Clones expressing the highest levels of RsTβRII-IFN-γ fusion protein were selected for production of the fusion protein. RsTβRII-IFN-γ fusion protein was purified from culture supernatant by protein A-Sepharose affinity chromatography under sterile and endotoxin free conditions. The protein is greater than 95% pure as judged by sodium dodecyl sulfate/polyacrylamide gel electrophoresis and contains less than 1 unit endotoxin per milligram protein.

Antiviral assay was performed as described by Pan et al. [²⁰] with minor modification. 1 × 10⁴ L929 cells (100 μl) per well were added to a 96-well microtiter tissue culture plate. The cells were incubated at 37°C in 5%CO₂ condition and 100 μl serial dilutions of RsTβRII-IFN-γ fusion protein in which RsTβRII activity has been blocked by its neutralizing antibody were added to the assay plate except for the normal and viral controls after the cells adhered to wells. After 3 hours, 25 μl of vesicular stomatitis virus (VSV) was added to the assay plate except for the normal controls. The cells were incubated for 24 hours totally and 10 μl MTT (5 mg/ml) were added to each well 6 hours before the end of the culture. After centrifugation at 1000 rpm for 5 minutes, supernatants were removed. The cells were then lysed in 150 μl dimethyl sulfoxide (DMSO). The absorbance at 570 nm was measured and percent protection of L929 cells from infection of VSV was calculated. Each assay was carried out in triplicate.

The Mv1Lu cell line, CCL64 cells, were maintained in minimal essential medium supplemented with 100 units/ml penicillin, 100 μg/ml streptomycin, and 10% fetal bovine serum. Serial dilutions of RsTβRII-IFN-γ fusion protein in which IFN-γ activity has been blocked by its neutralizing antibody were incubated with 800 pg/ml TGF-β1 for one hour in assay medium in a 96-well plate. 1 × 10⁴cells of CCL64 were added into each well. The cells were incubated at 37°C for 72 hours and 10 μl MTT (5 mg/ml) were added into each well 4 hours before the end of the culture. After centrifugation at 1000 rpm for 5 minutes, supernatants were removed. The cells were then lysed in 150 μl DMSO. The absorbance which reflects cell proliferation was measured at 570 nm. Each assay was carried out in triplicate.

RsTβRII genes were amplified from rat liver cDNA with specific primers as previously described [¹⁹]. The RsTβRII fragment was digested with HindIII/BamHI restriction enzymes, then cloned into the pSecTag2 expression vector. The ultimate construct was screened by ampicillin and identified by double enzyme digestion and DNA sequencing. Transfection of the pSecTag2/RsTβRII into CHO was conducted with liposomes and the stable expression strains were screened with Zeocin^®. The supernatants of transfected CHO were examined by Western blotting for RsTβRII expression. The purification and activity testing of the RsTβRII protein were performed as described for RsTβRII-IFN-γ fusion protein.

Sprague-Dawley rats (n = 210, provided by Zhejiang University Experimental Animal Center) weighing 135 g-150 g were used. Principles of laboratory animal care (NIH publication No. 86-23, revised 1985) were followed.

Hepatic fibrosis was induced by subcutaneous injection of CCl4 as described previously [²¹] with minor modification. Rats (n = 200, in addition, 10 rats were chosen as the normal control group, named group Nor.) were subcutaneously administered dissolved CCl4 in camellia oil (a proportion of 1:1) at 0.3 ml/100 g of body weight, twice a week for 8 weeks. Five rats were killed at the end of each week to examine pathological changes of livers.

According to histological examination, hepatic fibrosis appeared at the end of the eighth week from the start of the study. 156 rats with hepatic fibrosis could be used for the study. Model rats were randomly divided into five groups: 1) fibrotic model group (n = 10) (group E, each rat, 100 μl·of 0.9% NaCl day^-1, i.m.); 2) RsTβRII-IFN-γ fusion protein treatment group (group A, each rat, 0.136 mg (containing 7.5 MU IFN-γ and 0.048 mg RsTβRII)· day^-1, i.m.); 3) IFN-γ treatment group (group B, each rat, 7.5 MU· day^-1, i.m.); 4) RsTβRII protein treatment group (group C, each rat, 0.048 mg· day^-1, i.m.); and 5) mixture of IFN-γ and RsTβRII treatment group (group D, each rat, IFN-γ 7.5 MU· day^-1+ RsTβRII0.048 mg· day^-1, i.m.).

The latter four groups were given, every two days for totally 8 weeks, RsTβRII-IFN-γ fusion protein, IFN-γ, RsTβRII protein, mixture of IFN-γ and RsTβRII protein, respectively. Rats in the fibrotic model group were given 0.9% sodium chloride by intramuscular injection. Rats in the normal control group didn't receive any treatment.

Four weeks after the finish of treatment, rats of each group were killed, and histological sections of their livers were evaluated by hematoxylin and eosin and Sirius red stains. Expressions of collagen III (Col III), α-smooth muscle actin (α-SMA), TGF-β1 and TGF-βRII were examined with immunohistochemistry, Western blotting and real-time RT-PCR.

Livers were fixed in 10% neutral formalin, embedded in paraffin, sectioned, and stained with hematoxylin-eosin and Sirius red respectively. Sirius red staining method was used to specifically stain fibrous tissue components. For Sirius red staining, the sections were incubated for 30 minutes in 0.1% Sirius red F3B containing saturated picric acid and 0.1% Fast Green (Sigma Chemical Co., St. Louis, MO). After rinsing twice with distilled water, sections were briefly dehydrated with anhydrous alcohol and coverslipped. Collagen stained with Sirius red was quantitated in the sections that were randomly chosen (under 20× magnification; 10 fields each from 8 different rats for a total of 80 fields for each group) with Meta-View software (Universal Imaging Corp., Downington, PA). These sections were observed under polarization microscope (Leica DMLB, Leica Wetalar, Germany) to distinguish type I and III collagen [²²].

For measurement of Col III, α-SMA and TGF-β1, liver tissues were fixed in formalin, embedded in paraffin wax, and 5-μm sections were cut and stained using the immunohistochemical Streptavidin biotin-peroxidase complex (SABC) method as described previously with minor modification [²³,²⁴].

Frozen rat's liver tissue, stored at -80°C, were shattered in liquid nitrogen and total RNA was isolated by the RNAeasy isolation kit (Qiagen) according to the manufacturer's instructions. First-strand cDNA was synthesized. Real-time reverse transcriptase (RT)-PCR was performed as described by Li et al. [²⁵]. Briefly, a real-time PCR was performed using Light Cycler Fast-Start SYBR Green I Master (Roche) according to the manufacturer's instructions. Synthesized cDNA then served as a template in the following quantitative PCR reaction. Primer sequences were designed as follows: Col III (220 bp): forward 5'-TGGTCCTCAGGGTG TAAAGG-3', reverse 5'-GTCCAGCATCA CCTTT TGGT-3'; α-SMA (457 bp): forward 5'-AGGGAGTA ATGGTTGGAATGGG -3', reverse 5'-GGAGTACG GTACGC AGA-3'; TGF-β1 (404 bp): forward 5'-CA CCAT CCATGACATGAACC-3', reverse 5'-TCATG TTGGACAACTGCTCC-3'; and TGF-βRII (560 bp): forward 5'-CGTGTGGAGGAAGAACAACA-3', reverse 5'-TCTC AAACTGCTCTGAGGTG-3'; β-actin (619 bp): forward 5'-CGCTG CGCTGGTCGTC GA CA-3', reverse 5'-GTCACGCACGATTTCCCGCT-3'. β-actin was used as an internal standard. Each Light Cycler PCR reaction system (15 μl) included: 1.5 μl of Light Cycler FastStart DNA Master Mix, 1.8 μl of MgCl2 (25 mM), 0.4 μl of each primer (10 μM), 2 μl of 1:10 diluted cDNA, 9.3 μl of H₂O. The reactions started at 95°C for 10 minutes followed by 50 cycles at 94°C for 10 seconds, 65°C for 15 seconds, 72°C for 15 seconds. PCR products were denaturalized for 1 min at 95°C before rapidly chilling to 55°C for 30 seconds. Melting peaks of PCR products were determined by heat-denaturing them over a 40°C temperature gradient at 0.1°C/s from 55°C to 95°C. Reaction specificity was confirmed by 1.5% gel electrophoresis of products after real-time PCR and melting curve analysis. Ratios of Col III/β-actin, α-SMA/β-actin, TGF-β1/β-actin, TGF-βRII/β-actin, and mRNA were calculated for each sample and expressed as the means ± SEM.

Western blot analysis of Col III, α-SMA, TGF-β1, TGF-βRII and β-actin in the liver was performed as previously described [²⁶].

Anti-viral activity is one of the biological characteristics of IFN-γ [¹⁴,²⁷]. We tested if RsTβRII-IFN-γ fusion protein could protect L929 cells from infection of VSV by anti-viral assay. We found that L929 cells quickly exhibited typical signs of VSV infection in the absence of RsTβRII-IFN-γ, however, in the presence of RsTβRII-IFN-γ, L929 cells became insensitive to VSV challenge, and this protective effect is dose-dependent (Figure 1), indicating that RsTβRII-IFN-γ fusion protein retains the anti-viral activity of IFN-γ.

It was demonstrated that sTGF-βRII could antagonize the bioactivity of TGF-β1 to inhibit the proliferation of mink lung epithelial cells [¹⁴]. We tested if RsTβRII-IFN-γ has the bioactivity of sTGF-βRII by a mink lung epithelial cell proliferation assay. As shown in Figure 2, when TGF-β1 was added to the media, proliferation of CCL-64 was inhibited. While different concentration of RsTβRII-IFN-γ fusion protein and same quantity of TGF-β1 were incubated together with CCL-64 cells for 72 hours, TGF-β1 mediated inhibition of proliferation was blocked by RsTβRIIIFN-γ fusion protein in a dose-dependent manner, suggesting that RsTβRII-IFN-γ retains RsTβRII bioactivity.

To identify the bioactivity of the RsTβRII, we performed a mink lung epithelial cell proliferation assay. As shown in Figure 3, RsTβRII dose-dependently inhibited the growth-inhibitory effects of TGF-β1 on CCL64 cells, implying that RsTβRII we prepared could competitively bind to TGF-β1 with the membrane TGF-βRII, hereby inhibiting the bioactivity of TGF-β1.

Histological examination of livers from normal rats (group Nor) demonstrated that ECM deposition was present only in portal and central areas (Figure 4 Nor). Hematoxylin-eosin staining of liver sections from untreated rats with hepatic fibrosis (group E) exhibited periportal, pericentral, and central-central, portal-portal, and portal-central matrix deposition and numerous granulomas containing a diffuse array of inflammatory cells, within an invasive collagen network (Figure 4E). In contrast, livers from rats with hepatic fibrosis treated with RsTβRII-IFN-γ fusion protein, IFN-γ, RsTβRII, and mixture of IFN-γ and RsTβRII, respectively (group A, B, C and D) showed much less granulomas in which only a few cells were found, and markedly decreased collagen (Figure 4A, B, C and 4D).

Quantitative morphometric assessment of fibrosis (identified by Sirius red staining) showed a significant reduction in fibrogenesis after treatment of rats with RsTβRII-IFN-γ fusion protein, IFN-γ, RsTβRII, and mixture of IFN-γ and RsTβRII, respectively (Figures 4A, B, C and 4D). Sirius red staining showed that a little collagen deposition in the portal tracts and lobules was seen in livers from rats with hepatic fibrosis treated by RsTβRII-IFN-γ fusion protein. Furthermore, only collagen fibers around the terminal hepatic veins were observed (Figure 4A). Collagen depositions in liver sections from rats treated with IFN-γ, RsTβRII or RsTβRII+IFN-γ. respectively were particularly evident, but only thin bands of collagen which formed short, incomplete septum could be seen (Figures 4B, C and 4D).

Taken together, the most effective anti-fibrotic effect was achieved in model rats treated with RsTβRIIIFN-γ fusion protein.

Immunohistochemistry analyses of the liver further revealed the striking differences. As shown in Figures 5A, in the normal control group (Group Nor), α-SMA positive cells were detected mainly in the portal space as elements of vascular walls or as fibroblastlike cells scattering in the connective tissue or near bile ductules. In lobule, α-SMA positive cells were present on the walls of large and medium sized terminal hepatic veins. The pattern of distribution of α-SMA positive cells was modified in liver injured rats as compared with the normal control rats. In rats with hepatic fibrosis (group E), α-SMA positive cells were detected in portal space, sinusoid, lobule and areas where fibrotic septum appeared. Distributions of α-SMA positive cells in liver from rats with hepatic fibrosis treated with RsTβRIIIFN-γ fusion protein, IFN-γ, RsTβRII or RsTβRII +IFN-γ respectively were all reversed when compared with those of hepatic fibrosis model rats. After 8 weeks of RsTβRII-IFN-γ fusion protein treatment (group A), only thin and incomplete parenchymal α-SMA positive septum joining thickened centrilobular veins were observed in liver sections. α-SMA positive cells were mainly found in portal space and areas around fibrotic septum. Few α-SMA positive cells were present in sinusoid and lobule (Figure 5A). Moreover, it was obvious that the greatest anti-fibrotic effect was achieved by treatment with RsTβRII-IFN-γ fusion protein in accordance with the histopathological analysis (Figure 4A). The deposition and distribution range of positive staining for Col III in livers from rats with hepatic fibrosis treated by RsTβRII-IFN-γ fusion protein mirrored the same trend as α-SMA (Figure 5B).

As TGF-β1 is an profibrogenic cytokine that signals through a receptor consisting of type I and type II components [²⁸], the expression of TGF-β1 in livers from different groups was also examined by immunohistochemistry. As shown in Figure 5C, normal rat livers showed light, positive staining which was scattered in the perivascular spaces, around portal tracts. Livers from rats with hepatic fibrosis showed a marked increase in positive staining for TGF-β1, with the reactivity primarily occurring in the portal tract area and granulomas. However, livers from rats with hepatic fibrosis treated with RsTβRII-IFN-γ fusion protein, IFN-γ, RsTβRII, and mixture of IFN-γ and RsTβRII respectively (group A, B, C and D) showed decreased expression of TGF-β1. Moreover, liver sections from rats with hepatic fibrosis treated with RsTβRII-IFN-γ fusion protein (group A) showed marked decrease in expression of TGF-β1 when compared with those from rats with hepatic fibrosis treated with IFN-γ, RsTβRII and mixture of IFN-γ and RsTβRII, respectively (group B, C and D) although the distribution pattern of positive cells was similar to that of the hepatic fibrotic model group (group E).

Western blotting analyses of collagen III, α-SMA, TGF-β1 and TGF-βRII content in livers of different groups also substantiated the histological impression, showing decreased contents of Col III, α-SMA, TGFβ1 and TGF-βRII in livers from rats with hepatic fibrosis treated by RsTβRII-IFN-γ fusion protein, IFNγ, RsTβRII, and mixture of IFN-γ and RsTβRII, respectively (Figure 6). However, the most effective therapeutic effect was achieved again by treatment with RsTβRII-IFN-γ fusion protein.

Finally, we investigated the mRNA expression of Col III, α-SMA, TGF-β1 and TGF-βRII in livers of different groups. As shown in Figure 7, 4 weeks after the stop of treatment, there was an increase in TGFβ1 mRNA expression in livers from rats with hepatic fibrosis, while only detectable level of TGF-β1 mRNA was shown in livers from normal control rats. This increase in TGF-β1 steady-state mRNA levels approximately paralleled the increase in Col III, αSMA and TGF-βRII mRNA levels. Moreover, livers from rats with hepatic fibrosis treated with RsTβRIIIFN-γ fusion protein (group A) showed markedly decreased mRNA expression of Col III, α-SMA, TGFβ1 and TGF-βRII compared with untreated, hepatic fibrotic rats. In contrast, RsTβRII, IFN-γ or RsTβRII+IFN-γ treatment did not significantly alter steady-state mRNA levels for TGF-β1 and TGF-βRII (Figure 7A and 7B). These data substantiated again the histological analyses, and demonstrated that RsTβRIIIFN-γ fusion protein treatment did inhibit collagen deposition.