Targeted Recombinant Fusion Proteins of IFNγ and Mimetic IFNγ with PDGFβR Bicyclic Peptide Inhibits Liver Fibrogenesis In Vivo

Hepatic stellate cells (HSCs), following transdifferentiation to myofibroblasts plays a key role in liver fibrosis. Therefore, attempts to attenuate this myofibroblastic phenotype would be a promising therapeutic approach. Interferon gamma (IFNγ) is a potent anti-fibrotic cytokine, but its pleiotropic receptor expression leading to severe adverse effects has limited its clinical application. Since, activated HSC express high-level of platelet derived growth factor beta receptor (PDGFβR), we investigated the potential of PDGFβR-specific targeting of IFNγ and its signaling peptide that lacks IFNγR binding site (mimetic IFNγ or mimIFNγ) in liver fibrosis. We prepared DNA constructs expressing IFNγ, mimIFNγ or BiPPB (PDGFβR-specific bicyclic peptide)-IFNγ, BiPPB-mimIFNγ fusion proteins. Both chimeric proteins alongwith IFNγ and mimIFNγ were produced in E.coli. The expressed proteins were purified and analyzed for PDGFβR-specific binding and in vitro effects. Subsequently, these recombinant proteins were investigated for the liver uptake (pSTAT1α signaling pathway), for anti-fibrotic effects and adverse effects (platelet counts) in CCl4-induced liver fibrogenesis in mice. The purified HSC-targeted IFNγ and mimIFNγ fusion proteins showed PDGFβR-specific binding and significantly reduced TGFβ-induced collagen-I expression in human HSC (LX2 cells), while mouse IFNγ and mimIFNγ did not show any effect. Conversely, mouse IFNγ and BiPPB-IFNγ induced activation and dose-dependent nitric oxide release in mouse macrophages (express IFNγR while lack PDGFβR), which was not observed with mimIFNγ and BiPPB-mimIFNγ, due to the lack of IFNγR binding sites. In vivo, targeted BiPPB-IFNγ and BiPPB-mimIFNγ significantly activated intrahepatic IFNγ-signaling pathway compared to IFNγ and mimIFNγ suggesting increased liver accumulation. Furthermore, the targeted fusion proteins ameliorated liver fibrogenesis in mice by significantly reducing collagen and α-SMA expression and potentiating collagen degradation. IFNγ also induced reduction in fibrogenesis but showed significant decrease in platelet counts, which was restored with targeted proteins. These results suggest that these rationally designed proteins can be further developed as novel anti-fibrotic therapeutics.


Introduction
IFNc is a pleiotropic homodimeric Th1 cytokine mainly produced by activated inflammatory cells and has been documented to be highly effective in viral, immunological and malignant diseases [1,2]. IFNc (Interferon gamma-1b) has also been explored in clinical trials in patients suffering from liver fibrosis, renal fibrosis or idiopathic pulmonary fibrosis [3][4][5][6]. However, despite its promising effects in vivo, all clinical trials failed due to a lack of efficacy and unfavorable adverse effects [4,7,8].
The clinical application of this potent cytokine is nowadays limited to only a very few diseases. Many attempts have been made to prolong the half-life of IFNc by PEGylation or by increasing its activity through slow release by incorporation in nanoparticles, elastomers, microspheres or liposomes [9][10][11]. These approaches have shown to be beneficial, but adverse effects due to the longer exposure of IFNc to non-target tissues can still be detrimental. Therefore, targeted approaches leading to an increased therapeu-tic efficacy without eliciting adverse effects would be ideal to treat slowly progressing chronic diseases [12]. Liver fibrosis, induced by viral infections (e.g hepatitis B and C), alcohol abuse, metabolic syndrome or genetic disorders, is characterized by an excessive accumulation of matrix proteins in the liver [13,14]. Worldwide millions of people, suffering from one of these disorders, are at risk of developing liver fibrosis. Currently, there are no effective and clinically approved anti-fibrotic therapy available and the treatment is mainly based on the removal of the underlying cause of the disease [15,16]. Liver transplantation is the only option for the patients suffering from advanced liver fibrosis or end stage liver cirrhosis.
Hepatic stellate cells (HSC) are the key pathogenic cells involved in the progression of liver fibrosis. These cells gets activated following release of growth factors from damaged hepatocytes, kupffer cells and infiltrating inflammatory cells. These activated HSC are then transformed into proliferative and contractile myofibroblast-like cells that produce large amounts of extracellular matrix (ECM) proteins leading to impairment of the structure and function of liver [17,18]. Among the potent anti-fibrotic therapeutic cytokines, Interferon gamma (IFNc) is shown to be highly efficacious in vitro and in vivo in liver fibrosis models [19], but it failed in clinical trials due to reduced efficacy and unwanted systemic effects [8].
Others and we have shown that activated HSC abundantly express the platelet derived growth factor receptor (PDGFbR) during liver fibrosis, while its expression is relatively weak on other cells and normal tissues [20][21][22]. Recently, we have shown that using PDGFbR-specific delivery of IFNc to activated HSC; acute and advanced liver fibrosis in vivo could be significantly inhibited with minimal adverse effects [20,23]. The results of these chimerical constructs of IFNc and PDGFbR binding moieties were remarkably potent and encouraged us to pursue this strategy and prepare a targeted fusion proteins that are inexpensive and can be feasibly applied in clinical trials.
To that end, we have now produced the recombinant proteins containing a bicyclic PDGFbR-recognizing peptide (BiPPB) fused to IFNc to synthesize BiPPB-IFNc or to the signaling moiety of IFNc (mimetic IFNc or mimIFNc) lacking extracellular IFNcR binding site [24,25] to generate BiPPB-mimIFNc in E.coli. IFNc and mimetic IFNc were also expressed in parallel as control proteins. These proteins were analyzed in vitro in human HSC cells and in vivo in acute liver fibrogenesis mouse model. Encouragingly, we found that the targeted fusion proteins (BiPPB-IFNc and BiPPB-mimIFNc) specifically bound to PDGFbR-expressing human HSC and induced significant reduction in major ECM production in vitro (collagen). In vivo, in the CCl 4 mouse model, targeted proteins significantly stimulated the IFNc-mediated pSTAT1a-signaling pathway, inhibited collagen accumulation, HSC activation and induced fibrolysis. Apart from therapeutic effects, BiPPB-mimIFNc and to a lesser extent BiPPB-IFNc did not affect circulating platelets counts as observed by untargeted IFNc.

Materials and Methods
Plasmids, Bacterial Strains and Cell Culture pET42a and pET39b protein expression vectors were purchased from Novagen (Madison, WI, USA). All the restriction enzymes were procured from New England Biolabs (Beverly, MA).
Escherichia coli strain JM109 was used for plasmid propagation and cloning. Strain BL21 (DE3) (Novagen) was used as a host for the production of recombinant proteins. Human hepatic stellate cells, LX2 were kindly provided by Prof. Scott Friedman (Mount Sinai Hospital, New York). LX2 cell line is a well-established human HSC cell line [26]. LX2 cells were cultured in DMEM-Glutamax (Invitrogen, Carlsbad, CA) supplemented with 10% FBS and antibiotics (50 U/ml penicillin and 50 ng/ml streptomycin). Mouse spleen cells, freshly isolated from healthy C57BL/6 mice, were grown in DMEM cell culture medium.

Plasmid Construction
(a) Preparation of pET42a-IFNc and pET42a-mimIFNc. Splenocytes (freshly isolated from the spleen of healthy C57BL/6 mice) were seeded in the presence of 20 mg/ml phytohemagglutinin (PHA) for 24 h. RNA was isolated from spleen cells was used for PCR amplification of mouse IFNc and mimetic IFNc using gene specific primers (primer 1, primer 2 for IFNc and primer 3, primer 4 for mimIFNc as listed in Table 1). The PCR product was purified and digested with Eco RI. The digested gene was inserted into Psh A1/Eco RI digested pET42a vector to produce pET42a-IFNc or pET42a-mim IFNc ( Figure  S1 and S2).

Expression and Purification of Recombinant Proteins
DNA constructs were transformed into CaCl 2 competent BL21 (DE3) and then induced with 1 mM IPTG at 37uC for 4 h.
For purification of IFNc and mimIFNc, the cell pellets were washed and suspended in binding buffer (500 mM NaCl, 20 mM sodium phosphate buffer, 20 mM Imidazole pH7.4), followed by enzymatic lysis (0.2 mg/ml lysozyme, 20 mg/ml DNase, 10 mM PMSF) for 30 min at 4uC and sonication for 10 min (10-15 short bursts). The cell debris was removed by centrifugation at 12,000 g for 30 min at 4uC. The supernatant was extensively dialyzed against binding buffer.
For purification of periplasmic proteins BiPPB-IFNc and BiPPB-mimIFNc, the cell pellets were suspended in 0.03 mol/l Tris.HCl, 20% sucrose, 0.001 mol/l EDTA, pH8.0. The cells were incubated on ice for 5-10 min followed by centrifugation at 8000 g for 20 min at 4uC. The pellet was then resuspended in icecold 5 mM MgSO 4 and stirred for 10 min on ice bath. The cell debris was removed by centrifugation at 8000 g for 20 min at 4uC. The supernatant was extensively dialyzed against binding buffer.
The supernatants were then applied to Ni-charged chelating-Sepharose HisTrap High performance column (Pharmacia Biotech, Uppsala, Sweden). After being washed with binding  buffer, the protein was eluted with the elution buffer containing 500 mM imidazole and then dialysed against PBS buffer overnight.

Dot Blot Immunoassay
Purified proteins were applied on dehydrated PVDF membrane (Roche, Mannheim, Germany) in dot-blot apparatus (Bio-Rad, Hercules, CA, USA). The wells were then incubated with 200 ml blocking solution (1% BSA in TBS) for 1 h. After washing in TBST (0.05% tween 20 in TBS), the membrane was incubated for 1 h with either anti-IFNc antibody (1:2000; Abcam, Cambridge, UK) or anti-PPB antibody (1:1000, custom-made, Harlan) followed by 1 h incubation with horseradish peroxidase-labeled secondary antibody. The membrane was washed again with TBST subsequent with incubation in substrate solution (0.06% diaminobenzidine (DAB), 0.08% hydrogen peroxide in TBS) for color development. The membranes were finally rinsed in distilled water and air-dried.

In vitro Effects of Recombinant Proteins in Human HSC
Cells were plated in 24 well (30,000 cells/well) and 12 well (75,000 cells/well) culture plates, grown overnight and were starved with 0.5% containing medium for 24 h. Starved cells were then incubated with medium alone, recombinant IFNc, mi-mIFNc, BiPPB-IFNc or BiPPB-mimIFNc (equivalent to 1 mg/ml) plus 5 ng/ml of human recombinant TGFb1 (Roche) for 48 h. Subsequently, cells (24 well plates) were fixed and stained for collagen I (1:100, Southern Biotech, Birmingham, AL). In addition, cells (12 well plates) were lysed with 5x SDS-PAGE sample buffer constituted with b-mercaptoethanol (Stratagene, La Jolla, CA) to perform western blot analysis for collagen I (1:250; Southern Biotech) and b-actin (house-keeping gene) as per standard protocols.

Nitric Oxide (NO) Release Bioassay in Mouse RAW Macrophages
The biological activity of recombinant proteins were assessed by measuring the accumulation of nitrite NO2, a stable nitric oxide (NO) metabolite produced by RAW macrophages as described previously [27]. Briefly, cells (1610 5 cells/200 ml/well) were seeded in 96-well plates and grown overnight at 37uC/5%CO2. Then cells were incubated with either medium alone or recombinant proteins at different concentrations (10, 25, 50, 100 and 200 ng/ ml) together with 100 ng/ml lipopolysaccharides (LPS from E. coli 055:B5, Sigma). After 24 h, the secreted nitrite was measured using Griess reagent (1% sulfanilamide; 0.1% naphthylethylendiamine dihydrochloride; 3% H3PO4). The absorbance was determined at 550 nm using an ELISA plate reader.

Animal Experiments: Ethics Statement
All the animal experiments were performed in strict accordance with the guidelines and regulations for the Care and Use of Laboratory Animals, University of Groningen, The Netherlands. The protocols were approved by the Institutional Animal Ethics Committee of the University of Groningen, The Netherlands (Permit Number 5429A). Male 6-to 8-week old C57BL/6 mice were purchased from Harlan (Zeist, Netherlands) and kept at 12 h light/12 h dark cycles with ad libitum normal diet.

Immunohistochemistry and Immunofluorescence
The cells were fixed with acetone:methanol (1:1), dried and stored until immunostaining. Livers were harvested and transferred to Tissue-Tek OCT embedding medium, and snap frozen in isopentane chilled in a dry ice. Cryosections (4 mm) were cut using a Leica CM 3050 cryostat (Leica Microsystems, Nussloch, Germany). The sections were allowed to adhere to Superfrost microscopic glass slides (Menzel-Glä ser, Braunschweig, Germany), air-dried and fixed with acetone for 10 min. Cells or tissue sections were rehydrated with PBS and incubated with the collagen I antibody (1:100; Southern Biotech), a-SMA antibody (1:600, Sigma) or PPB (1:100; custom-made, Harlan) for 1 h. Thereafter, cells or sections were washed thrice with PBS and incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 30 min. Cells or sections were washed again and further incubated with HRP-conjugated tertiary antibody for 30 min. Thereafter, peroxidase activity was developed with 3-amino-9-ethyl carbazole (Sigma, St. Louis,MO) for 20 min and nuclei were counterstained with hematoxylin (Fluka Chemie, Buchs, Switzerland). Cells or sections were then mounted with Kaiser's gelatin (Darmstadt, Germany), visualized and photographed using a light microscope (Olympus UK Ltd., Essex, UK).

Quantitative Real-time PCR
Total RNA from liver tissues was isolated using RNeasy mini kit (Qiagen, Hilden, Germany) according to manufacturer's instructions. RNA concentrations were quantitated by a Nanodrop UV spectrophotometer (NanoDrop Technologies, Wilmington, DE). Total RNA (1.6 mg) was reverse transcribed in a volume of 50 ml using cDNA synthesis kit (Promega, Madison, WI). All primers were purchased from Sigma-Genosys (Haverhill, UK). Following primers were used, MMP13 forward: CCAGAACTTCCCAAC-CATGT; MMP13 reverse: GTCTTCCCCGTGTTCTCAAA; TIMP1 forward: ATCAGTGCCTGCAGCTTCTT; TIMP1 reverse: TGACGGCTCTGGTAGTCCTC; GAPDH forward: ACAGTCCATGCCATCACTGC; GAPDH reverse: GATC-CACGACGGACACATTG. The reactions were performed with 20 ng cDNA using SYBR green PCR master mix (Applied Biosystems, Foster City, CA) according to manufacturer's instructions and were analyzed by ABI7900HT sequence detection system (Applied Biosystems). Finally, the threshold cycles (Ct) were calculated and relative gene expression was normalized with GAPDH (for mouse) as housekeeping gene.

Construction and Expression of Recombinant Proteins
We synthesized four recombinant proteins: IFNc, mimetic IFNc (mimIFNc), BiPPB-IFNc and BiPPB-mimetic IFNc (BiPPB-mimIFNc) (Figure 1). IFNc and mimetic IFNc were expressed using pET42a derived bacterial expression vector to achieve cytoplasmic expression of the proteins ( Figure 1A, 1B and Figure S1, S2). While, BiPPB-IFNc and BiPPB-mimIFNc fusion proteins were expressed using pET39b (+) expression vector ( Figure 1C, 1D and Figure S4, S5) to yield periplasmic expression, required for their proper folding. DNA constructs were extensively analysed for the presence of inserts using specific restriction enzyme digestions and by automated DNA sequencing. Thereafter, constructs were transformed into E. coli strain BL21 (DE3) and protein expression was induced by 1 mM IPTG. Proteins of interest with the expected molecular weight were found in total cell pellets. Soluble proteins were purified from the supernatant of lysed bacterial pellets through Ni-NTA affinity columns. Purified proteins were further dialysed against PBS and concentrated by ultrafiltration. The presence of IFNc moieties or/ and PPB peptides in the prepared fusion proteins was confirmed in dot blots using anti-IFNc and anti-PPB antibodies (Figure 2A).

In vitro Binding of Recombinant Proteins in Human HSC
The binding of IFNc to its receptor is strictly species-specific, whereas PDGFbR binding is not. In order to demonstrate PDGFbR-specific binding of BiPPB containing fusion proteins, we used human LX2 hepatic stellate cells that are known to express PDGFbR. The results confirmed the species specificity of IFNc; mouse IFNc and mimetic IFNc (lacking extracellular IFNcR binding site) did not show any binding to human HSC ( Figure 2B). However, BiPPB-fused mouse IFNc or mimetic IFNc proteins showed high binding to human cells ( Figure 2B).

In vitro Effects of Recombinant Proteins in Human HSC and Mouse Macrophages
Following the binding studies, we investigated the anti-fibrotic effects of the recombinant proteins in human HSC after their activation with TGFb. In corroboration with the binding studies in human LX2 cells, TGFb-induced collagen expression was strongly inhibited by treatment with the PDGFR-specific BiPPB-IFNc and -mimIFNc fusion proteins ( Figure 3A, 3B) as analyzed by immunostaining and western blot analysis. While both mouse IFNc and mimIFNc did not induce any effect in human cells due to species differences and the absence of a receptor-binding site respectively. On the other hand, mouse IFNc and BiPPB-IFNc strongly potentiated LPS-induced activation and dose-dependent release of nitric oxide in mouse RAW macrophages due to presence of mouse IFNcR ( Figure 3C, 3D). However, mimIFNc lacking the IFNcR binding site and mimIFNc-BiPPB directed to PDGFbR did not induce any effect due to lack of PDGFbR expression on macrophages ( Figure 3C, 3D).

In vivo Liver Uptake of BiPPB-modified Fusion Proteins in Acute CCl 4 -induced Liver Fibrosis in Mice
Others and we have earlier demonstrated that PDGFbR is highly upregulated on activated HSC during acute and advanced liver fibrosis [20,21]. Our group has extensively studied the biodistribution of proteins modified with PDGFbR-recognizing cyclic peptides (PPB), using radiolabeled and imaging studies, and showed high distribution of PPB-modified proteins to PDGFbRexpressing hepatic stellate cells in fibrotic livers [23]. In the present study, we assessed the intracellular activation of the pSTAT1a signaling pathway after 24 h of IFNc, mimIFNc, BiPPB-IFNc and BiPPB-mimIFNc administration in CCl 4 -induced liver fibrosis in mice. We found a significant increase in pSTAT1a activation after 24 h of treatment with BiPPB-IFNc (p,0.05) and BiPPB-mimIFNc (p,0.01) compared to PBS, IFNc and mimIFNc treatments indicating enhanced liver uptake of the fusion proteins due to high PDGFbR expression on activated HSC in the fibrotic livers ( Figure 4 and Figure S6).

In vivo Effects of BiPPB-modified Fusion Proteins in Acute CCl 4 -induced Liver Fibrosis Model in Mice
A single CCl 4 administration induced acute liver injury in mice, characterized by an increased intrahepatic expression of HSC activation marker (a-smooth muscle actin) and an increased deposition of extracellular matrix molecule (collagen I) as shown in Figure 5A-5C. IFNc, mimIFNc, BiPPB-IFNc and BiPPB-mimIFNc proteins were examined for their anti-fibrotic effects in this experimental model in mice. After two intravenous injections, a strong reduction in collagen I expression was observed with IFNc (<45% reduction, p,0.05), HSC-targeted BiPPB-IFNc (<50% reduction, p,0.05) and BiPPB-mimIFNc (<60% reduction, p,0.01) as analyzed by immunostaining and quantitative western blot analysis ( Figure 5A, 5B and Figure S7). Furthermore, a-SMA expression was attenuated by HSC-targeted BiPPB-IFNc administration and more strongly by BiPPB-mimIFNc treatment as shown in Figure 5A, while IFNc had only little effect and mimIFNc did not induce any reduction in a-SMA expression levels ( Figure 5A). Apart from collagen expression and deposition, the balance between collagen degrading matrix metalloproteinases-13 (MMP-13) and their major endogenous inhibitor, tissue inhibitor of metalloproteinases-1 (TIMP-1), is an important determinant of progression or reversal of fibrosis [28,29]. A significant increase in MMP13/TIMP1 transcript ratio was observed after treatment with IFNc (p,0.05), BiPPB-IFNc (p,0.01) and BiPPB-mimIFNc (p,0.01) suggesting activation of fibrolysis and induction of reversal of fibrosis ( Figure 5C). Of note, no effect was observed with mimetic IFNc alone due to lack of an IFNcR binding site while the inhibitory effects of HSC-targeted BiPPB-modified fusion proteins on collagen deposition were higher compared to non-targeted IFNc ( Figure 5).
The main hurdles in IFNc-based therapies are the adverse effects that led to the failure of clinical trials. IFN-mediated reduction in blood platelets is clinically relevant and a well-known side effect [30]. Indeed, we observed a significant reduction in the platelet counts following two intravenous treatments with IFNc (p,0.001), which was significantly improved following treatment with BiPPB-mimIFNc and to a lesser extent with BiPPB-IFNc ( Figure 6). MimIFNc did not induce a significant change in platelets counts due to lack of binding to IFNcR or PDGFbR.

Discussion
Hepatocellular damage, hepatic inflammatory cell infiltration and extensive tissue remodeling ultimately culminate into the development of progressive fibrosis and cirrhosis [31]. During liver fibrosis, the interplay between growth factors and cytokines produced by the damaged hepatocytes; inflammatory cells and non-parenchymal cells activate quiescent hepatic stellate cells (HSC). These activated HSC in turn proliferate and accumulate in the injured liver producing large amounts of extracellular matrix proteins [32]. Therefore, therapeutic approaches to silence these activated HSC would be appropriate to inhibit or reverse liver fibrosis [12].
Fibrosis is an heterogeneous process e.g. CCl 4 intoxication results in hepatocyte damage, necrosis, inflammation, and fibrosis, Figure 6. Effects of recombinant proteins on IFNc-related adverse effect in the acute CCl 4 model. Graph represents platelet counts measured in CCl 4 -treated mice receiving PBS (n = 6) or the different recombinant proteins IFNc (n = 6), MimIFNc (n = 5), BiPPB-IFNc (n = 6), BiPPB-mimIFNc (n = 6). Bars represent mean 6 SEM of 5-6 mice per group. The results showed a significant reduction in platelet counts following two intravenous administration of IFNc, which was significantly improved following treatment with BiPPB-mimIFNc and to a lesser extent with BiPPB-IFNc. MimIFNc did not show significant change in platelets counts due to lack of binding to IFNcR or PDGFbR. *P,0.05, **P,0.01 denotes significance versus PBS treated CCl 4 mice. #P,0.05 denotes the significance versus IFNc-treated CCl 4 mice. doi:10.1371/journal.pone.0089878.g006 leading to portal fibrosis or cirrhosis. While bile duct ligation (BDL) stimulates the proliferation of biliary epithelial cells and oval cells, resulting in proliferating bile ductules accompanied by portal inflammation leading to biliary cirrhosis (PBC). As documented in fibrotic animal models (BDL or CCL 4 etc.) and diseased human livers, PDGFbR is expressed abundantly on activated HSC or portal fibroblasts (collectively referred as myofibroblasts-like cells) [21,22]. The expression was found to be very weak in normal healthy tissues or cells (pericytes or smooth muscle cells) relative to expression levels in the fibrotic livers, which motivated us to examine the potential of using PDGFbR for a targeted delivery of anti-fibrotic or apoptotic drugs to the fibrotic livers [12]. In the present study, we have directed IFNc or mimetic IFNc (signaling moiety of IFNc lacking extracellular IFNcR binding site while retaining activities of IFNc) [33] to PDGFbR-expressing activated HSC using bicyclic PDGFbR peptides to inhibit liver fibrosis with reduced adverse effects.
IFNc or mimetic IFNc modified with bicyclic PDGFbRrecognizing peptide were generated via recombinant technology in E.coli. Unmodified IFNc or mimetic IFNc were also produced in parallel to serve as respective controls. BiPPB-IFNc and BiPPB-mimIFNc were produced from a periplasmic vector (pET39b), since this vector contains the DsbA tag that exports the expressed proteins to the periplasmic space (space between plasma membrane and outer membrane). DsbA, periplasmic enzyme catalyzes the sequential formation of disulfide bonds therefore favors proper folding of the fusion proteins containing bicyclic peptide (cyclized via bisulfide bonds) at the N-terminus of IFNc and mimetic IFNc. Unmodified IFNc and mimetic IFNc were produced using pET42a vector as cytoplasmic proteins. These proteins were expressed, purified and analyzed using immune dot-blots where the presence of IFNc, mimIFNc, BiPPB-IFNc and BiPPB-mimIFNc was confirmed using anti-IFNc and anti-PPB antibodies.
The BiPPB-modified proteins were investigated for PDGFbRspecific binding in human HSC. BiPPB-IFNc and BiPPB-mimIFNc showed specific binding to PDGFbR-expressing human LX2 cells whereas unmodified mouse derived IFNc and mimIFNc did not show binding to these human cells due to species restrictive interaction of mouse IFNc to mouse IFNcR, while PDGFbR interaction is not species-specific. BiPPB-IFNc and BiPPB-mimIFNc also induced a strong reduction in TGFb-induced collagen I expression in human LX2 cells while IFNc and mimIFNc did not affect the collagen expression in these cells corroborating with the binding studies. We did not observe any difference in a-SMA expression in vitro after 48 hrs of treatment (data not shown) while significant reduction is observed in collagen expression ( Figure 3A, 3B), since IFNc has shown to have direct effect on collagen expression by directly acting on C/EBPbeta signaling pathway [34] while longer treatments might lead to inhibitory effects on a-SMA expression. These studies clearly demonstrate that directing IFNc or mimetic IFNc to the accessory PDGFbR can transform mouse (whose activity is restricted to mouse cells) and mim-IFNc (unable to enter any cell), into highly active proteins in human LX2 cells. On the other hand, we studied effect of these recombinant proteins in mouse macrophages expressing IFNcR and lacking PDGFbR. Results showed that IFNc or BiPPB-modified IFNc containing IFNcR binding site activated mouse RAW cells and induced NO release while both mimIFNc (lacking IFNcR binding site) or BiPPB-modified mimIFNc (containing PDGFbR binding site and lacking IFNcR binding site) did not show any effect in macrophages. These results indicate that targeted mimIFNc will not influence cell types (especially macrophages which are known to be highly influenced by IFNc) other than PDGFbR-expressing cells and therefore will not induce adverse effects in other normal cells or tissues.
In the past years, our group has extensively demonstrated the higher liver accumulation and HSC-specific distribution of PPBmodified proteins [35]. Here we examined the intra-hepatic activation of IFNc signaling pathway in vivo in treated livers to assess the increased accumulation of our targeted proteins in fibrotic livers. IFNc internalization results in the activation of the JAK-STAT pathway and subsequent phosphorylation of signal transducers and activators of transcription (STAT1) that binds to unique gamma-activated sequence (GAS) regulating IFNc-responsive genes [36]. Both BiPPB-modified IFNc and mimIFNc induced a significant increase in intrahepatic pSTAT1a signalling compared to unmodified IFNc and mimiFNc implicating increased liver accumulation of targeted proteins.
Thereafter, we studied the anti-fibrotic effects of recombinant proteins in the CCl 4 -induced liver fibrogenesis model in mice. This model is associated with HSC activation, enhanced PDGFR expression and ECM deposition, the key parameters of fibrogenesis. Two subsequent intravenous injections of IFNc, BiPPB-IFNc or BiPPB-mimIFNc led to the highly significant reduction in collagen I expression (major extracellular matrix protein) and a-SMA (HSC activation marker) expression. Additionally, IFNc, BiPPB-IFNc and BiPPB-mimIFNc enhanced the MMP13/ TIMP1 transcripts ratio implying activation of fibrolysis. In this study, we observed anti-fibrotic effects with IFNc in comparison with our previously reported studies [20,23], is attributed to the increased dose (5 mg) used here, as compared to 2.5 mg dose used earlier. MimIFNc that cannot be internalized due to lack of receptor binding sites did not show any effect on these fibrotic PPB-modified proteins can block the PDGFbR in vitro [35,37] and this might also account for the observed antifibrotic effects. However, previous studies with PPB-peptides coupled to albumin have shown that this effect does not occur in vivo at the doses used [20,23]. To further reinforce this hypothesis, we used synthetic BiPPB as a control but did not observed any effect on the fibrotic parameters (data not shown).
Liver fibrosis or cirrhosis is a slowly progressing disease that develops over many years, therefore patients are treated for longer periods to cure or reverse the disease. Therefore, therapies with improved therapeutic efficacy and preferably without provoking off-target systemic effects would be highly favourable. In clinical trials with IFNc, patients suffered from mild to severe adverse effects; therefore we investigated one of the well-known IFNmediated adverse effects, significant reduction in circulating blood platelets (mainly produced by bone marrow) can lead to fatal disorders and also found to be associated with the serotonin levels associated to depression [30]. We found that only two intravenous injections of IFNc led to a highly significant reduction in platelet counts (p,0.001) in comparison to PBS-treated mice. Furthermore, BiPPB-IFNc that can still interact with IFNcR receptor showed a slight reduction in platelet counts (p,0.05), but mimIFNc (without receptor binding sites) and BiPPB-mimIFNc (only specific to PDGFbR) did not induce a reduction in platelet counts. Further studies in long-term fibrotic models are important to examine the adverse effects of targeted constructs due to the long-term administration. Since platelets are produced by bone marrow, it is also important to study the effect of these constructs on the bone marrow as these effects may influence other crucial biological processes.
In conclusion, the results in this paper clearly demonstrate that HSC-targeted mimIFNc is capable of accumulating in PDGFbRexpressing fibrotic livers and exerting an inhibitory effect on fibrotic parameters, which makes this recombinant protein a highly attractive candidate to explore for the treatment of liver fibrosis. The recombinant synthesis of this chimeric compound will facilitate further translational research with this compound. Figure S1 Schematic representation depicting cloning strategy for preparation of pET42a-IFNc. PCR amplified and EcoRI digested mouse IFNc gene fragment was cloned in pET42a (+) prokaryotic vector at Psh A1/Eco RI site. (TIF) Figure S2 Schematic representation depicting cloning strategy for preparation of pET42a-mimIFNc. PCR amplified and EcoRI digested mouse mimetic IFNc gene fragment was cloned in pET42a (+) prokaryotic vector at Psh A1/Eco RI site. (TIF) Figure S3 Schematic representation depicting cloning strategy for preparation of pET39b-BiPPB. BiPPB was prepared by annealing and PCR extension of 2 sets of primers (refer to methods), which were linked together via Bam HI site and cloned in Sca I/Not I digested pET39b (+) prokaryotic vector. (TIF) Figure S4 Schematic representation depicting cloning strategy for preparation of pET39b-BiPPB-IFNc. PCR amplified IFNc prepared for fusion with BiPPB was digested with Not I/Xho I was cloned in pET39b vector at Not I/Xho I site.