p70 Ribosomal Protein S6 Kinase Is a Checkpoint of Human Hepatic Stellate Cell Activation and Liver Fibrosis in Mice

Background & Aims Progression of chronic liver disease (CLD) to liver cirrhosis and liver cancer is a major global cause of morbidity and mortality. Treatment options capable of inhibiting progression of liver fibrosis when etiological treatment of CLD is not available or fails have yet to be established. We investigated the role of serine/threonine kinase p70 ribosomal protein S6 kinase (p70S6K) as checkpoint of fibrogenesis in hepatic stellate cells (HSCs) and as target for the treatment of liver fibrosis. Approach & Results Immunohistochemistry was used to assess p70S6K expression in liver resection specimen. Primary human or murine HSCs from wild-type or p70S6K–/– mice as well as LX-2 cells were used for in vitro experiments. Specific small interfering RNA or CEP-1347 were used to silence or inhibit p70S6K and assess its functional relevance in viability, contraction and migration assays, fluorescence-activated cell sorting, and Western blot. These results were validated in vivo by a chemical model of fibrogenesis using wild-type and p70S6K–/– mice. Expression of p70S6K was significantly increased in human cirrhotic vs noncirrhotic liver-tissue and progressively increased in vitro through activation of primary human HSCs. Conversely, p70S6K induced fibrogenic activation of HSCs in different models, including the small interfering RNA–based silencing of p70S6K in HSC lines, experiments with p70S6K–/– cells, and the pharmacological inhibition of p70S6K by CEP-1347. These findings were validated in vivo as p70S6K–/– mice developed significantly less fibrosis upon exposure to CCl4. Conclusions We establish p70S6K as a checkpoint of fibrogenesis in vitro and in vivo and CEP-1347 as potential treatment option that can safely be used for long-term treatment.


SUMMARY
Targets and medical approaches to prevent progression of chronic liver disease to liver fibrosis or cirrhosis are urgently needed. We describe p70S6K being dysregulated in human liver fibrosis and define this protein as checkpoint of human hepatic stellate cell activation and liver fibrosis in vivo. Furthermore, we provide a clinical viable approach to inhibit this protein by the employment of CEP-1347.
BACKGROUND & AIMS: Progression of chronic liver disease (CLD) to liver cirrhosis and liver cancer is a major global cause of morbidity and mortality. Treatment options capable of inhibiting progression of liver fibrosis when etiological treatment of CLD is not available or fails have yet to be established. We investigated the role of serine/threonine kinase p70 ribosomal protein S6 kinase (p70S6K) as checkpoint of fibrogenesis in hepatic stellate cells (HSCs) and as target for the treatment of liver fibrosis.
APPROACH & RESULTS: Immunohistochemistry was used to assess p70S6K expression in liver resection specimen. Primary human or murine HSCs from wild-type or p70S6K -/mice as well as LX-2 cells were used for in vitro experiments. Specific small interfering RNA or CEP-1347 were used to silence or inhibit p70S6K and assess its functional relevance in viability, contraction and migration assays, fluorescence-activated cell sorting, and Western blot. These results were validated in vivo by a chemical model of fibrogenesis using wild-type and p70S6K -/mice. Expression of p70S6K was significantly increased in human cirrhotic vs noncirrhotic liver-tissue and progressively increased in vitro through activation of primary human HSCs. Conversely, p70S6K induced fibrogenic activation of HSCs in different models, including the small interfering RNA-based silencing of p70S6K in HSC lines, experiments with p70S6K -/cells, and the pharmacological inhibition of p70S6K by CEP-1347. These findings were validated in vivo as p70S6K -/mice developed significantly less fibrosis upon exposure to CCl 4 . C hronic liver disease (CLD) with its associated mortality and morbidity represents a considerable and increasing global socioeconomic problem. 1,2 Independently of its underlying etiology, CLD can cause liver fibrosis, eventually progressing to liver cirrhosis and its complications. Once liver cirrhosis becomes clinically apparent, 5year mortality without liver transplantation amounts to 85% 3 owing to the clinical complication such as variceal bleeding, ascites, kidney failure, and the occurrence of hepatocellular carcinoma, currently the most common cause of death in patients with liver cirrhosis and the fourth most common cause of tumor-related death worldwide. [3][4][5][6][7] Several advances have been made in the therapy of CLD, the latest and most remarkable perhaps being the development of direct-acting antiviral agents for the treatment of chronic hepatitis C. 8 However, for other CLD, an effective medical treatment is still lacking, only partially effective, or effective only in a subgroup of patients.
Establishing a medical treatment capable of directly targeting fibrogenesis and delaying or abolishing the progression of fibrosis will be indispensable to reduce CLDrelated morbidity and mortality until causal treatment options are established. The relevance of this issue is highlighted by the epidemiological trend of nonalcoholic steatohepatitis (NASH), a condition poorly responsive to treatment and due to become the leading cause of CLD and mortality globally. 9,10 Unfortunately, several promising agents, such as simtuzumab or selonsertib, failed to meet their primary endpoint of inhibiting fibrosis in the treatment of CLD. [11][12][13] The lack of treatment options in liver fibrosis to complement existing etiological treatment therefore represents an unmet medical need for an ever-increasing number of patients worldwide.
On the mechanistic level, liver fibrosis is thought to be caused mainly by hepatic stellate cells (HSCs) that, upon activation, transdifferentiate into myofibroblasts, cause deposition of extracellular matrix, and induce portal hypertension by contractility. [14][15][16] Profibrotic activation of HSCs can, in turn, be initiated at different sites, eg, the activation of transforming growth factor-b (TGF-b) or of platelet-derived growth factor BB (PDGF-BB), 2 molecules that trigger intracellular signaling converging to activate the PI3K-AKT-mTOR signaling. [17][18][19] However, attempts at blocking these receptors or the PI3K-AKT-mTOR axis by currently available agents poses several safety issues concerning long-term tolerability of their antiproliferative and immunosuppressive effects.
To establish a potentially accessible molecular target for antifibrotic treatment, we chose to investigate the serine/ threonine kinase p70 ribosomal protein S6 kinase (p70S6K), a molecule phosphorylated downstream of mTOR and possibly representing a distal effector of fibrogenesis, with the rationale that its pharmacological inhibition could be safer and more tolerable than that of its upstream targets. To this regard, we also assessed the potential use of CEP-1347 as p70S6K inhibitor, an agent for which extensive clinical research has proved to be safe and very well tolerated for long-term treatment in a previous clinical phase III trial for Parkinson's disease. 20 Results p70S6K expression is significantly increased in human liver resection specimens from cirrhotic patients compared with nonfibrotic liver tissue, and expression of p70S6K increases during the activation of primary human HSCs To assess whether p70S6K is differentially expressed in liver tissue with advanced stages of fibrosis or cirrhosis in comparison to tissue from healthy liver, immunohistochemical staining for p70S6K, along with immunohistochemical staining of a-smooth muscle actin (a-SMA), an established marker of HSC activation, was performed and quantified in surgical liver tissue specimens from patients with and without liver cirrhosis who had undergone partial hepatectomy. Staining for a-SMA was significantly higher in cirrhotic livers and was detectable predominantly along the fibrotic septa. Staining for p70S6K was also approximately 6-fold higher in cirrhotic liver compared with healthy liver tissue ( Figure 1A and B). p70S6K was costained with a-SMA or glial fibrillary acidic protein (GFAP) in mice treated with CCl 4 but not in vehicle-treated animals (nonfibrotic controls). This supports the observation that p70S6K is specifically increased in HSCs (Figures 1C and D and 2C). However, expression of p70S6K was not restricted to HSCs because p70S6K staining was observed in association with non-HSC markers such as F4/80 and cytokeratin 19 (CK19) (Figure 2A and B).
Because HSCs are thought to be the central regulators of fibrogenesis, in a next set of experiments we aimed to assess the role of p70S6K on HSC activation. For this purpose, we used a surrogate model of HSC activation by seeding primary *Authors share co-senior authorship.
Abbreviations used in this paper: a-SMA, a-smooth muscle actin; CLD, chronic liver disease; GFAP, glial fibrillary acidic protein; HSC, hepatic stellate cell; pmHSC, murine hepatic stellate cell; NASH, nonalcoholic steatohepatitis; p70S6K, p70 ribosomal protein S6 kinase; PDGF, platelet-derived growth factor; phHSC, primary human hepatic stellate cell; siRNA, small interfering RNA; TGF-b, transforming growth factorb; WST, water-soluble tetrazolium; Wt, wild-type. human HSCs (phHSCs) on uncovered plastic dishes for 13 days to induce their spontaneous transdifferention to activated myofibroblasts. 21 We observed that the expected increase of a-SMA, a hallmark of activation in these cells over 3 different time points (on days 2, 7, and 13 after isolation) was accompanied by an increased expression of p70S6K, as shown by Western blot (Figure 3). In summary, both the histological examination and the progressive increase of p70S6K during HSC activation support the hypothesis that this molecule plays a causal role in the process of fibrogenesis.
Silencing of p70S6K counteracts TGFb-dependent activation of LX-2 cells and spontaneous activation of isolated primary cells To assess whether the observed increase of p70S6K is an actual cause of fibrogenic HSC activation, p70S6K was silenced by small interfering RNA (siRNA) both in LX-2 cells and in isolated phHSCs. As shown in Figure 4A, silencing of p70S6K inhibited the activation of HSCs, as judged by the lack of a-SMA expression upon coincubation with TGF-b in LX-2 cells. Similar results were obtained with phHSCs plated on uncovered plastic wells and harvested after undergoing incubation for 7 days with siRNA targeting p70S6K or control siRNA ( Figure 4B).
To further validate this observation, we assessed the activation of primary murine HSCs (pmHSCs) isolated from p70S6K -/mice or syngeneic age-matched wild-type (Wt) control mice. While assessing the phenotype of spontaneous ageing of Wt vs p70S6K -/mice, we did not detect differences in body to liver or spleen weight, or in liver serum biochemistry between the 2 strains ( Figure 5A and B). Also, there were no differences in liver histology and  fibrosis markers as judged by Western blot or Sirius red staining ( Figure 5C-E). However, when we examined the ex vivo activation of pmHSCs isolated from p70S6K -/mice vs Wt mice, we observed a significantly lower increase of a-SMA expression in pmHSCs isolated from p70S6K -/mice in comparison with Wt mice throughout a 13-day time period ( Figure 4C). This suggests that p70S6K regulates HSC response to activation by external fibrogenetic stimuli. The findings on p70S6K activation in these 3 model systems pointed to a causal role of p70S6K in determining activation of HSC and a-SMA expression in both human and murine cells.

p70S6K knockout confers protection from liver fibrosis in mice
To validate these results, an in vivo model was employed, with the hypothesis that a lesser degree of fibrosis should be expected in p70S6K -/mice upon exposure to chemically induced fibrogenetic stimuli. Male Wt and p70S6K -/mice were thus challenged for 12 weeks with CCl 4 ( Figure 6A). In agreement with the in vitro experiments, we observed that p70S6K wt mice exhibited a much higher extent of fibrosis vs their p70S6K -/counterpart, as measured by Sirius red and a-SMA staining of liver sections from the respective strains ( Figure 6B). This was accompanied by numerically lower (though not significant) serum bilirubin values and hepatic Acta2 and collagen 1a1 messenger RNA expression ( Figure 7).

The effects of p70S6K silencing can be reproduced by CEP-1347 in vitro
In a next set of experiments, we aimed at assessing whether p70S6K can be inhibited pharmacologically. To this purpose, we used CEP-1347, a kinase inhibitor which we established as an effective inhibitor of p70S6K ( Figure 8A-C). When LX-2 cells were incubated with TGF-b in the presence or absence of CEP-1347, the activation of this HSC cell line was almost completely inhibited. Specifically, CEP-1347 strongly inhibited the induction of a-SMA, PDGF receptor-b, and collagen 1a1 protein-expression ( Figure 9A-E) as well as the migratory capacity and the contractility of HSCs ( Figure 10A-D). The capability of CEP-1347 to inhibit p70S6K was demonstrated by a significantly diminished phosphorylation of p70S6K when cells were incubated with TGF-b in the presence of CEP-1347 ( Figure 8A). Similar results were obtained when PDGF-BB (a well-established inducer of the PI3K-AKT-mTOR signaling in HSCs) 22,23 was used to activate HSCs ( Figure 8B and C). In comparison, CEP-1347 did not show any relevant effect on other kinases known to affect fibrogenesis ( Figure 8A).
Interestingly, CEP-1347 induced a significant G2 cell cycle arrest (Figure 11), an effect that might contribute to decreased fibrogenesis of HSCs. However, the possibility that this occurred solely as a consequence of a loss of cell viability caused by CEP-1347, rather than as a consequence of a direct antifibrotic effect of this drug, was excluded by viability assays, Hoechst staining and fluorescence-activated cell sorting analysis, which failed to show a relevant decrease in cell The illustrations and the quantitative assessment shown are representative of at least 4 separate experiments (*P < .05; t test). (C) pmHSCs were isolated from age-matched Wt and p70S6K -/mice. Protein expression of a-SMA was measured by Western blot and expressed as foldincrease vs baseline levels on day 2. pmHSCs from p70S6K -/mice showed a smaller increase in a-SMA protein expression over time (n ¼ 3; *P < .05; analysis of variance). viability or apoptosis within the time frame in which antifibrotic effects were observed (Figures 11 and 12). These results were further validated in experiments with phHSCs, showing that CEP-1347 inhibits p70S6K along with collagen 1a1 expression ( Figure 13A and B). These results show that CEP-1347 effectively inhibits p70S6K, hereby counteracting the profibrotic effects of HSC activation in different models.

Discussion
Progression of liver disease to fibrosis and cirrhosis as unmet medical challenge Many advances have been made in the prevention and causal treatment of CLD, including the widespread use of anti-Hepatitis B virus vaccines and antiviral drugs, with the latest and most significant achievement being the  development of direct-acting antiviral agents for the treatment of chronic hepatitis C. 24 For other CLDs, however, causal treatment is still lacking or poorly effective. Until individual curative etiological treatment options are established for all etiologies of CLD, the development of directacting agents capable of preventing progression of CLD to liver cirrhosis and its complications represents an important strategy to reduce CLD-related mortality. The importance of this approach is exemplified by NASH, a condition poorly responsive to treatment and due to become the leading cause of CLD globally. 25,26 However, attempts at establishing clinically successful primary antifibrotic agents have so far failed. [11][12][13] While setting out to investigate alternative molecular targets for antifibrogenic treatment, we chose to focus on p70S6K, as this molecule is located downstream of the PI3K-AKT-mTOR signaling, where several pathways of fibrogenesis converge. We speculated that p70S6K might be a potential downstream effector of fibrogenesis. Thus, we expect that the effects of its pharmacological inhibition should be restricted to the prevention of collagen deposition, limiting the possible spectrum of adverse effects common to upstream regulators of this axis. Three topics describe the main findings of our article and their clinical significance.

P70S6K is a central mediator of fibrogenesis downstream of PI3K-AKT-mTOR signaling
Firstly, we provide novel mechanistic data on fibrogenesis by describing p70S6K as a distal effector of fibrosis downstream of mTOR activation by consistent results obtained in cell lines, isolated primary cells, human tissues, and p70S6K -/mice. The several-fold increase of in p70S6K staining observed in cirrhotic human tissue suggested that p70S6K plays an important role in fibrogenesis, an indication supported by the fact that p70S6K also progressively increased during the process of activation of isolated phHSCs in vitro. The causal role of p70S6K in inducing liver fibrosis was confirmed by experiments showing that (1) silencing of p70S6K abolishes activation in human HSCs ( Figure 4A and B); (2) p70S6K -/-HSCs are protected from fibrogenic activation ( Figure 4C); (3) the effects of p70S6K silencing could be reproduced by CEP-1347, a clinically viable drug that we established as an effective inhibitor of p70S6K ( Figure 8A-C); and (4) when exposed to CCl 4 , p70S6K -/mice exhibited decreased liver fibrosis compared with Wt mice (Figure 6). These results are in line with the body of evidence that established the importance of the PI3K-AKT-mTOR axis in determining liver fibrosis and with 2 previous reports assessing the role of p70S6K in a rodent NASH model 27 and in bile acid-induced activation of HSCs. 28 We are aware of the limitation of our study that we cannot totally rule out that p70S6K expressed in other cell types (including F4/80 and CK19-positive cells) might also contribute to fibrogenesis.
Staining of p70S6K was not limited to HSCs. Future studies using cell-specific knockout models may shed more light in the role of p70S6K for hepatic fibrogenesis in specific cell compartments. Nevertheless, the current knowledge on the role of HSCs in fibrogenesis supports a likely function of p70S6K as a checkpoint of liver fibrogenesis by HSC activation.

Clinical implications
However, our results are not only relevant in terms of the definition of the mechanistic role of p70S6K within the pathways regulating fibrogenesis; our finding defining CEP-1347 as an effective inhibitor of this molecule has important potential clinical implications. Owing to the role of PI3K-AKT-mTOR signaling in the regulation of immune response and cell proliferation, several inhibitors of PI3K (comprising copanlisib, duvelisib, or idelalisib) and mTOR (sirolimus, everolimus) have been approved as anticancer or immunosuppressive agents. 29,30 However, the use of these substances to prevent the progression of liver fibrosis does not seem to be feasible in the setting of long-term treatment required to treat CLD, owing to potential adverse events anticipated by the well-known antiproliferative or immunosuppressive effects of these agents.
Targeting p70S6K would restrict the blockage of PI3K-AKT-mTOR axis to its effect on liver fibrosis without interfering with the actions of the upstream members of this signaling pathway, hereby representing a potentially safer and more tolerable therapeutic strategy. However, to our knowledge, no clinically viable p70S6K inhibitor had been previously described.

Clinical use of CEP-1347
Here, we suggest for the first time the kinase inhibitor CEP-1347 as an effective inhibitor of p70S6K. CEP-1347 was developed to treat Parkinson's disease and has proved to be safe, well tolerated, and devoid of long-term unwanted effects in a large cohort of patients (cumulatively, 1467 patient-years). 20,31 The extensive clinical data already available and the excellent safety profile of CEP-1347 fulfils a stringent prerequisite for its clinical assessment as a therapeutic agent for CLD and liver fibrosis, alone or in combination with established causal treatment of CLDs.

Conclusion
In summary, we describe p70S6K as a checkpoint of fibrogenesis and as an actionable target for CEP-1347, a clinically viable and available compound that can safely be administered for long-term treatment. Further studies should be conducted to investigate p70S6K as a therapeutic target to treat CLD.

Cell lines
LX-2 cells were purchased from Merck Millipore (Darmstadt, Germany). Their authenticity was assessed by the Leibniz-Institut (DSMZ-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH), and confirmed based on DNA-fingerprinting.

Cell culture
Cells were kept in Iscove Basal Medium (Merck Millipore; phHSCs) or in Dulbecco's modified Eagle medium (Sigma-Aldrich, Darmstadt, Germany; pmHSCs and LX-2 cells) containing 2% (LX-2) or 10% (phHSCs and pmHSCs) fetal bovine serum (Merck Millipore) and antibiotics (Sigma-Aldrich). All cells were cultivated in a humidified atmosphere with 5% CO 2 and 21% O 2 at 37 C. LX-2 cells were stimulated with 10 ng/mL TGF-b (PeproTech, Hamburg, Germany) or 30 ng/mL PDGF-BB (Biomol, Hamburg, Germany) and co-incubated with 1000 nM CEP-1347 (Tocris, Bristol, United Kingdom) for 1 hour, 24 hours, or 48 hours, where indicated. phHSCs and pmHSCs were cultured on uncovered plastic dishes for up to 13 days to induce a spontaneous activation. For these long-term experiments, culture media in phHSCs were replaced with fresh media and CEP-1347 3 times a week.

Isolation of phHSCs
A liver cell suspension was prepared using a 2-step collagenase perfusion technique with modifications as described previously. 33,34 In brief, hepatocytes were removed by multiple cycles of low-speed centrifugation, which resulted in a supernatant that contained the nonparenchymal cells. phHSCs were subsequently purified from the nonparenchymal cell fraction by discontinuous density centrifugation using Percoll. 33 Quality of isolation was checked using immunocytomicrocopy against a-SMA (Sigma-Aldrich), Desmin (Sigma-Aldrich), GFAP (Sigma-Aldrich), and CK-19 (Merck Millipore) ( Figure 14).

Isolation of pmHSCs
Isolation of pmHSCs was performed by pronasecollagenase perfusion followed by density gradient centrifugation in 13.2% Nycodenz (Axis-Shield PoC, Oslo, Norway) as described previously. 21

Viability assays
The effect of CEP-1347 on cell viability was assessed by 2 kinds of viability assays-PicoGreen (Invitrogen, Waltham, MA) and by water-soluble tetrazolium assay (Promega, Walldorf, Germany)-according to the instructions provided by the manufacturers.

Real-time polymerase chain reaction
Real-time polymerase chain reaction was performed in a SYBR Green system (QuantiTect SYBR Green PCR Kit; Qiagen, Venlo, the Netherlands) using a LightCycler 96 (Roche, Penzberg, Germany). Expression was calculated according to the DDCt method with Gapdh and 36b4 as the housekeeping genes and normalized to the means of the controls.

Sirius red staining
Liver samples were fixed using 4% formaldehyde. After embedding in paraffin, 4-mm sections were stained with Sirius red according to standard protocols.

Quantification of positive stained area
Slides from human or murine liver tissue were scanned as indicated after Sirius red or after immunhistochemical staining of p70S6K, a-SMA, and collagen 1a1 using a Pannoramic MIDI II digital slide scanner from 3D-Histech (Sysmex, Norderstedt, Germany). The stained fibrotic area was quantified via QuPath software (https://qupath.github. io/) and ImageJ2 software (National Institutes of Health, Bethesda, MD) as described previously. 35 Si-RNA silencing p70S6K siRNA was purchased from Qiagen; b-galactosidase siRNA, which was used as a control, was purchased from Dharmacon (Lafayette, CO). Lipofectamin 2000 (Invitrogen) was used as transfectant for siRNA experiments.  LX-2 cells were treated 48 hours after transfection with 10 ng/mL TGF-b for 24 hours. For siRNA transfection in phHSCs, cells were cultured on uncovered plastic for 7 days. Transfection with the respective p70S6K and control siRNA was performed on day 2 and day 5.

Fluorescence-activated cell sorting
For analysis of cell cycle, cells were stimulated with PDGF-BB, CEP-1347 or their combination for 24 hours before being harvested and undergoing propidium iodide staining (Sigma, Germany) as described previously. Fluorescence-activated cell sorting analysis was performed by using the Accuri C6 flow cytometer (BD Biosciences, Franklin Lakes, NJ) and its built-in software.

Gel-contraction assay
The gel-contraction assay was performed according to the method described by Ikenaga et al. 36 After polymerization of the collagen, the gel was mobilized from the surface using a pipette tip. Medium containing stimulating agents was added to the gels. After 24 hours of incubation, the gel area was measured using ImageJ software. The ratio of the well area to the gel area was calculated.

Migration assay
LX-2 cells were seeded in 6-well plates. When cells reached confluence, a horizontal scratch was made with the tip of a 100-mL pipette. Afterward, the wells were washed once with cell culture medium and stimulated with the different agents for 24 hours. Images were acquired before and after incubation with these agents to determine (ImageJ software) the reduction of the initial cell-free area as marker of cell migration activity.

Animal experiments
Liver fibrosis was induced using CCl 4 (Sigma-Aldrich) dissolved in corn oil (Sigma-Aldrich) as described previously. 35 Male C57BL/6 p70S6K þ/þ (Wt) vs p70S6K -/mice were bred from p70S6K þ/mice (purchased from RIKEN [Tokyo, Japan]) and challenged 3 times weekly with intraperitoneal injections of CCl 4 . A first dose of 0.25-mL CCl 4 per kg body weight was administered when animals reached the eighth week of age and was followed by injections of 0.5-mL CCl 4 per kg body weight for altogether 3 months. 35 Animals were kept on a 12-hour light/dark cycle, fed ad libitum, and received human care in compliance to the standard local and international regulations outlined in the Guide for the Care and Use of Laboratory Animals. All experiments were approved by the local authorities (Regierung von Oberbayern). All experiments conforms to the ARRIVE guidelines.

Statistical analysis
Statistical calculations were performed by using SPSS 25 (IBM, Armonk, NY) or GraphPad Prism 7 using analysis of variance, t test, Mann-Whitney U test, or Wilcoxon or Kruskal-Wallis test as indicated in the presentation of the specific experiments. When a relevant influence of an experiment was observed, univariate analysis of variance was performed and degrees of freedom, means of squares, and F values were calculated. P values lower than .05 were referred to as statistically significant. All data are presented as mean ± SEM.