Adenoviral expression of a transforming growth factor-β1 antisense mRNA is effective in preventing liver fibrosis in bile-duct ligated rats

Hepatic stellate cells (HSC) were isolated from male Sprague-Dawley rats following a standard procedure with slight modifications [18, 19]. Briefly, livers were perfused with pronase and collagenase and the resulting cell suspensions were filtered through a nylon mesh, centrifuged and washed in ice cold Hanks buffered standard saline (HBSS; PAA Laboratories GmbH, Linz, Austria) containing 0.25% (w/v) BSA. HSC were further purified by a single-step density gradient centrifugation with 8.25% (w/v) Nycodenz^® (Nycomed Pharma, Oslo, Norway) as described in detail elsewhere [20, 21] and seeded in Dulbeccos's modified Eagle medium (DMEM; Bio Whittaker Europe, Verviers, Belgium) supplemented with 10% (v/v) fetal calf serum (FCS; Seromed, Biochrom KG, Berlin, Germany), and 4 mM L-glutamine (ICN Biomedicals Inc., Aurora, Ohio). Additionally, the culture medium was supplemented with penicillin (100 IU/ml) and streptomycin (100 μg/ml).

Isolation and Northern blot analysis of total cellular RNA from HSC was carried out as described previously [19]. Briefly, equal amounts (5 μg) of total RNA were separated by electrophoresis on a 1.2% (w/v) denaturing agarose gel, transferred to a Hybond-N membrane (Amersham Pharmacia, Braunschweig, Germany), and fixed by baking for 2 hours at 80°C. Blots were hybridized with [α-³²P]-dCTP-labeled random primed probes (Amersham) and autoradiographs were exposed to Kodak X-OMAT AR films using intensifying screens. As an internal standard for equal gel loading the blots were re-hybridized with a cDNA specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

Whole cell protein extracts were prepared in RIPA buffer [20 mM Tris-HCl (pH 7.2), 0.15 M NaCl, 2% (v/v) NP-40, 0.1% (w/v) SDS, 0.5% sodium deoxycholate] and concentrations were quantified using the Micro BCA protein assay reagent kit (Pierce, Rockford, IL). Equal amounts of protein were resolved by reducing SDS-PAGE (Novex, Groningen, The Netherlands). For immunoblotting, proteins were electroblotted onto nitrocellulose membranes. Membrane blocking and incubation with antibodies were performed as described previously [17]. α-SMA was detected with the monoclonal mouse antibody clone asm-1 (Roche Diagnostics, Mannheim, Germany) followed by incubation with horseradish peroxidase (HRP)-conjugated anti-mouse IgG (Santa Cruz Biotech., Santa Cruz, CA) and the supersignal chemiluminescent substrate (Pierce).

The recombinant E1-deleted adenoviral vectors constitutively expressing the enhanced green fluorescent protein (EGFP) or rat antisense TGF-β1 were generated as described previously [17, 22]. Adenoviral particles were purified by a standard CsCl density protocol followed by a membrane based ion exchange chromatography for purification of adenoviral particles (BD Bioscience, CLONTECH, Palo Alto, CA). Adenoviral titers were spectrometrically determined and appropriate aliquots were stored at -80°C until use.

Male Sprague-Dawley rats (245.6 ± 10.8 g) were injected twice (at day 0 and day 7) with recombinant adenoviruses (1 × 10¹⁰ pfu/kg) or phosphate buffered saline via the tail vein. At day two, the common bile ducts were double ligated under halothane anesthesia. Rats were sacrificed after 12 days, and pieces of the livers were fixed in 10% formalin for histological examination or frozen immediately in liquid nitrogen and stored at -80°C for RNA isolation. Measurements of AST, ALT, and bilirubin were performed from blood samples following standard protocols. The study as presented is in compliance with the German Animal Protection Act, and was approved by the local committee for care and use of laboratory animals at the RWTH-University Hospital. Experiments applying adenoviral constructs to cells or animals are covered by permission of the Landesumweltamt Nordrhein-Westfalen (Az. 521-K-1.59/99).

Morphological evaluation of induced liver fibrogenesis was performed using the semi-quantitative fibrotic focus score proposed for staging and grading of histopathological lesion of chronic hepatitis in humans [23]. Briefly, pieces of left liver lobules were fixed in 10% formalin and stained with hematoxylin/eosine for identification of lesions. In the stained sections, the lesions were defined as follows: absence of lesions = 0; occasional small localized lesions with fibrous expansion of some portal areas = 1; thickening of liver septa with fibrous expansion of most portal areas = 2; and thickened continuous fibrous areas with periportal rounding and occasional portal to portal bridging = 3. The liver sections were coded and independently examined in a blinded manner by two different pathologists.

Liver tissues were fixed in 4% paraformaldehyde and embedded in paraffin and 1.5-μm sections were prepared. For immunohistochemistry, the sections were treated with xylene and rehydrated with decreasing graded ethanol washes. For α-SMA staining, the slides were treated with a polyclonal anti-rat α-SMA (Roche Diagnostics) followed by HRP-conjugated secondary anti-rabbit antibody (DAKO, Hamburg, Germany). α-SMA staining was detected using the DAB substrate (Vector Laboratories, Burlingame, CA) and sections were counterstained with eosine.

Collagen staining was performed as described previously [24]. Briefly, liver sections were deparaffinized and the slides were incubated for 30 min in a solution of saturated picric acid containing 0.1% direct red 80 and 0.1% Fast Green FCF. Stained slides were washed in running distilled water, dehydrated, mounted, and examined by light microscopy.

We cloned the adenoviral construct Ad5-CMV-AS-TGF-β1 expressing an antisense (AS) complementary to the 3'-portion of rat TGF-β1 mRNA and a control virus expressing EGFP [17, 22]. Both transgenes are driven by the human cytomegalovirus (CMV) immediate-early gene 1 enhancer/promoter and are fused to the SV40 early mRNA polyadenylation signal (Fig. 1A). In culture-activated HSC, Ad5-CMV-AS-TGF-β1 is able to direct high-level expression of the antisense mRNA (Fig. 1B, top). In agreement with the functionality of the transgene the level of colα1(I) mRNA, a classical marker of fibrosis, was significantly reduced in transdifferentiating HSC infected with the transgene (Fig. 1B, middle). The reduction of colα1(I) mRNA was specific for the expressed transgene and was not observed in cells infected with other recombinant adenoviruses. To assess the relative abundance of the expressed AS-mRNA over endogenous TGF-β1 transcripts we established SYBR Green I-based real-time quantitative PCR procedures. The assays were conducted on a LightCycler (LC) and amplification of transcripts for TGF-β or both, TGF-β- and AS-TGF-β, were performed by specific primer combinations (Fig. 1C). Compared to mock- or control virus-infected cells, the obtained fluorescent signals were higher in HSC infected with Ad5-CMV-AS-TGF-β1. Typically, the logarithmic amplification phase in samples taken from these cells were reached two cycles earlier suggesting that endogenous TGF-β transcripts were approximately four fold more abundant in these cells (Fig. 1D). This finding is consistent with our previous observation that the endogenous TGF-β1 mRNA is stabilized by the transgene most likely because under cellular environment mRNA/mRNA hybrids are resistant to the attack of ribonuclease H, the key enzyme of ribonucleic acid degradation. Simultaneous amplification of both transcripts (TGF-β and AS-TGF-β) revealed that equal fluorescent signals were obtained approximately 13 cycles earlier in samples of Ad5-CMV-AS-TGF-β1-infected cells than in the controls (untreated HSC or HSC infected with Ad5-CMV-EGFP). Provided that the PCR efficiencies during the logarithmic amplification phases in each probe equals two, this points to an 8000-fold excess of the AS-mRNA relative to endogenous TGF-β1 mRNA in these cells. With the knowledge about the high abundance of the antisense and the recent demonstration that the transgene is able to block synthesis of TGF-β1 in vitro [17], we next asked whether the transgene is useful for the treatment of hepatic fibrosis.

Hepatic fibrosis in rats can be experimentally induced by double ligature of the common bile duct (BDL) for several weeks or months resulting in morphological changes in liver tissue comparable to those seen in human biliary fibrosis [25‐27]. Since the first reports of this experimental model for liver injury it has been used in many laboratories to study the cellular and biological changes during liver fibrogenesis. The prolonged bile duct obstruction has many advantages over the experimentally induced cirrhosis generated by multiple doses of the CCl₄ or dimethylnitrosamine (DMN), because the individual response of animals to these hepatoxins is variable, the mortality is high, and it takes a relatively long time to produce persistent fibrosis or cirrhosis. However, to validate new therapeutic strategies in this obstructive model and to allow reproducibility between different laboratories it is necessary to follow standard operating procedures. In our laboratory we use male rats of the Sprague-Dawley strain weighed approximately 250 g that are fed with a standard chow and having free access to trap water. Surgeries for ligature of the common bile duct are performed under deep Ketamin-Rompun anesthesia. Following our protocol, the abdominal wall of the animals is first shaved and the abdomen is cut with a scalpel through a median line (Fig. 2A). Then, the abdominal cavity is opened, and selvages are stretched and fixed with two sterile high-grade steel tweezers (Fig. 2B). Afterwards, the liver lobules are turned down and intestines are carefully pulled out to uncover the bile duct (Fig. 2C). The extrahepatic common bile duct is pulled out, fixed with forceps, and obstructed with a piece of string (Fig. 2D). A second ligature is introduced in a distance of approximately 1 cm (Fig. 2E). Liver is replaced into its correct position within the abdominal cavity and the abdomen is closed with sutures (Fig. 2F). Sham-operated rats are treated in the same manner except that the bile duct is not ligated. After 12 days, tremendous differences between sham-operated and BDL-ligated rats in regard to their liver appearance are obvious (compare Figs. 2G, 2H). In general, the obstruction induces bile duct proliferation and inflammatory/fibrotic reactions with concomitant formation of oedema and fibrotic nodules on the surface of corresponding livers. To monitor ongoing fibrogenesis we routinely determine the concentrations of bilirubin, and the activities of serum AST and ALT from blood samples aspirated from the tail vein. In normal controls (n = 65) levels for individual parameters were measured as 0.14 ± 0.05 mg/dl (bilirubin), 53.49 ± 17.31 U/l (AST), and 28.18 ± 7.84 U/l (ALT), respectively (mean ± SD). In the course of obstructive cholestasis, the values of individual animals were highest at day 2 after ligature and were still increased by factors of approximately 100 (bilirubin), 6 (AST), and 5 (ALT) at day 12 compared to sham-operated controls.

Recent reports describe that the transduction efficiency by adenoviral mediated gene transfer is strongly reduced after liver injury, but remained significant [28, 29]. Therefore, we decided to administrate the adenoviruses two days prior setting of the injury followed by a booster injection at day 7 (5 days after BDL surgery). For histology, blood sampling, and RNA/protein extraction animals were sacrificed at day 14 corresponding to day 12 after surgery (Fig. 3A). As documented in previous studies, the administration of adenoviruses is an effective means to direct transgene expression in liver as assessed by visual detection of transduced EGFP under UV-light (not shown). However, compared to normal controls, transgene expression as assessed by delivery of an EGFP-reporter was drastically reduced in the BDL cirrhotic rats (Fig. 3B). To confirm successful transduction into liver and proper expression of the antisense we performed PCR allowing the amplification of a 225-bp fusion containing sequences specific for TGF-β1 and the SV40-polyadenylation signal (Fig. 3C). We next tested whether Ad5-CMV-AS-TGF-β1 infection was able to suppress liver fibrogenesis in BDL-treated rats. Therefore, we first quantified the static levels of collagen by Sirius red staining (Fig. 4). Impressively, the deposition of fibrillar collagen was substantially decreased in response to administration of the antisense. This inhibitory effect was observed in all areas of the liver, with no significant difference noticed between the different lobules (not shown) while no reduction in fibrosis was observed in Ad5-CMV-EGFP-treated rats. Also the tendency for proliferation of biliary ductules in portal areas was significantly reduced in injured livers expressing the antisense (compare Figs. 4C and 4E). The morphological evaluation of representative liver sections by two independent pathologists revealed fibrosis scores with increasing severity ranging from 0 (no BDL, saline), 1 (BDL, Ad5-CMV-AS-TGF-β1), and 3 (BDL, Ad5-CMV-EGFP), respectively. Additionally, we demonstrate the antifibrotic capacity of the transgene by its inhibitory effects on α-SMA expression. In livers of BDL-treated rats, the immunoreactivity for α-SMA as assessed by immunohistochemistry was significantly reduced by treatment with the TGF-β1 antisense construct (Figs. 5A,5B,5C,5D). To confirm and demonstrate the lowering in α-SMA expression more quantitatively, we further performed western blot analysis showing strong suppression of α-SMA in animals treated with the transgene (Fig. 5E). Again, these results indicate that Ad5-CMV-AS-TGF-β1 specifically inhibits ongoing fibrosis in liver.

In previous reports, the elimination of TGF-β signaling by adenovirus-mediated local expression of a dominant-negative type II TGF-β receptor in the liver of rats treated with dimethylnitrosamine was associated with serum levels of liver specific transaminases comparable to normal controls [11]. When we measured AST and ALT in the serum of sacrificed rats which received BDL and gene transfer, the levels of these hepatic enzymes were the same regardless of whether the BDL-treated rats were infused with saline, Ad5-CMV-EGFP, or Ad5-CMV-AS-TGF-β1 indicating that the hepatic injury induced by BDL per se is not influenced by TGF-β under the chosen experimental conditions.