Advertisement

Stellate cell expression of SPARC-related modular calcium-binding protein 2 is associated with human non-alcoholic fatty liver disease severity

  • Frederik T. Larsen
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Daniel Hansen
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Mike K. Terkelsen
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Sofie M. Bendixen
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Fabio Avolio
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Charlotte W. Wernberg
    Affiliations
    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark

    Department of Gastroenterology and Hepatology, University Hospital of Southern Denmark, Esbjerg, Denmark

    Center for Liver Research (FLASH), Department of Gastroenterology and Hepatology, Odense University Hospital, Odense C, Denmark
    Search for articles by this author
  • Mette E.M. Lauridsen
    Affiliations
    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark

    Department of Gastroenterology and Hepatology, University Hospital of Southern Denmark, Esbjerg, Denmark
    Search for articles by this author
  • Lea L. Grønkjaer
    Affiliations
    Department of Gastroenterology and Hepatology, University Hospital of Southern Denmark, Esbjerg, Denmark
    Search for articles by this author
  • Birgitte G. Jacobsen
    Affiliations
    Department of Gastroenterology and Hepatology, University Hospital of Southern Denmark, Esbjerg, Denmark
    Search for articles by this author
  • Ellen G. Klinggaard
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Susanne Mandrup
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Tina Di Caterino
    Affiliations
    Department of Pathology, Odense University Hospital, Odense C, Denmark
    Search for articles by this author
  • Majken S. Siersbæk
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Vineesh Indira Chandran
    Affiliations
    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark

    Department of Molecular Medicine, University of Southern Denmark, Odense C, Denmark
    Search for articles by this author
  • Jonas H. Graversen
    Affiliations
    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark

    Department of Molecular Medicine, University of Southern Denmark, Odense C, Denmark
    Search for articles by this author
  • Aleksander A. Krag
    Affiliations
    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark

    Center for Liver Research (FLASH), Department of Gastroenterology and Hepatology, Odense University Hospital, Odense C, Denmark
    Search for articles by this author
  • Lars Grøntved
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
  • Kim Ravnskjaer
    Correspondence
    Corresponding author. Kim Ravnskjaer, Department of Biochemistry and Molecular Biology, Center for Functional Genomics and Tissue Plasticity (ATLAS), Campusvej 55, 5230 Odense M, Denmark. Tel.: +45 65508906/+45 93979317, .
    Affiliations
    Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense M, Denmark

    Center for Functional Genomics and Tissue Plasticity (ATLAS), University of Southern Denmark, Odense M, Denmark
    Search for articles by this author
Open AccessPublished:October 27, 2022DOI:https://doi.org/10.1016/j.jhepr.2022.100615

      Highlights

      • Hepatic expression of SMOC2 correlates with NAFLD severity
      • SMOC2 is secreted by activated hepatic stellate cells/fibroblasts in the NASH liver
      • Hepatic expression of SMOC2 shows good predictive performance of NAFLD severity
      • SMOC2 is elevated in plasma from NASH patients
      • Plasma SMOC2 shows excellent predictive performance of NAFLD severity

      Abstract

      Background & Aims

      Non-alcoholic fatty liver disease (NAFLD) and its progressive form, non-alcoholic steatohepatitis (NASH), are the hepatic manifestations of metabolic syndrome. Histological assessment of liver biopsies is the gold standard for diagnosis of NASH. A Liver biopsy is resource heavy, can lead to complications such as bleeding, and does not fully capture tissue heterogeneity of the fibrotic liver. Therefore, non-invasive biomarkers that can reflect an integrated state of the liver are highly needed to improve diagnosis and sampling bias. Hepatic stellate cells (HSCs) are central in development of hepatic fibrosis, a hallmark of NASH. Secreted HSC-specific proteins may, therefore, reflect disease state in the NASH liver and serve as non-invasive diagnostic biomarkers.

      Methods

      We performed RNA-sequencing on liver biopsies from a histological characterised cohort of obese patients (n = 30, body mass index > 35 kg/m2) to identify and evaluate HSC-specific genes encoding secreted proteins. Bioinformatics was used to identify potential biomarkers and their expression at single-cell resolution. We validated our findings by single-molecule fluorescence in situ hybridisation (smFISH) and ELISA to detect mRNA in liver tissue and protein levels in plasma, respectively.

      Results

      Hepatic expression of SPARC-related modular calcium-binding protein 2 (SMOC2) was increased in NASH compared no-NAFLD (p.adj < 0.001). Single-cell RNA-sequencing data indicated SMOC2 expression by HSCs, which was validated using smFISH. Finally, plasma SMOC2 was elevated in NASH compared to no-NAFLD (p < 0.001) with a predictive accuracy of AUROC 0.88.

      Conclusions

      We propose increased SMOC2 in plasma reflects HSC activation, a key cellular event associated with NASH progression, and may serve as a non-invasive biomarker of NASH.

      Lay summary

      Non-alcoholic fatty liver disease (NAFLD) and its progressive form, non-alcoholic steatohepatitis (NASH), are the most common form of chronic liver diseases. Currently, liver biopsies are the gold standard for diagnosing NAFLD. Blood-based biomarkers to substitute liver biopsies for diagnosis of NAFLD are required. We found activated hepatic stellate cells, a central cell type in NAFLD pathogenesis, to up regulate expression of the secreted protein SPARC-related modular calcium-binding protein 2 (SMOC2). SMOC2 was elevated in blood samples from NASH patients and may hold promise as a blood-based biomarker for NAFLD diagnosis.

      Graphical abstract

      Keywords

      Introduction

      Obesity is a fast-evolving pandemic driven by a sedentary lifestyle and high-calorie diet, and genetic risk factors. In Europe, 23% of the population has a body mass index of 30 kg/m2 and the prevalence is increasing

      Organization WH. WHO European Regional Obesity Report 2022. Copenhagen: WHO Regional Office for Europe; 2022: World Health Organization; 2022.

      . Consequently, non-alcoholic fatty liver (NAFL) is the dominant cause of chronic liver disease
      • Younossi Z.M.
      • Golabi P.
      • de Avila L.
      • Paik J.M.
      • Srishord M.
      • Fukui N.
      • et al.
      The global epidemiology of NAFLD and NASH in patients with type 2 diabetes: A systematic review and meta-analysis.
      . Thus, the healthcare and associated economic burden of NAFLD is expected to increase dramatically with increasing rates of obesity
      • Hagstrom H.
      • Nasr P.
      • Ekstedt M.
      • Hammar U.
      • Widman L.
      • Stal P.
      • et al.
      Health Care Costs of Patients With Biopsy-Confirmed Nonalcoholic Fatty Liver Disease Are Nearly Twice Those of Matched Controls.
      ,
      • O'Hara J.
      • Finnegan A.
      • Dhillon H.
      • Ruiz-Casas L.
      • Pedra G.
      • Franks B.
      • et al.
      Cost of non-alcoholic steatohepatitis in Europe and the USA: The GAIN study.
      . NASH can be accompanied by different degrees of fibrosis. If uncontrolled, NASH may progress to liver cirrhosis and hepatocellular carcinoma
      • Matteoni C.A.
      • Younossi Z.M.
      • Gramlich T.
      • Boparai N.
      • Liu Y.C.
      • McCullough A.J.
      Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity.
      .
      Liver biopsy readings are the gold standard for diagnosis and staging of NASH and, therefore, pivotal in NASH trials
      • Campos G.M.
      • Bambha K.
      • Vittinghoff E.
      • Rabl C.
      • Posselt A.M.
      • Ciovica R.
      • et al.
      A clinical scoring system for predicting nonalcoholic steatohepatitis in morbidly obese patients.
      ,
      • Kleiner D.E.
      • Brunt E.M.
      • Van Natta M.
      • Behling C.
      • Contos M.J.
      • Cummings O.W.
      • et al.
      Design and validation of a histological scoring system for nonalcoholic fatty liver disease.
      . Liver biopsies are accompanied by a risk of complications, such as bleeding and pain and is subject to sampling bias
      • McGill D.B.
      • Rakela J.
      • Zinsmeister A.R.
      • Ott B.J.
      A 21-year experience with major hemorrhage after percutaneous liver biopsy.
      . Thus, it may not capture the heterogenous distribution of hepatic fibrosis. In addition, the risk of interobserver variability in histologic scores complicates diagnosis and prognosis of NAFLD severity
      • Vuppalanchi R.
      • Unalp A.
      • Van Natta M.L.
      • Cummings O.W.
      • Sandrasegaran K.E.
      • Hameed T.
      • et al.
      Effects of liver biopsy sample length and number of readings on sampling variability in nonalcoholic Fatty liver disease.
      • Ratziu V.
      • Charlotte F.
      • Heurtier A.
      • Gombert S.
      • Giral P.
      • Bruckert E.
      • et al.
      Sampling variability of liver biopsy in nonalcoholic fatty liver disease.
      • Davison B.A.
      • Harrison S.A.
      • Cotter G.
      • Alkhouri N.
      • Sanyal A.
      • Edwards C.
      • et al.
      Suboptimal reliability of liver biopsy evaluation has implications for randomized clinical trials.
      . Several non-invasive biochemical and imaging-based methods exist for diagnostic evaluation of NAFLD
      • Wong V.W.
      • Adams L.A.
      • de Ledinghen V.
      • Wong G.L.
      • Sookoian S.
      Noninvasive biomarkers in NAFLD and NASH - current progress and future promise.
      . Most non-invasive biochemical methods, however, exhibit modest accuracy in independent validation. Imaging-based methods, while having moderate to high accuracy, have limited utility due to cost and require well-equipped centres
      • Wong V.W.
      • Adams L.A.
      • de Ledinghen V.
      • Wong G.L.
      • Sookoian S.
      Noninvasive biomarkers in NAFLD and NASH - current progress and future promise.
      . Thus, non-invasive methods that fully capture NAFLD severity and dynamics are required to substitute liver biopsies for diagnostic and prognostic evaluation of NASH.
      A solid understanding of cellular changes underlying NAFLD progression is of major importance for improvement of diagnostic and prognostic tools as well as development of future treatment regimens. A consequential event in NASH development is the activation of hepatic stellate cells (HSCs). Activation of quiescent HSCs (qHSCs) is orchestrated by a complex series of cellular events initiated by lipotoxicity-induced necroptosis of hepatocytes
      • Tsuchida T.
      • Friedman S.L.
      Mechanisms of hepatic stellate cell activation.
      . The resulting proinflammatory milieu and infiltrating immune cells directly activate qHSCs, which transdifferentiate into fibrogenic myofibroblasts referred to as activated HSCs (aHSCs)
      • Tsuchida T.
      • Friedman S.L.
      Mechanisms of hepatic stellate cell activation.
      . As extracellular matrix (ECM) and matricellular proteins produced by HSCs mirror hepatocellular changes in NAFLD, monitoring their production is clinically relevant. Promising, blood biomarkers hence include proteins involved in ECM remodelling such as tissue inhibitor of metalloproteinase-1 (TIMP1), microfibrillar-associated protein 4 (MFAP4), and pro-peptide of type III collagen (PRO-C3)
      • Abdelaziz R.
      • Elbasel M.
      • Esmat S.
      • Essam K.
      • Abdelaaty S.
      Tissue Inhibitors of Metalloproteinase-1 and 2 and Obesity Related Non-Alcoholic Fatty Liver Disease: Is There a Relationship.
      • Boyle M.
      • Tiniakos D.
      • Schattenberg J.M.
      • Ratziu V.
      • Bugianessi E.
      • Petta S.
      • et al.
      Performance of the PRO-C3 collagen neo-epitope biomarker in non-alcoholic fatty liver disease.
      • Luo Y.
      • Oseini A.
      • Gagnon R.
      • Charles E.D.
      • Sidik K.
      • Vincent R.
      • et al.
      An Evaluation of the Collagen Fragments Related to Fibrogenesis and Fibrolysis in Nonalcoholic Steatohepatitis.
      • Saekmose S.G.
      • Mossner B.
      • Christensen P.B.
      • Lindvig K.
      • Schlosser A.
      • Holst R.
      • et al.
      Microfibrillar-Associated Protein 4: A Potential Biomarker for Screening for Liver Fibrosis in a Mixed Patient Cohort.
      .
      The aim of this study was to identify HSC-expressed secreted proteins such as ECM and matricellular proteins that accurately reflect NAFLD severity. This we did by transcriptomic analysis of human liver biopsies from severely obese patients populating the NAFLD disease spectrum followed by confirmation of cell type-specificity and validation at the protein level.

      Materials and Methods

      Study design and participants

      Liver and blood samples were obtained from participants enrolled in an ongoing prospective interventional case-control study, PROMETHEUS. The study is a liver biopsy controlled, single-center study from Denmark. Inclusion criteria is age 18-70 years and a body mass index ≥ 35 kg/m2. Exclusion criteria are overuse of alcohol (>12 g for women and >24 g for men per day), other known (or discovered) chronic liver disease, use of hepatotoxic medication (glucocorticoids, tamoxifen, amiodarone), short life expectancy, or contraindication towards liver biopsy.
      PROMETHEUS is registered at OPEN.rsyd.dk (OP-551, Odense Patient data Explorative Network) and at ClinicalTrial.gov (NCT03535142). The regional committee on health research ethics approved the study and all participant information (S-20170210). All participants gave written informed consent before study participation.
      Study data, such as biometrics (incl. height, weight, transient elastography, and body mass index), anthropometrics, and pharmacological treatment data were collected prospectively and managed using Research Electronic Data Capture (REDCap) tools hosted at OPEN.rsyd.dk. REDCap is a secure web-based software platform designed to support data capture for research studies (https://www.sdu.dk/en/forskning/open).

      Tissue and blood sampling

      Liver biopsies were taken before blood samples and elastography scans. Scans were performed and samples taken at the same day with participants being in a 12 h fasting state.
      Liver biopsies were sampled under sterile conditions by two trained clinicians from the right liver lobe with a 16-18G Menghini suction needle (Hepafix, Braun, Germany). Samples were immediately released into sterile saline water. A minimum of 15 mm was used for formaldehyde storage and liver histology. Subsequently, remaining tissue was divided into smaller pieces of 5-10 mm and preserved immediately in RNAlater (Sigma-Aldrich, St. Louis, MO) or snap frozen in liquid nitrogen. Blood was drawn by an experienced lab technician. Biochemical analyses were done according to standard regional protocols and using commercially available kits. All samples were handled by specialised research biochemical technicians and stored at -80 °C.

      Histology and staging of NAFLD

      All liver biopsies were staged and evaluated by one trained radiologist (T.D.C) blinded to all other data. Scores agreed to NASH Clinical Research Network (NAS-CRN) classification system for NAFLD: steatosis (0-3), lobular inflammation (0-3) and ballooning (0-2). NAFLD activity score (NAS 0-8) is the sum of these three assessments. Fibrosis was evaluated according to the Kleiner classification
      • Kleiner D.E.
      • Brunt E.M.
      • Van Natta M.
      • Behling C.
      • Contos M.J.
      • Cummings O.W.
      • et al.
      Design and validation of a histological scoring system for nonalcoholic fatty liver disease.
      , no fibrosis (F0), portal or periportal (F1A-C), perisinusoidal fibrosis in combination with portal and periportal fibrosis (F2), bridging fibrosis (F3), and cirrhosis (F4). Finally, NASH was evaluated according to the steatosis, activity, and fibrosis (SAF) scoring system, no NAFL (SAF 1, steatosis < 1), NAFL (SAF 2, steatosis ≥ 1), and NASH (SAF = 3, steatosis ≥ 5, ballooning ≥ 1)
      • Nascimbeni F.
      • Bedossa P.
      • Fedchuk L.
      • Pais R.
      • Charlotte F.
      • Lebray P.
      • et al.
      Clinical validation of the FLIP algorithm and the SAF score in patients with non-alcoholic fatty liver disease.
      . From liver biopsies, we defined patients with SAF 1 as no-NAFLD, patients with SAF 2 as NAFL, and patients with SAF 3 as NASH. For predictive modelling we, furthermore, defined patients with NAS ≥ 4 as severe NAFLD and patients with Kleiner fibrosis grade ≥ 2 as having fibrosis.

      RNA-sequencing and data analysis

      Needle biopsies of liver tissue were homogenised using FastPrep-24™ (MP biomedicals, Irvine, CA) and RNA purified using TRIzol-RNA lysis reagent (#T9424, Thermo Fisher, Waltham, MA) according to manufacturer’s instructions. Purified RNA was quantified using Qubit 3.0 Fluorometer (Thermo Fisher Scientific) and RNA quality was assessed using Fragment Analyzer 5200 (Agilent, Santa Clara, CA). NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs, San Diego, CA) were used for construction of libraries according to manufacture's protocol. RNA was paired-end sequenced using the NovaSeqTM 6000 platform (Illumina, San Diego, CA). Reads were aligned with STAR (v.2.7.8a)
      • Dobin A.
      • Davis C.A.
      • Schlesinger F.
      • Drenkow J.
      • Zaleski C.
      • Jha S.
      • et al.
      STAR: ultrafast universal RNA-seq aligner.
      to the human genome assembly (GRCh38, Ensembl release 101). FeatureCounts (v.2.0) was employed for exon read quantification
      • Liao Y.
      • Smyth G.K.
      • Shi W.
      featureCounts: an efficient general purpose program for assigning sequence reads to genomic features.
      . Quality of raw sequencing was assessed using FastQC (v.0.11.9) and MultiQC (v. v1.10.1)
      • Ewels P.
      • Magnusson M.
      • Lundin S.
      • Kaller M.
      MultiQC: summarize analysis results for multiple tools and samples in a single report.
      . Genetic variants were removed from raw sequencing reads using BAMboozle (v.0.5.0)
      • Ziegenhain C.
      • Sandberg R.
      BAMboozle removes genetic variation from human sequence data for open data sharing.
      . Sanitised reads have been deposited in the NCBI Gene Expression Omnibus repository and are accessible through GEO Series accession number GSE207310 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc= GSE207310).

      Public single cell integration and annotation

      For deconvolution of bulk RNA-seq, three independent public human single-cell RNA-sequencing (scRNA-seq) datasets were retrieved from GEO repositories GSE136103
      • Ramachandran P.
      • Dobie R.
      • Wilson-Kanamori J.R.
      • Dora E.F.
      • Henderson B.E.P.
      • Luu N.T.
      • et al.
      Resolving the fibrotic niche of human liver cirrhosis at single-cell level.
      , GSE115469
      • MacParland S.A.
      • Liu J.C.
      • Ma X.Z.
      • Innes B.T.
      • Bartczak A.M.
      • Gage B.K.
      • et al.
      Single cell RNA sequencing of human liver reveals distinct intrahepatic macrophage populations.
      , and GSE158723
      • Payen V.L.
      • Lavergne A.
      • Alevra Sarika N.
      • Colonval M.
      • Karim L.
      • Deckers M.
      • et al.
      Single-cell RNA sequencing of human liver reveals hepatic stellate cell heterogeneity.
      . Each dataset was initially processed with Seurat (v.4.0.3)
      • Hao Y.
      • Hao S.
      • Andersen-Nissen E.
      • Mauck 3rd, W.M.
      • Zheng S.
      • Butler A.
      • et al.
      Integrated analysis of multimodal single-cell data.
      to remove low quality cells (200 < n < 3000 genes, mitochondrial gene contributions < 20 %). Moreover, genes expressed in fewer than 50 cells were excluded to remove zero count genes. Following cell removal, normalisation, scaling, and dimensional reduction were performed. Predicted doublets were identified and removed using DoubletFinder (v.2.0.3)
      • McGinnis C.S.
      • Murrow L.M.
      • Gartner Z.J.
      DoubletFinder: Doublet Detection in Single-Cell RNA Sequencing Data Using Artificial Nearest Neighbors.
      . Integration was carried out by merging the three processed datasets and correction of the principal component analysis (PCA) embeddings using Harmony (v.0.1.0)
      • Korsunsky I.
      • Millard N.
      • Fan J.
      • Slowikowski K.
      • Zhang F.
      • Wei K.
      • et al.
      Fast, sensitive and accurate integration of single-cell data with Harmony.
      . Automated cell type annotation using CellTypist (v.0.1.4)
      • Dominguez Conde C.
      • Xu C.
      • Jarvis L.B.
      • Rainbow D.B.
      • Wells S.B.
      • Gomes T.
      • et al.
      Cross-tissue immune cell analysis reveals tissue-specific features in humans.
      was employed for annotation of the complete dataset (trained model reference = Immune All Low). Manual correction was done to increase annotation resolution for hepatocytes, cholangiocytes, liver sinusoidal endothelial cells, liver endothelial cells, aHSCs, qHSCs, and vascular smooth muscle cells (VSMCs).

      Single-molecule fluorescence in situ hybridisation

      Single-molecule fluorescence in situ hybridisation (smFISH) was performed using the RNAscope Multiplex Fluorescent Reagent Kit v2 assay (#323110, Advanced Cell Diagnostics [ACD], Newark, CA) according to manufacturer’s instructions. In short, formalin-fixed paraffin-embedded liver needle biopsies (n = 6) from obese patients (body mass index ≥ 35 kg/m2) histological graded as no-NASH (n = 3, 2 × NAS = 0 and 1 × NAS = 1) and NASH (n = 3, 2 × NAS = 7 and 1 × NAS = 8) were sectioned at 3 μm. Tissue sections were deparaffinised using histology-graded xylene and 100% ethanol followed by blockage of endogenous peroxidase using hydrogen peroxide (#322381, ACD). Antigen retrieval (HIER) was performed in 100°C 1x co-detection target retrieval solution (#323165, ACD) for 30 min followed by protease plus (#322331, ACD) treatment for 40 min. Hs-SMOC2 (NM_001166412.1, #522921, ACD), Hs-RGS5 (NM_003617.3, #533421-C2, ACD), and Hs-LUM (NM_002345.3, #494761-C4, ACD) probes were then hybridised to the tissue. SMOC2 was detected with OpalTM 570 (1:1000, #FP1488001KT, Akoya Biosciences, Marlborough, MA), RGS5 was detected with OpalTM 690 (1:1000, # FP1497001KT, Akoya Biosciences), and LUM was detected with OpalTM 520 fluorescent dye (#FP1487001KT, 1:750, Akoya Biosciences). For FBLN detection, sections from no-NASH (n = 2, 2 × NAS = 0) and NASH (n = 2, 1 × NAS = 7 and 1 × NAS = 8) were used with a Hs-FBLN2 probe (NM_001165035.2, #822761-C2) detected with OpalTM 690 as described above. Sections were counterstained using DAPI (#D9542, stock: 0.5 mg/ml, 1:500, Sigma) and slides were subsequently mounted using Prolong® Diamond Antifade Mountant (#P36961, Thermo Fisher Scientific). Images were acquired on a Nikon confocal A1 microscope (Nikon, Japan) at 20x magnification using NIS-Elements ER version 5.21.03 acquisition software.

      Statistical analysis

      Mann-Whitney U-test was used for multiple comparison of ELISA, normalised count data, and fraction of SMOC2+ cells. Bonferroni Hochberg correction was employed to adjust for α-error accumulation. Correlations between two groups (gene expression and clinical variables) were computed using Pearson correlation coefficient (two-tailed p-value). Poisson regression was employed to model smFISH transcript count data. Predictive modelling of Histological grades were calculated by area under the receiver operating characteristics curve (AUROC)
      • Robin X.
      • Turck N.
      • Hainard A.
      • Tiberti N.
      • Lisacek F.
      • Sanchez J.C.
      • et al.
      pROC: an open-source package for R and S+ to analyze and compare ROC curves.
      . Optimal cut-off points were estimated using the Youden index. For multiple comparisons, a nominal p value ≤ 0.05 was considered statistically significant. All statistical analyses were performed in R (v. 4.0.3).

      Results

      Patient characteristics

      We generated RNA-seq data from liver needle biopsies obtained from 30 severely obese patients (body mass index > 35 kg/m2). The cohort consisted of no-NAFLD (n = 5), NAFL (n = 15), and NASH (n = 10) patients. Clinical biometric and biochemistry features are shown in Table 1. Histological grading of the cohort is shown in Table 2.
      Table 1Biometric and biochemistry variables of patient cohort in mean ± SD.
      GroupnSex (M/F)Age (yr)BMI (kg/m2)ALT (U/L)AST (U/L)C-peptide (pmol/L)Cholesterol (mmol/L)Triglycerides (mmol/L)APRI (IU/L)FIB-4HOMA-IRLSM (kPa)eGFR (ml/min/1.73 m2)CRP (mg/L)
      No NAFLD50/537±1246.4±2.820.0±6.819.6±4.51058.8±365.83.7±0.71.0±0.40.2±0.10.6±0.33.8±1.85.24±1.588.3±17.717.6±7.0
      NAFL151/1446±1243.5±4.338.9±26.629.3±17.71461.6±467.55.1±1.11.8±0.70.3±0.20.7±0.48.4±4.99.5±7.097.6±18.414.3±15.9
      NASH102/847±1345.3±6.175.1±50.375.6±50.31865.6±616.94.8±1.02.4±1.10.8±0.61.8±1.325.1±26.622.9±15.296.7±19.913.4±11.0
      M, male, F, female, BMI, body mass index, ALT, alanine aminotransferase, AST, aspartate aminotransferase, APRI, AST to Platelet Ratio Index, FIB-4, fibrosis-4, HOMA-IR, homeostatic model assessment for insulin resistance, LSM, liver stiffness measure, eGFR, estimated glomerular filtration rate , CRP, C-reactive protein
      Table 2Histological grading of used obese patient cohort grouped by steatosis, activity, and fibrosis (SAF) score.
      GroupSAFnSteatosisLobular inflammationBallooningKleiner fibrosis grade
      0123012301201234
      No NAFLD15541541
      NAFL215753132153102
      NASH31034344282451
      PCA was conducted based on expression of all protein-coding genes and the first six components were used to stratify the patient cohort. One-way ANOVA showed an effect on variance from gender and tissue-preservation method (Fig. S1A), which we thus corrected for in the analysis. PCA showed a separation of transcriptional profiles from patients with NAS ≥ 5 and patients with NAS ≤ 2 (Fig. S1B). By comparing patients scored with NAS ≥ 5 and NAS ≤ 2, we identified 1078 differentially expressed genes (DEGs, q ≤ 0.05) (Table S1) of which 507 were downregulated and 571 were upregulated. PCA based on these DEGs showed a progressive transcriptional change from no-NAFLD to NASH patients (Fig. S1C). Moreover, the NASH core transcriptomic signature proposed by Govaere et al. (Fig. S2A) and Terkelsen et al. (Fig. S2B) was, in agreement, significantly different in NASH compared to no-NAFLD
      • Govaere O.
      • Cockell S.
      • Tiniakos D.
      • Queen R.
      • Younes R.
      • Vacca M.
      • et al.
      Transcriptomic profiling across the nonalcoholic fatty liver disease spectrum reveals gene signatures for steatohepatitis and fibrosis.
      ,
      • Terkelsen M.K.
      • Bendixen S.M.
      • Hansen D.
      • Scott E.A.H.
      • Moeller A.F.
      • Nielsen R.
      • et al.
      Transcriptional Dynamics of Hepatic Sinusoid-Associated Cells After Liver Injury.
      . Together, the transcriptomic profiles stratified the cohort by NAFLD severity, which led us to further analyse fibrogenesis-related NASH signature transcripts.

      WGCNA identifies modules of co-expressed genes associated with NAFLD progression

      We employed WGCNA for hepatic transcriptome profiling and identification of fibrogenesis-related NASH signature transcripts (Fig. 1). A total of 27 modules of co-expressed genes were identified, which we merged to 25 modules (Table S2). A significant correlation was found between 12 of the modules and at least 1 clinical variable (pearson r ≥ 0.5, p ≤ 0.05) (Fig. 1A). High correlation was generally found between modules and histological gradings of biopsies as well as the non-invasive measures LSM and diagnostic biochemical parameters alanine aminotransferase (ALT), aspartate aminotransferase (AST), and aspartate aminotransferase-to-platelet ratio index (APRI) (Table S3). GO analysis revealed significant enrichment of pathways in 8 of the modules including sterol biosynthesis process, carboxylic acid catabolic process, and extracellular structure organisation (Fig. 1B, Table S4). We identified Module XII to exhibit strongest correlations with histological gradings and diagnostic biochemical parameters (Table S3). Module XII was moreover enriched for genes associated with fibrogenesis pathways (Table S3). Of the 116 module-XII genes best correlating with NAS (r ≥ 0.6, p ≤ 0.0001) (Fig. 1C), 41 overlapped with the human secretome (Fig. 1D). Although a recent study has shown GPNMB to be secreted
      • Gong X.M.
      • Li Y.F.
      • Luo J.
      • Wang J.Q.
      • Wei J.
      • Wang J.Q.
      • et al.
      Gpnmb secreted from liver promotes lipogenesis in white adipose tissue and aggravates obesity and insulin resistance.
      , genes such as GPNMB, ADGRB2, and CD24 encode proteins typically not secreted indicating the human secretome database may contain false positives. Nevertheless, of the 41 genes overlapping with the human secretome, 39 were differentially expressed (q ≤ 0.05) in NAS ≥ 5 compared to NAS ≤ 2 (Table S5) and expression of the 41 genes stratified patients by NAFLD disease status (Fig. 1E). Top 6 Module XII genes correlating with NAS, COMP, CCL20, LPL, TREM2, SMOC2, and SPP1, also present in the human secretome, were significantly induced (p.adj. < 0.05) in NASH compared to no-NAFLD (Fig. 1F, Table S5). Using WGCNA analysis, we thus identified fibrogenesis-related NASH signature transcripts encoding secreted proteins. We next wanted to resolve expression of the differentially expressed Module XII-genes into individual cell types.
      Figure thumbnail gr1
      Fig. 1Hepatic transcriptome profiling and identification of fibrogenesis-related NASH signature transcripts. (A) Pearson correlation (r) between weighted gene co-expression network analysis (WGCNA) identified module eigengenes and biometric, blood chemistry, and histopathologic grades. Size of dots are proportional to the pearson correlation coefficient. (B) Enriched (p < 0.05) and representative (p ≤ 0.05) GO-slim categories for each module. (C) Pearson correlation of between NAFLD activity score (NAS) and hepatic expression of transcripts in Module XII in the 30 patients. Transcripts showing high correlation (n = 116, r ≥ 0.6, -log10(p.adj.) ≤ 4) are shown in red. (D) Venn diagram showing overlap of transcripts highly correlating with NAS and the human secretome obtained from SignalP. (E) Hierarchical clustering of Z-scores of transcripts highly correlating with NAS and overlapping with the human secretome (n = 41). Transcripts found in the human secretome from SignalP are shown on left and NAFLD status of patients (n = 30) is shown as colors on top. (F) Expression of Module XII top 6 transcripts correlating with NAS and overlapping with the human secretome. Hepatic expression of Module XII top 6 transcripts is visualized as boxplots with dots representing biological replicates. Mann-Whitney U test was employed to test difference in distribution between groups with Holm-corrected p values. Significance levels are * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p ≤ 0.0001.

      Hepatic mesenchymal cells express SMOC2

      HSCs are the major source of ECM deposition in the NASH liver and, therefore, key in development of fibrosis. We thus analysed public scRNA-seq data for HSC-specific expression of Module XII secretome genes. Three public human scRNA-seq datasets were integrated and reannotated for identification of cell type-specific expression of Module-XII secretome genes (Fig. 2 and Fig. S3A). By automated and manual annotation, based on expression of lineage markers, 25 distinct cell types were identified (Fig. 2A, Fig. S3A). Abundance of seven cell types (cell type abundance > 1 %) in no-NAFLD, NAFL, and NASH patients were estimated from bulk RNA-seq data (Fig. 2B). The estimated proportions were higher for aHSCs (p = 0.004) and lower for Kupffer cells (KCs) (p = 0.01) in NASH compared to no-NAFLD patients (Fig. 2C). From the 41 Module-XII secretome genes, we ascertained clear cell type enrichment for 11 genes which had a log2FC > 2 (Fig. 2D). LOXL1 was exclusively expressed by aHSCs while SMOC2 was expressed by qHSCs, aHSCs, and VSMCs. LOXL1 and SMOC2 were upregulated (q < 0.05) in NAS ≥ 5 compared to NAS ≤ 2 (Table S2). We detected the highest increase in gene expression for SMOC2 (log2FC = 3.6, q = 0.0001). This prompted us to resolve cell type-specific expression of SMOC2 by subsetting the mesenchymal cell population (Fig. 2E, Figs. S3B–C). Leiden clustering did not separate VSMCs and qHSCs into distinct clusters but subclustering pointed to HSCs as the main SMOC2 expressing cell type (Fig. 2F). Furthermore, the major hepatic cell types, liver sinusoidal endothelial cells (LSECs), hepatocytes, cholangiocytes, monocytes, macrophages, and KCs did not express SMOC2 (Fig. 2G).
      Figure thumbnail gr2
      Fig. 2Identification of cell type-specific gene expression of Module XII secreted proteins. (A) UMAP of human scRNA-seq integrated datasets GSE136103, GSE158723, and GSE115469 (n = 75,632 cells). (B) Estimated abundance of cell types (n = 7, estimated abundance > 0.01) from the patient cohort (n = 30) RNA-seq data. Estimated abundance is normalised to total estimated abundance for each cell type. Results are represented as stacked barplots and show mean estimated abundance for no-NAFLD, NAFL, and NASH patients. (C) Estimated cell type abundance of activated hepatic stellate cells (aHSCs) and Kupffer cells (KCs) shown as proportions of all estimated cell types. Mann-Whitney U test was employed to test difference in distribution between groups with Holm-corrected p values. (D) Cell type-resolved expression of Module XII genes encoding secreted proteins (n = 11, log2FC > 2, expression > 5%) shown by dotplot. (E) UMAP showing Leiden clustering of qHSCs, aHSCs, and VSMCs (n =1767 cells). Right panel shows the UMAP of the representation of the different human scRNA-seq datasets and treatment groups within each dataset. (F) UMAP showing normalised log2-expression of SMOC2 in qHSCs, aHSCs, and VSMCs. (G) Normalised log2-expression of SMOC2 in the major hepatic cell types represented as violin plots.

      SMOC2 expression by human hepatic stellate cells detected by single-molecule fluorescence in-situ hybridisation

      As both SMOC2 expression and the proportion of aHSCs followed NAFLD severity, and since scRNA-seq identified mesenchymal cells as the SMOC2-expressing cell population, we next sought to establish aHSC/fibroblast-expression of SMOC2 in situ. We validated SMOC2 expression in histologically graded liver needle biopsies from no-NASH (Fig. S4) and NASH using a triplex probe smFISH assay (Fig. 3). RGS5 was chosen as a marker for qHSCs while LUM and FBLN2 were chosen as markers for and aHSCs/fibroblasts (Fig. 3A and 3B)
      • Ramachandran P.
      • Dobie R.
      • Wilson-Kanamori J.R.
      • Dora E.F.
      • Henderson B.E.P.
      • Luu N.T.
      • et al.
      Resolving the fibrotic niche of human liver cirrhosis at single-cell level.
      ,
      • Wang Z.Y.
      • Keogh A.
      • Waldt A.
      • Cuttat R.
      • Neri M.
      • Zhu S.
      • et al.
      Single-cell and bulk transcriptomics of the liver reveals potential targets of NASH with fibrosis.
      ,
      • Dobie R.
      • Wilson-Kanamori J.R.
      • Henderson B.E.P.
      • Smith J.R.
      • Matchett K.P.
      • Portman J.R.
      • et al.
      Single-Cell Transcriptomics Uncovers Zonation of Function in the Mesenchyme during Liver Fibrosis.
      . We observed distinct signals upon confocal microscopy analysis for RGS5, LUM, and FBLN2 transcripts in biopsies from both no-NASH (Figs. S4A and S4B) and NASH (Fig. 3A and 3B). Moreover, we found distinct signals of FBLN2 in cells lining vessel walls (Figs. S4A–B) and in cells in no-NASH biopsies (Fig. S4A). In NASH biopsies, FBLN2 was found co-expressed with LUM and SMOC2 is cells throughout the hepatic parenchyma as indicated by white arrows (Fig. 3B). SMOC2, RGS5, and LUM were similarly detected in cells throughout the hepatic parenchyma as indicated by white arrows. SMOC2 was localised in proximity to larger vessels or lining the vessel wall in only few cells (Figs. S4C–E). We found no significant differences between biopsies from no-NASH and NASH in fractions for SMOC2-single positive cells (Fig. 3C-D) or SMOC2+RGS5+ cells (Fig. 3C). The fractions, however, were lower for SMOC2+FBLN2+ cells (p < 0.05) in biopsies from NASH compared to no-NASH (Fig. 3D). Moreover, the fractions were higher for SMOC2+LUM+ (p < 0.001), SMOC2+RGS5+LUM+ cells (p < 0.0001), and SMOC2+LUM+FBLN2+ (p < 0.001) in biopsies from NASH compared to no-NASH. Finally, by quantifying transcripts/cell, we found the number of SMOC2 transcripts to be higher in SMOC2+RGS5+LUM+ (p < 0.0001) and SMOC2+LUM+ (p < 0.001) compared to SMOC2 single-positive cells (Fig 3D). Taken together, we validated HSC/fibroblast expression of SMOC2 by smFISH and, furthermore, found that RGS5+LUM+ and LUM+FBLN2+ aHSCs/fibroblasts are the main SMOC2-expressing cells in the NASH liver.
      Figure thumbnail gr3
      Fig. 3Single-cell resolution of SMOC2, RGS5, and LUM transcripts show SMOC2 expression by HSCs in human liver. Confocal images of a human liver needle biopsies from severely obese patients histological graded as NASH (n = 2-3) showing SMOC2 (orange) co-localised with RGS5 (magenta) and/or LUM (green) (A) and co-localised with FBLN2 and/or LUM (B). Scale bars; upper panel = 50 μm and lower panel = 10 μm. (C) Fraction of total cells/image being SMOC2+, SMOC2+RGS5+, SMOC2+LUM+, and SMOC2+RGS5+LUM+. (D) Fraction of total cells/image being SMOC2+, SMOC2+FBLN2+, SMOC2+LUM+, and SMOC2+LUM+FBLN2+. (E) Quantification of SMOC2 transcripts in SMOC2+, SMOC2+RGS5+, SMOC2+LUM+, and SMOC2+RGS5+LUM+ cells. (F) Quantification of SMOC2 transcripts in SMOC2+, SMOC2+ FBLN2+, SMOC2+LUM+, and SMOC2+LUM+FBLN2+ cells. QuPath was employed to detect and quantify single-, double-, and triple-positive stained cells. Cells with ≥ 2 SMOC2 transcripts/cell were considered SMOC2-positive cells. Fractions of positive cells are shown as mean + SE (n = 6-12 images/biological replicate) and SMOC2 transcripts/cell are shown as boxplots (n = 6-12 images/biological replicate). Mann-Whitney U test with Holm-correction was employed to test difference between cell fractions. Poisson regression with robust standard errors and p value calculation was employed to test difference between estimated transcripts/cell. Significance levels are * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p ≤ 0.0001.

      Hepatic expression of SMOC2 discriminates NASH from no-NASH

      To evaluate the diagnostic potential of SMOC2 as a biomarker for NASH, we employed predictive modelling of histological grades NAS, Kleiner fibrosis, or SAF scores using hepatic expression of SMOC2 (Fig. 4). To validate predictive accuracy of SMOC2 in our cohort, we included RNA-seq data from the, currently, most comprehensive multi-centre NAFLD cohort published (GSE135251, n = 206 NAFLD patients)
      • Govaere O.
      • Cockell S.
      • Tiniakos D.
      • Queen R.
      • Younes R.
      • Vacca M.
      • et al.
      Transcriptomic profiling across the nonalcoholic fatty liver disease spectrum reveals gene signatures for steatohepatitis and fibrosis.
      . Moreover, using this multi-centre NAFLD cohort, we benchmarked the performance of SMOC2 against recently proposed single-gene biomarkers of NASH and hepatic fibrosis (TREM2, AKR1B10, MFAP4, and GDF15) and determined predictive performance of gene combinations (Table S6)
      • Govaere O.
      • Cockell S.
      • Tiniakos D.
      • Queen R.
      • Younes R.
      • Vacca M.
      • et al.
      Transcriptomic profiling across the nonalcoholic fatty liver disease spectrum reveals gene signatures for steatohepatitis and fibrosis.
      ,
      • Bracht T.
      • Schweinsberg V.
      • Trippler M.
      • Kohl M.
      • Ahrens M.
      • Padden J.
      • et al.
      Analysis of disease-associated protein expression using quantitative proteomics-fibulin-5 is expressed in association with hepatic fibrosis.

      Chandran VI, Wernberg CW, Lauridsen MM, Skytthe MK, Bendixen SM, Larsen FT, et al. Circulating TREM2 as a non-invasive diagnostic biomarker for NASH in patients with elevated liver stiffness. Hepatology;n/a.

      Hendrikx T, Porsch F, Kiss MG, Rajcic D, Papac-Miličević N, Hoebinger C, et al. Soluble TREM2 levels reflect the recruitment and expansion of TREM2+ macrophages that localize to fibrotic areas and limit NASH. Journal of Hepatology.

      . Patients were grouped into severe NAFLD (NAS ≥ 4) and mild NAFLD (NAS < 4), fibrosis (Kleiner fibrosis grade ≥ 2) and mild fibrosis (Kleiner fibrosis grade < 2), and finally NASH (SAF = 3) and no-NASH (SAF ≤ 2) (Table S7). We sequenced 3 male needle biopsies, which was insufficient to determine gender-specific differences in hepatic SMOC2 expression. Thus, we segmented the multi-centre NAFLD cohort into predicted females and males by XIST expression (Fig. S5). No significant differences were found between gender in the segmented groups. In patients with severe NAFLD compared to mild NAFLD, SMOC2 expression was elevated (patient cohort; p < 0.001, multi-centre NAFLD cohort; p < 0.0001) with a predictive accuracy for severe NAFLD of AUROC 0.89 (sen. 0.69, spe. 1) in our cohort and AUROC 0.7 (sen. 0.84, spe. 0.57) in the multi-centre cohort (Fig. 4A, 4D, and 4G). In patients with fibrosis compared to mild fibrosis, expression of SMOC2 was elevated (patient cohort; p < 0.01, multi-centre cohort; p < 0.0001) with a predictive accuracy for fibrosis of AUROC 0.75 (sen. 0.25, spe. 1) in our cohort and AUROC 0.75 (sen. 0.86, spe. 0.51) in the multi-centre cohort (Fig. 4B, 4E, and 4G). Finally, in NASH compared to no-NASH patients, expression of SMOC2 was elevated (patient cohort; p < 0.001, multi-centre cohort; p < 0.0001) with a predictive accuracy of NASH of AUROC 0.9 (sen. 0.8, spe. 95) in our cohort and AUROC 0.70 (sen. 0.86, spe. 0.51) in the multi-centre cohort (Fig. 4C, 4F, and 4G). Predictive performance of hepatic SMOC2 expression was in an overall similar range as TREM2, AKR1B10, MFAP4, GDF15 (Fig. 4D-F).
      Figure thumbnail gr4
      Fig. 4Predictive modelling of histological grades by hepatic expression of SMOC2. (A - C) Association of SMOC2 expression with NAFLD progression in the patient cohort RNA-seq data (n = 30). (D - F) Association of SMOC2 expression and previously proposed biomarkers of NAFLD (TREM2, AKR1B10, MFAP4, and GDF15) with NAFLD progression in previously described RNA-seq data from a NAFLD multi-centre cohort (GSE135251, n = 206). NALFD patients are segmented by (A and D) severe NAFLD (NAS ≥ 4), (B and E) fibrosis (Kleiner fibrosis grade ≥ 2), and (C and F) NASH (SAF > 2). Performance of SMOC2 expression was evaluated using area under the receiver operating characteristic (AUROC). Sensitivity and specificity were determined from optimal cut-off points using the Youden index. SMOC2 expression in the patient cohort RNA-seq segmented groups are visualized as boxplots with dots representing biological replicates. (G) Expression of SMOC2, TREM2, AKR1B10, MFAP4, and GDF15 in the multi-centre NAFLD cohort are visualized as mean differences between segmented groups with dots representing the mean difference and whiskers representing 95% confidence intervals. Significance levels are ** p < 0.01 and *** p < 0.001 (Mann-Whitney U-test).
      The high predictive accuracy of hepatic SMOC2 expression for histological grades in both our cohort and the multi-centre NAFLD cohort suggested SMOC2 as a potential diagnostic biomarker of NASH.

      Plasma SMOC2 levels are associated with NASH severity

      Next, we investigated if the NASH-induced elevation of SMOC2 expression translated into increased SMOC2 protein in plasma. To assess the potential contribution of SMOC2 expression in adipose tissue we first quantified SMOC2 expression by RT-qPCR in subcutaneous adipose tissue from our cohort (Fig. S6). We found no effect of NAFLD status on the variance in SMOC2 expression in subcutaneous adipose tissue (p = 0.33). This prompted us to quantify levels of SMOC2 in plasma from a histological characterised severely obese cohort (n = 35, body mass index > 35 kg/m2). This cohort was segmented as described above (Table S8). We did not find any significant differences in plasma SMOC2 levels between genders in the segmented groups (Fig. S7). Plasma SMOC2 levels were elevated in severe NAFLD compared to mild NAFLD (p < 0.0001) with a predictive accuracy of severe NAFLD of AUROC 0.89 (sen. 0.89, spe. 0.81) (Fig. 5A). Plasma SMOC2 levels were elevated in patients with fibrosis compared to mild fibrosis (p < 0.01) with a predictive accuracy of fibrosis of AUROC 0.80 (sen. 0.96, spe. 0.42) (Fig. 5B). Finally, plasma SMOC2 levels were elevated in patients with NASH compared to no-NASH (p < 0.0001) with a predictive accuracy of NASH of AUROC 0.88 (sen. 0.85, spe. 0.80) (Fig. 5C). Our findings strongly suggest that circulating SMOC2 protein could be a good diagnostic biomarker for NASH in obese patients.
      Figure thumbnail gr5
      Fig. 5Predictive modelling of histological grades by plasma SMOC2. (A) Plasma SMOC2 levels in severe NAFLD (NAS ≥ 4, n = 19) and in mild NAFLD (NAS < 4, n = 16) patients (left panel) and predictive accuracy of severe NAFLD (right panel). (B) Plasma SMOC2 levels in patients with fibrosis (Kl. fibrosis ≥ 2, n = 12) and with mild fibrosis (Kl. fibrosis < 2, n = 23) (left panel) and predictive accuracy of fibrosis (right panel). (C) Plasma SMOC2 levels in patients with NASH (SAF > 2, n = 20) and no-NASH (SAF < 2, n = 15) (left panel) and predictive accuracy of NASH (right panel). Plasma SMOC2 performance was evaluated using area under the receiver operating characteristic (AUROCs). Sensitivity and specificity were determined from optimal cut-off points using the Youden index. Significance levels are ** p < 0.01 and **** p ≤ 0.0001 (Mann-Whitney U-test).

      Discussion

      Histological grading of liver biopsies is the gold standard for diagnosis and prognosis of NASH
      • Campos G.M.
      • Bambha K.
      • Vittinghoff E.
      • Rabl C.
      • Posselt A.M.
      • Ciovica R.
      • et al.
      A clinical scoring system for predicting nonalcoholic steatohepatitis in morbidly obese patients.
      ,
      • Kleiner D.E.
      • Brunt E.M.
      • Van Natta M.
      • Behling C.
      • Contos M.J.
      • Cummings O.W.
      • et al.
      Design and validation of a histological scoring system for nonalcoholic fatty liver disease.
      . Proteins involved in liver ECM remodelling are promising non-invasive biomarkers of fibrosis
      • Abdelaziz R.
      • Elbasel M.
      • Esmat S.
      • Essam K.
      • Abdelaaty S.
      Tissue Inhibitors of Metalloproteinase-1 and 2 and Obesity Related Non-Alcoholic Fatty Liver Disease: Is There a Relationship.
      ,
      • Boyle M.
      • Tiniakos D.
      • Schattenberg J.M.
      • Ratziu V.
      • Bugianessi E.
      • Petta S.
      • et al.
      Performance of the PRO-C3 collagen neo-epitope biomarker in non-alcoholic fatty liver disease.
      ,
      • Saekmose S.G.
      • Mossner B.
      • Christensen P.B.
      • Lindvig K.
      • Schlosser A.
      • Holst R.
      • et al.
      Microfibrillar-Associated Protein 4: A Potential Biomarker for Screening for Liver Fibrosis in a Mixed Patient Cohort.
      . Activated HSCs are believed to be the main contributors of ECM deposition in the NASH liver. Our aim, therefore, was to identify HSC-expressed ECM proteins related to NAFLD progression by transcriptomic analysis of human liver needle biopsies from severely obese patients populating the NAFLD disease spectrum. Using the powerful WGCNA approach, we identified a module of co-expressed genes relating to fibrogenesis, which correlated with histological grades. Within this module, SMOC2 was found to encode a secreted protein and upregulated in livers of patients with advanced NAFLD. We integrated publicly available scRNA-seq data to deconvolve our bulk RNA-seq data and resolve SMOC2 expression in individual cell types. We found SMOC2 to be expressed exclusively in mesenchymal cells. Using smFISH on human liver needle biopsies, we validated SMOC2 expression in HSCs and found aHSCs/fibroblasts to be the main source of SMOC2 in the NASH liver. Finally, we found elevated plasma protein levels of SMOC2 in NASH patients compared to no-NASH, pointing to plasma SMOC2 a potential non-invasive biomarker of NASH.
      SMOC2 encodes a matricellular protein (MCP) of the secreted protein acidic and cysteine-rich (SPARC) family of MCPs
      • Vannahme C.
      • Gosling S.
      • Paulsson M.
      • Maurer P.
      • Hartmann U.
      Characterization of SMOC-2, a modular extracellular calcium-binding protein.
      . MCPs are non-structural components of the ECM, which bind growth factors, cytokines, and chemokines thereby playing pivotal roles in ECM-cell signal transduction
      • Bornstein P.
      Matricellular proteins: an overview.
      . Following injury, MCPs are secreted into the ECM to facilitate cell signalling, migration, and adhesion
      • Bornstein P.
      • Sage E.H.
      Matricellular proteins: extracellular modulators of cell function.
      . SMOC2 is associated with fibrosis, inflammation, and cell growth acting downstream of TGF-β1 signalling
      • Luo L.
      • Wang C.C.
      • Song X.P.
      • Wang H.M.
      • Zhou H.
      • Sun Y.
      • et al.
      Suppression of SMOC2 reduces bleomycin (BLM)-induced pulmonary fibrosis by inhibition of TGF-beta1/SMADs pathway.
      • Su J.R.
      • Kuai J.H.
      • Li Y.Q.
      Smoc2 potentiates proliferation of hepatocellular carcinoma cells via promotion of cell cycle progression.
      • Liu P.
      • Lu J.
      • Cardoso W.V.
      • Vaziri C.
      The SPARC-related factor SMOC-2 promotes growth factor-induced cyclin D1 expression and DNA synthesis via integrin-linked kinase.
      . In renal fibrosis, SMOC2 has been shown to play an important role and suggested as a potential biomarker
      • Gerarduzzi C.
      • Kumar R.K.
      • Trivedi P.
      • Ajay A.K.
      • Iyer A.
      • Boswell S.
      • et al.
      Silencing SMOC2 ameliorates kidney fibrosis by inhibiting fibroblast to myofibroblast transformation.
      ,
      • Schmidt I.M.
      • Colona M.R.
      • Kestenbaum B.R.
      • Alexopoulos L.G.
      • Palsson R.
      • Srivastava A.
      • et al.
      Cadherin-11, Sparc-related modular calcium binding protein-2, and Pigment epithelium-derived factor are promising non-invasive biomarkers of kidney fibrosis.
      . Whole-body Smoc2 ablation in mice ameliorated diet-induced obesity, hepatic steatosis, and fibrosis
      • Yuting Y.
      • Lifeng F.
      • Qiwei H.
      Secreted modular calcium-binding protein 2 promotes high fat diet (HFD)-induced hepatic steatosis through enhancing lipid deposition, fibrosis and inflammation via targeting TGF-beta1.
      . The study further showed elevated SMOC2 mRNA and protein levels in human steatotic liver biopsies and proposed a fibrogenic role for hepatocyte-derived SMOC2 through direct interaction with TGF-β1. Using publicly available scRNA-seq data, we identified HSCs and a proportion of VSMCs as the SMOC2-expressing cell types of the human liver. Using smFISH, we pinpointed aHSCs/fibroblasts as the main SMOC2-expressing cell type of the human liver. Expression of SMOC2 increased with NAFLD severity in our cohort and we identified LUM+ and LUM+FBLN2+ aHSCs/fibroblasts to be the main source of this increase in the NASH liver. The fraction of SMOC2+FBLN2+ decreased significantly in NASH compared to no-NASH, which indicates portal fibroblasts express SMOC2 in the healthy human liver as suspected. In contrast, the fraction of SMOC2+LUM+FBLN2+ cells increased significantly in NASH compared to no-NASH indicating aHSCs/fibroblasts express SMOC2 in the human NASH liver. In the cirrhotic human liver, aHSCs are the main fibrogenic cells
      • Ramachandran P.
      • Dobie R.
      • Wilson-Kanamori J.R.
      • Dora E.F.
      • Henderson B.E.P.
      • Luu N.T.
      • et al.
      Resolving the fibrotic niche of human liver cirrhosis at single-cell level.
      , which in different murine models of fibrosis account for 80% to 95% of the myofibroblasts
      • Mederacke I.
      • Hsu C.C.
      • Troeger J.S.
      • Huebener P.
      • Mu X.
      • Dapito D.H.
      • et al.
      Fate tracing reveals hepatic stellate cells as dominant contributors to liver fibrosis independent of its aetiology.
      . This suggests that increased plasma levels of SMOC2 reflect an expansion of aHSCs during NAFLD progression rather than SMOC2 expression in hepatocytes or portal fibroblasts. Still, aHSC/fibroblast-derived SMOC2 could contribute to NAFLD progression. Recently, SMOC2 was shown to promote lung fibroblast-to-myofibroblast transformation in vitro through activation of ERK and AKT pathways
      • Wang Y.
      • Yang H.
      • Su X.
      • Cao A.
      • Chen F.
      • Chen P.
      • et al.
      TGF-beta1/SMOC2/AKT and ERK axis regulates proliferation, migration, and fibroblast to myofibroblast transformation in lung fibroblast, contributing with the asthma progression.
      . Through similar mechanisms, SMOC2 may play a role in also priming activation and transdifferentiation of HSCs during NAFLD progression, which we previously demonstrated, depends on MEK-ERK signalling
      • Marcher A.B.
      • Bendixen S.M.
      • Terkelsen M.K.
      • Hohmann S.S.
      • Hansen M.H.
      • Larsen B.D.
      • et al.
      Transcriptional regulation of Hepatic Stellate Cell activation in NASH.
      .
      We tested performance of SMOC2 expression in discriminating NASH from no-NASH in our cohort and benchmarked predictive performance of SMOC2 against previously proposed biomarkers using a comprehensive public RNA-seq study from a multi-centre NAFLD cohort
      • Govaere O.
      • Cockell S.
      • Tiniakos D.
      • Queen R.
      • Younes R.
      • Vacca M.
      • et al.
      Transcriptomic profiling across the nonalcoholic fatty liver disease spectrum reveals gene signatures for steatohepatitis and fibrosis.
      . In our cohort, SMOC2 expression exhibited good performance (AUROC 0.79-0.90) in discriminating NAFLD severity defined by NAS score, Kleiner fibrosis grade, and SAF score. In the multi-centre cohort, SMOC2 expression exhibited modest-to-good performance (AUROC 0.67-0.83), which was comparable to previously proposed biomarkers. Measuring non-invasive biomarkers in liquid biopsies, however, is of high value. In discriminating NASH from no-NASH, we found plasma SMOC2 exhibited good performance in predicting fibrosis (AUROC 0.80, Kleiner fibrosis grade ≥ 2) and excellent performance in predicting severe NAFLD (AUROC 0.89, NAS ≥ 4) and NASH (AUROC 0.88, SAF = 3). Variations in specificity ranges (0.80-0.89) and high variations in sensitivity ranges (0.42-0.81) were observed. Restriction on tissue volume in liver biopsies and subsequent underestimation of disease severity as well as sampling bias is a known problem
      • Ratziu V.
      • Charlotte F.
      • Heurtier A.
      • Gombert S.
      • Giral P.
      • Bruckert E.
      • et al.
      Sampling variability of liver biopsy in nonalcoholic fatty liver disease.
      . Predictive performance of plasma SMOC2 may, consequently, be restricted by a discordance with histological scores. Of note, plasma SMOC2 had a low specificity (0.42) in predicting fibrosis. Several patients (n = 8) in the mild fibrosis group were graded with steatosis ≥ 2, lobular inflammation ≥ 2, and ballooning ≥ 1. Thus, in line with a potential role of priming activation of HSCs, plasma SMOC2 may reflect cellular changes related to active fibrogenesis rather than fibrosis. Our evaluation of plasma SMOC2 as a non-invasive biomarker is, moreover, restricted by size of our cohort and distribution of histological grades and gender. Thus, a larger cohort is required for further analysis and establishment of plasma SMOC2 as a non-invasive biomarker for diagnosis of NASH and, moreover, exclude potential gender-specific differences in plasma SMOC2 levels.
      NASH co-morbidities and overall low tissue-specificity of SMOC2 may compromise the specificity of SMOC2 as a NASH biomarker. Dyslipidaemia-related cardiovascular disease is the leading cause of death in patients with NASH
      • Ekstedt M.
      • Hagstrom H.
      • Nasr P.
      • Fredrikson M.
      • Stal P.
      • Kechagias S.
      • et al.
      Fibrosis stage is the strongest predictor for disease-specific mortality in NAFLD after up to 33 years of follow-up.
      ,
      • Samuel V.T.
      • Shulman G.I.
      Nonalcoholic Fatty Liver Disease as a Nexus of Metabolic and Hepatic Diseases.
      and atherosclerosis-related fibrosis
      • Lan T.H.
      • Huang X.Q.
      • Tan H.M.
      Vascular fibrosis in atherosclerosis.
      may contribute to elevated plasma SMOC2 levels in NASH. An aspect of SMOC2 biology that deserves further investigation. The core fibrosis signalling pathway induced by persistent injury-induced inflammation involves activation and transdifferentiation of mesenchymal cells into scar-forming myofibroblasts
      • Henderson N.C.
      • Rieder F.
      • Wynn T.A.
      Fibrosis: from mechanisms to medicines.
      . Murine SMOC2 has been implicated in the development of renal
      • Gerarduzzi C.
      • Kumar R.K.
      • Trivedi P.
      • Ajay A.K.
      • Iyer A.
      • Boswell S.
      • et al.
      Silencing SMOC2 ameliorates kidney fibrosis by inhibiting fibroblast to myofibroblast transformation.
      , skeletal muscle
      • Schuler S.C.
      • Kirkpatrick J.M.
      • Schmidt M.
      • Santinha D.
      • Koch P.
      • Di Sanzo S.
      • et al.
      Extensive remodeling of the extracellular matrix during aging contributes to age-dependent impairments of muscle stem cell functionality.
      , and pulmonary fibrosis
      • Luo L.
      • Wang C.C.
      • Song X.P.
      • Wang H.M.
      • Zhou H.
      • Sun Y.
      • et al.
      Suppression of SMOC2 reduces bleomycin (BLM)-induced pulmonary fibrosis by inhibition of TGF-beta1/SMADs pathway.
      suggesting that at least in mice, SMOC2 is part of a core regenerative signalling pathway. Human SMOC2 may have a similar function in tissue regeneration and elevated plasma SMOC2 levels in NASH patients may, thus, derive from fibrosis in other tissues. In the current study, the estimate glomerular filtration rate of the cohort did not indicate chronic kidney disease and the CRP levels did not differ between groups. Moreover, we excluded adipose tissue expression of SMOC2 as a source of elevated plasma SMOC2.

      Conclusion

      In conclusion, we have identified increased hepatic SMOC2 expression and concomitant elevated plasma SMOC2 level as a novel biomarker for diagnosis of NASH. We described cell-type specific expression of SMOC2 by HSCs/fibroblasts thereby linking SMOC2 to a key cell type in NAFLD and fibrosis progression. Plasma SMOC2 in severely obese may hence reflect liver fibrogenesis and be useful as diagnostic tool in stratification of patients for further examination and treatment of NASH. Combination of SMOC2 with other proposed biomarkers such as TREM2, may aid in diagnosis and prognosis of NAFLD patients and deserves further investigation.

      Conflict of interest

      All authors declare that they have no conflicts of interest.

      Author Contributions

      Study conceptualization: A.A.K, M.E.M.L, C.W.W., L.G., and K.R.; Manuscript preparation: F.T.L. and K.R.; NGS data analysis: F.T.L., M.K.T., and K.R.; Statistical analysis: F.T.L.; Imaging data analysis: F.T.L. and D.H.; Laboratory and practical work: F.T.L., D.H., S.M.B., V.I.H., J.H.G., M.E.M.L, C.W.W., M.S., E.G., L.L.G, and B.G.J. Critical review during manuscript preparation: K.R., D.H., V.I.H., J.H.G., M.E.M.L, L.G., M.S., E.G., C.W.W., and A.A.K. K.R. supervised, managed, and coordinated the study.

      Financial support

      Danish National Research Foundation (grant DNRF141); Danish Diabetes Academy fellowship (D.H.) NNF17SA0031406

      Data availability

      Data are accessible through GEO Series accession number GSE207310.

      Acknowledgment:

      We thank Tenna P. Mortensen, Maibrith Wishoff, and Signe Marie Andersen for expert technical assistance. We thank Casimiro Gerarduzzi (affiliation) for valuable intellectual input. Bioimaging was performed at DaMBIC, a bioimaging research core facility at University of Southern Denmark (SDU), established by an equipment grant from the Danish Agency for Science, Technology and Innovation and internal funding from SDU.
      This work was supported by the Danish National Research Foundation (grant DNRF141) to Center for Functional Genomics and Tissue Plasticity (ATLAS), and a Fellowship (D.H.) from the Danish Diabetes Academy, which is funded by the Novo Nordisk Foundation, grant number NNF17SA0031406.

      Appendix A. Supplementary data

      References

      Author names in bold designate shared co-first authorship.
      1. Organization WH. WHO European Regional Obesity Report 2022. Copenhagen: WHO Regional Office for Europe; 2022: World Health Organization; 2022.

        • Younossi Z.M.
        • Golabi P.
        • de Avila L.
        • Paik J.M.
        • Srishord M.
        • Fukui N.
        • et al.
        The global epidemiology of NAFLD and NASH in patients with type 2 diabetes: A systematic review and meta-analysis.
        J Hepatol. 2019; 71: 793-801
        • Hagstrom H.
        • Nasr P.
        • Ekstedt M.
        • Hammar U.
        • Widman L.
        • Stal P.
        • et al.
        Health Care Costs of Patients With Biopsy-Confirmed Nonalcoholic Fatty Liver Disease Are Nearly Twice Those of Matched Controls.
        Clin Gastroenterol Hepatol. 2020; 18 (e1598): 1592-1599
        • O'Hara J.
        • Finnegan A.
        • Dhillon H.
        • Ruiz-Casas L.
        • Pedra G.
        • Franks B.
        • et al.
        Cost of non-alcoholic steatohepatitis in Europe and the USA: The GAIN study.
        JHEP Rep. 2020; 2100142
        • Matteoni C.A.
        • Younossi Z.M.
        • Gramlich T.
        • Boparai N.
        • Liu Y.C.
        • McCullough A.J.
        Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity.
        Gastroenterology. 1999; 116: 1413-1419
        • Campos G.M.
        • Bambha K.
        • Vittinghoff E.
        • Rabl C.
        • Posselt A.M.
        • Ciovica R.
        • et al.
        A clinical scoring system for predicting nonalcoholic steatohepatitis in morbidly obese patients.
        Hepatology. 2008; 47: 1916-1923
        • Kleiner D.E.
        • Brunt E.M.
        • Van Natta M.
        • Behling C.
        • Contos M.J.
        • Cummings O.W.
        • et al.
        Design and validation of a histological scoring system for nonalcoholic fatty liver disease.
        Hepatology. 2005; 41: 1313-1321
        • McGill D.B.
        • Rakela J.
        • Zinsmeister A.R.
        • Ott B.J.
        A 21-year experience with major hemorrhage after percutaneous liver biopsy.
        Gastroenterology. 1990; 99: 1396-1400
        • Vuppalanchi R.
        • Unalp A.
        • Van Natta M.L.
        • Cummings O.W.
        • Sandrasegaran K.E.
        • Hameed T.
        • et al.
        Effects of liver biopsy sample length and number of readings on sampling variability in nonalcoholic Fatty liver disease.
        Clin Gastroenterol Hepatol. 2009; 7: 481-486
        • Ratziu V.
        • Charlotte F.
        • Heurtier A.
        • Gombert S.
        • Giral P.
        • Bruckert E.
        • et al.
        Sampling variability of liver biopsy in nonalcoholic fatty liver disease.
        Gastroenterology. 2005; 128: 1898-1906
        • Davison B.A.
        • Harrison S.A.
        • Cotter G.
        • Alkhouri N.
        • Sanyal A.
        • Edwards C.
        • et al.
        Suboptimal reliability of liver biopsy evaluation has implications for randomized clinical trials.
        Journal of Hepatology. 2020; 73: 1322-1332
        • Wong V.W.
        • Adams L.A.
        • de Ledinghen V.
        • Wong G.L.
        • Sookoian S.
        Noninvasive biomarkers in NAFLD and NASH - current progress and future promise.
        Nat Rev Gastroenterol Hepatol. 2018; 15: 461-478
        • Tsuchida T.
        • Friedman S.L.
        Mechanisms of hepatic stellate cell activation.
        Nat Rev Gastroenterol Hepatol. 2017; 14: 397-411
        • Abdelaziz R.
        • Elbasel M.
        • Esmat S.
        • Essam K.
        • Abdelaaty S.
        Tissue Inhibitors of Metalloproteinase-1 and 2 and Obesity Related Non-Alcoholic Fatty Liver Disease: Is There a Relationship.
        Digestion. 2015; 92: 130-137
        • Boyle M.
        • Tiniakos D.
        • Schattenberg J.M.
        • Ratziu V.
        • Bugianessi E.
        • Petta S.
        • et al.
        Performance of the PRO-C3 collagen neo-epitope biomarker in non-alcoholic fatty liver disease.
        JHEP Rep. 2019; 1: 188-198
        • Luo Y.
        • Oseini A.
        • Gagnon R.
        • Charles E.D.
        • Sidik K.
        • Vincent R.
        • et al.
        An Evaluation of the Collagen Fragments Related to Fibrogenesis and Fibrolysis in Nonalcoholic Steatohepatitis.
        Sci Rep. 2018; 812414
        • Saekmose S.G.
        • Mossner B.
        • Christensen P.B.
        • Lindvig K.
        • Schlosser A.
        • Holst R.
        • et al.
        Microfibrillar-Associated Protein 4: A Potential Biomarker for Screening for Liver Fibrosis in a Mixed Patient Cohort.
        PLoS One. 2015; 10e0140418
        • Nascimbeni F.
        • Bedossa P.
        • Fedchuk L.
        • Pais R.
        • Charlotte F.
        • Lebray P.
        • et al.
        Clinical validation of the FLIP algorithm and the SAF score in patients with non-alcoholic fatty liver disease.
        J Hepatol. 2020; 72: 828-838
        • Dobin A.
        • Davis C.A.
        • Schlesinger F.
        • Drenkow J.
        • Zaleski C.
        • Jha S.
        • et al.
        STAR: ultrafast universal RNA-seq aligner.
        Bioinformatics. 2013; 29: 15-21
        • Liao Y.
        • Smyth G.K.
        • Shi W.
        featureCounts: an efficient general purpose program for assigning sequence reads to genomic features.
        Bioinformatics. 2014; 30: 923-930
      2. Andrews S. FastQC. 2019 [cited; Available from: https://www.bioinformatics.babraham.ac.uk/projects/fastqc/

        • Ewels P.
        • Magnusson M.
        • Lundin S.
        • Kaller M.
        MultiQC: summarize analysis results for multiple tools and samples in a single report.
        Bioinformatics. 2016; 32: 3047-3048
        • Ziegenhain C.
        • Sandberg R.
        BAMboozle removes genetic variation from human sequence data for open data sharing.
        Nat Commun. 2021; 12: 6216
        • Ramachandran P.
        • Dobie R.
        • Wilson-Kanamori J.R.
        • Dora E.F.
        • Henderson B.E.P.
        • Luu N.T.
        • et al.
        Resolving the fibrotic niche of human liver cirrhosis at single-cell level.
        Nature. 2019; 575: 512-518
        • MacParland S.A.
        • Liu J.C.
        • Ma X.Z.
        • Innes B.T.
        • Bartczak A.M.
        • Gage B.K.
        • et al.
        Single cell RNA sequencing of human liver reveals distinct intrahepatic macrophage populations.
        Nat Commun. 2018; 9: 4383
        • Payen V.L.
        • Lavergne A.
        • Alevra Sarika N.
        • Colonval M.
        • Karim L.
        • Deckers M.
        • et al.
        Single-cell RNA sequencing of human liver reveals hepatic stellate cell heterogeneity.
        JHEP Rep. 2021; 3100278
        • Hao Y.
        • Hao S.
        • Andersen-Nissen E.
        • Mauck 3rd, W.M.
        • Zheng S.
        • Butler A.
        • et al.
        Integrated analysis of multimodal single-cell data.
        Cell. 2021; 184: 3573-3587 e3529
        • McGinnis C.S.
        • Murrow L.M.
        • Gartner Z.J.
        DoubletFinder: Doublet Detection in Single-Cell RNA Sequencing Data Using Artificial Nearest Neighbors.
        Cell Syst. 2019; 8: 329-337 e324
        • Korsunsky I.
        • Millard N.
        • Fan J.
        • Slowikowski K.
        • Zhang F.
        • Wei K.
        • et al.
        Fast, sensitive and accurate integration of single-cell data with Harmony.
        Nat Methods. 2019; 16: 1289-1296
        • Dominguez Conde C.
        • Xu C.
        • Jarvis L.B.
        • Rainbow D.B.
        • Wells S.B.
        • Gomes T.
        • et al.
        Cross-tissue immune cell analysis reveals tissue-specific features in humans.
        Science. 2022; 376eabl5197
        • Robin X.
        • Turck N.
        • Hainard A.
        • Tiberti N.
        • Lisacek F.
        • Sanchez J.C.
        • et al.
        pROC: an open-source package for R and S+ to analyze and compare ROC curves.
        BMC Bioinformatics. 2011; 12: 77
        • Govaere O.
        • Cockell S.
        • Tiniakos D.
        • Queen R.
        • Younes R.
        • Vacca M.
        • et al.
        Transcriptomic profiling across the nonalcoholic fatty liver disease spectrum reveals gene signatures for steatohepatitis and fibrosis.
        Sci Transl Med. 2020; 12
        • Terkelsen M.K.
        • Bendixen S.M.
        • Hansen D.
        • Scott E.A.H.
        • Moeller A.F.
        • Nielsen R.
        • et al.
        Transcriptional Dynamics of Hepatic Sinusoid-Associated Cells After Liver Injury.
        Hepatology. 2020; 72: 2119-2133
        • Gong X.M.
        • Li Y.F.
        • Luo J.
        • Wang J.Q.
        • Wei J.
        • Wang J.Q.
        • et al.
        Gpnmb secreted from liver promotes lipogenesis in white adipose tissue and aggravates obesity and insulin resistance.
        Nat Metab. 2019; 1: 570-583
        • Wang Z.Y.
        • Keogh A.
        • Waldt A.
        • Cuttat R.
        • Neri M.
        • Zhu S.
        • et al.
        Single-cell and bulk transcriptomics of the liver reveals potential targets of NASH with fibrosis.
        Sci Rep. 2021; 1119396
        • Dobie R.
        • Wilson-Kanamori J.R.
        • Henderson B.E.P.
        • Smith J.R.
        • Matchett K.P.
        • Portman J.R.
        • et al.
        Single-Cell Transcriptomics Uncovers Zonation of Function in the Mesenchyme during Liver Fibrosis.
        Cell Rep. 2019; 29: 1832-1847 e1838
        • Bracht T.
        • Schweinsberg V.
        • Trippler M.
        • Kohl M.
        • Ahrens M.
        • Padden J.
        • et al.
        Analysis of disease-associated protein expression using quantitative proteomics-fibulin-5 is expressed in association with hepatic fibrosis.
        J Proteome Res. 2015; 14: 2278-2286
      3. Chandran VI, Wernberg CW, Lauridsen MM, Skytthe MK, Bendixen SM, Larsen FT, et al. Circulating TREM2 as a non-invasive diagnostic biomarker for NASH in patients with elevated liver stiffness. Hepatology;n/a.

      4. Hendrikx T, Porsch F, Kiss MG, Rajcic D, Papac-Miličević N, Hoebinger C, et al. Soluble TREM2 levels reflect the recruitment and expansion of TREM2+ macrophages that localize to fibrotic areas and limit NASH. Journal of Hepatology.

        • Vannahme C.
        • Gosling S.
        • Paulsson M.
        • Maurer P.
        • Hartmann U.
        Characterization of SMOC-2, a modular extracellular calcium-binding protein.
        Biochem J. 2003; 373: 805-814
        • Bornstein P.
        Matricellular proteins: an overview.
        J Cell Commun Signal. 2009; 3: 163-165
        • Bornstein P.
        • Sage E.H.
        Matricellular proteins: extracellular modulators of cell function.
        Curr Opin Cell Biol. 2002; 14: 608-616
        • Luo L.
        • Wang C.C.
        • Song X.P.
        • Wang H.M.
        • Zhou H.
        • Sun Y.
        • et al.
        Suppression of SMOC2 reduces bleomycin (BLM)-induced pulmonary fibrosis by inhibition of TGF-beta1/SMADs pathway.
        Biomed Pharmacother. 2018; 105: 841-847
        • Su J.R.
        • Kuai J.H.
        • Li Y.Q.
        Smoc2 potentiates proliferation of hepatocellular carcinoma cells via promotion of cell cycle progression.
        World J Gastroenterol. 2016; 22: 10053-10063
        • Liu P.
        • Lu J.
        • Cardoso W.V.
        • Vaziri C.
        The SPARC-related factor SMOC-2 promotes growth factor-induced cyclin D1 expression and DNA synthesis via integrin-linked kinase.
        Mol Biol Cell. 2008; 19: 248-261
        • Gerarduzzi C.
        • Kumar R.K.
        • Trivedi P.
        • Ajay A.K.
        • Iyer A.
        • Boswell S.
        • et al.
        Silencing SMOC2 ameliorates kidney fibrosis by inhibiting fibroblast to myofibroblast transformation.
        JCI Insight. 2017; 2
        • Schmidt I.M.
        • Colona M.R.
        • Kestenbaum B.R.
        • Alexopoulos L.G.
        • Palsson R.
        • Srivastava A.
        • et al.
        Cadherin-11, Sparc-related modular calcium binding protein-2, and Pigment epithelium-derived factor are promising non-invasive biomarkers of kidney fibrosis.
        Kidney Int. 2021; 100: 672-683
        • Yuting Y.
        • Lifeng F.
        • Qiwei H.
        Secreted modular calcium-binding protein 2 promotes high fat diet (HFD)-induced hepatic steatosis through enhancing lipid deposition, fibrosis and inflammation via targeting TGF-beta1.
        Biochem Biophys Res Commun. 2019; 509: 48-55
        • Mederacke I.
        • Hsu C.C.
        • Troeger J.S.
        • Huebener P.
        • Mu X.
        • Dapito D.H.
        • et al.
        Fate tracing reveals hepatic stellate cells as dominant contributors to liver fibrosis independent of its aetiology.
        Nat Commun. 2013; 4: 2823
        • Wang Y.
        • Yang H.
        • Su X.
        • Cao A.
        • Chen F.
        • Chen P.
        • et al.
        TGF-beta1/SMOC2/AKT and ERK axis regulates proliferation, migration, and fibroblast to myofibroblast transformation in lung fibroblast, contributing with the asthma progression.
        Hereditas. 2021; 158: 47
        • Marcher A.B.
        • Bendixen S.M.
        • Terkelsen M.K.
        • Hohmann S.S.
        • Hansen M.H.
        • Larsen B.D.
        • et al.
        Transcriptional regulation of Hepatic Stellate Cell activation in NASH.
        Sci Rep. 2019; 9: 2324
        • Ekstedt M.
        • Hagstrom H.
        • Nasr P.
        • Fredrikson M.
        • Stal P.
        • Kechagias S.
        • et al.
        Fibrosis stage is the strongest predictor for disease-specific mortality in NAFLD after up to 33 years of follow-up.
        Hepatology. 2015; 61: 1547-1554
        • Samuel V.T.
        • Shulman G.I.
        Nonalcoholic Fatty Liver Disease as a Nexus of Metabolic and Hepatic Diseases.
        Cell Metab. 2018; 27: 22-41
        • Lan T.H.
        • Huang X.Q.
        • Tan H.M.
        Vascular fibrosis in atherosclerosis.
        Cardiovasc Pathol. 2013; 22: 401-407
        • Henderson N.C.
        • Rieder F.
        • Wynn T.A.
        Fibrosis: from mechanisms to medicines.
        Nature. 2020; 587: 555-566
        • Schuler S.C.
        • Kirkpatrick J.M.
        • Schmidt M.
        • Santinha D.
        • Koch P.
        • Di Sanzo S.
        • et al.
        Extensive remodeling of the extracellular matrix during aging contributes to age-dependent impairments of muscle stem cell functionality.
        Cell Rep. 2021; 35109223