Lenvatinib prevents liver fibrosis by inhibiting hepatic stellate cell activation and sinusoidal capillarization in experimental liver fibrosis

Abstract Molecular targeted agents are pharmacologically used to treat liver fibrosis and have gained increased attention. The present study examined the preventive effect of lenvatinib on experimental liver fibrosis and sinusoidal capillarization as well as the in vitro phenotypes of hepatic stellate cells. LX‐2, a human stellate cell line, was used for in vitro studies. In vivo liver fibrosis was induced in F344 rats using carbon tetrachloride by intraperitoneal injection for 8 weeks, and oral administration of lenvatinib was started two weeks after initial injection of carbon tetrachloride. Lenvatinib restrained proliferation and promoted apoptosis of LX‐2 with suppressed phosphorylation of extracellular signal‐regulated kinase 1/2 and AKT. It also down‐regulated COL1A1, ACTA2 and TGFB1 expressions by inhibiting the transforming growth factor‐β1/Smad2/3 pathway. Treatment with lenvatinib also suppressed platelet‐derived growth factor‐BB‐stimulated proliferation, chemotaxis and vascular endothelial growth factor‐A production, as well as basic fibroblast growth factor‐induced LX‐2 proliferation. In vivo study showed that lenvatinib attenuated liver fibrosis development with reduction in activated hepatic stellate cells and mRNA expression of profibrogenic markers. Intrahepatic neovascularization was ameliorated with reduced hepatic expressions of Vegf1, Vegf2 and Vegfa in lenvatinib‐treated rats. Collectively, these results suggest the potential use of lenvatinib as a novel therapeutic strategy for liver fibrosis.

ultimately results in liver cirrhosis and hepatocellular carcinoma (HCC). 3,4 Activation of hepatic stellate cells (HSCs) is commonly recognized as a key step in hepatic fibrogenesis. 5 In normal liver tissue, quiescent HSCs store vitamin A to maintain retinoid homeostasis. 6 However, persistent liver injury can cause HSCs to acquire an activated phenotype and transdifferentiate into myofibroblast-like cells that synthesize extracellular matrix components and produce the profibrogenic mediators. 7 Thus, suppression of HSC activation appears to be a powerful therapeutic strategy for liver fibrosis development.
Inhibition of tyrosine kinases have attracted attention to therapeutically reduce fibrogenesis around for more than 20 years, and several tyrosine kinases have been recognized to regulate the activation of HSCs. 8 For example, platelet-derived growth factor receptor (PDGFR) and vascular endothelial growth factor receptor (VEGFR) were markedly increased during the development of liver fibrosis 9 and were shown to activate multiple downstream signalling pathways, such as MEK/ERK and PI3K/Akt pathways, during HSC activation. 10 These critical roles of tyrosine kinases in hepatic fibrogenesis have suggested the possibility that inhibition of tyrosine kinases could be beneficial as antifibrotic treatment. 11 Several molecular targeted agents (MTAs) including tyrosine kinase inhibitors could exert efficient suppressive effects on proliferation, chemotaxis and collagen synthesis pathways in HSCs. 12 Furthermore, beneficial outcomes by MTAs in liver fibrosis have been investigated in preclinical experimental animal models. 13 For example, sorafenib, a MTA used to treat advanced HCC by targeting the Raf/ERK, VEGFR and PDGFRβ pathways, has been reported to show antifibrotic effects with inhibition of HSC activation in several experimental fibrotic rodent models. 14 Moreover, early clinical trials showed that cirrhotic patients who received treatment with sorafenib exhibited a significant improvement of portal hypertension. 15 Alterations to the hepatic vasculature have also been defined as key components in the process of liver fibrogenesis. 16 Emerging evidence has indicated that aberrant architecture of micro-vessels exacerbates portal hypertension and liver fibrosis progression. 17 In parallel with sinusoidal capillarization, intrahepatic angiogenesis triggering shunt formation leads to increased portal vascular resistance and decreased effective parenchymal perfusion. 18,19 VEGFRs are the most crucial tyrosine kinases that participate in angiogenesis during liver fibrosis development, and studies have suggested that VEGFRtargeting treatment using MTAs significantly attenuates liver fibrosis progression and decreases sinusoidal capillarization. 20,21 Lenvatinib is currently clinically approved in Japan, the United States and the European Union for use as first-line treatment of unresectable HCC. 22 It is an oral, small-molecule MTA that targets VEGFR1-3, fibroblast growth factor (FGF) receptor (FGFR)1-4, PDGFRα/β, KIT and RET, and differs from sorafenib in that it targets FGF signalling pathways in HCC. 23,24 Lenvatinib demonstrated similar rates of severe toxicity and delayed decline in health-related quality of life compared with sorafenib. 25 These pharmacological targets of lenvatinib highlight its potential use as an antifibrotic agent, similar to sorafenib.
The present study examined the preventive effects of lenvatinib on in vitro activation of HSCs and in vivo liver fibrosis development and intrahepatic angiogenesis in CCl4-induced rat fibrotic models.

| Cell viability assay
LX-2 cells were seeded in 96-well plates with DMEM which included 10% FBS for 24 hours. Then, the cells were exposed to different concentration of lenvatinib (0-400 nmol/L) for 24 hours. The Premix WST-1 Cell Proliferation Assay system (Takara Bio) was used to assess cell viability according to the manufacturer's protocol.
Moreover, for other set of experiments, the cells were pre-treated with PDGF-BB (50 ng/mL) or bFGF (10 ng/mL) for 2 hours and then were treated with lenvatinib (0-400 nmol/L) for 24 hours. The BrdU Cell Proliferation ELISA (Cosmo Bio, Tokyo, Japan) was used to evaluate cell proliferation according to the manufacturer's protocol.

| Cell chemotaxis assay
Cell migration of LX-2 cells was determined using the CytoSelect 24-Well Cell Migration Assay (8 µm, Colorimetric Format) (Cell Biolabs, Inc, San Diego, CA, USA) according the manufacturer's instructions.

| Measurement of VEGFA levels
VEGFA concentration in the cultured media from LX-2 cells was measured using RayBio Human VEGFA ELISA Kit (RayBiotech, Inc) according to the manufacturer's instructions. A total of 1 × 10 6 LX-2 cells were pre-treated with PDGF-BB (50 ng/mL) for 2 hours and then were treated with lenvatinib (0-100 nmol/L) for 6 hours following overnight starvation.

| Protein extraction and Western blotting
Proteins were extracted from 10 6 cultured LX-2 cells. For this purpose, T-PER Tissue Protein Extraction Reagent as lysis buffer supplemented with proteinase and phosphatase inhibitors (Thermo Fisher Scientific) was used. The protein concentration was measured by protein assay (Bio-Rad), and all samples were normalized to 100 μg. Western blotting was performed as described previously. 27 The membranes were incubated overnight with antibodies against

| Animals and experimental protocol
Six-week-old male Fisher 344 rats (Japan SLC) were treated twice a week for eight weeks with intraperitoneal injections of 0.5 mL/kg chronic carbon tetrachloride (CCl4) (diluted 1:10, Nacalai Tesque) or corn oil only as described previously. 28 After 2 weeks of initial injections with CCl4 or corn oil, the administration of vehicle or lenvatinib was started. First, to optimize in vivo dose of lenvatinib, we administered the different doses of lenvatinib (0.4, 0.8, 1.2, 1.6, 3.2, 6.4 and 9.6 mg/kg) to CCl4-mediated rats (n = 5) based on the previous reports to evaluate its anticancer effect in HCC xenograft model. 29,30 Next, the rats were divided into four groups (n = 10) according to treatment as follows: corn oil injections and vehicle administration (C/O group); CCl4 injections and vehicle administration (CCl4 group); CCl4 injections and low-dose (0.4 mg/kg) lenvatinib (Ld group); and CCl4 injections and high-dose (0.8 mg/kg) lenvatinib (Hd group). Because of incomplete water solubility of lenvatinib agent, we used water containing 10% carboxymethyl cellulose to suspend lenvatinib for oral administration to rats. Rats in the Ld and Hd groups received oral administration of lenvatinib by gavage daily throughout the experimental period. Water containing 10% carboxymethyl cellulose was given as vehicle. Rats were killed at the end of the eight-week experimental period, then body and liver weights were measured and blood was collected from

| Histological and immunohistochemical analyses
Liver specimens were fixed in 10% formalin and embedded in paraffin. Sections of 5μm thickness were stained with haematoxylin and eosin (H&E) and Sirius-Red. Primary antibodies, including alphasmooth muscle actin (α-SMA) (ab124964, AbCam), CD34 (ab81289, AbCam), Desmin (413651; Nichirei Biosciences) and GFAP (ab7260, Abcam), were used, and staining was performed according to the manufacturer's instructions. A goat anti-rabbit biotinylated secondary antibody was used and visualized using a horseradish peroxidase (HRP)-conjugated ABC system (Vector Laboratories). DAB was used as the chromogen. Immunofluorescence for glutamine synthetase (GS) (ab176562, AbCam) and Ki67 (ab15580, AbCam) was performed on paraffin-embedded intestinal sections using monoclonal antibodies against GS to assess hepatocyte expression and polyclonal antibodies against Ki67 to assess cell proliferation. Detection of the primary antibodies was performed with Alexa Fluor-conjugated secondary antibodies (Invitrogen). Images were captured using a BX53 (Olympus) for histology and immunohistochemistry and a BZ-X700 (Keyence) for immunofluorescence. Semi-quantitative analysis was performed with ImageJ software version 64.

| RNA extraction and reverse transcriptionquantitative polymerase chain reaction (RT-qPCR)
Total RNA was isolated from liver tissues and 10 6 cultured LX-2 cells using the RNeasy Mini Kit (Qiagen). The resulting RNA concentrations were determined using a NanoDropTM 2000c Spectrophotometer (Thermo Fisher Scientific Inc). High-capacity RNA-to-cDNA kit (Applied Biosystems) was used for reverse transcription to generate cDNA. Quantitative RT-PCR (qRT-PCR) was performed with the primer pairs described in Table S1 using a SYBR™ Green PCR Master

Mix (Applied Biosystems) and an Applied Biosystems StepOnePlus™
Real-Time PCR ® system (Applied Biosystems). Relative expression levels were normalized to GAPDH/Gapdh expression and estimated using the 2 −ΔΔCT method and presented as fold changes relative to controls.

| Statistical analyses
Statistical analyses were performed with Prism, version 9 (GraphPad Software). Data are expressed as the mean ± standard deviation.
Statistical variance between each experimental group was analysed using an analysis of variance test. Bartlett's test was used to determine homogeneity of variances. All tests were two-tailed, and Pvalues < 0.05 were considered statistically significant. We calculated the overall survival of rats in vivo experiments for lenvatinib dose F I G U R E 1 Effects of lenvatinib on in vitro proliferation, apoptosis and TGF-β1-induced activation of LX-2 cells. A, Cell viability of LX-2 incubated with lenvatinib (Lenv) (0-400 nmol/L) for 24 h by WST-1 assay. Cell viability was indicated as ratio to the value in the group of Lenv (0 nmol/L). B, Cell proliferation of LX-2 incubated with Lenv (0-200 nmol/L) for 24 h by BrdU assay. Cell proliferation was indicated as fold changes to the value in the group of Lenv (0 nmol/L). C, Time-dependent effects of Lenv on cell proliferation in LX-2 cells. The cells were cultured for 24, 48 and 72 h at 0 or 100 nmol/L of Lenv. D, Western blots (WBs) of whole cell lysates for total-and phospho-ERK1/2 and AKT on LX-2 cultured with Lenv (0-100 nmol/L) for 6 h. Semi-quantification of phosphorylation rate as ratio of p-ERK1/2 to t-ERK1/2 and p-AKT to t-AKT, respectively. E, The levels of cleaved caspase-3 and cleaved PARP1 in LX-2 cell culture extract assessed by ELISA. The levels were indicated as fold changes to the value in the group of Lenv (0 nmol/L). F, WBs of whole cell lysates for total-and phospho-SMAD2/3 on LX-2 cultured with Lenv (0-100 nmol/L) for 3 h. Semi-quantification of phosphorylation rate as ratio of p-SMAD2/3 to t-SMAD2/3. G, Relative mRNA expression levels of profibrogenic markers in LX-2 cells. The mRNA expression levels were measured by qRT-PCR, and GAPDH was used as internal control. Quantitative values are relatively indicated as fold changes to the values of non-treatment groups. Actin was used as the loading control for WBs (D and F). Cells were pre-treated with TGF-β1 (10 ng/mL) 3 h before Lenv treatment (F and G). Data are mean ± SD (n = 3 independent experiments with n = 12 (A-C,G) n = 3 (D, F) or n = 6 (E) samples per condition). *P < 0.05; **P < 0.01, indicating a significant difference compared with non-treatment groups (B-G). † P < 0.05, indicating a significant difference compared with TGF-β1(+)/Lenv(−) groups (F and G) optimization using the Kaplan-Meier method. Analyses were conducted using EZR (Saitama Medical Center, Jichi Medical University), a graphical user interface of R version 2.13.0 (The R Foundation for Statistical Computing) and a modified version of R commander (version 1.6-3) that includes statistical functions that are frequently used in biostatistics. 31

| Lenvatinib suppressed the proliferative capacity and TGF-β1-induced activation of human HSCs
Initially, we examined the impact of lenvatinib at different doses on cell viability of LX-2 to optimize the concentrations of lenvatinib used for in vitro studies. As shown in Figure 1A Interestingly, TGF-β1-stimulated phosphorylation of Smad 2/3 was attenuated by treatment with lenvatinib of LX-2 cells ( Figure 1F). In accordance with interference of Smad 2/3 activation, lenvatinib significantly reduced mRNA expression levels of COL1A1, ACTA2 and TGFB1, which was up-regulated by TGF-β1 stimulation ( Figure 1G).

| Lenvatinib suppressed bFGF-stimulated cell growth in human HSCs
Lenvatinib also inhibits FGFRs; therefore, we next evaluated its ef-

| Lenvatinib attenuated hepatic fibrosis development in CCl4-treated rats
Given the cell-based actions of lenvatinib on HSCs, we examined the effect on hepatic fibrosis development using a CCl4-induced liver fibrosis rat model ( Figure 4A). Initially, we optimized the dose of lenvatinib by evaluating the overall survival of rats receiving the different doses (0.4, 0.8, 1.2, 1.6, 3.2, 6.4 and 9.6 mg/kg) to CCl4-mediated rats. As shown in Figure S1A, the rats which were administered with more than 3.2 mg/kg were all dead within 4 weeks from start of lenvatinib administration as a result of the impairment of detoxification capacity to lenvatinib in rats by CCl4 administration. Moreover, 80% of rats which were administered with 1.2 or 1.6 mg/kg were also dead within 4 weeks from start of administration. By contrast, all of rats which were administered with 0.4 or 0.8 mg/kg were alive at the end of experiment. Consequently, we defined that these doses were suitable for CCl4-mediated rats. Systemic analysis showed that body and liver weights remained unchanged by oral administration of lenvatinib at both low (0.4 mg/kg) and high (0.8 mg /kg) doses in this experimental model, indicating that these doses led to low toxicity ( Figure S1B). Serum levels of AST and ALT were not elevated, but declined following treatment with lenvatinib in CCl4-treated rats ( Figure 4B). These results suggest that lenvatinib shows hepatoprotective properties against chemically induced liver injury.
The fibrotic areas stained by Sirius-Red were mildly increased by CCl4 administration for two weeks at the start of the lenvatinib treatments ( Figure S2). Continuous administration of CCl4 for another six weeks significantly developed liver fibrosis which were remarkably suppressed by following treatment with lenvatinib ( Figure 4C). Semiquantitative analysis demonstrated that the fibrosis degree relative to the start of the lenvatinib treatment was suppressed by >50% in CCl4-treated rats treated with lenvatinib, especially at higher doses ( Figure 4D). Along with liver fibrosis, lenvatinib treatment attenuated

| Lenvatinib suppressed intrahepatic capillarization with decreased transcription of growth factors and their receptors in CCl4-treated rats
Intrahepatic capillarization is reported to exacerbate liver fibrosis development. Therefore, we also evaluated changes in angiogenic status to determine the relevance of antiangiogenic activity in lenvatinib on attenuated liver fibrosis in experimental rat livers. Newly formed CD34-positive intrahepatic vessels were mildly increased by CCl4 administration for two weeks at the start of the lenvatinib treatment along with liver fibrosis ( Figure 5A). Continuous administration of CCl4 for another six weeks progressively enhanced intrahepatic capillarization which were remarkably suppressed in lenvatinibmediated rats ( Figure 5A). Semi-quantitative analysis showed a profound reduction in the capillarization degree relative to the start of the lenvatinib treatment in the Hd group to <25% ( Figure 5B). This finding coincided with the decreased hepatic mRNA levels of Cd31 observed in both Ld and Hd groups ( Figure 5C). As the major targets of lenvatinib, including VEGF/VEGFR, PDGF/PDGFR and FGF/ FGFR, are known to act as angiogenic regulators, we next assessed hepatic mRNA expression of these growth factors and receptors.
Hepatic mRNA levels of Vegfa, Vegfr1 and Vegfr2 were increased in accordance with CCl4-induced liver fibrosis and were significantly reduced in response to treatment with lenvatinib ( Figure 5D).
Lenvatinib treatment also led to decreased mRNA levels of Pdgfrb, which was increased in CCl4-induced fibrotic liver, whereas Pdgfb  Figure 5F). These results suggest that the antiangiogenic effects of lenvatinib contribute to the amelioration of CCl4-induced hepatic fibrosis development.

| Lenvatinib treatment showed no detrimental effect on hepatocyte proliferation
Finally, we examined whether lenvatinib affected hepatocytes in pericentral zone 3 injured by CCl4 administration as well as proliferation of hepatocytes in periportal zones 1 and 2. Serum albumin levels were not changed by two weeks of CCl4 administration but were decreased by eight weeks administration. In CCl4-treated rats, administration of lenvatinib at low and high doses did not alter serum albumin levels, indicating that these doses did not extensively damage hepatocytes ( Figure 6A). Furthermore, treatment with lenvatinib did not change serum levels of ALP and bilirubin, indicating low biliary epithelial damage ( Figure 6B). Moreover, immunostaining for GS demonstrated that the number of hepatocytes in pericentral zone 3 remained unchanged by treatment with lenvatinib compared with vehicle treatment in CCl4-treated rats ( Figure 6C,D). Additionally, the numbers of Ki67-positive proliferative hepatocytes in zones 1 and 2 were not altered by lenvatinib-treated rats compared with vehicle-treated rats ( Figure 6C,E). These results suggest low detrimental effects of lenvatinib on hepatic regeneration via proliferation of hepatocytes in CCl4-induced liver injury.

F I G U R E 4
Effects of lenvatinib on CCl4-induced liver fibrosis in rats. A, Experimental protocol. B, Serum levels of aspartate transaminase (AST) and alanine aminotransferase (ALT) in the experimental groups. C, Representative microphotographs of Sirius-Red, α-SMA, Desmin and GFAP staining in the experimental groups. Scale bar; 100 μm. pv; portal vein. D, Semi-quantitation of Sirius-Red-stained fibrotic area and α-SMA immuno-positive area in high-power field (HPF) by ImageJ software. E, Relative mRNA expression levels of Col1A1, Ctgf and Tgfb1 in the experimental groups. The mRNA expression levels were measured by qRT-PCR, and Gapdh was used as internal control. (F,G) Semi-quantitation of Desmin (F) and GFAP (G) immuno-positive area in high-power field (HPF) by ImageJ software. C/O, Corn oiladministered group; CCl4, CCl4 with vehicle-administered group, Ld; CCl4 with low dose of lenvatinib-administered group, Hd; CCl4 with high dose of lenvatinib-administered group. Data are mean ± SD (n = 10). Histochemical quantitative analyses included five fields per section (D,F,G). Quantitative values are relatively indicated as fold changes to the values of CCl4 at day 14 (D-G). *P < 0.05; **P < 0.01, indicating a significant difference between groups (B,D-G). NS, not significant

| D ISCUSS I ON
The present study revealed that lenvatinib efficiently attenuated progression of CCl4-induced liver fibrosis in rats. The observed antifibrotic effects mediated by lenvatinib may potentially be involved in several underlying mechanisms as a result of its pharmacological properties as a MTA. 32 We confirmed the effects of lenvatinib on in vitro phenotypes of Quantitative analyses included five fields per section. Ct, control; Ld, low dose of lenvatinib; Hd, high dose of lenvatinib; Veh; vehicle. Data are mean ± SD (n = 10). *P < 0.05; indicating a significant difference between groups (A,B,D and E). NS, not significant up-regulating gene expression of CYGB, a reactive oxygen species scavenger. 46 Pan et al suggested that low-and high-molecularweight bFGF play opposing roles. 47 In the present study, stimulation of bFGF did not alter mRNA levels of profibrogenic markers in LX-2 cells, and lenvatinib did not affect the expression of these markers in bFGF-treated LX-2 cells. However, further studies are required to elucidate the functional mechanisms involved and the role of bFGF in HSCs using bFGF isoforms, as well as the effects of lenvatinib.
Chronic administration of lenvatinib significantly reduced intrahepatic angiogenesis along with attenuation of CCl4-induced liver fibrosis in rats. In all lenvatinib-treated groups, hepatic expression of VEGFR1 and VEGFR2 were markedly decreased compared with the vehicle-treated group, in parallel with CD34-positive neovascularization. These results indicate that lenvatinib may exert its effects on sinusoidal capillarization via the VEGFR signalling pathway.
Furthermore, the present study elucidated that lenvatinib significantly decreased PDGF-BB-stimulated VEGFA production in LX-2 cells and reduced mRNA levels of VEGFA in CCl4-induced fibrotic livers. As PDGF is reported to stimulate HSCs to acquire an angiogenic phenotype via modification of HSC-based vascular formation, these findings suggest that the antiangiogenic effects of lenvatinib can be attributed to PDGFR signalling blockade in HSCs as well as VEGFR signalling on sinusoidal endothelial cells. 48,49 The present results reinforce the functional connectivity between intrahepatic angiogenesis and liver fibrosis development. Moreover, these findings support paracrine signalling between sinusoidal endothelial cells as well as the orchestration of fibrogenesis, angiogenesis and portal hypertension by HSCs. 50 The present study has several considerable limitations. As lenvatinib is a MTA used in the treatment of HCC, it is clinically administered to patients who tend to have established liver fibrosis. 51 In the present study, oral administration of lenvatinib began two weeks after the initiation of CCl4-induced liver fibrosis. This period is estimated as the point before the loss of hepatic functional reserve because lenvatinib is not available for patients with decompensated cirrhosis in the clinical settings. Further studies are required to evaluate the effect of lenvatinib in more advanced fibrotic models to verify whether it is clinically beneficial for liver fibrosis. Additionally, this study evaluated the effect of lenvatinib on in vivo liver fibrogenesis only in a single animal model. To reinforce translational impact, it is necessary to assess the effects in an additional model of liver fibrosis (eg a NASH model). Finally, results of this study found that administration of lenvatinib did not significantly affect in vivo proliferation of hepatocytes during CCl4-induced chronic liver injury.
This agent suppressed in vitro proliferation of human liver cancer cells, whereas we speculate that this potentially facilitates hepatic regeneration by inhibiting sinusoidal capillarization to some extent y. 52 Thus, there may be a possibility that these opposing effects of lenvatinib on hepatocyte proliferation offset each other in the current model.
Collectively, the present study is the first to report that lenvatinib prevents liver fibrosis progression using an experimental model and that this effect is based on the suppression of HSCs proliferation, migration and profibrogenic activity as well as attenuation of sinusoidal capillarization. It is important to emphasize that these actions of lenvatinib were achieved using a pharmacological dose without toxic hepatocytes damage. Although lenvatinib has been approved for the treatment of HCC, our findings indicate that this drug could be used as a novel therapeutic option for liver fibrosis development.

ACK N OWLED G M ENTS
The authors would like to thank Enago (www.enago.jp) for the English language review.

CO N FLI C T O F I NTE R E S T
No potential conflicts of interest were disclosed by all authors.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.