RAR- is downregulated in HCC & cirrhosis and its expression inhibits myosin-driven activation and durotaxis in hepatic stellate cells

Hepatic stellate cells (HSCs) are essential perisinusoidal cells in the healthy and diseased liver. HSCs modulate extracellular matrix (ECM) homeostasis when quiescent, but in liver fibrosis, HSCs become activated and promote excess deposition of ECM molecules and tissue stiffening via force generation and mechanosensing. In hepatocellular carcinoma (HCC), activated HSCs infiltrate the stroma and migrate to the tumor core to facilitate paracrine signalling with cancer cells. Since the function of HSCs is known to be modulated by retinoids, we investigated the expression profile of retinoic acid receptor beta (RAR- in cirrhotic and HCC patients, as well as the effects of RAR- activation in HSCs. We found that RAR-β expression is significantly reduced in cirrhotic and HCC tissues. Using a A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. comprehensive set of biophysical methods combined with cellular and molecular biology, we have elucidated the biomechanical mechanism by which all trans-retinoic acid (ATRA) promotes HSC deactivation via RAR-β-dependent transcriptional downregulation of myosin light chain 2 (MLC-2) expression. Furthermore, this also abrogated mechanically driven migration towards stiffer substrates. Conclusion: Targeting mechanotransduction in HSCs at the transcriptional level may offer new therapeutic options for a range of liver diseases. Hepatocellular carcinoma (HCC) is the most relevant form of liver cancer and the second most deadly cancer globally. This disease develops most commonly in individuals with sustained liver fibrosis or cirrhosis, following chronic inflammation of the liver (1). The development of fibrosis involves the deposition of high levels of extracellular matrix (ECM) proteins and increased tissue tension, which together produce a stiff desmoplasia that promotes progression to cirrhosis (2). Non-alcoholic fatty liver disease (NAFLD), a growing worldwide epidemic (3), also involves liver fibrosis in a large number of cases (4). The role of fibrosis in the early and late stages of liver diseases highlights the urgent need to find therapies that target this desmoplastic reaction. Hepatic stellate cells (HSCs) are perisinusoidal cells, residing in the liver (5), with roles in retinoid storage and force-mediated maintenance of the extracellular matrix (6). In HCC, HSCs transition from a quiescent to a contractile activated state, and promote tumor development through production of a fibrotic stroma (7). HSCs also engage in cross-talk with cancer cells in positive-feedback loops that maintain the activated phenotype of HSCs as well as promoting the growth of cancer cells (8). Furthermore, the elevated deposition of fibronectin from activated HSCs drives the formation of the pre-metastatic niche that precedes liver colonization by pancreatic ductal adenocarcinoma (PDAC) cells (9). A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. HSCs are generated in the bone marrow along with pancreatic stellate cells (PSCs) (10), and share many equivalent functions with PSCs, such as storage of vitamin A droplets and ECM homeostasis. When activated, both cell types lose their vitamin A droplets and show high levels of stromal remodelling through their contractile myofibroblast-like phenotype (11, 12). Cancer associated fibroblasts (CAFs), like activated HSCs and PSCs, are myofibroblast-like cells and correspondingly drive tumor progression through restructuring of the tumor microenvironment (13). The similarity between these cell types highlights the importance of cell mechanics in tumor malignancy, and suggests that targeting the mechanical phenotype of these cells would be pertinent. Activated HSCs, activated PSCs, and CAFs perpetuate fibrosis and a tumor-permissive microenvironment through multiple positive-feedback loops that increase the active population of these cells. The actomyosin-dependent mechanical loop remodels and stiffens the ECM to a fibrotic state, and then responds to external rigidity by increasing cell contractility though myosin light chain 2 (MLC-2) regulation and inducing pro-fibrotic signalling pathways (6, 12-14). In addition, HSCs have been shown to infiltrate into the tumor stroma and localise around the cancer cells, facilitating paracrine signalling (15). Durotaxis, the process of directed migration based on substrate rigidity, has been observed in HSCs and is suggested to lead to accumulation of HSCs and further potentiation of fibrosis as migrating HSCs become activated by the stiff environment (16). The retinoic acid receptor family is known to modulate the function of HSCs, though the molecular mechanisms behind this regulation are not well understood. Elucidating the molecular events underlying HSC activation holds the key to blocking progression of early A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. and advanced liver diseases. Here we show that expression of RAR-β is significantly downregulated in HCC and cirrhotic patients in comparison with healthy individuals. We observed that ATRA (all trans-retinoic acid) induces the RAR-β-mediated negative transcriptional regulation of MLC-2, and consequently reduces force generation and mechanosensing. These processes sustain the stiff microenvironment needed for the activated phenotype in HSCs. Furthermore, deactivated HSCs have an impaired ability to remodel the matrix and do not display directed migration toward stiffer substrates. Experimental Procedures Tissues microarrays Tissue microarray (TMA) blocks were constructed as previously described (17) using 20 archival paraffin-embedded HCC tissue blocks and 10 healthy controls retrieved from the Imperial College London Tissue Bank (Ethical Approval nr.R15058). All experiments were performed in accordance with the Imperial College guidelines and regulations. All experimental protocols have been approved by the Research Ethics Committee Wales approval 12/WA/0196, license number 12275. Written consent was obtained from all subjects. A consultant histopathologist reviewed all the materials prior to inclusion on freshly cut hematoxylin & eosin (H&E) slides. We constructed TMA blocks using an MTA-1 Microarrayer (Mitogen, UK) following H&E-slide guided microdissection of target tumor and surrounding non-tumorous areas. We obtained triplicates of 1 mm cores from separate central and peripheral areas of tumor and matching surrounding liver. Adequate sampling of target tissues was confirmed on a freshly cut H&E section from the recipient TMA block before downstream analysis. For quantification of the levels of RAR-b, MLC-2, and pMLC-2 in human liver tissues from healthy, cirrhosis, and HCC, the TMAs were obtained from Biomax (Catalog number: BC03117). A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. Murine fibrosis model All research using live animals was approved by local ethics committees (Imperial College Central Animal Welfare Ethical Review Board (AWERB)) and was carried out under Home Office supervision in accordance with the European Directive 2010/63/EU. All animals received humane care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication 86-23 revised 1985) Mice were housed in specific pathogen free conditions in individually ventilated cages (Techniplast UK Ltd). The environment was controlled with a 12 hour light/dark cycle, ambient temperature of 21°C (+/-2°C) and humidity of 55% (+/-10%). Mice had access to irradiated diet (Special Diet Services UK) and water (25ppm chlorine) ad libitum. Five male C57BL/6 mice, aged 8-12 weeks, were dosed twice weekly for six weeks with 0.4μl/g CCl4 diluted 1 in 3 in olive oil, administered via the intraperitoneal route. Three male age and strain-matched co-housed ‘vehicle only’ control mice were dosed to the same regimen with the same by weight volume of olive oil. Mice were culled 72 hours following the last injection of CCl4. To validate the induction of fibrosis, Sirius Red staining was performed (by the Department of Pathology, St Mary’s Hospital Imperial College London). The collagen proportionate area (CPA) was calculated by quantification of the area of collagen (Sirius red stained tissue) as a proportion of the total tissue area using Image J. Eight representative fields at x10 magnification were assessed for each mouse. Cell culture and antibodies Primary, culture-activated human hepatic stellate cells (HSCs), passage 3-6, (HHStec 5300; ScienCell, Carlsbad, CA, USA) were cultured in phenol red medium (DMEM-F12 HAM, Sigma Aldrich, D6434) supplemented with 10% Foetal Bovine Serum (Gibco, 10500-064), A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. 1% Penicillin/Streptomycin (Sigma Aldrich, P4333), and 1% Fungizone R Amhotenicin (Gibco, 15290-026). HSCs were treated with 1 μM ATRA dissolved in absolute ethanol (Fisher, E/0650DF/17) for 10 days. Control HSCs were culture with an equivalent amount of 0.1% ethanol in media as vehicle control. Medium for both control and ATRA treated HSCs was changed every 24 h and changing of media was performed in subdued light. Both control and ATRA treated HSCs were incubated with phenol red culture medium supplemented with 2% Foetal Bovine Serum. HSCs were exposed to 1 μM of RAR-β agonist (cd2314, Tocris, 3824) for 72 hours. RAR-β antagonist (cd2665, Tocris, 3800) used at 1 M. All cells were tested for mycoplasma contamination. Antibodies: RAR- receptor (Abcam, ab53161), MLC-2 (Millipore, MABT180, WB 1/100 and IF 1/200), pMLC-2 /Thr18/Ser19 (Cell Signalling, 3674, WB 1/100, IF 1/200), αSMA (Abcam, ab7817, IF 1/200), Vimentin (M0725 DAKO, IF 1/200), collagen-I (abcam, ab34710, IF 1/200), fibronectin (abcam, ab2413, IF 1/100),Anti-Mouse HRP (Invitrogen,