**1. Introduction**

Liver fibrosis and subsequent cirrhosis result in over one million deaths per year worldwide, making it the eleventh most common cause of death in adults [1]. The three most important causes for the development of liver fibrosis are chronic alcohol abuse, chronic infection with the hepatitis B (HBV) or C (HCV) virus, and metabolic syndrome, which can result in non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) [2]. Upon the chronic presence of such liver injury-causing circumstances, inflammatory signals within the liver will induce hepatocyte-damage and the activation of liver-resident hepatic stellate cells (HSCs) towards a myofibroblastic phenotype.

This activation process is marked by an excessive production and deposition of extracellular matrix [3]. The early development of fibrosis, and its progression towards cirrhosis, can be halted and even reverted upon suitable treatment, such as the administration of anti-viral drugs, or significant life-style changes [4]. Importantly, overall disease outcome improves when the treatment is started as early as possible in the disease process [5].

To date, liver biopsy remains the gold standard for grading and staging liver fibrosis. Unfortunately, this technique is associated with sampling and interpretation variability [6], doubtable cost–benefit ratios [7], and a risk of pain and bleeding [8]. These drawbacks limit the use of liver biopsy as a tool for screening or follow-up. Various non-invasive diagnostic tools have been developed, of which some have made their way into clinical practice, including serological scoring tools, such as the enhanced liver fibrosis (ELF) test, aspartate aminotransferase/alanine aminotransferase (AST/ALT) ratio, Fibrosis-4 (Fib-4) score, and the AST to platelet ratio index (APRI), and imaging-based tools, such as transient elastography (FibroScan®), shear wave elastography (SWE), and acoustic radiation force impulse (ARFI) [9]. However, these non-invasive techniques have not yet led to full redundancy of the liver biopsy, especially due to their limited sensitivity and specificity for the detection of early stages of liver fibrosis.

MicroRNAs (miRNAs) are non-coding, single-stranded RNA structures of approximately 22 nucleotides long, and function as posttranscriptional gene regulators. More specifically, through binding to the 3' untranslated regions of target messenger RNAs (mRNA), they can induce their cleavage, or prevent their translation into proteins [10]. Some miRNAs are known to be expressed in a cell- or tissue-specific manner. miRNAs are found in almost all body fluids, where they obtain stability by packaging into extracellular vesicles, or association to Argonaute2 or high-density lipoproteins [11]. Recent research has identified the potential of miRNAs to be used as diagnostic tools for specific subsets of liver disease, often focusing on the diagnosis of liver cirrhosis and hepatocellular carcinoma (HCC) [12,13]. However, the diagnostic value of individual miRNAs, or miRNA-panels, for the identification of early stages of liver fibrosis in heterogeneous patient populations remains to be proven.

In this study, we aimed to investigate the diagnostic value of miRNAs for the identification of significant liver fibrosis in a heterogeneous patient cohort suffering from chronic liver disease. We found that plasma levels of several individual miRNAs associated with HSC activation can distinguish between no or mild fibrosis (F0–1) and significant liver fibrosis (F ≥ 2). More importantly, the combination of the plasma levels of five miRNAs and PDGFRβ protein levels into the miRFIBP-score increased the predictive capacity for the diagnosis of significant liver fibrosis and outperformed clinical scores, such as Fib-4, the AST/ALT ratio, and APRI.

### **2. Materials and Methods**

### *2.1. Animal Studies*

The use and care of animals was reviewed and approved by the Ethical Committee of Animal Experimentation of the Vrije Universiteit Brussel (Brussels, Belgium) in project 16-212-2, and was carried out in accordance with European Guidelines for the Care and Use of Laboratory Animals. All mice were housed in a controlled environment with free access to water and food. Primary HSCs were isolated from male Balb/c mice aged 25 to 30 weeks (Charles River Laboratories, L'Arbresle, France), as described earlier [14,15]. Briefly, murine livers were digested by enzymatic solutions consisting of collagenase (Roche diagnostics, Mannheim, Germany) and pronase E (Merck, Darmstadt, Germany). The resulting cell suspension was centrifuged at low speed to remove hepatocytes. Hepatic stellate cells (HSCs) were purified from the non-parenchymal fraction based on their buoyancy, using an 8% Nycodenz (Axis-shield PoC AS, Dundee, Scotland) solution. Isolated HSCs were cultured on regular tissue culture dishes (Greiner Bio-One, Vilvoorde, Belgium), in Dulbecco's modified Eagle's medium (Lonza, Verviers, Belgium) supplemented with 10% foetal bovine serum (Lonza, Verviers, Belgium), 2 mM L-glutamine (Ultraglutamine 1®) (Lonza), 100 U/mL penicillin, and 100 μg/mL streptomycin

(Pen-Strep®) (Lonza), inducing in vitro myofibroblastic transdifferentiation. Cell purity was confirmed by the presence of lipid droplets and staining for HSC-specific markers.

For the in vivo induction of liver fibrosis, 10-week old mice received eight intraperitoneal injections of 15 μL carbon tetrachloride (CCl4) diluted in 85μL mineral oil (Sigma-Aldrich, St. Louis, MO, USA) per 30 g bodyweight over a period of four weeks. Mice were sacrificed 24 h after the last injection.
