HSC activation occurring in liver slices in response to incubation with CCl4 likely results from a multicellular process. This suggests that, in contrast to current in vitro models, the liver slice system can be used to study multicellular toxicity leading to activation of HSC in a physiologic milieu. However, although there are no indications that CCl4 exerts direct effects on HSC, this has not been ruled out yet. 4.3. Applications The aim of the present study was to evaluate the applicability of human liver slices for the study of early
68
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
Fig. 4. (A) mRNA expression of synaptophysin in human liver slices after 3 and 16 h of incubation relative to expression levels in slices directly after preincubation (0). (B) mRNA expression of synaptophysin in human liver slices after 3 and 16 h of incubation with 0–15 l CCl4 relative to slices incubated without CCl4 (0). The mean of at least six independent experiments ± S.E.M. is shown.
HSC activation. To address this issue we studied whether the liver slices remained viable during incubation and whether HSC activation in the slices could be induced by incubation with the fibrogenic compound CCl4 . Our results show persistent viability of human liver slices during control incubation and give a clear indication that CCl4 -induced toxicity results in HSC activation in the liver slices. Importantly, the response of human liver slices to incubation with CCl4 was reproducible and independent of the origin of human liver tissue, of the age of the donor or the patient, and of the ATP-content in the liver slices after 3 h of incubation, indicating the general applicability of human liver tissue for this type of research. Due to the relatively short incubation time used in our experiments we could not yet detect fibrogenesis, i.e. increased production of collagens, or complete transformation of activated HSC into myofibroblasts as reflected by increased ␣SMA expression. However, we did observe a clear increase in the expression of two early markers for HSC activation and fibrogenesis in response to incubation with CCl4 . It is well known that activation of HSC is the key event in the natural process of wound healing and scar tissue formation as well as in
the development of liver fibrosis. Therefore, we argue that the CCl4 -induced early HSC activation in human liver slices does reflect the onset stage of fibrogenesis. The present model could provide a powerful tool for the study of processes and pathways involved in early HSC activation preceding toxicity-induced fibrogenesis in human liver. In principle it can be employed to compare early responses of human livers to known (non)fibrogenic toxic compounds, among others by using genomic and proteomics tools. It remains to be elucidated whether prolongation of the incubation times also provides a useful system to test the effects of anti-fibrotic drugs on toxicity-induced fibrosis in human liver. 4.4. In conclusion We have developed a technique to reproducibly induce early HSC activation in precision-cut human liver slices. This model provides a multicellular, physiological milieu to study the mechanisms underlying toxicity-induced early HSC activation in human liver tissue. The development of this model may also contribute substantially to the reduction, refinement, and possible
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
replacement of animal experiments. Finally, the use of human liver tissue in this in vitro model may improve its predictive value for the human condition.
[9]
Acknowledgments M. van de Bovenkamp was supported by grants of ZON/Mw, and the Johns Hopkins Center for Alternatives to Animal Testing; P. Olinga was supported by Organon NV.
[10]
[11]
References [1] S.L. Friedman, Liver fibrosis—from bench to bedside, J. Hepatol. 38 (Suppl. 1) (2003) S38–S53. [2] A.Y. Hui, S.L. Friedman, Molecular basis of hepatic fibrosis, Exp. Rev. Mol. Med. 5 (2003), www.expertreviews.org/ 03005684h.htm. [3] M. Van de Bovenkamp, G.M.M. Groothuis, A.L. Draaisma, M.T. Merema, J.I. Bezuijen, M.J. Van Gils, D.K.F. Meijer, S.L. Friedman, P. Olinga, Precision-cut liver slices as a new model to study toxicity-induced hepatic stellate cell activation in a physiologic milieu, Toxicol. Sci. 85 (2005) 632–638. [4] P. Olinga, G.M.M. Groothuis, Use of human tissue slices in drug targeting research, in: G. Molema, D.K.F. Meijer (Eds.), Drug Targeting: Organ Specific Strategies, Wiley-VCH, Weinheim, 2001, pp. 309–331. [5] S. Basu, Carbon tetrachloride-induced lipid peroxidation: eicosanoid formation and their regulation by antioxidant nutrients, Toxicology 189 (2003) 113–127. [6] K. Ikejima, H. Honda, M. Yoshikawa, M. Hirose, T. Kitamura, Y. Takei, N. Sato, Leptin augments inflammatory and profibrogenic responses in the murine liver induced by hepatotoxic chemicals, Hepatology 34 (2001) 288–297. [7] D. Cassiman, J. van Pelt, R. De Vos, F. Van Lommel, V. Desmet, S.H. Yap, T. Roskams, Synaptophysin: a novel marker for human and rat hepatic stellate cells, Am. J. Pathol. 155 (1999) 1831–1839. [8] M.A. Brouwers, P.M. Peeters, K.P. De Jong, E.B. Haagsma, I.J. Klompmaker, C.M. Bijleveld, J.H. Zwaveling, M.J. Slooff, Sur-
[12] [13]
[14]
[15]
[16]
[17]
[18]
69
gical treatment of giant haemangioma of the liver, Br. J. Surg. 84 (1997) 314–316. P. Olinga, I.H. Hof, M.T. Merema, M. Smit, M.H. de Jager, P.J. Swart, M.J.H. Slooff, D.K.F. Meijer, G.M.M. Groothuis, The applicability of rat and human liver slices to the study of mechanisms of hepatic drug uptake, J. Pharmacol. Toxicol. Methods 45 (2001) 55–63. C.L. Krumdieck, J.E. dos Santos, K.J. Ho, A new instrument for the rapid preparation of tissue slices, Anal. Biochem. 104 (1980) 118–123. M.G.L. Elferink, P. Olinga, A.L. Draaisma, M.T. Merema, K.N. Faber, M.J.H. Slooff, D.K.F. Meijer, G.M.M. Groothuis, LPSinduced downregulation of MRP2 and BSEP in human liver is due to a posttranscriptional process, Am. J. Physiol. Gastrointest. Liver Physiol. 287 (2004) G1008–G1016. S. Azri, A.J. Gandolfi, K. Brendel, Carbon tetrachloride toxicity in precision-cut rat liver slices, In Vitro Toxicol. 3 (1990) 127–138. K.E. Brown, K.A. Broadhurst, M.M. Mathahs, E.M. Brunt, W.N. Schmidt, Expression of HSP47, a collagen-specific chaperone, in normal and diseased human liver, Lab. Invest. 85 (2005) 789–797. D. Cassiman, T. Roskams, J. van Pelt, L. Libbrecht, P. Aertsen, T. Crabbe, H. Vankelecom, C. Denef, Alpha B-crystallin expression in human and rat hepatic stellate cells, J. Hepatol. 35 (2001) 200–207. N. Kawada, T. Kuroki, K. Kobayashi, M. Inoue, K. Nakatani, K. Kaneda, K. Nagata, Expression of heat-shock protein 47 in mouse liver, Cell Tissue Res. 284 (1996) 341–346. A. Lang, L.W. Schrum, R. Schoonhoven, S. Tuvia, J.A. SolisHerruzo, H. Tsukamoto, D.A. Brenner, R.A. Rippe, Expression of small heat shock protein alphaB-crystallin is induced after hepatic stellate cell activation, Am. J. Physiol. Gastrointest. Liver Physiol. 279 (2000) G1333–G1342. P. Mathurin, S. Xiong, K.K. Kharbanda, N. Veal, T. Miyahara, K. Motomura, R.A. Rippe, M.G. Bachem, H. Tsukamoto, IL10 receptor and coreceptor expression in quiescent and activated hepatic stellate cells, Am. J. Physiol. Gastrointest. Liver Physiol. 282 (2002) G981–G990. K. Uchio, B. Tuchweber, N. Manabe, G. Gabbiani, J. Rosenbaum, A. Desmouliere, Cellular retinol-binding protein-1 expression and modulation during in vivo and in vitro myofibroblastic differentiation of rat hepatic stellate cells and portal fibroblasts, Lab. Invest. 82 (2002) 619–628.
Human liver slices as an in vitro model to study toxicity-induced hepatic stellate cell activation in a multicellular milieu M. van de Bovenkamp a,∗ , G.M.M. Groothuis a , D.K.F. Meijer a , M.J.H. Slooff b , P. Olinga a a
Department of Pharmacokinetics and Drug Delivery, University Center for Pharmacy, Groningen, The Netherlands b Department of Surgery, Division of Hepatobiliary Surgery and Liver Transplantation, University Hospital Groningen, The Netherlands Received 18 April 2006; accepted 11 May 2006 Available online 17 May 2006
Abstract Introduction: Hepatic stellate cell (HSC) activation is a key event in wound healing as well as in fibrosis development in the liver. Previously we developed a technique to induce HSC activation in slices from rat liver. Although this model provides a physiologic, multicellular milieu that is not present in current in vitro models it might still be of limited predictive value for the human situation due to species-differences. Therefore, we now aimed to evaluate the applicability of human liver slices for the study of HSC activation. Method: Liver slices (8 mm diameter, 250 m thickness) were generated from human liver tissue and incubated for 3 or 16 h with 0–15 l of carbon tetrachloride (CCl4 ) after which ATP-content and expression levels of HSC (activation) markers was determined. Results: Human liver slices remained viable during incubation as shown by constant ATP levels. Incubation with CCl4 caused a dose-dependent decrease in viability and an increase in mRNA expression of the early HSC activation markers HSP47 and ␣Bcrystallin, but not the late markers for HSC activation, ␣SMA and pro-collagen 1a1. Synaptophysin mRNA expression remained constant during incubation with or without CCl4 , indicating a constant number of HSC in the liver slices. Conclusion: We developed a technique to induce early toxicity-induced HSC activation in human liver slices. This in vitro model provides a multicellular, physiologic milieu to study mechanisms underlying toxicity-induced HSC activation in human liver tissue. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Hepatic stellate cells; Carbon tetrachloride; Precision-cut human liver slices
1. Introduction Liver fibrosis due to viral or metabolic liver injury is one of the leading causes of death worldwide [1]. ∗
Correspondence to: University of Groningen, Department of Pharmacokinetics and Drug Delivery, A. Deusinglaan 1, 9713 AV Groningen, The Netherlands. Tel.: +31 50 3637565; fax: +31 50 3633247. E-mail address: [email protected] (M. van de Bovenkamp).
It is well known that the injury-induced activation of hepatic stellate cells (HSC) is a key event in the natural process of scar tissue formation as well as excessive fibrogenesis in the liver [1,2]. However, to date no curative treatment for liver fibrosis is available and patients are dependent on liver transplantations. In order to develop effective anti-fibrotic therapies it is of importance to elucidate the mechanisms underlying initial injury-induced HSC activation and subsequent fibrosis development in human liver. Current in vivo and
0009-2797/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.cbi.2006.05.006
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
in vitro models contribute significantly to the understanding of HSC biology. However, experimental animal models for liver fibrosis are difficult to extrapolate to human disease and in vitro cell culture models cannot incorporate cell–cell and cell–extracellular matrix interactions that play an important role in the development of fibrosis. Thus, there persists a need for an in vitro model that can mimic the in vivo situation more closely. Previously we showed that precision-cut rat liver slices could be used to study early activation of HSC in a physiological milieu [3]. The main advantage of this system over current available in vitro models for the study of HSC activation is that in the liver slice all liver cell types are present in a physiological extracellular matrix. This feature of liver slices provides the opportunity to study HSC activation while preserving cell–cell and cell–extracellular matrix interactions [4]. However, when using rodent liver tissue for this purpose, the model may still be of limited predictive value for the human situation. Therefore, the aim of our present study was to evaluate the applicability of precision-cut liver slices derived from human liver tissue for the study of HSC activation. The design of the study was two-fold. Firstly, we evaluated the viability of human liver slices during incubation and determined whether responses of the liver slices were dependent on the origin of the liver tissue used. Secondly, we determined if HSC activation could be induced in the liver slices. For this purpose, we incubated the slices with carbon tetrachloride (CCl4 ), a known fibrogenic compound whose toxicity and fibrogenic activity requires conversion into a free radical by hepatocytes. Its fibrogenic activity, therefore, has been linked to oxidative damage to hepatocytes, yielding lipid peroxides and other mediators that activate HSC [5,6]. To assess HSC activation in the liver slices heat shock protein 47 (HSP47) and ␣B-crystallin expression were used as early marker for HSC activation, and mRNA expression of ␣ smooth muscle actin (␣SMA) and pro-collagen 1a1 as late marker for HSC activation and fibrogenesis. Synaptophysin mRNA, which is expressed in quiescent as well as activated HSC [7] was used as a marker for the amount of HSC present. The development of this in vitro model enables the study of human HSC activation in a physiological milieu and may contribute to the understanding of processes underlying early HSC activation in human liver. In addition, the model may contribute substantially to the reduction, refinement, and potential replacement of animal experiments.
63
2. Materials and methods 2.1. Human liver tissue Human liver tissue was obtained from multi-organ donors (Tx-livers) or from patients after partial hepatectomy because of metastasis of colorectal carcinoma (PH-livers). Consent from the legal authorities and from the families concerned was obtained for the use of Txlivers for transplantation-related research. The Tx-liver was perfused with cold University of Wisconsin organ storage solution (UW, DuPont Critical Care, Waukegan, IL, USA) in situ before explantation and stored in cold UW until the liver was reduced in order to perform reduced-size or split liver transplantation. During reduction the liver was immersed in UW cooled with ice slush. The liver tissue remaining after bipartition was stored in cold UW solution until the start of the slicing procedure. In case of PH-livers, consent to use liver tissue for research purposes was obtained from the patients concerned. The technique of partial hepatectomy was performed as described earlier [8] after which a wedge of non-cancerous liver tissue was cut from the resected liver lobe at distance from the metastases. Immediately after excision the biopsy wedge was perfused with cold UW and transported to the laboratory were the slicing procedure was started within 30 min. The medical ethical committee of the University of Groningen approved the research protocols. 2.2. Preparation of human liver slices Precision-cut liver slices (8 mm diameter, 250 m thickness) were prepared in ice-cold Krebs–Henseleit buffer saturated with carbogen (95% O2 /5% CO2 ) and containing 25 mM glucose (Merck, Darmstadt, Germany), 25 mM NaHCO3 (Merck), and 10 mM HEPES (ICN Biomedicals Inc. Aurora, OH, USA) using the Krumdieck tissue slicer [9,10]. Slices were stored on melting ice in UW until the start of the experiments. 2.3. Incubation of human liver slices Slices were preincubated for 1 h in 3.2 ml Williams Medium E with glutamax-I (Gibco, Paisly, Scotland) supplemented with 25 mM d-glucose and 50 g/ml gentamycin (Gibco) (WEGG) under carbogen atmosphere at 37 ◦ C in six well culture plates (Greiner bio-one, Frickenhausen, Germany) while gently shaken. Preincubation enables the slices to restore their ATP-levels [4]. After preincubation slices were transferred to 25 ml Erlenmeyer flasks containing 5 ml carbogen-saturated WEGG
64
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
and incubated at 37 ◦ C for 3 or 16 h, while gently shaken. At the start of the incubation period 0, 5, 10 or 15 l carbon tetrachloride (CCl4 , Fluka Chemie, Steinheim, Switzerland) was added in the 20 ml headspace of the flask to a paper attached to the stopper. During incubation CCl4 evaporates and equilibrium is reached between the gas phase and the medium. 2.4. Viability Each slice was transferred to a sonication solution containing 70% ethanol and 2 mM EDTA, snap frozen in liquid nitrogen, and stored at −80 ◦ C until ATP determination. ATP was determined in the supernatant that was obtained after sonication the samples 15 s and centrifuging the homogenate 2 at 13,000 rpm using the ATP Bioluminescence assay kit CLSII (Roche diagnostics, Mannheim, Germany). 2.5. Real-time PCR RNA was isolated from three snap-frozen slices using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). cDNA was synthesized from 2 g total RNA, using the Promega Reverse Transcription System (Promega, Madison, WI, USA). 0.63 l cDNA was used in real-time PCR reactions using Taqman reaction mixture (Applied Biosystems, Warrington, UK), and the appropriate primers and probes (Assays-On-Demand, Applied Biosystems, Foster City, CA, USA). The comparative threshold cycle (CT ) method was used for relative quantification. CT is inversely related to the abundance of mRNA transcripts in the initial sample. Mean CT of duplicate measurements was used to calculate the difference in CT for target and reference GAPDH gene (CT ), which was compared to the corresponding CT of the control (CT ). Data are expressed as fold-induction or repression of the gene of interest according to the formula 2(−CT ) . 2.6. Western blot analysis Total protein was isolated from three snap-frozen liver slices by homogenizing in RIPA (50 mM Tris (Roche), 1% Igepal, 0.5% Na-deoxycholate, 0.1% SDS (all from Sigma–Aldrich, Steinheim, Germany), and 150 mM NaCl) containing protease inhibitors. Liver slice homogenates (20 g protein) were electrophoresed in a 10% acrylamide gel and transferred electrophoretically to a polyvinylidene difluoride membrane (Roche). The membrane was blocked with 4% bovine serum albumin (ICN Biomedicals Inc.) and subsequently incubated overnight with the primary antibody against HSP47
(StressGen Biotechnology, Victoria, BC, Canada, 1:500 dilution) or ␣B-crystallin (StressGen, 1:2000 dilution). Afterwards the membrane was washed with PBSTween (Sigma–Aldrich) and incubated for 3 h with peroxidase-conjugated secondary antibody (rabbit–antimouse IgG, Dako, Denmark, 1:100 dilution). After washing, 0.4 mg/ml Luminol (Sigma–Aldrich) and 1 mg/ml enhancer (4-iodophenol, Acros Organics, Fair Lawn, NJ, USA) were added and signals were visualized on a Kodak Biomax Light film. Optical density × mm2 of the bands was measured using Quantity One (Bio-Rad Laboratories). 2.7. Statistics Experiments were performed with at least six livers using slices in triplicate from each liver. Data were compared using a two-tailed unpaired Students-t-test. Results obtained with real-time PCR were compared using mean CT values, and are presented as mean fold induction 2(−CT ) . A P-value < 0.05 indicated significance. Data are presented as mean ± standard error of the mean (S.E.M.). 3. Results 3.1. Characteristics of the human liver tissue used The characteristics of the human liver tissue used for the experiments with respect to age of the donor or patient, and the ATP-content of the liver slices are indicated in Table 1. No significant differences were observed in these parameters when comparing the different liver origins, although the age of the patients donating liver tissue after partial hepatectomy (PH) tended to be higher than the age of multi-organ donors (Tx). Importantly, neither the marker expression in the liver slices in response to incubation with CCl4 nor the toxic effect of Table 1 Characteristics of the human liver tissue used for the experiments Origin
Age mean ± S.E.M. [range]
ATP-content mean nmol/slice ± S.E.M.
Tx (N = 17) PH (N = 11) P-value PH + Tx
44 ± 3 [22–59] (N = 15) 51 ± 3 [33–57] (N = 8) 0.08 47 ± 2 [22–59]
6.0 ± 0.8 (N = 17) 5.2 ± 1.4 (N = 10) 0.59 5.7 ± 0.7
Origin of the liver tissue, age of the donors or patients as well as ATPcontent of the generated liver slices after 3 h of incubation are indicated. PH: partial hepatectomy, Tx: transplantation liver. In some cases, the age of the donor was not known and/or the ATP-content of the slices was not measured after 3 h of incubation (see N-values).
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
65
Fig. 1. (A) ATP-content of human liver slices after different incubation intervals relative to slices directly after preincubation (0). (B and C) ATPcontent of human liver slices after 3 h (B) and 16 h (C) of incubation with increasing doses of CCl4 relative to slices incubated without CCl4 (0). The mean of at least nine independent experiments ± S.E.M. is shown. * P < 0.05 compared to the control.
CCl4 showed a correlation with the origin of the liver, the age of the donor or patient, or the ATP-content of the liver slices after 3 h of incubation (data not shown). Therefore, in the present study all experiments were analyzed together. 3.2. Viability of human liver slices during incubation Human liver slices remained viable in the incubation system used during the incubation time of the experiment, as shown by constant ATP levels (Fig. 1A). Addition of 5–15 l CCl4 during 3 or 16 h of incubation caused a dose-dependent decrease in ATP-content of the slices (Fig. 1B and C). 3.3. Induction of HSC activation in human liver slices by CCl4 To assess whether the CCl4 -induced toxicity results in activation of HSC in the human liver slices, mRNA expression of markers for HSC-activation and fibrosis was measured in the liver slices after 3 and 16 h of incubation with 0–15 l CCl4 . Incubation of human liver slices in the presence of CCl4 resulted in a dosedependent increase in mRNA expression of the early markers HSP47 and ␣B-crystallin (Fig. 2A and B). This increase was observed as early as 3 h after addition of CCl4 and maintained after 16 h of incubation. A small, but non-significant increase in protein expression of these markers was observed (Fig. 2C). In contrast, mRNA expression of the late markers for HSC activation, ␣SMA and pro-collagen 1a1, was not increased (Fig. 3A and B). mRNA expression of synaptophysin in the liver slices remained constant during control incubation and during incubation with CCl4 , indicating a constant number of HSC in the liver slices (Fig. 4A and B).
4. Discussion Previously we have developed a technique to induce activation of HSC in rat liver slices, which provides the opportunity to study early toxicity-induced HSC activation in a physiologic, multicellular milieu [3]. However, due to species-differences this model may still be of limited predictive value for human disease. In the present study, we therefore evaluated the applicability of liver slices derived from human liver tissue for the study of HSC activation. 4.1. Viability of human liver slices during culture Human liver slices remained viable during the incubation period of 16 h as determined by ATP-content of the liver slices, a parameter generally used as indicator of viability [4]. The sustained mRNA expression of the HSC-specific marker synaptophysin during incubation (Fig. 4) indicates that HSC in the liver slices remain viable. In addition, as will be discussed below, the increased expression of HSP47 and ␣B-crystallin in response to incubation with CCl4 (Fig. 2) indicates activation of HSC, implying that these cells remain functionally active during preparation and subsequent culture of human liver slices. Previously we showed that hepatocytes, Kupffer cells, and endothelial cells remain viable and functional in human liver slices during culture up to 24 h [9,11]. Since HSC activation in vivo is a multicellular process the presence and functionality of all liver cell types could offer a major advantage over current in vitro models for human HSC activation, which are mainly single-cell cultures. 4.2. Toxicity-induced HSC-activation To evaluate the use of human liver slices for the study of toxicity-induced HSC activation, the slices were incu-
66
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
Fig. 2. (A and B) mRNA expression of HSP47, and ␣B-crystallin, respectively, in human liver slices after 3 and 16 h of incubation with 0–15 l CCl4 ; (C) protein expression of ␣B-crystallin and HSP47, in human liver slices after 16 h of incubation with 0 or 10 l CCl4 . The mean of at least four independent experiments ± S.E.M. is shown. Data are expressed relative to slices incubated without CCl4 (0). * P < 0.05 compared to slices incubated without CCl4 .
bated for 3 or 16 h with the fibrogenic compound CCl4 after which ATP-content of the liver slices and mRNA expression of markers for HSC activation and fibrosis were determined and compared to those in unstimulated slices. CCl4 was added in the headspace of the culture flasks to a paper attached to the stopper. During incu-
bation CCl4 evaporates and dissolves into the medium until equilibrium is reached between the gas phase and the medium. This way of administration yields relative constant concentrations of CCl4 in the culture medium and the liver slice and is more reproducible than administering CCl4 directly into the medium [12].
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
67
Fig. 3. mRNA expression of ␣SMA (A), and pro-collagen 1a1 (B) in human liver slices after 3 and 16 h of incubation with 0–15 l CCl4 . The mean of six independent experiments ± S.E.M. is shown. Data are expressed relative to slices incubated without CCl4 (0).
Incubation of human liver slices in the presence of CCl4 resulted in a dose-dependent decrease in ATPcontent, indicating toxicity. In vivo liver toxicity of CCl4 is caused by cytochrome P450 mediated conversion of CCl4 into free radicals in the hepatocytes, which yields lipid peroxides and other mediators that can activate HSC [5,6]. In accordance, CCl4 -induced toxicity in human liver slices resulted in a dose-dependent increase of HSC activation in the liver slices, as reflected by the increased mRNA expression of HSP47 and ␣Bcrystallin, two early markers for HSC activation and fibrogenesis [13,14]. We were unable to measure significant changes in protein expression of these markers in human liver slices after incubation with CCl4 . This could be ascribed to differences in sensitivity between real-time PCR and Western blot analysis or to the relatively short incubation times used. The constant expression of synaptophysin and ␣SMA in human liver slices incubated with CCl4 indicates that the observed changes in HSP47 and ␣B-crystallin expression are due to HSC activation, rather than to an increase in the amount of (activated) HSC present in the
liver slices. Increased mRNA expression of HSP47, and ␣B-crystallin is also observed within the first 24 h of spontaneous HSC activation in in vitro cell culture models [15,16]. In contrast, pro-collagen 1a1 and ␣SMA are expressed later during activation of HSC [17,18]. In line with these studies, no differences in mRNA expression of these late markers for HSC-activation were observed in human liver slices after 16 h of incubation with CCl4 . Given the mechanism of action of CCl4 -induced fibrosis, HSC activation occurring in liver slices in response to incubation with CCl4 likely results from a multicellular process. This suggests that, in contrast to current in vitro models, the liver slice system can be used to study multicellular toxicity leading to activation of HSC in a physiologic milieu. However, although there are no indications that CCl4 exerts direct effects on HSC, this has not been ruled out yet. 4.3. Applications The aim of the present study was to evaluate the applicability of human liver slices for the study of early
68
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
Fig. 4. (A) mRNA expression of synaptophysin in human liver slices after 3 and 16 h of incubation relative to expression levels in slices directly after preincubation (0). (B) mRNA expression of synaptophysin in human liver slices after 3 and 16 h of incubation with 0–15 l CCl4 relative to slices incubated without CCl4 (0). The mean of at least six independent experiments ± S.E.M. is shown.
HSC activation. To address this issue we studied whether the liver slices remained viable during incubation and whether HSC activation in the slices could be induced by incubation with the fibrogenic compound CCl4 . Our results show persistent viability of human liver slices during control incubation and give a clear indication that CCl4 -induced toxicity results in HSC activation in the liver slices. Importantly, the response of human liver slices to incubation with CCl4 was reproducible and independent of the origin of human liver tissue, of the age of the donor or the patient, and of the ATP-content in the liver slices after 3 h of incubation, indicating the general applicability of human liver tissue for this type of research. Due to the relatively short incubation time used in our experiments we could not yet detect fibrogenesis, i.e. increased production of collagens, or complete transformation of activated HSC into myofibroblasts as reflected by increased ␣SMA expression. However, we did observe a clear increase in the expression of two early markers for HSC activation and fibrogenesis in response to incubation with CCl4 . It is well known that activation of HSC is the key event in the natural process of wound healing and scar tissue formation as well as in
the development of liver fibrosis. Therefore, we argue that the CCl4 -induced early HSC activation in human liver slices does reflect the onset stage of fibrogenesis. The present model could provide a powerful tool for the study of processes and pathways involved in early HSC activation preceding toxicity-induced fibrogenesis in human liver. In principle it can be employed to compare early responses of human livers to known (non)fibrogenic toxic compounds, among others by using genomic and proteomics tools. It remains to be elucidated whether prolongation of the incubation times also provides a useful system to test the effects of anti-fibrotic drugs on toxicity-induced fibrosis in human liver. 4.4. In conclusion We have developed a technique to reproducibly induce early HSC activation in precision-cut human liver slices. This model provides a multicellular, physiological milieu to study the mechanisms underlying toxicity-induced early HSC activation in human liver tissue. The development of this model may also contribute substantially to the reduction, refinement, and possible
M. van de Bovenkamp et al. / Chemico-Biological Interactions 162 (2006) 62–69
replacement of animal experiments. Finally, the use of human liver tissue in this in vitro model may improve its predictive value for the human condition.
[9]
Acknowledgments M. van de Bovenkamp was supported by grants of ZON/Mw, and the Johns Hopkins Center for Alternatives to Animal Testing; P. Olinga was supported by Organon NV.
[10]
[11]
References [1] S.L. Friedman, Liver fibrosis—from bench to bedside, J. Hepatol. 38 (Suppl. 1) (2003) S38–S53. [2] A.Y. Hui, S.L. Friedman, Molecular basis of hepatic fibrosis, Exp. Rev. Mol. Med. 5 (2003), www.expertreviews.org/ 03005684h.htm. [3] M. Van de Bovenkamp, G.M.M. Groothuis, A.L. Draaisma, M.T. Merema, J.I. Bezuijen, M.J. Van Gils, D.K.F. Meijer, S.L. Friedman, P. Olinga, Precision-cut liver slices as a new model to study toxicity-induced hepatic stellate cell activation in a physiologic milieu, Toxicol. Sci. 85 (2005) 632–638. [4] P. Olinga, G.M.M. Groothuis, Use of human tissue slices in drug targeting research, in: G. Molema, D.K.F. Meijer (Eds.), Drug Targeting: Organ Specific Strategies, Wiley-VCH, Weinheim, 2001, pp. 309–331. [5] S. Basu, Carbon tetrachloride-induced lipid peroxidation: eicosanoid formation and their regulation by antioxidant nutrients, Toxicology 189 (2003) 113–127. [6] K. Ikejima, H. Honda, M. Yoshikawa, M. Hirose, T. Kitamura, Y. Takei, N. Sato, Leptin augments inflammatory and profibrogenic responses in the murine liver induced by hepatotoxic chemicals, Hepatology 34 (2001) 288–297. [7] D. Cassiman, J. van Pelt, R. De Vos, F. Van Lommel, V. Desmet, S.H. Yap, T. Roskams, Synaptophysin: a novel marker for human and rat hepatic stellate cells, Am. J. Pathol. 155 (1999) 1831–1839. [8] M.A. Brouwers, P.M. Peeters, K.P. De Jong, E.B. Haagsma, I.J. Klompmaker, C.M. Bijleveld, J.H. Zwaveling, M.J. Slooff, Sur-
[12] [13]
[14]
[15]
[16]
[17]
[18]
69
gical treatment of giant haemangioma of the liver, Br. J. Surg. 84 (1997) 314–316. P. Olinga, I.H. Hof, M.T. Merema, M. Smit, M.H. de Jager, P.J. Swart, M.J.H. Slooff, D.K.F. Meijer, G.M.M. Groothuis, The applicability of rat and human liver slices to the study of mechanisms of hepatic drug uptake, J. Pharmacol. Toxicol. Methods 45 (2001) 55–63. C.L. Krumdieck, J.E. dos Santos, K.J. Ho, A new instrument for the rapid preparation of tissue slices, Anal. Biochem. 104 (1980) 118–123. M.G.L. Elferink, P. Olinga, A.L. Draaisma, M.T. Merema, K.N. Faber, M.J.H. Slooff, D.K.F. Meijer, G.M.M. Groothuis, LPSinduced downregulation of MRP2 and BSEP in human liver is due to a posttranscriptional process, Am. J. Physiol. Gastrointest. Liver Physiol. 287 (2004) G1008–G1016. S. Azri, A.J. Gandolfi, K. Brendel, Carbon tetrachloride toxicity in precision-cut rat liver slices, In Vitro Toxicol. 3 (1990) 127–138. K.E. Brown, K.A. Broadhurst, M.M. Mathahs, E.M. Brunt, W.N. Schmidt, Expression of HSP47, a collagen-specific chaperone, in normal and diseased human liver, Lab. Invest. 85 (2005) 789–797. D. Cassiman, T. Roskams, J. van Pelt, L. Libbrecht, P. Aertsen, T. Crabbe, H. Vankelecom, C. Denef, Alpha B-crystallin expression in human and rat hepatic stellate cells, J. Hepatol. 35 (2001) 200–207. N. Kawada, T. Kuroki, K. Kobayashi, M. Inoue, K. Nakatani, K. Kaneda, K. Nagata, Expression of heat-shock protein 47 in mouse liver, Cell Tissue Res. 284 (1996) 341–346. A. Lang, L.W. Schrum, R. Schoonhoven, S. Tuvia, J.A. SolisHerruzo, H. Tsukamoto, D.A. Brenner, R.A. Rippe, Expression of small heat shock protein alphaB-crystallin is induced after hepatic stellate cell activation, Am. J. Physiol. Gastrointest. Liver Physiol. 279 (2000) G1333–G1342. P. Mathurin, S. Xiong, K.K. Kharbanda, N. Veal, T. Miyahara, K. Motomura, R.A. Rippe, M.G. Bachem, H. Tsukamoto, IL10 receptor and coreceptor expression in quiescent and activated hepatic stellate cells, Am. J. Physiol. Gastrointest. Liver Physiol. 282 (2002) G981–G990. K. Uchio, B. Tuchweber, N. Manabe, G. Gabbiani, J. Rosenbaum, A. Desmouliere, Cellular retinol-binding protein-1 expression and modulation during in vivo and in vitro myofibroblastic differentiation of rat hepatic stellate cells and portal fibroblasts, Lab. Invest. 82 (2002) 619–628.