Epigenetic therapy based on micrornas inhibition for the amelioration of masld phenotype

UNIVERSIDAD DE SEVILLA FACULTAD DE MEDICINA Departamento de Medicina “EPIGENETIC THERAPY BASED ON MICRORNAS INHIBITION FOR THE AMELIORATION OF MASLD PHENOTYPE” TESIS DOCTORAL Rocío Montero Vallejo SeLiver Group. UGC Aparato Digestivo. Instituto de Biomedicina de Sevilla (IBiS). Hospital Universitario Virgen del Rocío. Sevilla, 2024 2 3 D. MANUEL ROMERO GÓMEZ, CATEDRÁTICO DE UNIVERSIDAD DEL DEPARTAMENTO DE MEDICINA DE LA UNIVERSIDAD DE SEVILLA, CERTIFICA: Que la Tesis Doctoral que lleva por título: “EPIGENETIC THERAPY BASED ON MICRORNAS INHIBITION FOR THE AMELIORATION OF NAFLD PHENOTYPE”, presentada por Dña. Rocío Montero Vallejo para optar al grado de Doctor con Mención Internacional, ha sido realizada en el Departamento de Medicina de la Facultad de Medicina de la Universidad de Sevilla. Revisado el texto, doy mi conformidad para su presentación upregulated defensa para optar al grado de Doctor con Mención Internacional por la Universidad de Sevilla. Fdo. Prof. Dr. Manuel Romero Gómez Director up-regulated Tutor de la Tesis Doctoral 30 de enero de 2024 4 5 DÑA. ROCÍO GALLEGO DURÁN, INVESTIGADORA POSTDOCTORAL DEL INSTITUTO DE BIOMEDICINA DE SEVILLA, CERTIFICA: Que la Tesis Doctoral que lleva por título: “EPIGENETIC THERAPY BASED ON MICRORNAS INHIBITION FOR THE AMELIORATION OF NAFLD PHENOTYPE”, presentada por Dña. Rocío Montero Vallejo para optar al grado de Doctor con Mención Internacional, ha sido realizada en el Departamento de Medicina de la Facultad de Medicina de la Universidad de Sevilla. Revisado el texto, doy mi conformidad para su presentación upregulated defensa para optar al grado de Doctor con Mención Internacional por la Universidad de Sevilla. Fdo. Dra. Rocío Gallego Durán Co-directora de la Tesis Doctoral 30 de enero de 2024 6 7 ABBREVIATIONS .............................................. 11 ABSTRACT ......................................................... 15 1. INTRODUCTION ......................................... 17 1.1 MASLD: definition, incidence, prevalence and risk factors ............................................................... 17 1.2 MASLD: Diagnosis ........................................... 25 1.2.1 Liver biopsy procedures .......................................... 25 1.3 MASLD therapeutic approaches ....................... 29 1.4 Endothelial dysfunction MASLD-related ......... 31 1.5 MicroRNAs ....................................................... 35 1.5.1 MicroRNAS: definition, mechanism of action and link with MASLD. ............................................................ 35 1.5.2 MicroRNAs: epigenetic therapies ........................... 40 1.6 Preclinical models in MASLD .......................... 41 2. HYPOTHESIS ............................................... 49 3. OBJECTIVES ................................................ 51 4. MATERIALS AND METHODS .................. 53 4.1 Animals: Overall design .................................... 53 4.1.1 Ethics ....................................................................... 54 4.1.2 Experimental design: groups of study ..................... 54 4.1.3 Dietary regimen ....................................................... 57 4.2 Experimental procedures ................................... 61 4.2.1 Treatment: antisense therapy with morpholinos ...... 63 4.2.1.1 Groups of study ......................................... 63 4.2.1.2 Treatment administration ........................... 65 4.2.2 MiRNAs isolation .................................................... 66 4.2.2.1 Hepatic miRNAs ........................................ 66 4.2.2.2 Circulating miRNAs .................................. 66 4.2.3 RT-qPCR ................................................................. 67 4.2.4 Histological evaluation ............................................ 72 4.2.5 GO and KEGG analysis ........................................... 74 4.2.6 Study design for the clinical settings ....................... 74 8 4.2.6.1 Measurements including clinical, anthropometric and biochemical parameters ......... 75 4.2.6.2 Blood collection and processing procedures ............................................................................... 76 4.2.6.3 Liver biopsy and patient classification ...... 76 4.2.7 Enzyme-linked immunosorbent assay (ELISA) ...... 77 4.2.8 Immunohistochemistry ............................................ 78 4.2.9 Statistical analysis ................................................... 79 5 RESULTS .......................................................... 81 5.1 MicroRNAs in MASLD .................................... 81 5.1.1 Pilot study ................................................................ 81 5.1.1.1 To assess the suitability of the selected animal model prior to the morpholino intervention at the hepatic and metabolic levels. ................................. 81 5.1.2 Exploratory study .................................................... 91 5.1.2.1 To analyse whether morpholino injection improves the metabolic phenotype and other surrogate markers of liver disease ......................... 91 5.1.2.2 To determine whether microRNAs are overexpressed in the context of a diet inducing NAFLD. ............................................................... 103 5.1.2.3 To evaluate the impact of the use of vehicle in these animals ................................................... 105 5.1.2.4 To evaluate the successful rate of the experimental treatment based on morpholinos injection ............................................................... 107 5.1.2.5 To determine the effectiveness of the treatment to ameliorate the hepatic injury or to revert the phenotype ....................................................... 111 5.1.3 Validation study ..................................................... 125 5.1.3.1 To analyse whether morpholino injection improves the metabolic phenotype and other surrogate markers of liver disease ....................... 125 9 5.1.3.2 To determine whether the miRNA-200b was overexpressed in MASLD mice .......................... 134 5.1.3.3 To evaluate the impact of the vehicle in disease progression .............................................. 135 5.1.3.4 To assess whether the morpholino injection has been successful .............................................. 136 5.1.3.5 To determine the effectiveness of morpholino injections at the hepatic level ............................... 138 5.1.3.6 Other microRNAs of study ...................... 142 5.1.4 The role of miRNAs in biopsy-proven MASLD patients ............................................................................ 144 5.1.4.1 Baseline patients´ characteristics ............. 144 5.1.4.2 miRNAs expression ................................. 145 5.2 Endothelial dysfunction MASLD related ....... 147 5.2.1 Mice study ............................................................. 147 5.2.1.1 Genes associated with endothelial dysfunction. ......................................................... 147 5.2.1.2 Hepatic protein levels of ED markers ...... 148 5.2.2 Clinical study ......................................................... 148 5.2.2.1 ELISAS .................................................... 148 5.2.2.2 Hepatic expression of endothelial dysfunction biomarkers ........................................................... 154 6. DISCUSSION .............................................. 159 7. CONCLUSION ............................................ 171 REFERENCES ................................................... 173 ANNEXES .......................................................... 185 16 on MASLD, and to assess their inhibition. After extensive screenings, three microRNAs, miR-200b, miR-409 and miR- 146b, among others, were found up-regulated in mice fed with choline-deficient, L-amino acid-defined, high-fat diet (CDAHFD), and their expression was blocked by using in vivo morpholinos. After several analyses, performed on explorative sets of animals, microRNA 200b was selected over the other two and validated in a further cohort of mice. In summary, this study led us to the conclusion that miR-200b was found to be increased in mice fed with CDA-HFD, and that after inhibition of the expression of this microRNA with vivo morpholino injection, an improvement in liver fibrosis was observed. However, it was not sufficient to reverse the liver damage. Further studies are necessary in another animal model but this microRNA could be a potential therapeutic target for MASLD. 17 1. INTRODUCTION 1.1 MASLD: definition, incidence, prevalence and risk factors Metabolic dysfunction-associated steatotic liver disease (MASLD), formerly known as non-alcoholic fatty liver disease or NAFLD1 is the most prevalent liver disorder worldwide. Based on a recent meta-analysis, MASLD prevalence rates have increased from 25.3% during 1990-2006 to 38.2% during 2016- 2019, representing an increase of 50.4% in the prevalence of MASLD over approximately three decades. This meta-analysis also showed that the highest prevalence of MASLD was found in Latin America (44.4%), and about 25% was found in Europe2. MASLD clinical spectrum is broad, ranging from simple steatosis characterized by more than 5% fat in the hepatocytes, initially benign, to metabolic dysfunction-associated steatohepatitis (MASH) formerly known as non-alcoholic steatohepatitis (NASH), which is characterized by inflammation and ballooning degeneration. Patients may experience various stages of liver fibrosis, which is currently considered the best predictor of 18 mortality among these patients. Fibrosis levels may lead to cirrhosis and cause hepatic decompensations, and eventually, hepatocellular carcinoma3 (Please refer to Figure 1). Formerly, it was believed that 20-30% of patients with steatosis would advance to MASH and only 2-5% eventually will develop cirrhosis and hepatocellular carcinoma. However, recent studies have shown that this disease usually follows a non-linear pattern of progression and regression, even in the case of fibrosis4. Figure 1. Metabolic dysfunction-associated steatotic liver disease (MASLD) development. MASH, metabolic dysfunction-associated steatohepatitis. Image created with BioRender.com3,5. MASLD is strongly associated with risk factors such as obesity, type 2 diabetes mellitus (T2DM) and metabolic syndrome and 19 therefore it can be considered the hepatic manifestation of the metabolic syndrome, strongly related to cardiovascular diseases6,7. Metabolic syndrome was defined by the National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III). It is diagnosed when an individual has three or more of the following criteria: increased waist circumference, blood pressure exceeding 130/85 mmHg, hypertriglyceridemia, highdensity lipoprotein (HDL) cholesterol level less than 40 mg/dl (men) or 50 mg/dl (women) and fasting blood sugar over 100 mg/dl8. However, metabolic alterations can be both, the cause or the consequence of liver diseases. On one hand, suffering from MASLD predisposes to metabolic comorbidities, and having metabolic comorbidities also increases the risk of MASLD9,10. Currently, the 'multiple hit' hypothesis is the most widely supported explanation for the pathophysiology of MASLD. This theory suggests that genetically predisposed individuals suffer from a mixture of insults that combine together to cause MASLD, such as environmental factors, insulin resistance, adipose tissue 20 secreted hormones, nutritional factors, gut microbiota and genetic and epigenetic factors11. For instance, insulin resistance plays a crucial role in the onset of steatosis/NASH, resulting in increased hepatic de novo lipogenesis (DNL) and diminished control over adipose tissue lipolysis inhibition. Consequently, there is an increased flow of fatty acids towards the liver12. Triglycerides accumulate in the liver along with elevated levels of free fatty acids, free cholesterol, and other lipid metabolites, resulting in heightened lipotoxicity. It is important to note that this is an objective evaluation of the effects of triglyceride accumulation in the liver. This, in turn, triggers mitochondrial dysfunction, oxidative stress, and the generation of reactive oxygen species (ROS), as well as mechanisms associated with endoplasmic reticulum (ER) stress11. Changes in gut microbiota result in additional production of fatty acids within the bowel, leading to the increased permeability of the barrier, increasing fatty acid absorption and raised circulating levels of molecules that contribute to the initiation of inflammatory pathways. This results in the release of 21 proinflammatory cytokines, including IL-6 and TNF-α13 (Please refer to Figure 2). Figure 2. The multiple-hit pathogenesis of MASLD (referred to as NALFD in the image initiates with the accumulation of lipids in the liver and insulin resistance. It advances to MASH through the collective impact of various factors, including inflammation, endotoxin, adipokines and hepatokines and oxidative stress14. In individuals with a predisposition to suffer from MASLD due to genetic factors or epigenetic modifications, all these factors collectively influence the fat content in hepatocytes and the inflammatory environment within the liver. This results in a condition of persistent hepatic inflammation through diverse pathways of hepatocellular damage, potentially progressing to hepatocellular death induced by both direct toxicity and apoptosis activation mechanisms. Additionally, there is the activation of 22 hepatic stellate cells and the deposition of a fibrous matrix, leading to the so-called fibrosis11!. Studies have shown that MASH can be the initial liver lesion, challenging the belief that steatosis is the first step in the disease spectrum. The presence of steatosis or MASH is determined by hepatic insults, such as the timing and interplay of genetic, environmental and intracellular events11. Genome-wide association studies (GWAS) have consistently shown that genetic variations in several genes, including patatinlike phospholipase domain-containing 3 (PNPLA3), transmembrane 6 superfamily member 2 (TM6SF2), membrane bound O-acyltransferase domain-containing 7 (MBOAT7), glucokinase regulator (GCKR) and hydroxysteroid 17-beta dehydrogenase 13 (HSD17B13) are strongly associated with the natural history of MASLD15. A single nucleotide polymorphism (SNP) is a variation in the DNA sequence that affects a single base in a genome sequence15. The most studied and relevant SNP in this disease is the rs738409 23 C>G variant in PNPLA3 gene. The PNPLA3 protein with wildtype characteristics demonstrates hydrolase activity towards triglycerides and retinyl esters16, and this SNP has proven its impact in facilitating the accumulation of fat in the liver, increasing MASLD susceptibility. Multiple studies have verified that the PNPLA3 variant is associated with the full spectrum of MASLD17,18 . In addition, the PNPLA3 variant was related with hepatic inflammation18. Substitution at position 148M results in a loss of enzymatic activity, leading to the sequestration of triglycerides and retinyl esters within lipid droplets in hepatocytes and hepatic stellate cells. This process may cause liver damage and prevent the release of extracellular proteins that normally counteract the fibrosis process16. Further, TM6SF2 is involved in the pathway of very low-density lipoprotein secretion from the hepatocyte19. The rs58542926 C>T TM6SF2 variant causes a loss of function resulting in increased liver triglyceride content and reduced circulating lipoproteins. Therefore, this SNP is strongly related with liver damage, promoting conditions such as MASH and fibrosis 15,16. 24 MBOAT7 is a protein involved in the remodeling of phosphatidylinositol with arachidonic acid in the Lands cycle16.! The protein expression reduction caused by the MBOAT variant rs641738 C>T has been linked to an increase in liver fat content and the development of MASLD, which includes inflammation, fibrosis, and HCC20. Besides, glucokinase regulates the storage and disposal of glucose in the liver. Its activity is regulated by the glucokinase regulatory protein (GKRP), GCKR gene21. The GCKR variant rs1260326 C>T encodes for a protein associated with steatosis and risk of development of MASLD. On the other hand, the rs780094 SNP C>T was linked with elevated triglyceride levels, the development of MASLD and liver fibrosis 15,22. Finally, HSD17B13 gene located in lipid droplets has three different SNPs with a protective role in liver damage, rs72613567, rs143404524, and rs62305723. This role is associated with the reduction of transaminases and a reduction of 25 liver injury in the setting of a fatty liver but, however, further studies are required to confirm the protective role of this SNP. Gil-Gomez et al. found no evidence of a protective role in cirrhotic patients23. On the contrary, the variant rs72613567: TA showed an increase in the risk of decompensation and death in cirrhotic patients23. 1.2 MASLD: Diagnosis 1.2.1 Liver biopsy procedures Currently, the liver biopsy is considered the gold standard technique for the diagnosis of MASLD, being therefore considered the method of reference. However, this technique has some pitfalls, such as invasiveness24, cost25 and variability among both inter- and intraobservers26. Furthermore, liver biopsy can result in severe complications, including mortality, bleeding, or pain24. Over the years, several scores have been developed for the standardization of biopsy readouts. So far, the most validated has been the NAFLD Activity Score (NAS Score), in combination 32 and the alteration of microcirculation. At the same time, the liver releases inflammatory molecules that are considered proatherogenic and potentially playing a role in vascular endothelial dysfunction leading to the development of atherosclerosis and cardiovascular disease43. Furthermore, researchers have demonstrated that LSECs capillarization precedes MASLD and play a role in the advancement and perpetuation of chronic liver injury in NASH44. Characterizing the dysfunctional paracrine communication between liver sinusoidal endothelial and parenchymal cells is a crucial objective in the development of upcoming treatments for MASLD and MASH44. Several genes encoding adhesion molecules expressed in endothelial cells, such as Icam-1 and Vcam-1, could be then reliable biomarkers for endothelial dysfunction. During the inflammatory response, monocyte adhesion to LSECs is a crucial step in directing circulating monocytes to the liver. Vascular cell adhesion molecule-1 (VCAM-1) belongs to the immunoglobulin superfamily of cell adhesion molecules and is found on the surface of endothelial cells, where it regulates 33 leukocyte adhesion45. Murine and human studies have shown the relation between VCAM-1 and MASH 45. Furthermore, recent studies have linked the presence of VCAM-1 in serum of MASLD-fibrotic patients, positioning VCAM-1 as a potential contributor to MASH progression46. Duan et al. showed in a meta-analysis that intercellular adhesion molecule-1 (ICAM-1) levels were associated with increased risk of MASLD in serum samples47 and also MASH48. Studies have shown that in MASH, there is an elevation in the production of TNF-α and IL-1 from macrophages which induce the expression of ICAM-1 from the hepatocytes and/or LESCs48. Angiogenesis is a convoluted process that results in the formation of new blood vessels from pre-existing blood vessels. This process is mainly initiated by hypoxia (due to an increased resistance to blood flow and oxygen delivery from the sinusoids to the parenchyma) and inflammation, as well as cytokines such as vascular endothelial growth factor (VEGF)49. Although antiangiogenic treatments are not specific for MASLD, their antifibrotic effects are currently under investigation in liver 34 diseases49. Yang et al. demonstrated that VEGF promotes fibrogenesis in a mouse model of liver fibrosis. However, VEGF also played a role in hepatic tissue reparation and resolution of fibrosis regulating the permeability of blood vessels, the infiltration of monocytes and the scar formation of macrophages50. Previous investigations suggest that the renin-angiotensin system (RAS) plays a complex role in MASLD development. The reduction in angiotensinogen (AGT), the exclusive precursor of the RAS, significantly affects the activity of this system51. Serpin Family E Member 1/Plasminogen activator inhibitor-1 (Serpin-E1/PAI-1) has several functions, including serving as a marker and critical mediator of cellular senescence. The incidence of several liver diseases, specially MASLD, tends to increase with age and recent studies indicate that the accumulation of senescent cells may be a contributing factor of the disease52,53. Alsharoh et al. demonstrated this association 35 between elevated serum PAI-1 levels and MASLD patients in their study53. Finally, disintegrin and metalloproteinase 12 (ADAM12) is related with several tumors such as colorectal or breast cancer. Du et al. have shown the relation between ADAM 12 and liver cancer. Their results indicated a poor prognosis for patients with primary liver cancer and additionally, the presence of ADAM12 may serve as a potential diagnostic biomarker for liver cancer54. Further studies are necessary to establish the relationship between ADAM12 and MASLD. 1.5 MicroRNAs 1.5.1 MicroRNAS: definition, mechanism of action and link with MASLD. MiRNAs are small non-coding RNAs (sncRNAs) of 20-24 nucleotides in length, discovered in 1993, which regulate gene expression post-transcriptionally by targeting messenger RNAs, either preventing their translation or facilitating their degradation, affecting a wide range of physiological and pathological functions55,56. In addition, miRNAs are stable molecules and are 36 detectable in tissues and fluids, including serum, plasma and urine and even in paraffined tissue samples57. The formation of mature single-stranded miRNAs is characterized by a process where initially, primary miRNA transcripts, which include hairpin structures encoding mature miRNAs, are transcribed within the nucleus. Drosha, which belongs to the RNase III family, removes these hairpin structures, creating pre-miRNA intermediates that are approximately 70 nucleotides in length. Subsequently, these intermediates are transported out of the nucleus to the cytoplasm where premiRNAs are processed by the double-stranded endoribonuclease Dicer. After this process, the results are imperfect miRNA duplexes, consisting of mature miRNAs that are around 21–24 nucleotides in length. The mature miRNA becomes a part of the miRNA-induced silencing complex (RISC), guiding the miRNA to its mRNA target in a sequence-specific manner to induce either translational inhibition or mRNA degradation (Please refer to Figure 5). In animals, the miRNAs commonly forms an imperfectly complementary binding with the 3'UTR of the target gene. A 37 single mRNA can be influenced by multiple miRNAs, and each miRNA may regulate several genes58. Figure 5. This figure summarises the microRNA biogenesis, which involves the processing of miRNAs from transcription to mature miRNA RISCs. MiRNAs are accounted on several pathways such as cell differentiation, proliferation and apoptosis, and consequently, they have been linked to several diseases such as cancer, viral pathogenesis or metabolic diseases55,59. Over the last three decades, there has been extensive research into studies of the involvement of hepatic and circulating miRNAs in liver inflammation, fibrosis, and cirrhosis55,60. Recent studies have 38 identified several miRNAs associated with the liver, specifically, the most common MASLD-related miRNAs are miR-122, miR- 21, miR-34a, miR-192, miR-155 and miR-22360,61. The miR-122 is a highly abundant and liver- and hepatocyte-specific miRNA, comprising 70% of the hepatic miRNome in mice and 52% in humans with minimal manifestation in other cell types. It is involved in many aspects of liver physiology62. Studies have demonstrated that hepatic miR-122 is down-regulated in MASLD patients but surprisingly, circulating miR-122 is up-regulated in these patients compared to healthy controls55,60,63. Furthermore, miR-122 plays a crucial role in cancer as a tumor suppressor, particularly in liver cancer as it is involved in several pathways related to cancer such as apoptosis, cell metabolism, immune system response, migration, invasion and signaling pathways64. MiR-122 exhibits elevated expression levels in liver tissue but is down-regulated in hepatocellular carcinoma (HCC). This reduction is strongly associated with the initiation of hepatocarcinogenesis, unfavorable prognoses, and the occurrence of metastasis in HCC. Conversely, the restoration of miR-122 has been shown to inhibit the progression of the HCC and enhance 39 the sensitivity of HCC to chemotherapeutic agents61. Apart from acting as a potential biomarker exerted by miR-122, this microRNA has an important role as therapeutic target playing a crucial function in the cancer development. Bai et al. demonstrated the effect of miR-122 on HCC cells when treated with a chemotherapy agent. The use of sorafenib, which is clinically effective against HCC, significantly decreased the growth and survival of HCC cells expressing miR-12265. These findings suggest that miR-122 deficiency contributes to HCC cell malignancy. Hsu et al. created a new type of lipid-based nanoparticles for the treatment of the liver tumor cells with miR- 12266." The injection of nanoparticle-conjugated miR-122 into HCC xenografts established in nude mice resulted in tumor inhibition within 30 days of miR-122 delivery 66. Nowadays, miR-122 is being investigated for its potential in treating HCV infection and high cholesterol64. 40 1.5.2 MicroRNAs: epigenetic therapies MirRNAs expression levels can be artificially increased (mainly through the use of miRNA mimics) or decreased (by using antimiR techniques)58. Designed synthetic oligonucleotides have the capability to disrupt miRNA function at various points of their life cycle67. Morpholino is a form of antisense RNA inhibition. The interference of miRNA function through morpholinos can be designed to influence various stages in the life cycle of miRNAs. Morpholinos oligonucleotides are uncharged, exhibit stability and possess water solubility68. They are used to block complementary RNA sequences, which prevents processing, readthrough, or protein binding at targeted sites68. Morpholinos constitute stable base pairings with complementary nucleic acid sequences, however, their binding to proteins is not significant. Proteins do not recognize them, and they do not undergo catalysis mediated by proteins. Furthermore, they do not facilitate RNA cleavage through either RNase UP-REGULATED or the RISC complex68. 41 Thanks to the information in the databases, such as miRbase or miRDB, it is possible to create morpholinos that inhibit the small RNA itself by targeting the mature strand, its processing, or the recognition of mRNA targets67. While morpholinos have several advantages, it is crucial to be careful regarding their potential toxicity. Ferguson et al. hypothesized that the deaths of the preclinical animal models were a result of oligonucleotide hybridization, which led to an increased cationic charge associated with the dendrimer delivery component of the vivo-morpholino. The increase in cationic charge resulted in the formation of blood clots in whole blood exposed to vivo-morpholinos. This suggests that clotting may have contributed to the cardiac arrest in the dead mice69. 1.6 Preclinical models in MASLD Selecting preclinical models is a difficult task in which the ability of models to recapitulate clinically relevant endpoints needs to be appropriately balanced to other issues like ethics (e.g. number of experimental animals required to achieve statistical significance) or budget (diet duration, mouse colony maintenance). In fact, the 48 49 2. HYPOTHESIS Considering that the incidence of MASLD has been dramatically increasing over the last years worldwide, it is likely that this pathology will become the leading cause of liver disease in the near future. There is growing body of evidence pointing out that microRNAs have a direct impact on disease pathogenesis suggesting their potential as non-invasive biomarkers and therapeutic targets. This would facilitate the early diagnosis and would optimize the available resources in the current healthcare system. Since microRNAs are considered genetic markers of the affected tissue that release them, they would be useful tools to understand the biology of the disease. Furthermore, their involvement in cellular communication processes suggests that they can function similarly to hormones by modulating cellular phenotypes at a distance from the emitting cells. Finally, considering cardiovascular diseases as the first comorbidity in MASLD, opening new insights in endothelial dysfunction and searching for novel biomarkers in MASLD patients might be of interest for the scientific community. 50 51 3. OBJECTIVES Overall aim: The main aim was to discover and validate a differential profile of microRNAs in biopsy-proven MASLD patients, and to evaluate the potential therapeutic application of antisense probes for microRNAs in MASLD, MASH and fibrosis in a preclinical model. Specific aims: Ø To evaluate the differential profile of liver and circulating microRNAs as non-invasive disease biomarkers in preclinical models and clinical settings. Ø To define the therapeutic application of antisense probes for specific microRNAs for MASH and liver fibrosis treatment using an animal model mirroring this disease. Ø To evaluate both the metabolic and hepatic effects of this proposed therapeutic intervention in the preclinical model. Ø To explore and validate these findings in a MASLD- biopsy proven cohort of patients. 52 Ø To asses a panel of endothelial dysfunction biomarkers in serum and liver tissue in mice and in biopsy-proven MASLD patients. 53 4. MATERIALS AND METHODS 4.1 Animals: Overall design Male C57BL6J mice, aged five weeks, were acquired from Charles River (Cedex, France) and used for the entire duration of the research project. Mice were bred and maintained in the specific pathogen-free facilities of the Central Animal House at the Andalusian Center of Molecular Biology and Regenerative Medicine (CABIMER). All animal care and experimental protocols were executed following the guidelines for the care and use of laboratory animals of the Institution, together with the European Community Policy for Experimental Animal Studies. The study was approved by the Institutional Animal Care Committee of CABIMER (permission numbers 06-10-14-138 and 19/02/2016/023). Mice were maintained in a temperaturecontrolled environment at 23 °C with a constant humidity and a light-dark cycle of 12 to 12 hours per day. After one week of adaptation, mice were fed with control diet or CDA-HFD. Mice were provided with free ad libitum access to fresh water and a 54 fixed amount of food three times per week throughout the entire study. 4.1.1 Ethics The study was approved by the Institutional Animal Care Committee of CABIMER (permission numbers 06-10-14-138 and 19/02/2016/023), and performed according to the Spanish law on animal use RD 53/2013. 4.1.2 Experimental design: groups of study Three different conditions were analysed throughout the study. The first pilot study, named model establishment, was intended to fulfil a dual purpose: (i) to confirm the stability and suitability of our model for NAFLD, and (ii) to select the microRNAs with potential impact in the development and progression of the disease (Please refer to Figure 7). Figure 7. Experimental design: Control and CDA-HFD mice of the pilot study. 55 Further, the second set of animals, the so-called exploratory study (or 100-experiment), aimed to (i) validate the overexpression of the selected microRNAs and (ii) test the efficacy and safety of the treatment with morpholinos aimed to inhibit the expression of the three miRNAs of interest selected in the pilot study under different dietary regimes (control and CDAHFD) (Please refer to Figure 8 and Figure 9). Figure 8. Experimental design: Control mice of the exploratory study (or 100- experiment). 56 Figure 9. Experimental design: CDA-HFD and Reversion mice of the exploratory study (or 100-experiment). Finally, the last batch of animals, the so-called validation study (or the 45-experiment), was aimed to (i) select and finally validate the best miRNA chosen from the exploratory experiment under the same dietary regimen but with a shorter period of time (Please refer to Figure 10). 57 Figure 10. Experimental design: Control and CDA-HFD mice of the validation study (or 45-experiment). 4.1.3 Dietary regimen Mice were fed with two different diets during 16 weeks in both the pilot and the validation studies. Further, in the exploratory study, the diet was maintained for a total of 28 weeks. The control group was provided with chow diet constituted by a low-fat diet composed by 13% kcal from fat, 20% kcal from protein and 67% kcal from carbohydrate (2014; Harlan-Teklad Custom 64 oligonucleotides that were the mirror images of the target miRNA (Please refer to Table 3)67,68,85. A standard control was used in this study as a negative control called “Vivo standard control”. This molecule has no completely matching target in the mouse transcriptome. Thus, this control specifically assesses the impact of the delivery moiety without considering any antisense-related effects. The groups with our microRNA of interest “in vivo morpholino miR-200b-3p”, “in vivo morpholino miR-409-3p” and “in vivo morpholino miR-146- 5p” were studied compared to “Vivo standard control” and also the specific controls of the study called “in vivo morpholino miR- 200b-5p”, “in vivo morpholino miR-409-5p” and “in vivo morpholino 146b-3p” were included in the study. Table 3. Morpholinos sequences used to inhibit the miRNAs of the study (mmu-miR- 200b-3p, mmu-miR-146b-5p and mmu-miR-409-3p) and the specific controls (mmumiR-200b-5p, mmu-miR-146b-3p and mmu-miR-409-5p). Oligo name Sequence Concentration mmu-miR-200b- 3p-Drosha GCCGTCATCATTACCAGGCAGTATT (25) 0.5 mM mmu-miR-200b- 5p-Dicer CTATCCAATGCTGCCCAGTAAGATG (25) 0.5 mM mmu-mir-146b- 5p-Drosha ATTCAGTTCTCAGAGCCAGGTGGCC (25) 0.5 mM mmu-mir-146b- 3p-Drosha CCGGGCACCAGAACTGAGTCCCTAG (25) 0.5 mM 65 4.2.1.2 Treatment administration Prior to the administration to the groups described above, a confirmation process was study to make sure our treatment effectively reached the liver. To this aim, we prepared a 0.5% sterile solution of Evans Blue (E2129, Sigma-Aldrich®) diluted in PBS. Thus, 200 µl was slowly injected (needle with small gauge, 27-30) into the mice with the needle within a 10–15 degree angle through the tail vein. After 30 minutes of observation, the mice were sacrificed through cervical dislocation and several pictures from both the abdominal and thoracic cavity were taken to observe the differences in Evans Blue extravasation. Once this was confirmed, in order to facilitate the injection of morpholino through the tail vein, the tail of the mouse was heated using an infrared lamp to dilate the vein. After this, the tail was wiped with alcohol and the needle was inserted and 0.25 mg of morpholino was delivered. mmu-mir-409-3p- Dicer GGGTTCACCGAGCAACATTCGTCC (24) 0.5 mM mmu-mir-409-5p- Dicer TCCTCCAGATGCAAAGTTGCTCGGG (25) 0.5 mM Vivo standard control CCTCTTACCTCAGTTACAATTTATA(25) 0.5 mM 66 4.2.2 MiRNAs isolation 4.2.2.1 Hepatic miRNAs Total RNA was isolated using a mirVana miRNA isolation kit (Life Technologies, Grand Island, NY). The tissue was disintegrated using an autoclaved pistil and lysis/binding buffer (100 µl/sample) and it was homogenized adding 1/10 of lysis/binding volume (10 µl/sample) of miRNA Homogenate Additive. With the purpose of isolate the RNA, phenolchloroform was used (100 µl/sample) and after the centrifugation, we maintain carefully the aqueous phase and 1.25 parts of 100% ethanol was added. Afterwards, a series of centrifugations were carried out onto the collection tubes with filter cartridges along with some solutions from the kit (miRNA Wash Solution 1 upregulated miRNA Wash Solution 2/3). Finally, we transferred the filter cartridge to a fresh collection tube and we added the elution solution (100 µl/sample) which facilitated the extraction. 4.2.2.2 Circulating miRNAs The miRNEasy serum/plasma kit (Qiagen, Hilden, Germany) was used to obtain circulating miRNAs, following the 67 manufacturer’s instructions. QIAzol lysis Reagent (5 volumes) was added to 200 μL of plasma-EDTA samples. After incubation, an equal volume of chloroform was added. The upper aqueous layer was transferred to a new microtube and mixed with 1.5 volumes of 100% ethanol. The liquid was transferred to a RNeasy MinElute spin column (Qiagen, Hilden, Germany) and underwent centrifugation and washing. Finally, 80% ethanol was added, and miRNAs were eluted in 14 μL of RNase-free water. 4.2.3 RT-qPCR In the exploratory analysis in the preclinical and clnical studies, RNA quantification was performed with Qubitâ. cDNA synthesis was performed using miScript II RT Kit (Qiagen) to convert mature miRNAs into cDNA. The master mix was prepared with all necessary components except RNA: for a final volume of 20 μl, master mix was composed of 4 μl of miScript Buffer HiSpec 5x, 2 μl of Nucleics Mix 10x, 2 μl of miScript Reverse Transcriptase Mix and RNAse-free water. After this process, 1 µg of RNA was added per reaction. 68 The protocol for detecting mature miRNA using Real-Time PCR was performed with the LightCycler® 96. The master mix for a volume of 20 µL was prepared with 10 µL of QuantiTect SYBR® Green, 2 µL of miScript Universal Primer (reverse primer), 2 µL of miScript Primer Assay and RNAse-free water. Afterwards, 0.5 ng/µL of cDNA was added. The primer sequences (Qiagen (Valencia, CA)) were as follows: - mmu-miR-122-5p: UGGAGUGUGACAAUGGUGUUUG - mmu-miR-146b-5p: UGAGAACUGAAUUCCAUAGGCU - mmu-miR-200b-3p: UAAUACUGCCUGGUAAUGAUGA - mmu-miR-409-3p: GAAUGUUGCUCGGUGAACCCCU - hsa-miR-200b-3p: UAAUACUGCCUGGUAAUGAUGA - hsa-miR-224-5p: UCAAGUCACUAGUGGUUCCGUUUAG The reactions were incubated in a 96-well plate at 95ºC for 15 min, followed by 40 cycles of a denaturation at 94ºC for 15s, 69 annealing at 55ºC for 30 and an extension at 70ºC for 30s. All reactions were run in duplicate. The changes of differentially expressed miRNAs level were calculated with 2 delta-delta Ct. Finally, in the validation study, RNA quantification was performed with Qubitâ and the cDNA synthesis was performed using the TaqMan™ Advanced miRNA cDNA Synthesis Kit. In order to obtain the cDNA, the following reactions were carried out in the same reaction plate: poly(A) tailing reaction (0.5 μl of 10X Poly(A) Buffer, 0.5 μl of ATP, 0.3 μl of Poly(A) Enzyme and 1.7 μl of RNAse-free water. After this process, 10 ng of RNA was added per well. Once the incubation was completed, the second step was the adaptor ligation reaction (3μl of 5X DNA Ligase Buffer, 4.5 μl of 50% PEG 8000, 0.6 μl of 25X Ligation Adaptor, 1.5 μl of RNA Ligase and 0.4 μl of RNAse-free water) and a new incubation. The third step was the reverse transcription (RT) reaction (6 μl of 5X RT Buffer, 1.2 μl of dNTP Mix (25 mM each), 1.5 μl of 20X Universal RT Primer, 3 μl of 10X RT Enzyme Mix and 3.3 μl of RNase-free water. Once the incubation period ended, a preamplification of miRNAs with the 70 miR-Amp reaction (25 μl of 2X miR-Amp Master Mix, 2.5 μl of 20X miR-Amp Primer Mix and 17.5 μl of RNAse-free water) was performed. The incubation period of this reaction was performed under the following conditions: at 95ºC for 5 min, followed by 14 cycles of a denaturation at 95ºC for 3 s, annealing at 60ºC for 30 and stop reaction at 99ºC for 10min. All reactions were run in duplicate. The protocol for detecting mature miRNA using Real-Time PCR was performed with the QuantStudio 6 Flex™ and a dilution 1/10 of the cDNA. The master mix for a volume of 20 µL was prepared with 10 µL of TaqManTM Fast Advanced Master Mix (2X), 1 µL of TaqManTM Advanced miRNA Assay (20X) and 4 µL of RNAse-free water. Afterwards, 5 µL of the diluted cDNA was added. The primer sequences (TaqMan™ Advanced Assay/ TaqMan™ miRNA Assay) were as follows: - mmu-miR-200b-3p: UAAUACUGCCUGGUAAUGAUGA - mmu-miR-200b-5p: CAUCUUACUGGGCAGCAUUGGA - mmu-miR-224-5p: UAAGUCACUAGUGGUUCCGUU 71 - mmu-miR-423-5p: UGAGGGGCAGAGAGCGAGACUUU - mmu-miR-146b-5p: UGAGAACUGAAUUCCAUAGGCU - RNU6: GTGCTCGCTTCGGCAGCACATATACTAAAATTG GAACGATACAGAGAAGATTAGCATGGCCCCTGC GCAAGGATGACACGCAAATTCGTGAAGCGTTCC ATATTTT The reactions were incubated in a 384-well plate at 95ºC for 20s, followed by 40 cycles of a denaturation at 95ºC for 1s and an annealing/extension at 60ºC for 20s. The changes of differentially expressed miRNAs level were calculated with 2 delta-delta Ct. Finally, in the endothelial dysfunction study, RNA quantification was performed with Qubitâ and the reverse transcription was performed using the PrimeScriptTM RT Master Mix (Perfect Real Time) (v202008Da, Takara). The protocol for detecting the expression of our genes of interest using Real-Time PCR was performed with the QuantStudio 1™ and 200 ng/µl of cDNA. 72 The master mix for a volume of 10 µL was prepared with 5 µL of TaqManTM Fast Advanced Master Mix (2X), 0.5 µL of TaqManTM Assay (20X) and 3.5 µL of RNAse-free water. Afterwards, 1 µL of th cDNA was added. The primer TaqMan™ Advanced Assay were as follows: ICAM-1 (Mm00516023_m1, ThermoFisher), VCAM-1 (Mm01320970_m1, ThermoFisher), VEGF (Mm00437306_m1, ThermoFisher), SERPIN-E1 (Mm00435858_m, ThermoFisher) and GAPDH (Mm99999915_g1, ThermoFisher). The reactions were incubated in a 96-well plate at 95ºC for 20s, followed by 40 cycles of a denaturation at 95ºC for 1s and an annealing/extension at 60ºC for 20s. The changes of differentially expressed genes level were calculated with 2 delta-delta Ct and GAPDH was used as a control. 4.2.4 Histological evaluation Liver tissue samples were obtained, immediately fixed with 4% paraformaldehyde for six hours at 4 °C and then processed for paraffin embedding with a Leica ASP200S tissue processor (Leica, Wetzlar, Germany). Sections of 5 µm were cut on a Leica 73 DM6000B microtome and different staining techniques were used on the liver sections following the standard protocols: a) Sirius Red: to estimate the percentage of the surface area that stained positive for collagen; b) Oil Red O: to evaluate the tissue stained for the presence of lipids and triglycerides; c) Hematoxylin and eosin (UP-REGULATED&E) to calculate the percentage of the surface area that stained positive for steatosis. An experienced pathologist, unaware of the provenance of the samples and data, evaluated the samples to determine the NAFLD Activity Score (NAS Score), fibrosis stage according to the Kleiner Score together with the SAF Score27,31. In addition, fibrotic and steatotic areas were evaluated by direct pixel counting on binary images captured under 10x magnification using the ImageJ program, the IHC Profiler Plugin Average was calculated using 10 images per liver at 40x magnification. Quantification was performed using the IHC Profiler protocol, which was previously optimized as an Open Source Plugin. This plugin facilitates pixel-by-pixel quantification, and the results are subsequently presented as the 80 When comparing two groups with a normal distribution, the analysis was performed using the unpaired t-test and the Mann- Whitney test when the data were non-parametric. All values are expressed as the mean ± SEM and a p value of ≤ 0.05 was considered statistically significant. 81 5 RESULTS 5.1 MicroRNAs in MASLD 5.1.1 Pilot study 5.1.1.1 To assess the suitability of the selected animal model prior to the morpholino intervention at the hepatic and metabolic levels. After 16 weeks of diet, mice fed CDA-HFD diet showed a significant reduction in body weight when compared to control group (20.6±1.5 vs. 26.6±3 grams; p=0.0008) (Please refer to Figure 11). At the end of the study, biochemical parameters were analysed and we observed a reduction in glucose in CDA-HFD vs. control (73.67±8.56 vs. 105.7±3.34 mg/dL; p=0.0059) and total cholesterol (49.75±3.8 vs. 120.2±5.15 mg/dL; p<0.0001). Further, both ALT (182.5±26.4 vs. 27.6±2.9 IU/L; p=0.002) and AST levels (479.5±48.49 vs. 90.15±7.06 IU/L; p<0.0001) were found to be increased in CDA-HFD animals. Finally, no significant differences were found regarding triglycerides levels between both groups (67.01±3.6 vs. 64,73±4.01 mg/dL; p=ns) (Please refer to Figure 12). 82 Figure 11. Weight comparison in Control mice and CDA-HFD mice. Data shown as mean ± SEM (**p<0.01 and ***p<0.001). 83 Figure 12. a) ALT (U/L), b) AST (U/L), c) total cholesterol (mg/dL), d) triglycerides (mg/dL) and e) glucose (mg/dL) levels at final of the model establishment study (16 weeks) in CDA-HFD mice vs. control mice. Data shown as mean ± SEM (**p<0.01 and ****p<0.0001). The pathologist concluded that CDA-HFD mice showed features of both NASH and liver fibrosis at 16 weeks (Please, refer to Table 4). 84 Table 4. Histological findings regarding CDA-HFD animals versus controls. Control diet CDA-HFD p-value Steatosis, grade 0±0 2.8±0.4 p<0.0001 Ballooning, grade 0±0 0.2±0.4 p=ns Lobular inflammation, grade 1.6±0.5 3.0±0 p=0.005 NAS Score 1.6±0.5 6±0.7 p<0.0001 Liver fibrosis, stage 0±0 1.5±0.6 p=0.014 Transcriptomic analysis was performed in liver and plasma samples from mice. In this analysis, we found 21 microRNAs that were up- and down-regulated in both liver and plasma samples (Please refer to Figure 13). In addition, GO analysis gave us information about the processes in which the most related miRNAs were involved. Up-regulated liver miRNAs are involved in processes such as extrinsic apoptotic signaling, cell cycle, cell differentiation, transport, multicellular organism development, protein phosphorylation or transcriptional pathways. In plasma samples, up-regulated miRNAs are involved in processes such as cell migration, transcription, cell cycle regulation, negative regulation of focal adhesion assembly or transforming growth factor-beta receptor signaling pathways (Please refer to Figure 14). 85 Figure 13.Heatmaps processed by the analysis of circulating and hepatic microRNAs expression. 86 Figure 14. Biological processes obtained through GO analysis of the microRNAs involved in liver and plasma samples (up- and down-regulated). In the case of the liver, down-regulated liver-miRNAs are involved in pathways such as activin receptor signaling pathway, intracellular signal transduction, transcription, cell differentiation or multicellular organism development. Down-regulated plasmamiRNAs are involved in regulation of transcription, cell differentiation, extrinsic apoptotic signaling pathway in absence of ligand or protein transport pathways (Please refer to Figure 14). 87 KEGG analysis revealed the pathways involved in the microRNAs of the study. In the case of the liver, up-regulated miRNAs are involved in pathways such as T cell receptor signaling, cAMP signaling, estrogen signaling or VEGF signaling pathways. On the other hand, down-regulated livermiRNAs are involved in pathways such as mTOR signaling, Hepatitis B, insulin resistance or gastric acid secretion pathways. In plasma samples, up-regulated plasma-miRNAs are involved in Hepatitis B, estrogen signaling or focal adhesion pathways. Furthermore, down-regulated plasma-miRNAs are involved in colorectal cancer, hepatitis B or T cell receptor signaling pathways (Please refer to Figure 15). 88 Figure 15. Pathways affected by overexpressed and down-regulated microRNAs in liver and plasma samples.. Moreover, we also analysed the 21 miRNAs that are common to liver and plasma and we can observe the following microRNAs up-regulated in liver: miR-23a-5p, miR-6241, miR-200b-5p and down-regulated: miR-7093 and miR-16-2-3p. On the other hand, in plasma, we found the following miRNAs up-regulated: miR- 143-3p, miR-199a-3p, miR-214-3p, miR-200b-5p, miR-23a-5p and miR-6241. MiR-7093 and miR-16-2-3p were found downregulated in plasma samples (Please refer to Figure 16). 89 Figure 16. Heatmaps processed by Go Analysis of circulating and hepatic microRNAs expression. The GO analysis showed the relationship between the microRNAs and the biological processes of the 21 microRNAs in the study. The up-regulated microRNAs are involved in vesicle transport along microtubules, embryo development, biological process, cell differentiation, cellular protein modification process and cell death processes. The down-regulated microRNAs showed a relation with regulation of insulin secretion involved in cellular response to glucose stimulus, cell differentiation, biosynthesis, embryo development or regulation of cholesterol biosynthesis processes among others (Please refer to Figure 17). 96 Figure 23. ALT levels (UI/L) in Reversion and CDA-HFD groups before and after morpholinos intervention. The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM (**p<0.01). Further, similar AST levels were found in mice fed control diet throughout the whole study, without showing significant differences (Please refer to Figure 24). 97 Figure 24. AST levels (UI/L) in mice fed control diet before and after morpholinos intervention. The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM. In CDA-HFD mice at baseline and at the end of the study, AST levels were found decreased but did not reach statistical significance in any of the groups. However, at week 16 AST levels were increased in all groups (Please refer to Figure 25). 98 Figure 25. AST levels (UI/L) in CDA-HFD and Reversion mice groups before and after morpholino intervention. The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM. Regarding total cholesterol, no significant differences were observed between groups in mice fed with control diet at baseline and after 16 weeks. A slight decrease was found after morpholinos therapy, but it was not statistically significant (Please refer to Figure 26). 99 Figure 26. Total cholesterol (mg/dL) in control diet, mice groups before and after morpholino intervention. The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM. No differences were observed between groups in the CDA-HFD and Reversion. However, an increase in total cholesterol levels was observed after sacrifice, although there were no significant differences among them (Please refer to Figure 27). 100 Figure 27. Total cholesterol levels (mg/dL) in CDA-HFD and Reversion mice groups before and after morpholino intervention. The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM. A normal range of triglycerides was found in mice fed with control, CDA-HFD and in the Reversion group. Triglycerides were found to be increased after morpholino injection, but did not reach statistical significance (control diet 589.35 ± 283.4; p=ns; control Vivo standard control 726.2 ± 255.4; p=ns; control in vivo morph miR-200b-3p 635.5 ± 169.9; control in vivo morph miR-200b-5p 725.52 ± 253.67; p=ns; control in vivo morph miR- 146-3p 274.3 ± 80.4; p=ns; control in vivo morph miR-146-5p 160.64 ± 25.94; p=ns; control in vivo morph miR-409-3p 422.1 101 ± 168.68; p=ns; control in vivo morph miR-409-5p 505.68 ± 256.53; p=ns; CDA-HFD 499.48 ± 120.19; p=ns; Reversion 326.18 ± 81.8; p=ns; CDA-HFD Vivo standard control 618.56 ± 83.5; p=ns; CDA-HFD in vivo morph miR-200b-3p 731.57 ± 66.4; p=ns; CDA-HFD in vivo morph miR-200b-5p 1016 ± 196.55; p=ns; CDA-HFD in vivo morph miR-146-3p 649.75 ± 190.43; p=ns; CDA-HFD in vivo morph miR-146-5p 735.23 ± 131.79; p=ns; CDA-HFD in vivo morph miR-409-3p 922.33 ± 148.1; p=ns; CDA-HFD in vivo morph miR-409-5p 866.98 ± 323.04 mg/dL; p=ns) (Please refer to Figure 28 and Figure 29). Figure 28. Triglycerides (mg/dL) in control mice before and after morpholino intervention. The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM. 102 Figure 29. Triglycerides (mg/dL) in CDA-HFD and Reversion mice groups before and after morpholino intervention. The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM. Glucose levels were measured all over the study and remained stable in all mice groups on both diets. Similarly, no changes were detected in insulin levels in CDA-HFD mice or controls throughout the study. Finally, mice were subjected to the intraperitoneal glucose tolerance test (IPGTT) evaluation and as expected, normal values were obtained in all groups of animals. 103 5.1.2.2 To determine whether microRNAs are overexpressed in the context of a diet inducing NAFLD. As expected, selected miRNAs were up-regulated in CDA-HFD mice when compared to the control group (miRNA 200b-3p : Fold change (FC) 62.6 ± 20.75; vs. 1 ± 0.78; p=0.01 (Please refer to Figure 30); miRNA 409-3p: FC 18.7 ± 12.21 vs. 1 ± 0.95; p=0.008; ) (Please refer to Figure 31); miRNA 146b-5p: FC 271.3 ± 145.1 vs.1 ± 0.66; p=0.02) (Please refer to Figure 32) . After Reversion diet, microRNAs were found to be normalized almost reaching the level of expression in control mice but without significant differences (miRNA 200b-3p: FC 4.58 ± 3.7 vs. 1 ± 0.78; p=ns; (Please refer to Figure 30); miRNA 409-3p: FC 1.408 ± 0.8 vs. 1 ± 0.96 p=ns; (Please refer to Figure 31); miRNA 146b- 5p: FC 8.4 ± 5.4 vs. 1 ± 0.66; p=ns)(Please refer to Figure 32). MiRNA 146b-5p also showed significant differences between CDA-HFD and control vivo standard control (FC 271.3 ± 145.1 vs. 0.89 ± 0.06; p=0.04). 104 Figure 30. The relative expression levels (fold change) of miR-200b-3p in controls (CONTROL DIET), control mice treated with vehicle (CONTROL vivo standard control), CDA-HFD and Reversion groups. Data are shown as mean ± SEM. Figure 31. The relative expression levels (fold change) of miR-409-3p in control mice without morpholino therapy (CONTROL DIET), control mice with vehicle (CONTROL vivo standard control), CDA-HFD and Reversion mice. Data are shown as mean ± SEM (** p<0.01). 105 Figure 32. The relative expression levels (fold change) of miR-146b-5p in control mice without morpholino therapy (CONTROL DIET), control mice with vehicle (CONTROL vivo standard control), CDA-HFD and Reversion mice. Data are shown as mean ± SEM (* p<0.05). 5.1.2.3 To evaluate the impact of the use of vehicle in these animals To determine whether the vehicle has an impact on the animals, we compared the vivo standard control groups with the CDAHFD fed mice and controls. No differences were found between control vivo standard control mice and control diet in the panel of miRNAs under evaluation (miRNA 200b-3p: FC 1.63 ± 0.37 vs. 1 ± 0.78; p=ns; miRNA 409: FC 0.078 ± 0.04 vs. 1 ± 0.95; p=ns; miRNA 146 : FC 0.89 ± 0.06 vs.1 ± 0.66; p=ns) (Please refer to Figure 30, Figure 31 and Figure 32). 112 Figure 39. Quantification of lipid droplets in stained livers using Hematoxylin and Eosin with ImageJ. Dots represent individual values. Data are shown as mean ± SEM. Figure 40.Histology of all mice groups and their different diets. Hematoxylin and eosin staining of liver samples showed normal tissue (Control, Reversion, Controlvivo standard control, control in vivo morph miR-200b-3p and control in vivo morph miR- 200b-5p), steatosis (CDA-HFD, CDA-HFD vivo standard control, CDA-HFD in vivo 113 morph miR-200b-3p and CDA-HFD in vivo morph miR-200b-5p) and steatohepatitis (CDA-HFD in vivo morph miR-200b-5p and CDA-HFD). All specimens are presented in 40x. Oil Red O analysis revealed a triglycerides increase in CDA-HFD mice in relation to control diet (28.42 ± 2.29 vs. 3.47 ± 2.2; p<0.0001) and Reversion diet (28.42 ± 2.29 vs. 7.5 ±1.28; p<0.0001) and as expected, the same increase in the rest of CDAHFD mice was observed (CDA-HFD in vivo morph miR-200b- 5p vs. CDA-HFD vivo standard control (26.33 ± 2.03 vs. 26.17 ± 1.81; p=ns). However, a slight decrease was detected in CDAHFD in vivo morph miR-200b-3p in relation to CDA-HFD vivo standard control was detected (17.13 ± 0.89 vs. 26.17 ± 1.81; p=0.006) (Please refer to Figure 41 and Figure 42). 114 Figure 41. Quantification of triglycerides in stained livers using Oil red O with ImageJ. Dots represent individual values. Data are shown as mean ± SEM. Figure 42. Histology of all mice groups and their different diets. Oil red O staining of mice liver samples showing higher triglycerides expression in CDA-HFD, CDA-HFD vivo standard control, CDA-HFD in vivo morph miR-200b-3p and , CDA-HFD in vivo morph miR-200b-5p. All specimens are presented in 40x. 115 As expected, higher expression levels of Sirius red were detected in mice fed CDA-HFD. The % Area Threshold was increased in CDA-HFD compared to the Reversion group (14.7 ± 2.1 vs. 5.4 ± 0.52; p=0.04) and control diet (14.7 ± 2.1 vs. 0.66 ± 0.46; p=ns). Moreover, an increase of collagen was observed in the Reversion group in relation to control diet (5.4 ± 0.52 vs. 0.66 ± 0.46; p=0.03). CDA-HFD vivo standard control, CDA-HFD in vivo morph miR-200b-3p and CDA-HFD in vivo morph miR-200b-5p presented higher collagen levels without achieving statistical significance (Please refer to Figure 43 and Figure 44). Figure 43. Quantification of collagen in stained livers using Sirius red with ImageJ. Dots represent individual values. Data are shown as mean ± SEM. 116 Figure 44. Histology of all mice groups and their different diets. Sirius red staining of mice liver samples showing higher collagen expression in CDA-HFD, CDA-HFD vivo standard control, CDA-HFD in vivo morph miR-200b-3p, CDAHFD in vivo morph miR-200b-5p and Reversion groups. All specimens are presented in 40x. In the case of miR-409, same findings were observed in Control, CDA-HFD and Reversion diets. Higher expression of lipid droplets was found in CDA-HFD in vivo morph miR-409-3p and CDA-HFD in vivo morph miR-409-5p compared to their control CDA-HFD vivo standard control (13.22 ± 1.84; 20.62 ± 1.84 vs. 11.91 ± 2.89; p=ns), without statistical significance (Please refer to Figure 45 and Figure 46). 117 Figure 45. Quantification of lipid droplets in stained livers using Hematoxylin and Eosin with ImageJ. Dots represent individual values. Data are shown as mean ± SEM. Figure 46. Histology of all mice groups and their different diets. Hematoxylin and eosin staining of liver samples showed normal tissue (Control, Reversion, Controlvivo standard control, control in vivo morph miR-409-3p and control in vivo morph miR- 409-5p), steatosis (CDA-HFD, CDA-HFD vivo standard control, CDA-HFD in vivo morph miR-409-3p and CDA-HFD in vivo morph miR-409-5p) and steatohepatitis 118 (CDA-HFD in vivo morph miR-409-3p and CDA-HFD). All specimens are presented in 40x. All specimens are presented in 40x. Triglyceride levels were found to be high in the case of miR-409 mice groups without statistical significance. However, an increase in triglyceride expression was observed for CDA-HFD in vivo morph miR-409-5p compared to CDA-HFD vivo standard control (31.44 ± 1.71 vs. 26.17 ± 1.81; p=ns) while a slight decrease was found in CDA-HFD in vivo morph miR-409-3p in relation to the control group, CDA-HFD vivo standard control (19.93 ± 1.19 vs. 26.17± 1.81; p=ns) (Please refer to Figure 47 and Figure 48). Figure 47. Quantification of triglycerides in stained livers using Oil red O with ImageJ. Dots represent individual values. Data are shown as mean ± SEM. 119 Figure 48. Histology of all mice groups and their different diets. Oil red O staining of mice liver samples showing higher triglycerides expression in CDA-HFD, CDA-HFD vivo standard control, CDA-HFD in vivo morph miR-409-3p and CDA-HFD in vivo morph miR-409-5p. All specimens are presented in 40x. The miR-409 mice groups fed with CDA-HFD showed higher expression levels of collagen, similar to the trend observed in miR-200b compared to mice fed with control diet. CDA-HFD in vivo morph miR-409-3p shown a collagen increase in relation to the control group, CDA-HFD vivo standard control (13.99 ± 0.8 vs. 11.46 ± 0.91; p=ns). However, a slight decrease was observed in CDA-HFD in vivo morph miR-409-5p in relation to CDAHFD vivo standard control (9.24 ± 0.87 vs. 11.46 ± 0.91; p=ns) (Please refer to Figure 49 and Figure 50). 120 Figure 49. Quantification of collagen in stained livers using Sirius red with ImageJ. Dots represent individual values. Data are shown as mean ± SEM. Figure 50. Histology of all mice groups and their different diets. Sirius red staining of mice liver samples showing higher collagen expression in CDA-HFD, CDA-HFD vivo standard control, CDA-HFD in vivo morph miR-409-3p, CDA-HFD in vivo morph miR-409-5p and Reversion diet. All specimens are presented in 40x. 121 Finally, in the case of miR-146 mice groups, similar trend as described below was observed in the expression of lipid droplets. An increase of lipid droplets was found in mice fed with CDAHFD, for example in the case of CDA-HFD in vivo morph miR- 146-5p compared to CDA-HFD vivo standard control (12.56 ± 3.39 vs. 11.91 ± 2.89; p=ns) and a slight decrease was observed in CDA-HFD in vivo morph miR-146-3p in relation to CDAHFD vivo standard control (8.55 ± 1.90 vs. 11.91 ± 2.89; p=ns) (Please refer to Figure 51 and Figure 52). Figure 51. Quantification of lipid droplets in stained livers using Hematoxylin and Eosin with ImageJ. Dots represent individual values. Data are shown as mean ± SEM. 128 In addition, after morpholinos injection, AST levels were increased in all mice without any significant differences between the groups (Please refer to Figure 59). Figure 59. AST levels (UI/L) in control and CDA-HFD mice groups before and during morpholino intervention (from 16 week). The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM (**p<0.01). As in previous studies, mice fed with a CDA-HFD showed a reduction in total cholesterol compared to mice fed with control diet during the study. At the week 6, the CDA-HFD group had consistently lower HDL cholesterol levels than the controls and after morpholino injection 129 significant reduced levels were found in CDA-HFD mice following the same pattern. However, among them, HDL levels were increased in CDA-HFD vivo standard group compared with CDA-HFD (1.71 ± 0.05 vs. 0.76 ± 0.04 mmol/L; p<0.0001) and simultaneously a reduced HDL cholesterol in CDA-HFD when compared with control diet was observed (0.76 ± 0.04 vs. 2.066 ± 0.03 mmol/L; p<0.0001). Also, a decrease in HDL levels was found in CDA-HFD in vivo morph miR-200b-3p in relation to the control group, CDA-HFD vivo standard control (0.81 ± 0.03 vs. 1.706 ± 0.05 mmol/L; p<0.0001) (Please refer to Figure 60). Figure 60. HDL cholesterol (mmol/L) in control and CDA-HFD mice groups before and during morpholino intervention (from 16 week). The intervention with morpholinos 130 is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM (****p<0.0001). All groups exhibited low LDL values and there were no changes at baseline and week 6. After morpholinos intervention, reduced LDL levels were found in CDA-HFD and CDA-HFD in vivo morph miR-200b-3p when compared to CDA-HFD vivo standard control (0.14 ± 0.02;p=0.01 ; 0.110 ± 0.01 ;p=0.003 vs. 0.27 ± 0.02 mmol/L) (Please refer to Figure 61). Figure 61. LDL cholesterol (mg/dL) in control and CDA-HFD mice groups before and during morpholino intervention (from 16 week). The intervention with morpholinos is indicated with a dot line. Dots represent individual values. Data are shown as mean ± SEM (*p<0.05 and ** p<0-005). 131 As in our previous studies, triglyceride levels remained stable during this study in both groups of mice, regardless of whether they were fed the control diet or CDA-HFD. After sacrifice, the anatomopathological examination revealed significantly higher steatosis in CDA-HFD mice when compared to controls. No evidence of steatosis was found in mice fed with the control diet (Please refer to Table 5). On the other hand, steatosis was found to be decreased in mice fed with CDA-HFD and in vivo morpholino mmu-miR-200b-3p compared to their respective controls CDA-HFD vivo standard control and control diet respectively (Please refer to Table 6). 132 Table 5. Histological findings regarding control diet, control vivo standard control, control in vivo morpholino mmu-miR-200b-3p and control in vivo morpholino mmumiR-200b-5p mice versus their controls. Control diet Control Vivo Standard control Control in vivo morph mmu-miR- 200b-3p Control in vivo morph mmu-miR- 200b-5p Steatosis, grade 0 ± 0 0 ± 0; p=ns 0 ± 0; p=ns 0 ± 0; p=ns Ballooning, grade 0 ± 0 0 ± 0; p=ns 0 ± 0; p=ns 0 ± 0; p=ns Lobular inflammation, grade 1.2±0.2 2 ± 0; p=ns 1.8 ± 0.2; p=ns 1.8 ± 0.2; p=ns NAS Score 1.2±0.2 2 ± 0; p=ns 1.8 ± 0.2; p=ns 1.8 ± 0.2; p=ns Liver fibrosis 0 ± 0 0 ± 0; p=ns 0 ± 0; p=ns 0 ± 0; p=ns Table 6. Histological findings regarding CDA-HFD, CDA-HFD vivo standard control, CDA-HFD in vivo morpholino mmu-miR-200b-3p and CDA-HFD in vivo morpholino mmu-miR-200b-5p mice versus their controls. CDAHFD CDAHFD Vivo Standard control CDA-HFD in vivo morph mmu-miR- 200b-3p CDA-HFD in vivo morph mmu-miR- 200b-5p Steatosis, grade 3 ± 0; p=0.007 1 ± 0; p=ns 3 ± 0; p=ns 1 ± 0; p=ns Ballooning, grade 0 ± 0; p=ns 0 ± 0; p=ns 0 ± 0; p=ns 0 ± 0; p=ns Lobular inflammation, grade 3 ± 0; p=0.007 3 ± 0; p=ns 3 ± 0; p=ns 3 ± 0; p=ns NAS Score 6 ± 0; p=0.002 4 ± 0; p=ns 5.4 ± 0.2; p=ns 4 ± 0; p=ns Liver fibrosis 1 ± 0; p=0.05 1 ± 0; p=ns 1 ± 0; p=ns 1 ± 0; p=ns 133 On the other hand, all CDA-HFD mice showed 1 stage of fibrosis when compared to controls. Significant differences were found in CDA-HFD in relation to control diet mice on week 16. No differences were in both mice groups regarding ballooning. Finally, inflammation is an important feature of MASLD. After sacrifice, higher inflammation levels were observed in all CDAHFD groups compared to mice fed control diet. Significant findings were found in CDA-HFD in relation to control diet (Please refer to tables Table 5 and Table 6). Moreover, as expected, there was an increase in NAS score in CDA-HFD mice groups compared to controls (Please refer to Table 5 and Table 6) (Please refer to Figure 62). 134 Figure 62. NAS score calculated (0-8) in control and CDA-HFD mice groups after morpholino intervention. 5.1.3.2 To determine whether the miRNA-200b was overexpressed in MASLD mice MiRNA-200b-3p was significantly increased in CDA-HFD animals when compared to control diet (FC 19.84 ± 8.2 vs. 1 ± 0.28; p=0.04) and Control vivo standard control (FC 2.87 ± 0.5; p=ns) (Please refer to Figure 63). 135 Figure 63. The relative expression levels (fold change) of miR-200b-3p in control mice without morpholino therapy (Control diet), control mice with vehicle (Control vivo standard control) and CDA-HFD. Data are shown as mean ± SEM (*p<0.05). 5.1.3.3 To evaluate the impact of the vehicle in disease progression In CDA-HFD mice, a clear trend of higher miR-200b-3p expression was observed in CDA-HFD vivo standard control when compared to control group, CDA-HFD (FC 3.08 ± 0,7; vs. 1 ± 0.41; p=0.03) (Please refer to Figure 64). 136 Figure 64. The relative expression levels (fold change) of miR-200b-3p in CDA-HFD without morpholino therapy (CDA-HFD), and CDA-HFD with vehicle (CDA-HFD vivo standard control). Data are shown as mean ± SEM (*p<0.05). 5.1.3.4 To assess whether the morpholino injection has been successful No significant differences in miR-200b-3p expression were found in mice fed with control diet and mice fed with control diet with morpholino intervention (Please refer to Figure 65). 137 Figure 65. The relative expression levels (fold change) of miR-200b-3p in control mice with vehicle (Control vivo standard control), Control mice after morpholino injections (Control in vivo morph miR-200b-3p and Control in vivo morph miR-200b-5p). Data are shown as mean ± SEM. Our primary focus was the group of mice fed a CDA-HFD diet and injected with morpholinos. As expected, no significant differences between CDA-HFD in vivo morph miR-200b-3p and control group (CDA-HFD vivo standard control) (FC 0.94 ± 0.15 vs. 1 ± 0.41; p=ns) was observed after the administration of antimir 200b-3p. Further, miR-200b-3p expression was higher in CDA-HFD in vivo morph miR-200b-5p (FC 6.62 ± 0.91; p<0.0001) (Please refer to Figure 66). 144 Figure 73. Panel of miRNAs of interest, including miR-224-5p (a) and miR-423-5p (b). For miR-423-5p, the results revealed a similar expression in both groups of mice, mice fed with CDA-HFD and those on the control diet (FC 1.315 ± 0.12 vs. 1 ± 0.5; p=ns) (Please refer to Figure 73b). 5.1.4 The role of miRNAs in biopsy-proven MASLD patients 5.1.4.1 Baseline patients´ characteristics The study also investigated miRNAs of interest in sixty-six biopsy-proven patients with MASLD, 77.3% of them with steatosis and 53% of them with MASH. Half of the patients were men, with a mean age of 57 years and an average BMI of 42.26 kg/m2 (Please refer to Table 7). 145 Table 7. Main features of MASLD cohort employed for the study. Variable Patients (N=66) Age (years) 57.47 ± 12 Sex (males, %) 50 (33/66) BMI (kg/m2) 42.26 Diabetes mellitus (%, n/N) 48.5 (32/66) Arterial hypertension (%, n/N) 59.1 (39/66) ALT (IU/mL) 52.1 AST (IU/mL) 39.7 GGT (IU/mL) 58.4 MASH (%, n/N) 53, (35/66) Significant fibrosis (F2-F4, %) 28.8 (19/66) Steatosis, presence (%, n/N) 77.3 (51/66) Ballooning degeneration, presence (%, n/N) 74.2 (49/66) In this part of the study, all miRNAs expression analysis were performed comparing our study group with the steatosis group. 5.1.4.2 miRNAs expression miR-200b-3p expression was increased in MASH patients compared to steatosis, although results were not statistically significant (FC 1.48 ± 0.32 vs. 1 ± 0.2; p=0.2) (Please refer to Figure 74). 146 Figure 74.The relative expression levels (fold change) of miR-200b-3p in MASLD patients. Data are shown as mean ± SEM. The same trend was observed although the results were not statistically significant in miR-224-5p. The miRNA expression was higher in MASH patients compared to steatosis group (FC 2.5 ± 0.59 vs. 1 ± 0.2; p=0.3) (Please refer to Figure 75). Figure 75. The relative expression levels (fold change) of miR-224-5p in MASLD patients. Data are shown as mean ± SEM. 147 5.2 Endothelial dysfunction MASLD related 5.2.1 Mice study 5.2.1.1 Genes associated with endothelial dysfunction. In order to evaluate the same circulating biomarkers of endothelial dysfunction as in MASLD patients in five CDA-HFD mice, RT-qPCR was performed. ICAM-1, VCAM-1 and SERPIN-E1 gene expression was increased in CDA-HFD mice compared to mice fed a control diet (2.99 ± 1.4; p=0.3; 2.23 ± 0.37; p=0.1; 10.27 ± 2.73; p= 0.08; vs. 1 ± 0.4, respectively). Results were not statistically significant (Please refer to Figure 76). No significant results were found in VEGF. Figure 76. Relative expression levels (fold change) of (a) ICAM-1 (b) VCAM-1 and (c) SERPIN-E1 genes. Data are shown as mean ± SEM. 148 5.2.1.2 Hepatic protein levels of ED markers As a proof of concept, for the histological analysis of mice, we measured ICAM-1 in two mice fed with chow diet and CDAHFD for 16 weeks, and results shown that ICAM-1 immunoreactivity was detected in CDA-HFD group but not in the control group (Please refer to Figure 77). Figure 77. Immunohistochemistry of endothelial dysfunction marker ICAM-1 in CDAHFD mice. All specimens are presented in 40x. Hepatic sections for ICAM-1 were analysed with Image J (IHC Profiler) (control 1533 ± 16.95 vs. CDA-HFD 3120 ± 523.2). 5.2.2 Clinical study 5.2.2.1 ELISAS For the analysis of non-invasive endothelial dysfunction biomarkers in MASLD patients, a total of one hundred and fortyeight biopsy-proven MASLD patients at different stages of the disease were included in this study. 149 The baseline characteristics of patients of this study included 93.2% of patients with steatosis and 34.4% of them with MASH. Almost half of the patients were male, with a median age of 55 years and an average BMI of 34.04 kg/m2 (Please refer to Table 8). Table 8. Main features of MASLD cohort employed for the study. Variable Patients (N=148) Age (years) 55 ± 11.7 Sex (males, %) 46.6 (69/148) BMI (kg/m2) 34.04 Diabetes mellitus (%, n/N) 51.4 (75/146) Arterial hypertension (%, n/N) 63 (92/146) ALT (IU/mL) 57.2 AST (IU/mL) 46.4 GGT (IU/mL) 101.97 MASH (%, n/N) 34.4 (51/148) Significant fibrosis (F2-F4, %) 44.6 (66/148) Steatosis, presence (%, n/N) 93.2 (138/148) Ballooning degeneration, presence (%, n/N) 83.1 (123/148) First, circulating levels of ICAM-1 were associated with liver fibrosis (Steatosis 18.1 ± 1.01 ng/mL; p=ns; MASH F0 20.6 ± 1.48 ng/mL; p=ns; MASH F1-F2 22.5 ± 1.7 ng/mL; p=ns; and MASH F3-F4 25.5 ± 2.4 ng /mL; p=0.04; vs. control 17.8 ± 3.4 ng/mL) (Please refer to Figure 78). ICAM-1 levels were 150 associated with ballooning (mild 20.6 ± 0.77 ng/mL; p=0.02; and severe 28.6 ± 2.9 ng/mL; p= p<0.001 vs. none 16.9 ± 1.2 ng/mL;) (Please refer to Figure 79). Besides, ICAM-1 was positively correlated with NAS Score (r: 0.420; p< 0.001; n=138) and stiffness evaluated by transient elastography (TE) in kPa (r=0.437; p<0.001; n=120). Figure 78. Circulating levels of ICAM-1/CD54 in biopsy-proven MASLD patients (control (C), steatosis simple (steatosis), MASH without fibrosis (MASH F0), MASH with fibrosis F1/F2 (MASH F1-F2) and MASH with moderate fibrosis (F3-F4). Data are shown as mean ± SEM (*p<0.05). 151 Figure 79. Circulating levels of ICAM-1/CD54 associated with ballooning in biopsyproven MASLD patients. Data are shown as mean ± SEM. Further, circulating levels of VCAM-1 were associated with the different stages of liver fibrosis (F1 47.6 ± 4.1 ng/mL; p=ns; F2 60.4±5.6 ng/mL; p=0.18; F3 65.4 ± 5.9 ng/mL; p=0.005; and F4 63.7 ± 8.87 ng/mL; p=0.2; vs. F0: 44.5 ± 4 ng /mL) (Please refer to Figure 80). 152 Figure 80. VCAM-1/CD106 serum levels related to fibrosis in the disease. Data are shown as mean ± SEM. Serpin E1 was found to be increased together with the increase of hepatic steatosis (S1 7.3 ± 0.7 ng/mL; p=ns; vs. S2 9.9 ± 0.9 ng/mL; p= 0.04; and S3 10.5 ± 1.9 ng/mL; p=0.1; vs. S0 4.7 ± 1.3 ng/mL) (Please refer to Figure 81). Serpin E1 was also correlated with the Castelli index (r:0.248; p=0.007; n=117) and with the NAS Score (r:0.247; p=0.003; n=140). 153 Figure 81. Circulating levels of Serpin E1/PAI-1 in MASLD patients correlated to steatosis grades: S0 (<5%), S1 (5-33%), S2 (34-66%) and S3 (>66%). Data are shown as mean ± SEM. Finally, VEGF circulating levels were increased in patients with NAS Score > 5: 265.3 ± 14.7 pg/mL vs. NAS score ≤ 5 235.6 ± 17.2 pg/mL; p = 0.018) (Please refer to Figure 82). Liver protein levels (% of control) PGC1a SIRT1 0.0 0.5 1.0 1.5 2.0 *** * p = 0.06 0.0 0.5 1.0 1.5 2.0 2.5 ** * Oil Red O area % of control (A. U.) 3.0 3.5 4.0 4.5 5.0 5.5 Liver mass (g/100 g of BW) 0246810121416 0 50 100 150 200 250 300 shATG3 Weeks Cumulative food intake (g) 6 p AT-8V AA3α 0246810121416 0 2 4 6 8 10 SD diet AAV8-GFP + sh scrambled SD diet AAV8-TAp63α + sh scrambled SD diet AAV8-TAp63α+shATG3 shATG3 Weeks Body weight change (g) AAV8-TAp63α 0 1 2 3 4 TAp63α mRNA levels (fold change) ** 0 50 100 150 200 250 AST (U/L) ** Lentiv-shATG3 WT AAV8-TAp63α SDdiet FAS ACC p-ACC CPT1a LPL GAPDH 0 20 40 60 ALT (U/L) * p = 0.0513 Liver protein levels (% of control) FAS ACC p-ACC p-ACC/ACC CPT1a LPL 0 1 2 3 4 ** * ** STD diet AAV8-GFP sh scrambled STD diet AAV8-TAp63α sh scrambled STD diet AAV8-TAp63α shATG3 50μm50μm50μm 50μm50μm50μm HEOil Red O staining ATG3 GAPDH 0 2 4 6 ATG3 protein levels (% of control) *** *** GAPDH SIRT1 PGC1α 0.0 0.5 1.0 1.5 2.0 2.5 Triglycerides (mg TG/g of liver) ** GFP TAp63α GFP TAp63α TAp63α + shATG3 GFP TAp63α TAp63α + shATG3 GFP TAp63α TAp63α + shATG3 GFP TAp63α TAp63α + shATG3 GFP TAp63α TAp63α + shATG3 GFP TAp63α TAp63α + shATG3 AAV8-GFP sh scrambled AAV8-TAp63α sh scrambled AAV8-TAp63α shATG3 AAV8-GFP sh scrambled AAV8-TAp63α sh scrambled AAV8-TAp63α shATG3 AAV8-GFP sh scrambled AAV8-TAp63α sh scrambled AAV8-TAp63α shATG3 III-UQCRC2 I-NDUFB8 V-ATP5A IV-COX II II-SDH5 HSP70 Liver protein levels (% of control) 0.0 0.5 1.0 1.5 C-I C-II C-III C-IV C-V ** * ** ** ** ** ** ** AAV8-GFP sh scrambled AAV8-TAp63α sh scrambled AAV8-TAp63α shATG3 DCE G F H I A B Fig. 6. Liver-specific downregulation of ATG3 ameliorated NAFLD induced by TAp63 a .(A) Body weight change, (B) food intake, (C) hepatic levels of TAp63 a mRNA and ATG3 protein, and (D) liver mass of mice fed a SD for 16 weeks, with TVI of AAV8–TAp63 a or AAV8–GFP at week 1, and a second TVI of lentivirus encoding the shRNA-ATG3 or shRNA-scrambled control at week 8, as indicated. (E) Serum levels of AST and ALT. (F) Representative microphotographs of H&E (upper panel) and Oil Red O staining (lower panel) of liver sections. Oil Red O-stained sections were quantified using ImageJ. Hepatic triglyceride content was also directly measured. (G) Hepatic protein levels of lipid metabolism markers. (H) Liver protein levels of PGC1 a and SIRT1. (I) Liver protein levels of OXPHOS complex subunits I–V. Protein levels were normalized to GAPDH and HPS70, and mRNA levels, to HPRT. Dividing lines indicate splicing in the same gel. Data are presented as mean ± SEM; *p<0.05, **p<0.01, using one-way ANOVA followed by a Newman-Keuls Multiple comparison test (n = 7–10 per group). AAV, adeno-associated virus; NAFLD, non-alcoholic fatty liver disease; OXPHOS, oxidative phosphorylation; SD, standard diet; sh(RNA), short-hairpin (RNA); TA, transactivation domain; TVI, tail vein injection. Journal of Hepatology 2022 vol. 76 j11–24 19 0.0 0.5 1.0 1.5 2.0 ATG3 protein levels (% of control) ** SIRT1 mRNA levels (fold change) 0.0 0.5 1.0 1.5 2.0 *** Liver mass (g/100 g of BW) 2.5 3.0 3.5 4.0 4.5 5.0 5.5 AST (U/L) 0 10 20 30 40 * Lentiv-shATG3 WT Lentiv-shSIRT1 CDHFD FAS ACC p-ACC CPT1a LPL GAPDH sh scrambled shATG3 shSIRT1 shATG3 + shSIRT1 Weeks Body weight change (g) 012345678 0 3 6 9 12 15 18 CDHFD sh scrambled CDHFD shATG3 CDHFD shSIRT1 CDHFD shATG3 + shSIRT1 shATG3 shSIRT1 Weeks Cumulative food intake (g) 012345678 0 50 100 150 shATG3 shSIRT1 ALT (U/L) 0 20 40 60 80 * Complex II activity (SDH units/g tissuel) 0.00 0.05 0.10 0.15 ** *** ** Liver protein levels (% of control) C-I C-II C-III C-IV C-V 0 1 2 3 *** ***** Liver protein levels (% of control) FAS ACC p-ACC p-ACC/ACC CPT1a LPL 0 1 2 3 4 * ** *** CDHFD sh scrambled 50 μm 50 μm 50 μm50 μm 50 μm 50 μm 50 μm 50 μm CDHFD shSIRT1 CDHFD shATG3 CDHFD shATG3 + shSIRT1 HE Oil Red O Staining sh scrambled ATG3 GAPDH shATG3 shSIRT1 shATG3 + shSIRT1 V-ATP5A IV-COX II III-UQCRC2 II-SDH5 GAPDH I-NDUFB8 sh scrambled shATG3 shSIRT1 shATG3 + shSIRT1 0 2 4 6 8 10 Triglycerides (mg TG/g of liver) ** ** ** Oil Red O Area % of control (A. U.) 0.0 0.5 1.0 1.5 2.0 * Scrambled shATG3 shSIRT1 shATG3 + shSIRT1 Scrambled shATG3 shSIRT1 shATG3 + shSIRT1 Scrambled shATG3 shSIRT1 shATG3 + shSIRT1 Scrambled shATG3 Scrambled shATG3 shSIRT1 Scrambled shSIRT1 shATG3 + shSIRT1 Scrambled shATG3 shSIRT1 shATG3 + shSIRT1 Scrambled shATG3 shSIRT1 shATG3 + shSIRT1 A C B F D G HI E Fig. 7. Inhibition of SIRT1 blunts the effects of liver-specific downregulation of ATG3 in NAFLD. (A) Body weight change, (B) food intake, (C) hepatic levels of ATG3 protein and SIRT1 mRNA, and (D) liver mass in mice fed CDHFD for 8 weeks, with a lentivirus with shRNA-ATG3 or shRNA-scrambled at week 1, and a second TVI of lentivirus with shRNA-SIRT1 or shRNA-scrambled at week 4, as indicated. (E) Serum levels of AST and ALT. (F) Representative microphotographs of Oil Red O (upper panel) and H&E staining (lower panel) of liver sections. Oil Red O staining was quantified using ImageJ. Hepatic triglyceride content was also directly measured. (G) Hepatic protein levels of lipid metabolism markers. (H) Liver protein levels of OXPHOS complex subunits I–V. (I) Hepatic Complex II activity. Protein levels were normalized with GAPDH, and mRNA levels, with HPRT. Dividing lines indicate splicing in the same gel. *p<0.05, **p<0.01, ***p<0.001, using one-way ANOVA followed by a Newman-Keuls Multiple Comparison Test (n = 9–10 per group). CDHFD, choline-deficient high-fat diet; NAFLD, nonalcoholic fatty liver disease; OXPHOS, oxidative phosphorylation; sh(RNA), short-hairpin (RNA); TVI, tail vein injection. 20 Journal of Hepatology 2022 vol. 76 j11–24 Research Article NAFLD and Alcohol-Related Liver Diseases (Fig. 7D-7E). The reduced lipid content and hepatic triglycerides in mice with ATG3-knockdown was blunted when SIRT1 was subsequently inhibited (Fig. 7F). The knockdown of hepatic ATG3 caused an increase in CPT1a protein levels as well as a higher activity of complex II, but these effects were blocked by cosilencing ATG3 and SIRT1 (Fig. 7G-7I). On the other hand, the expression of fibrotic markers remained unchanged between the groups (Fig. S4C). ATG3 regulates protein levels of SIRT1 via JNK1 A key factor in regulating SIRT1 protein levels is c-Jun N-terminal kinase 1 (JNK1), whose activation induces SIRT1 protein degradation, 10 while JNK2 induces SIRT1 protein stability. 11 We measured JNK1 and JNK2 in the liver of mice fed a CDHFD where ATG3 was knocked down in the liver and detected reduced levels of JNK1 but not JNK2 (Fig. S10A). Consistent with in vivo data, silencing ATG3 in THLE2 cells reduced JNK1 (Fig. S10B), while the induction of ATG3 increased JNK1 (Fig. S10C). Silencing ATG3 in primary hepatocytes decreased JNK1 (Fig. S10D), while its overexpression increased JNK1 (Fig. S10E). Finally, treatment of isolated hepatocytes with the JNK1 inhibitor SP600125 blunted the ATG3-induced actions, including increased lipid content, reduced oxygen consumption rate and SIRT1 activity (Fig. S10F- 10H). Although a previous report found that modulation of ATG3 in the adipose tissue activates the Nrf2/Keap1 signaling pathway in the liver, 12 the expression of Nrf2 and Keap1 in the white adipose tissue and liver of mice fed a CDHFD where ATG3 was knocked down remained unaltered (Fig. S11A-11C). Similar results were found in the white adipose tissue of mice where Tap63 a was manipulated (Fig. S11D) and in isolated hepatocytes after silencing ATG3 (Fig. S11E). ATG3 requires CPT1a to modulate lipid metabolism and mitochondrial activity CPT1a floxed mice fed a CDHFD were injected with AAV8-GFP (control) or AAV8-Cre; at week 4, experimental groups were injected with either scrambled shRNA (control) or shRNA-ATG3, and mice were sacrificed at week 8. Body weight remained unchanged (Fig. S12A). CPT1a and ATG3 levels in livers were reduced after AAV8-Cre or shRNA-ATG3 treatment, respectively (Fig. S12B). Liver mass was not affected, but AST and ALT levels were lower after ATG3-knockdown; these levels returned to baseline when both ATG3 and CPT1a were silenced (Fig. S12C- 12D). The reduced lipid and hepatic triglyceride content and increased levels of complex II and IV in mice with ATG3 downregulation were blunted when CPT1a was also reduced (Fig. S12E-12F). The expression of fibrotic markers was not altered between these groups (Fig. S12G). ATG3 regulates hepatic lipid content in an autophagyindependent manner To further explore whether ATG3 exerts its hepatic actions via an autophagy-dependent process, we used both THLE2 and HepG2 hepatocyte cell lines to manipulate ATG3. Upon induction of autophagy, microtubule-associated protein 1 light chain 3 alpha (MAP1LC3A or LC3) is lipidated, and this LC3-phospholipid conjugate (LC3-II) is recruited to autophagosomal membranes and fuses with lysosomes to form autolysosomes, which degrade intra-autophagosomal components and LC3-II. 13,14 As expected, silencing ATG3 in HepG2 cells decreased LC3-II levels (Fig. 8A). We then monitored the autophagic flux by analyzing LC3-II turnover, in the presence or absence of the inhibitor of lysosome-mediated proteolysis, chloroquine (CQ). CQ administration induced the expected LC3-II accumulation (Fig. 8A). Hepatocytes treated with OA and transfected with siATG3 stored fewer lipids, and CQ treatment increased the lipid content (Fig. 8B). Strikingly, however, in cells overexpressing ATG3, inhibition of autophagy by CQ did not affect ATG3-induced lipid droplets in THLE2 or HepG2 cells (Fig. 8C-8D). Thus, these results suggested that the lipid accumulation induced by ATG3 is independent of autophagy. This was confirmed in vivo, as ATG3- knockdown in our 2 animal models of steatosis (induced by TAp63 a or CDHFD) reduced LC3-II accumulation but did not affect protein and mRNA levels of well-established markers of autophagy such as ATG5, ATG7, and p62 (Fig. 8E-8F). Overall, these data indicate that the effects of ATG3 on lipid content are independent of its autophagic action. This is also supported by the fact that autophagy contributes to the turnover of lipid droplets, and thereby to an increased fatty acid oxidation, a process named lipophagy. 15,16 Discussion In this work, we identified for first the time the role of ATG3 on fatty acid metabolism and its implications in the development of NAFLD. ATG3 is elevated in the liver of animal models as well as patients with NAFLD. Moreover, in vitro and in vivo genetic functional studies indicated that the overexpression of ATG3 favors lipid deposition, while its silencing alleviates steatosis induced by OA, p63, or diet. These novel findings expand our knowledge about the molecular mechanisms involved in liver steatosis. Autophagy has a well-established role in hepatic lipid metabolism, insulin sensitivity, and cellular injury, suggesting different potential mechanistic roles for autophagy in NAFLD. 17,18 Our initial hypothesis was that ATG3, an enzyme that catalyzes the LC3 lipidation process essential for autophagocytosis, 19 exerts its effects through an autophagic action, as p63 (as well as its family co-members p53 and p73) can induce autophagy. 20,21 Contrary to our expectations, we found that ATG3 modulates lipid metabolism via non-autophagic mechanisms but in a JNK1- and SIRT1-dependent manner. ATG3 increases JNK1 protein levels, which are known to induce SIRT1 protein degradation. 10 The lower protein levels of SIRT1, concomitant to its inhibited activity in human hepatocytes and murine liver reduced fatty acid oxidation and mitochondrial function, leading to a higher lipid load. In agreement with this, the inhibition of ATG3 decreased JNK1 and increased protein levels of SIRT1 and mitochondrial function, ultimately decreasing lipid content. Indeed, many reports have demonstrated a key role of SIRT1 in stimulating mitochondrial function 22 and fatty acid oxidation, 23,24 which contribute to attenuating hepatic steatosis. 25 Our findings show that the protection against steatosis induced by the inhibition of ATG3 was lost when SIRT1 and CPT1a were also suppressed in cells and mice, indicating a functional role for SIRT1 and CPT1a in the anti-steatotic action, which was not associated with changes in autophagic flux. Our results in the liver are opposite to the actions of ATG3 in adipose tissue, where depletion of ATG3 triggered the Nrf2/Keap1 pathway in the liver. 12 However, it is important to note that mice lacking ATG3 in adipose tissue displayed increased adipose tissue inflammation, systemic insulin resistance and accumulation of lipid peroxides, which are known to cause a marked impact on liver metabolism. Journal of Hepatology 2022 vol. 76 j11–24 21 A D E F C B LC3 Flux (% of control) 0.0 0.5 1.0 1.5 HepG2 OA siControl HepG2 OA siATG3 ** Oil Red O staining 50 μm 50 μm 50 μm 50 μm Veh CQ OA siATG3 OA siCTL LC3-II ATG5 ATG7 LC3-I CDHFD sh scrambled p62 GAPDH CDHFD shATG3 GAPDH LC3 II LC3 I OA siControl - - - +++ +++--- OA siATG3 CQ ATG5 ATG7 LC3-II LC3-I p62 GAPDH Oil Red O Area % of control (A. U.) 0 50 100 150 -+ -+ OA siControl OA siATG3 CQ ** * Liver protein levels (% of control) ATG5 ATG7 LC3 II-I p62 0 1 2 3 4CDHFD sh scrambled CDHFD shATG3 * GFP TAp63α TAp63α+ shATG3 GFP TAp63α TAp63α+ shATG3 0.0 0.5 1.0 1.5 2.0 ATG5 mRNA levels (fold change) 0 1 2 3 ATG7 mRNA levels (fold change) Scrambled shATG3 0.0 0.5 1.0 1.5 2.0 ATG5 mRNA levels (fold change) Scrambled shATG3 0.0 0.5 1.0 1.5 2.0 ATG7 mRNA levels (fold change) Liver protein levels (% of control) ATG5 ATG7 LC3 II-I p62 0.0 0.5 1.0 1.5 2.0 2.5 SD diet AAV8-GFP + sh scrambled SD diet AAV8-TAp63α + sh scrambled SD diet AAV8-TAp63α + shATG3 *** Oil Red O staining Veh CQ pATG3 Control Oil Red O staining Veh CQ pATG3 Control Oil Red O area % of control (A. U.) 0 100 200 300 400 -+ -+ pControl pATG3 CQ * * Oil Red O area % of control (A. U.) 0 100 200 300 -+ -+ pControl pATG3 CQ * * 50 μm 50 μm 50 μm 50 μm 50 μm 50 μm 50 μm 50 μm AAV8-GFP sh scrambled AAV8-TAp63α sh scrambled AAV8-TAp63α shATG3 Fig. 8. ATG3 regulates hepatic lipid accumulation in an autophagy-independent manner. HepG2 cells silencing ATG3 were treated for 24 hours with 1 mM OA and 60 l M CQ (added 18 h after OA) (n = 3-4 per group). (A) LC3-I and LC3-II protein levels. (B) Representative microphotographs of Oil Red O staining. Oil Red O staining was quantified using ImageJ and normalized to the total number of nuclei per field. (C) HepG2 cells and (D) THLE2 cells overexpressing ATG3 were 22 Journal of Hepatology 2022 vol. 76 j11–24 Research Article NAFLD and Alcohol-Related Liver Diseases On the other hand, it has also been reported that there is a tissue-specific autophagy response. 26–28 It is important to highlight that most, if not all, components of the molecular machinery for autophagy mediate effects that do not depend on lysosomal degradation of autophagy substrates. 29,30 Consequently, numerous, distinct non-autophagic biological functions for different ATG proteins have been reported, including cell survival, modulation of cellular transport, secretory processes, signaling, transcriptional/translational responses, and membrane reorganization. 29,30 More specifically, ATG3 has been shown to modulate in an autophagicindependent manner diverse cellular processes such as phagocytosis, secretion and exocytosis and cell proliferation. 30 Our study indicating that the effects of ATG3 are modulated by JNK1, SIRT1 and CPT1a points to another, previously undescribed nonautophagic action. Importantly, our results also propose that different ATGs have opposite actions in the liver, at least in terms of fatty liver disease. While we demonstrated that inhibition of ATG3 ameliorates NAFLD, liver-specific deletion of ATG5, ATG7, or ATG14 exacerbate the susceptibility of mice to develop NAFLD in response to a HFD. 31–34 However, the loss of autophagic factors does not always involve negative actions. For example, mice lacking ATG5 specifically in the liver are protected against acetaminophen-induced liver injury. 35 These paradoxical actions of autophagy might occur after long periods in response to cellular compensatory mechanisms. The ameliorated steatosis after ATG3-knockdown, and subsequent increased SIRT1 and CPT1a protein levels, was associated with stimulated mitochondrial function. The link between ATG3 and mitochondrial homeostasis has been studied before in the context of cell death, pluripotency acquirement, and maintenance of embryonic stem cells. 36,37 Although those studies suggested that the effects of ATG3 on mitochondria are dependent on autophagy, disruption of ATG3/ATG12 conjugation does not affect starvation-induced autophagy, 37 supporting an autophagic-independent role of ATG3. In addition to the unexpected mechanism used by ATG3, another important feature is that ATG3 levels are elevated in the livers of different animal models and patients with steatosis and NASH, while its genetic inhibition in mice only alleviated steatosis but did not alter fibrosis. Although ATG3 does not seem to be differentially expressed in the late stages of NAFLD compared to initial phases, it is important to consider that a fraction of people with NAFL will develop NASH or advanced liver fibrosis, 38–40 implying that specific signaling mechanisms are involved in liver disease progression. Considering that ATG3 overexpression is an early dysregulated mechanism and that levels of ATG3 remain high during advanced NASH, it could be of interest for diagnosis and/or as a therapeutic target to prevent the progression of the disease. In summary, our findings show for the first time that: a) ATG3 expression increased in the liver of mouse models and of patients with NAFLD, who also show a positive correlation between ATG3 with steatosis grade and NAFLD activity score; b) ATG3 overexpression induced lipid load, and its inhibition in the liver ameliorated TAp63 a - and diet-induced liver steatosis; c) this anti-steatotic action is mediated by reduction of JNK1 and increased levels of SIRT1, CPT1a, fatty acid oxidation, and mitochondrial function, in an autophagic-independent manner (Fig. S13). Overall, our results point towards ATG3 as a novel molecule implicated in the development of steatosis. Abbreviations AAV, adeno-associated virus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ATG3, autophagy-related gene 3; CDHFD, choline-deficient high-fat diet; CPT1a, carnitine palmitoyltransferase 1a; CQ, chloroquine; DIO, diet-induced obese; FAS, fatty acid synthase; HFD, high-fat diet; JNK1/2, c-Jun N- terminal protein kinase 1/2; LC3 (MAP1LC3A), microtubuleassociated protein 1 light chain 3 alpha; LC3-II, LC3- phospholipid conjugate; NAFL, non-alcoholic fatty liver; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; OA, oleic acid; PGC1 a , proliferator-activated receptor gamma coactivator 1 alpha; SD, standard diet; shRNA, shorthairpin RNA; siRNA, small-interfering RNA; SIRT1, sirtuin 1; TA, transactivation domain. Financial support This work has been supported by grants from FEDER/Ministerio de Ciencia, Innovación y Universidades-Agencia Estatal de Investigación (PA: RTI2018-095134-B-100; DS and LH: SAF2017- 83813-C3-1-R; MLMC: RTC2019-007125-1; CD: BFU2017-87721; ML: RTI2018–101840-B-I00; GS; PID2019-104399RB-I00; RN: RTI2018-099413-B-I00 and RED2018-102379-T; MLMC: SAF2017- 87301-R; TCD: RTI2018-096759-A-100), FEDER/Instituto de Salud Carlos III (AGR: PI19/00123), Xunta de Galicia (ML: 2016- PG068; RN: 2015-CP080 and 2016-PG057), Fundación BBVA (RN, GS and MLM), Proyectos Investigación en Salud (MLMC: DTS20/00138), Sistema Universitario Vasco (PA: IT971-16); Fundación Atresmedia (ML and RN), Fundación La Caixa (M.L., R.N. and M.C.), Gilead Sciences International Research Scholars Program in Liver Disease (MVR), Marató TV3 Foundation (DS: 201627), Government of Catalonia (DS: 2017SGR278) and European Foundation for the Study of Diabetes (RN and GS). This research also received funding from the European Community’s H2020 Framework Programme (ERC Synergy Grant-2019- WATCH- 810331, to RN, VP and MS). Centro de Investigación Biomédica en Red (CIBER) de Fisiopatología de la Obesidad y Nutrición (CIBERobn), Centro de Investigación Biomédica en Red (CIBER) de Enfermedades Hepáticas y Digestivas (CIBERehd) and CIBER de Diabetes y Enfermedades Metabólicas Asociadas (CIBERdem). CIBERobn, CIBERehd and CIBERdem are initiatives of the Instituto de Salud Carlos III (ISCIII) of Spain which is supported by FEDER funds. We thank MINECO for the Severo Ochoa Excellence Accreditation to CIC bioGUNE (SEV- 2016-0644). = treated for 24 hours with 1 mM OA and 60 l M CQ (added 18 h after OA) (n = 4-7 per group). Representative microphotographs of Oil Red O staining. The n represents the number of independent experiments. (E) Protein levels of autophagy markers and mRNA levels of ATG5 and ATG7 in the livers of mice fed a SD with a TVI of AAV8-TAp63 a or AAV8-GFP at week 1, and a second TVI of lentivirus with shRNA-ATG3 or shRNA-scrambled at week 8 (n=7-10 per group); and in liver of mice fed a CDHFD with hepatic ATG3 downregulation (n = 7-15 per group) (F). Protein levels were normalized with GAPDH, and mRNA levels, with HPRT. Dividing lines indicate splicing in the same gel. Data are presented as mean ± SEM; *p<0.05, **p<0.01, using a Student’sttest (A) (B) (C) (D) (F) or one-way ANOVA followed by a Newman-Keuls Multiple Comparison Test (E). AAV, adeno-associated virus; CDHFD, choline-deficient high-fat diet; CQ, chloroquine; OA, oleic acid; SD, standard diet; sh(RNA), short-hairpin (RNA); TVI, tail vein injection. Journal of Hepatology 2022 vol. 76 j11–24 23 Conflict of interest The authors declare that they have no conflicts of interest related to the study. Please refer to the accompanying ICMJE disclosure forms for further details. Authors’contributions N.D.S.L., M.F.F., E.N., X.B., S.G., M.J.G.-R., U.F., A.L., M.G-V., M.D.P. C- V., S.B.B., P.M., A. E., M.L., D.G., M.F., P.G., M.C., G.S., M. V-R., T.C.D., M.S., R.G.: study conception and design, data acquisition, and data analysis and interpretation; M.L., C.D., L.H., D.S., M.L.M.-C., M.R-G., V.P., C.G-C., A.G-R., P.A., R.N.: manuscript writing and final review. Data availability statement The data associated with this paper are available upon request to the corresponding author. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.jhep.2021.09.008. References Author names in bold designate shared co-first authorship [1] Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch V, et al. p63, a p53 homolog at 3q27-29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Mol Cell 1998;2:305–316. [2] Porteiro B, Fondevila MF, Delgado TC, Iglesias C, Imbernon M, Iruzubieta P, et al. Hepatic p63 regulates steatosis via IKKbeta/ER stress. Nat Commun 2017;8:15111. [3] Amir M, Czaja MJ. Autophagy in nonalcoholic steatohepatitis. Expert Rev Gastroenterol Hepatol 2011;5:159–166. [4] Singh R,Kaushik S, Wang Y, Xiang Y, Novak I, Komatsu M, et al. Autophagy regulates lipid metabolism. Nature 2009;458:1131–1135. [5] Yang L,Li P, Fu S, Calay ES, Hotamisligil GS. Defective hepatic autophagy in obesity promotes ER stress and causes insulin resistance. Cell Metab 2010;11:467–478. [6] Thoen LF, Guimaraes EL, Dolle L, Mannaerts I, Najimi M, Sokal E, et al. A role for autophagy during hepatic stellate cell activation. J Hepatol 2011;55:1353–1360. [7] Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005;41:1313–1321. [8] Gerhart-Hines Z, Rodgers JT, Bare O, Lerin C, Kim SH, Mostoslavsky R, et al. Metabolic control of muscle mitochondrial function and fatty acid oxidation through SIRT1/PGC-1alpha. EMBO J 2007;26:1913–1923. [9] Fondevila MF, Fernandez U, Gonzalez-Rellan MJ, Da Silva Lima N, Buque X, Gonzalez-Rodriguez A, et al. The L-alpha-lysophosphatidylinositol/GPR55 system induces the development of non-alcoholic steatosis and steatohepatitis. Hepatology 2020. [10] Gao Z, Zhang J, Kheterpal I, Kennedy N, Davis RJ, Ye J. Sirtuin 1 (SIRT1) protein degradation in response to persistent c-Jun N-terminal kinase 1 (JNK1) activation contributes to hepatic steatosis in obesity. J Biol Chem 2011;286:22227–22234. [11] Ford J, Ahmed S, Allison S, Jiang M, Milner J. JNK2-dependent regulation of SIRT1 protein stability. Cell Cycle 2008;7:3091–3097. [12] Cai J, Pires KM, Ferhat M, Chaurasia B, Buffolo MA, Smalling R, et al. Autophagy ablation in adipocytes induces insulin resistance and reveals roles for lipid peroxide and Nrf2 signaling in adipose-liver crosstalk. Cell Rep 2018;25:1708–1717 e1705. [13] Nakatogawa H. Mechanisms governing autophagosome biogenesis. Nat Rev Mol Cell Biol 2020;21:439–458. [14] Tsuboyama K,Koyama-Honda I, Sakamaki Y, Koike M, Morishita H, Mizushima N. The ATG conjugation systems are important for degradation of the inner autophagosomal membrane. Science 2016;354: 1036–1041. [15] Mizushima N, Levine B, Cuervo AM, Klionsky DJ. Autophagy fights disease through cellular self-digestion. Nature 2008;451:1069–1075. [16] Martinez-Lopez N, Singh R. Autophagy and lipid droplets in the liver. Annu Rev Nutr 2015;35:215–237. [17] Hazari Y, Bravo-San Pedro JM, Hetz C, Galluzzi L, Kroemer G. Autophagy in hepatic adaptation to stress. J Hepatol 2020;72:183–196. [18] Ueno T, Komatsu M. Autophagy in the liver: functions in health and disease. Nat Rev Gastroenterol Hepatol 2017;14:170–184. [19] Fujita N, Itoh T, Omori H, Fukuda M, Noda T, Yoshimori T. The Atg16L complex specifies the site of LC3 lipidation for membrane biogenesis in autophagy. Mol Biol Cell 2008;19:2092–2100. [20] Napoli M, Flores ER. The family that eats together stays together: new p53 family transcriptional targets in autophagy. Genes Dev 2013;27:971–974. [21] Kenzelmann Broz D, Spano Mello S, Bieging KT, Jiang D, Dusek RL, Brady CA, et al. Global genomic profiling reveals an extensive p53- regulated autophagy program contributing to key p53 responses. Genes Dev 2013;27:1016–1031. [22] Tang BL. Sirt1 and the mitochondria. Mol Cells 2016;39:87–95. [23] Gillum MP, Erion DM, Shulman GI. Sirtuin-1 regulation of mammalian metabolism. Trends Mol Med 2011;17:8–13. [24] Nogueiras R, Habegger KM, Chaudhary N, Finan B, Banks AS, Dietrich MO, et al. Sirtuin 1 and sirtuin 3: physiological modulators of metabolism. Physiol Rev 2012;92:1479–1514. [25] Li Y, Wong K, Giles A, Jiang J, Lee JW, Adams AC, et al. Hepatic SIRT1 attenuates hepatic steatosis and controls energy balance in mice by inducing fibroblast growth factor 21. Gastroenterology 2014;146:539– 549 e537. [26] Chapin HC, Okada M, Merz AJ, Miller DL. Tissue-specific autophagy responses to aging and stress in C. elegans. Aging (Albany NY) 2015;7:419–434. [27] Marino G, Salvador-Montoliu N, Fueyo A, Knecht E, Mizushima N, Lopez- Otin C. Tissue-specific autophagy alterations and increased tumorigenesis in mice deficient in Atg4C/autophagin-3. J Biol Chem 2007;282:18573–18583. [28] Singh R. Autophagy and regulation of lipid metabolism. Results Probl Cell Differ 2010;52:35–46. [29] Subramani S, Malhotra V. Non-autophagic roles of autophagy-related proteins. EMBO Rep 2013;14:143–151. [30] Galluzzi L, Green DR. Autophagy-independent functions of the autophagy machinery. Cell 2019;177:1682–1699. [31] Hammoutene A, Biquard L, Lasselin J, KheloufiM, Tanguy M, Vion AC, et al. A defect in endothelial autophagy occurs in patients with nonalcoholic steatohepatitis and promotes inflammation and fibrosis. J Hepatol 2020;72:528–538. [32] Xiong X, Tao R, DePinho RA, Dong XC. The autophagy-related gene 14 (Atg14) is regulated by forkhead box O transcription factors and circadian rhythms and plays a critical role in hepatic autophagy and lipid metabolism. J Biol Chem 2012;287:39107–39114. [33] Liu K, Zhao E, Ilyas G, Lalazar G, Lin Y, Haseeb M, et al. Impaired macrophage autophagy increases the immune response in obese mice by promoting proinflammatory macrophage polarization. Autophagy 2015;11: 271–284. [34] Ni HM, Woolbright BL, Williams J, Copple B, Cui W, Luyendyk JP, et al. Nrf2 promotes the development of fibrosis and tumorigenesis in mice with defective hepatic autophagy. J Hepatol 2014;61:617–625. [35] Ni HM, Boggess N, McGill MR, Lebofsky M, Borude P, Apte U, et al. Liverspecific loss of Atg5 causes persistent activation of Nrf2 and protects against acetaminophen-induced liver injury. Toxicol Sci 2012;127:438–450. [36] Liu K, Zhao Q, Liu P, Cao J, Gong J, Wang C, et al. ATG3-dependent autophagy mediates mitochondrial homeostasis in pluripotency acquirement and maintenance. Autophagy 2016;12:2000–2008. [37] Radoshevich L, Murrow L, Chen N, Fernandez E, Roy S, Fung C, et al. ATG12 conjugation to ATG3 regulates mitochondrial homeostasis and cell death. Cell 2010;142:590–600. [38] McPherson S, Hardy T, Henderson E, Burt AD, Day CP, Anstee QM. Evidence of NAFLD progression from steatosis to fibrosing-steatohepatitis using paired biopsies: implications for prognosis and clinical management. J Hepatol 2015;62:1148–1155. [39] Adams LA, Ratziu V. Non-alcoholic fatty liver - perhaps not so benign. J Hepatol 2015;62:1002–1004. [40] Pais R, Charlotte F, Fedchuk L, Bedossa P, Lebray P, Poynard T, et al. A systematic review of follow-up biopsies reveals disease progression in patients with non-alcoholic fatty liver. J Hepatol 2013;59:550–556. 24 Journal of Hepatology 2022 vol. 76 j11–24 Research Article NAFLD and Alcohol-Related Liver Diseases ORIGINAL ARTICLE—LIVER, PANCREAS, AND BILIARY TRACT Systematic review and meta-analysis: analysis of variables influencing the interpretation of clinical trial results in NAFLD Javier Ampuero 1,2,3 •Rocı ´o Gallego-Dura ´n 2,3 •Douglas Maya-Miles 2,3 • Rocı ´o Montero 2,3 •Sheila Gato 2,3 •A ´ngela Rojas 2,3 •Antonio Gil 2,3 • Rocı ´o Mun ˜oz 2,3 •Manuel Romero-Go ´mez 1,2,3 Received: 12 December 2021 / Accepted: 2 February 2022 / Published online: 24 March 2022 ÓThe Author(s) 2022 Abstract Background NAFLD clinical trials have shown suboptimal results, particularly for liver fibrosis, despite the robust preclinical drug development. We aimed to assess the histological response after the experimental treatment versus placebo by carrying out a meta-analysis of NAFLD clinical trials. Methods After a systematic review of NAFLD clinical trials to May 2021, applying strict selection criteria, the following primary outcomes were observed: (a) NASH resolution, with no worsening of fibrosis when available; (b) fibrosis improvement C1 stage, with no worsening of NAS when available; (c) worsening of NAS; (d) worsening of liver fibrosis C1 stage, including the progression to cirrhosis on histopathology. Other histological, clinical, and biochemical outcomes were considered secondary endpoints. Heterogeneity was explored by subgroup and sensitivity analyses, and univariable meta-regression. Results Twenty-seven randomized clinical trials were included. The pooled efficacy for NASH resolution receiving experimental therapy was 19% (95%CI 15–23; I 2 96.2%) compared with placebo 10% (95%CI 7–12; I 2 85.8%) (OR 1.66 (95%CI 1.24–2.21); I 2 57.8%), while it was 26% (95%CI 22–29); I 2 90%)) versus 18% (95%CI 15–21; I 2 59%)) for fibrosis improvement (OR 1.34 (95%CI 1.13–1.58); I 2 25.4%). For these outcomes, the therapy showed higher efficacy in trials longer than 48 weeks, with \60% of diabetic population, and when it targeted FXR, PPAR, and antidiabetic mechanisms, and with a NAS \5 for NASH resolution. Also, NASH (OR 0.57 (95%CI 0.39–0.84); I 2 67%) and fibrosis worsening (OR 0.65 (95%CI 0.46–0.92); I 2 61.9%) were prevented with the therapy. Conclusion This meta-analysis provides information about the efficacy of the therapy versus placebo by comparing different and combined trial outcomes such as NASH resolution, fibrosis improvement, and NAS and fibrosis worsening. Changes in the experimental design and selection criteria of the clinical trials might be suitable to increase the efficacy. Keywords NAFLD Fibrosis NASH Placebo Drug Abbreviations BMI Body mass index CI Confidence interval DNL De novo lipogenesis NAFL Non-alcoholic fatty liver NAFLD Non-alcoholic fatty liver disease NASH Non-alcoholic steatohepatitis NAS NAFLD activity score OR Odds ratio T2DM Type 2 diabetes mellitus Javier Ampuero and Rocı ´o Gallego-Dura ´n share co-first authorship. &Javier Ampuero [email protected] 1 Digestive Disease Department, Virgen del Rocio University Hospital, Avenida Manuel Siurot s/n, 41013 Sevilla, Spain 2 SeLiver Group, Instituto de Biomedicina de Sevilla/CSIC, Universidad de Sevilla, Sevilla, Spain 3 Hepatic and Digestive Diseases Networking Biomedical Research Centre (CIBERehd), Madrid, Spain 123 J Gastroenterol (2022) 57:357–371 https://doi.org/10.1007/s00535-022-01860-0 Introduction Within the past decade, the individual contribution of distinct etiologies to the burden of liver diseases has switched from viral hepatitis towards non-alcoholic fatty liver disease (NAFLD) [1]. NAFLD constitutes a complex metabolic disorder that manifests with fat accumulation in the cytoplasm of the hepatocyte in the absence of significant alcohol consumption or other causes of liver diseases [2]. Moreover, NAFLD severity ranges from simple steatosis to non-alcoholic steatohepatitis (NASH), usually accompanied by different stages of liver fibrosis, ballooning, and overall chronic inflammation status. This disease has surprisingly become a leading cause of liver cirrhosis and hepatocellular carcinoma worldwide. The risk of liverrelated mortality exponentially grows with an increase in fibrosis stages [3]; therefore, the diagnosis of liver fibrosis is one of the first steps when stratifying patients prior to the inclusion in clinical trials. Besides, NAFLD plays a catalytic role in the development of metabolic comorbidities in these multimorbid patients. Significant fibrosis but not simple steatosis or NASH predicts type 2 diabetes mellitus (T2DM) and arterial hypertension in these patients [4]. Despite its enormous prevalence, no regulatory-ap- proved therapeutic option has been authorized yet [5]; therefore. the cornerstone of NAFLD management still relies on lifestyle interventions [6]. In the context of clinical trials, the identification of patients at risk of suffering from liver-related and non-liver-related complications is tough. Besides, identifying the most appropriate therapy in NAFLD patients still remains a challenge. Recent studies have pointed out that the ideal treatment should address liver fibrosis and NASH in a joint fashion. To date, clinical trials have fallen short when testing the efficacy of novel molecular targets due to changes in some uncontrolled variables, such as dysmetabolic comorbidities and/or daily habits, since they are not reported nor adequately represented [7]. Considering hepatic fibrosis as a crucial factor of clinical prognosis and reinforcing the role of inflammation and disease activity as key players in the maintenance of chronicity in this disease, both factors should be taken into account as primary clinical trial endpoints. Data obtained from NAFLD clinical trials have shown suboptimal results, particularly for liver fibrosis, despite the robust preclinical development of the therapies. Therefore, in this setting, we carried out a meta-analysis to assess the histological response after the experimental treatment versus placebo (including NASH resolution and fibrosis improvement C1 stage) and, as a main novelty, the clinical benefits in delaying disease progression. Methods Study identification and selection We conducted our review according to the PRISMA reporting guideline for systematic reviews [8]. One of the reviewers (JA) with experience in database searches designed the search strategy, which was subsequently revised by other three investigators (RG, DM, AR). They independently searched MEDLINE (using PUBMED as the search engine), EMBASE, and Cochrane databases and collected all results separately. Disagreements between them were resolved by a third investigator (MRG) or by consensus. Databases were used to identify suitable studies that were published up to 1 May 2021. MeSH terms and keywords were used, and the search terms were as follows: NAFLD, MAFLD, NASH, non-alcoholic fatty liver disease, non-alcoholic steatohepatitis, fatty liver, liver fat, steatosis, clinical trial, treatment, therapy, drug, and a combination of those MeSH terms by using the appropriate Boolean logic. The searches were limited to English-lan- guage publications with human subjects. A manual search was conducted using the references listed in the original articles and review articles retrieved. Only fully published articles and oral presentations subjected to the same assessment as regular articles (AASLD and EASL meetings) were considered, so abstracts and posters were not considered. The inclusion criteria were as follows: (a) randomized clinical trial; (b) placebo-controlled clinical trial; (c) Phase II and Phase III clinical trial; (d) paired biopsy; (e) adults (C18 years old). The exclusion criteria were as follows: (a) duplicate reports; (b) case reports, comments, and letters to the editors; (c) systematic reviews or metaanalyses; (d) botanical products, herbal medicines, or antioxidants; (e) lifestyle intervention. Data extraction and quality assessment The following data were extracted: author, year, population selection criteria, sample size, experimental drug, histological endpoint (NASH, NAFLD Activity Score (NAS), fibrosis stage, steatosis, lobular inflammation, ballooning), biochemical response (AST, ALT), age, sex, body mass index (BMI), T2DM. When the same population was published in several journals, we retained only the most informative article or the most complete study to avoid duplication. We also asked the investigators for additional information, and if we received no answer, ‘‘unreported’’ items were treated as ‘‘unclear’’ or ‘‘not available’’. On the other hand, four investigators (AG, SG, RM, RM) independently assessed the quality of the studies using the ‘‘Quality in Prognostic Studies (QUIPS)’’ tool [9]. 123 358 J Gastroenterol (2022) 57:357–371 Outcome measures Given that the ultimate goal of NASH treatment is to slow the progress of, halt, or reverse disease progression and improve clinical outcomes, we selected the following the histological response after experimental treatment or placebo as the primary outcome: (a) NASH resolution, with no worsening of fibrosis when available; (b) fibrosis improvement C1 stage, with no worsening of NAS when available. On the other hand, as a clinical benefit can be verified by demonstrating superiority to placebo in delaying disease progression, we additionally considered: (a) worsening of NAS; (b) worsening of liver fibrosis C1 stage, including the progression to cirrhosis on histopathology. In addition, other histological outcomes were assessed as secondary endpointsasfollows: (a) NAS improvement [2 points, irrespective of fibrosis improvement; (b) improvement of steatosis, lobular inflammation, and ballooning. Also, the occurrence of cirrhosis complications was analyzed. Finally, the biochemical response (ALT, AST) was also assessed. Statistical analysis We used STATA version 16 (Stata Corp; College Station, TX). All statistical tests were two-sided, with P-val- ues B0.05 denoting statistical significance. Confidence intervals (CIs) of individual studies were determined from the available data. For the dichotomous variables, the effect denotes odds ratio (OR) and corresponding 95% CIs, while we used the difference in means to specifically provide measures of the absolute difference between the mean values of the explored variables. To estimate the pooled prevalence, the prevalence rates were combined in a random-effects meta-analysis. The assumption of heterogeneity was tested for each planned analysis using the Cochran-Q heterogeneity and I 2 statistics (significant heterogeneity according to I 2 value [50%) [10]. The random-effects model was applied to pool results from studies. We planned a priori subgroup analyses according to the following criteria: trials with cirrhotic versus non-cirrhotic population (three studies included 100% of cirrhotic patients, while other two included 50% and one 11%; however, they were considered as a cirrhotic population because separated information was not available), trials with C60% versus \60% of diabetic population, trials with mean NAS less than 5 versus 5 or greater, treatment duration 48 weeks or less versus greater than 48 weeks, and therapeutic class (Supplementary Table 1). Additionally, significant heterogeneity for primary outcomes was explored by univariable meta-regression, and a sensitivity analysis was performed to determine if there was any undue influence exerted by a single study on the results of the combined studies [11,12]. Finally, the potential publication bias was assessed by Egger’s test and graphically by a funnel plot when there was an adequate number of studies ([10 studies). Results Eligible study characteristics and quality assessment The flowchart diagram details the article selection process for this meta-analysis (Supplementary Fig. 1), which ended with 27 studies included. The characteristics of the eligible studies are listed in Table 1. Supplementary Table 2 shows the quality assessment of the clinical trials by QUIPS. Data analyses about NASH NASH resolution was assessed by 26 clinical trials (N= 7239 patients). The pooled efficacy for NASH resolution obtained by patients treated with any experimental drug was 19% (95%CI 15–23; I 2 96.2%) when compared with placebo 10% (95%CI 7–12; I 2 85.8%) (Supplementary Fig. 2). The treatment difference between receiving a therapy placebo was higher considering the studies evaluating additionally the lack of worsening of fibrosis (N= 17) (OR 2.32 (95%CI 1.67–3.23); I 2 4.9%) than considering the total of studies (N= 26) (OR 1.66 (95%CI 1.24–2.21); I 2 57.8%) (Fig. 1a). The subgroup analysis showed that NASH resolution was more difficult to achieve in cirrhotic in comparison with non-cirrhotic patients for both experimental therapy [(4% (95%CI 1–8; I 2 80.1%) versus 22% (95%CI 17–28; I 2 95.7%))] and placebo [(2% (95%CI 0–4; I 2 48%) versus 12% (95%CI 9–14; I 2 65.8%))] (Supplementary Fig. 2). In addition, the experimental drug showed higher efficacy in clinical trials with a mean NAS \5 versus NAS C5, with a duration [48 versus B48 weeks, with less than 60% of diabetic population, and when it was based on antimetabolic mechanisms, targeting de novo lipogenesis (DNL) and FXR agonist (Fig. 1). Improvement of NAS by C2 points is another typical endpoint of clinical trials. Up to 19 studies assessed this outcome (N= 3798 patients). Overall, receiving an experimental treatment increased the likelihood of achieving this outcome (37% (95%CI 0.28–0.46; I 2 96.1%)) compared to placebo (23% (95%CI 0.16–0.30; I 2 89%)) (OR 1.72 (95%CI 1.23–2.41); I 2 71%) (Supplementary Fig. 3). Similar to NASH resolution, the improvement of NAS C2 points was higher in non-cir- rhotic than in cirrhotic patients for the experimental (40% (95%CI 30–51; I 2 96.8%) and 24% (95%CI 14–34; I 2 57.4%), respectively) and placebo arms (26% (95%CI 123 J Gastroenterol (2022) 57:357–371 359 Table 1 Characteristics of the studies included in the meta-analysis Drug First author Year Phase Population Intervention Duration Histological endpoints (weeks) Aldafermin [14] Stephen A. Harrison 2021 IIb s=78 NAS (mean): 5.6 Fibrosis (%): F2 56, F3 44 Diabetes (%): 61.5% BMI (mean): 36.1 kg/m 2 Aldafermin 1 mg (n= 53) Placebo (n= 25) 24 NASH resolution & no worsening of fibrosis Improvement [2 NAS points & no worsening of fibrosis Fibrosis improvement C1 stage & no worsening of NAS Improvement of individual components of NAS Aramchol [35] Vlad Ratziu 2020 IIb N= 247 NAS (mean): 5.12 Fibrosis (%): F1 40, F2 20, F3 40 Diabetes (%): All diabetic or prediabetic BMI (mean): 32.7 kg/m 2 Aramchol 400 mg (n= 101) Aramchol 600 mg (n= 98) Placebo (n= 48) 52 NASH resolution & no worsening of fibrosis Fibrosis improvement C1 stage & no worsening of NAS Progression to cirrhosis Belapectin [15] Naga Chalasani 2020 IIb N= 162 NAS (mean): 4.2 Fibrosis (%): F4 100 Diabetes (%): 60.5 BMI (mean): 34.9 kg/m 2 Belapectin 2 mg/ kg (n= 54) Belapectin 8 mg/ kg (n= 54) Placebo (n= 54) 52 Fibrosis improvement C1 stage Cirrhosis complications Cenicriviroc [16] Scott L. Friedman 2018 IIb N= 289 NAS (mean): 5.3 Fibrosis (%): F1 33, F2 28, F3 38 Diabetes (%): 50.5 BMI (mean): 33.9 kg/m2 CVC 150 mg (n= 145) Placebo (n= 144) 52 NASH resolution & no worsening of fibrosis Improvement [2 NAS points & no worsening of fibrosis Fibrosis improvement C1 stage & no worsening of NAS Improvement of individual components of NAS Fibrosis worsening C1 stage Progression to cirrhosis Worsening of individual components of NAS 123 360 J Gastroenterol (2022) 57:357–371 (95%CI 25–34); I 2 61%) (OR 1.55 (95%CI 1.19–2.01); I 2 47.9%) (Fig. 4c). On the other hand, steatosis (OR 0.34 (95%CI 0.22–0.52); I 2 0%) had a lower likelihood to progress in patients receiving any experimental therapy than placebo, although this did not occur with ballooning (OR 0.87 (95%CI 0.45–1.67); I 2 69%) and lobular inflammation (OR 0.71 (95%CI 0.34–1.46); I 2 79.8%). Otherwise, the occurrence of cirrhosis complications was not prevented when using an experimental treatment (N= 3) (OR 1.41 (95%CI 0.86–2.32); I 2 29.4%). Finally, the necro-inflammatory activity also improved when taking an experimental therapy. AST levels were significantly decreased in these individuals compared to those receiving placebo (mean difference –10.1 IU/L (95%CI (–14.7 to –5.4); I 2 79.4%)) (Fig. 5a). Similarly, ALT levels were found to be diminished under treatment (mean difference –13.8 IU/L (95%CI (–23.5 to –4.1); I 2 92.3%)) (Fig. 5b). Heterogeneity assessment and publication bias The leave-one-out sensitivity analysis did not identify any single study that significantly contributed to the betweenstudies variability for NASH resolution (Supplementary Table 3), fibrosis improvement (Supplementary Table 4), and fibrosis progression (Supplementary Table 5). On the other hand, meta-regression showed no evidence of a differential effect of study-level characteristics on the impact of the outcomes, apart from baseline NAS for NASH resolution (P= 0.010) (Supplementary Table 6). Publication bias was conducted by Egger’s test and funnel plot asymmetry. There was no formal evidence of Fig. 4 The effect of the experimental drug on: ASteatosis; BBallooning; CLobular inflammation 123 J Gastroenterol (2022) 57:357–371 367 publication bias for NASH resolution (P= 0.453) (Supplementary Fig. 9), improvement of NAS by C2 points (P= 0.101) (Supplementary Fig. 10), fibrosis improvement C1 stage (P= 0.451) (Supplementary Fig. 11), fibrosis progression (P= 0.105) (Supplementary Fig. 12), steatosis improvement (P= 0.312), ballooning improvement (P= 0.496), lobular inflammation improvement (P= 0.232), and biochemical response (P= 0.812 and P= 0.957 for AST and ALT, respectively). Discussion Over the past five years, many clinical trials testing new drugs for NAFLD have been published [13–37. The scientific community has initially witnessed these preliminary data with enthusiasm and the final publication with a relative skepticism because of the suboptimal results, particularly for liver fibrosis improvement as an endpoint [38]. Our meta-analysis observed that taking an experimental therapy versus placebo increased the likelihood of resolute NASH and regress liver fibrosis. Despite the fact that only 5 of 76 studies demonstrated a beneficial effect of the therapy on NASH, the likelihood of NASH resolution was 60% higher than receiving placebo. In liver fibrosis, the likelihood of improving at least one stage was 30% higher with the therapy, although only 3 out of the 77 studies showed an individual benefit. Also, the individual components of NAS (steatosis, ballooning, and lobular inflammation), as well as the necro-inflammatory activity (evaluated by AST and ALT levels) significantly improved with the therapy. Of note, the percentage of NASH resolution and fibrosis regression for placebo was similar to that published in the literature [39], although recently it has Fig. 5 The effect of the experimental drug on: AAST levels; BALT levels 123 368 J Gastroenterol (2022) 57:357–371 been suggested a lower fibrosis progression rate in this group probably related to the number of patients without fibrosis [40]. Despite the global positive results, we found that the percentage of NASH resolution and fibrosis improvement was 19–28%, respectively, for experimental therapies based on biological plausibility, which are far from desirable. Due to its multifaceted nature [41], this fact mirrors the complexity of the underlying mechanisms of the pathogenesis of NAFLD. We found that some baseline variables and features related to the clinical trial design influenced the likelihood of achieving the outcomes. First, FXR agonists and antimetabolic drugs (including anti-diabetic therapies and PPAR agonists), and DNL-targeting therapies for NASH resolution, showed the highest efficacy for inflammation and fibrosis improvement. These findings have been documented in some studies [42,43], although they showed limitations such as not assessing the clinical benefit in delaying disease progression and the baseline features impacting the efficacy of the drugs. Second, NASH resolution was easier to achieve in trials with non-cirrhotic patients, with baseline NAS \5, with a low proportion of diabetic patients, and with a longer length of the therapy. These data did agree on variables associated with NASH- resolution after life-style intervention supporting a group of features defining more difficult-to-solve patients [44]. These characteristics were also more frequently associated with fibrosis improvement, with the exception of NAS. These results should make us to meditate on the design of NAFLD clinical trials and the adequacy of the endpoints to balance them with the prognostic relevance. On the one hand, the experimental therapy appeared to require at least 1 year to be effective. Thus, a longer duration than 48 weeks is preferred in NAFLD clinical trials. On the other hand, the drug effect was superior to placebo when achieving NASH resolution and fibrosis improvement, but it was not in trials including more severe patients. Therefore, NASH resolution should be required for the experimental therapy in non-cirrhotic patients and in those with a baseline NAS \5, but questionable for individuals showing a NAS C5 and, especially, for cirrhotic patients since most of them have lost some of the single components of NAS [45]. Similarly, although desirable, fibrosis improvement is not a realistic aim for cirrhotic patients using the current experimental therapies, according to our results. Instead, clinical trials on cirrhotic patients should focus on preventing portal hypertension, hepatocellular carcinoma occurrence, and mortality, extending the treatment course, rather than in the regression of liver disease. Therefore, we should make efforts to redirect the design and selection criteria of clinical trials because some potentially useful drugs could be discarded too early. NAFLD clinical trials should report a minimum of information about all relevant aspects that could impact on the efficacy of the experimental drug tested [7]. In this setting, results about efficacy tend to focus on the histological improvement (e.g., NASH or fibrosis), but frequently fail to mention data associated with the prevention of its progression. For example, in our meta-analysis, only 4 and 16 of 27 clinical trials reported information about worsening of NASH and fibrosis, respectively. Considering this, our results indicate that patients receiving therapy were protected against NASH worsening and/or fibrosis progression. In other liver diseases, the treatment aims mainly to eliminate (e.g., hepatitis C) or control the etiology (e.g., hepatitis B, autoimmune hepatitis) but does not reverse liver fibrosis, which is a consequence that requires an extended follow-up [46–48]. Instead, NAFLD clinical trials require an early resolution of NASH or fibrosis improvement to be considered a success. Given the nescience regarding in whom fibrosis regression can be expected and how quickly it occurs, we should consider halting the disease as a relevant outcome and, thus, complementary to the improvement of liver disease. Therefore, we encourage NAFLD clinical trials to report essential information about the progression of the disease to have an overall vision of the efficacy of experimental drugs. Beyond the strengths, our meta-analysis also has some limitations. First, the interpretation of some results could be challenging because of the different mechanisms of action of included drugs. However, this kind of approach has been done for other therapeutic areas (e.g., biologics in ulcerative colitis [49]) and could be interesting to provide additional data to guide the NAFLD drug pipeline properly. Second, studies reporting cirrhotic-related outcomes were scarce, precluding making robust analysis. Third, some baseline variables, such as T2DM or NAS, were categorized. This usual aspect allows making subgroup analyses or a meta-regression in the absence of the individual data but with a limitation in the interpretation. In conclusion, developing therapeutic strategies to revert or, at least, slow down steatohepatitis and fibrosis progression as much as possible in NAFLD is an unmet need. This meta-analysis provides information about the efficacy of the therapy versus placebo by comparing different and combined trial outcomes such as NASH resolution, fibrosis improvement, and NAS and fibrosis worsening. Given that novel pharmacological agents focused on NASH resolution and liver fibrosis regression are expected to be available in the upcoming years, changes in the experimental design and selection criteria of the clinical trials may increase the ability to demonstrate efficacy. 123 J Gastroenterol (2022) 57:357–371 369 Author statement Guarantor of the article: JA. Study design, statistical analyses, and interpretation: JA. Drafting the manuscript: JA, RGD, DMM, MRG. Data acquisition and critical review of the manuscript. All authors approved the final version of the article, including the authorship list. Funding Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This project has been partially funded by the ‘‘Consejerı ´a de Salud de la Junta de Andalucı ´a’’ (PI- 0075–2014), the ‘‘Spanish Ministry of Economy, Innovation and Competition, Instituto de Salud Carlos III’’ (PI16/01842, PI17/00535, PI19/01404, PI19/00589, IFI18/00041 and GLD19/00100).*The funders have not had any role in the design, analysis, writing, or interpretation of this project. Declarations Potential conflict of interest None. Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s00535- 022-01860-0. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/. References 1. Flemming JA, Djerboua M, Groome PA, et al. NAFLD and alcohol-related liver disease will be responsible for almost all new diagnoses of cirrhosis in Canada by 2040. Hepatology 2021. hep.32032. 2. Eslam M, Sanyal AJ, George J, et al. MAFLD: A consensusdriven proposed nomenclature for metabolic associated fatty liver disease. Gastroenterology. Published Online First: 7 February 2020. 3. Vilar-Gomez E, Calzadilla-Bertot L, Wai-Sun Wong V, et al. Fibrosis Severity as a Determinant of Cause-Specific Mortality in Patients With Advanced Nonalcoholic Fatty Liver Disease: A Multi-National Cohort Study. Gastroenterology. 2018;155:443- 457.e17. 4. Ampuero J, Aller R, Gallego-Dura ´n R, et al. Significant fibrosis predicts new-onset diabetes mellitus and arterial hypertension in patients with NASH. J Hepatol. 2020;73(1):17–25. 5. Ampuero J, Sa ´nchez-Torrijos Y, Aguilera V, et al. New therapeutic perspectives in non-alcoholic steatohepatitis. Gastroenterol Hepatol. 2018;41(2):128–42. 6. Vilar-Gomez E, Martinez-Perez Y, Calzadilla-Bertot L, et al. Weight loss through lifestyle modification significantly reduces features of nonalcoholic steatohepatitis. Gastroenterology. 2015;149:367-378.e5. 7. Ampuero J, Romero-Gomez M. Stratification of patients in NASH clinical trials: A pitfall for trial success. JHEP Rep. 2020;2(5):100148 8. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009;151:264–9. 9. Hayden JA, Co ˆte ´P, Bombardier C. Evaluation of the quality of prognosis studies in systematic reviews. Ann Intern Med. 2006;144:427–37. 10. Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. Br Med J. 2003;327:557–60. 11. Copas JB, Shi JQ. A sensitivity analysis for publication bias in systematic reviews. Stat Methods Med Res. 2001;10:251–65. 12. Petitti DB. Approaches to heterogeneity in meta-analysis. Stat Med. 2001;20:3625–33. 13. Albhaisi SAMM, Sanyal AJ. New drugs for NASH. Liver Int. 2021;41:112–8. 14. Harrison SA, Neff G, Guy CD, et al. Efficacy and safety of Aldafermin, an engineered FGF19 analog, in a randomized, double-blind, placebo-controlled trial of patients with nonalcoholic steatohepatitis. Gastroenterology. 2021;160:219-231.e1. 15. Chalasani N, Abdelmalek MF, Garcia-Tsao G, et al. Effects of belapectin, an inhibitor of galectin-3, in patients with nonalcoholic steatohepatitis with cirrhosis and portal hypertension. Gastroenterology. 2020;158:1334-1345.e5. 16. Friedman SL, Ratziu V, Harrison SA, et al. A randomized, placebo-controlled trial of cenicriviroc for treatment of nonalcoholic steatohepatitis with fibrosis. Hepatology. 2018;67:1754–67. 17. Loomba R, Noureddin M, Kowdley KV, et al. Combination therapies including cilofexor and firsocostat for bridging fibrosis and cirrhosis attributable to NASH. Hepatology. 2021;73:625–43. 18. Harrison SA, Ruane PJ, Freilich BL, et al. Efruxifermin in nonalcoholic steatohepatitis: a randomized, double-blind, placebocontrolled, phase 2a trial. Nat Med. 2021;27:1262–71. 19. Ratziu V, Harrison SA, Francque S, et al. Elafibranor, an agonist of the peroxisome proliferator-activated receptor-aand -d, induces resolution of nonalcoholic steatohepatitis without fibrosis worsening. Gastroenterology. 2016;150:1147-1159e5. 20. Harrison SA, Goodman Z, Jabbar A, et al. A randomized, placebo-controlled trial of emricasan in patients with NASH and F1– F3 fibrosis. J Hepatol. 2020;72:816–27. 21. Armstrong MJ, Gaunt P, Aithal GP, et al. Liraglutide safety and efficacy in patients with non-alcoholic steatohepatitis (LEAN): a multicentre, double-blind, randomised, placebo-controlled phase 2 study. Lancet (London, England). 2016;387:679–90. 22. Neuschwander-Tetri BA, Loomba R, Sanyal AJ, et al. Farnesoid X nuclear receptor ligand obeticholic acid for non-cirrhotic, nonalcoholic steatohepatitis (FLINT): a multicentre, randomised, placebo-controlled trial. Lancet. 2015;385:956–65. 23. Younossi ZM, Ratziu V, Loomba R, et al. Obeticholic acid for the treatment of non-alcoholic steatohepatitis: interim analysis from a multicentre, randomised, placebo-controlled phase 3 trial. Lancet. 2019;394:2184–96. 24. Sanyal. Pioglitazone, vitamin E, or placebo for nonalcoholic steatohepatitis [1]. N Engl J Med 2010;363:1185–6. 25. Cusi K, Orsak B, Bril F, et al. Long-term pioglitazone treatment for patients with nonalcoholic steatohepatitis and prediabetes or type 2 diabetes mellitus a randomized trial. Ann Intern Med. 2016;165:305–15. 26. Harrison SA, Bashir MR, Guy CD, et al. Resmetirom (MGL- 3196) for the treatment of non-alcoholic steatohepatitis: a multicentre, randomised, double-blind, placebo-controlled, phase 2 trial. Lancet. 2019;394:2012–24. 27. Loomba R, Lawitz E, Mantry PS, et al. The ASK1 inhibitor selonsertib in patients with nonalcoholic steatohepatitis: A randomized, phase 2 trial. Hepatology. 2018;67:549–59. 123 370 J Gastroenterol (2022) 57:357–371 28. Harrison SA, Wong VWS, Okanoue T, et al. Selonsertib for patients with bridging fibrosis or compensated cirrhosis due to NASH: results from randomized phase III STELLAR trials. J Hepatol. 2020;73:26–39. 29. Newsome PN, Buchholtz K, Cusi K, et al. A placebo-controlled trial of subcutaneous semaglutide in nonalcoholic steatohepatitis. N Engl J Med. 2021;384:1113–24. 30. Harrison SA, Abdelmalek MF, Caldwell S, et al. Simtuzumab is ineffective for patients with bridging fibrosis or compensated cirrhosis caused by nonalcoholic steatohepatitis. Gastroenterology. 2018;155:1140–53. 31. Newsome PN, Palmer M, Freilich B, et al. Volixibat in adults with non-alcoholic steatohepatitis: 24-week interim analysis from a randomized, phase II study. J Hepatol. 2020;73:231–40. 32. Harrison SA, Alkhouri N, Davison BA, et al. Insulin sensitizer MSDC-0602K in non-alcoholic steatohepatitis: a randomized, double-blind, placebo-controlled phase IIb study. J Hepatol. 2020;72:613–26. 33. Lucas KJ, Lopez P, Lawitz EJ, et al. Safety and efficacy of tropifexor in patients with fibrotic nonalcoholic steatohepatitis: 48-week results from part C of the phase 2 FLIGHT-FXR study. Hepatology 2020;72. 34. Harrison SA, Gunn NT, Khazanchi A, et al. A 52-week multicenter double-blind randomized phase 2 study of seladelpar, a potent and selective peroxisome proliferator-activated receptor delta (PPAR-delta) agonist, in patients with nonalcoholic steatohepatitis (NASH). Hepatology 2020;72. 35. Ratziu V, Guevara L de, Safadi R, et al. One-year results of the global phase 2b randomized placebocontrolled ARREST trial of aramchol, a stearoyl CoA desaturase modulator in NASH patients. Hepatology 2018. 36. Harrison SA, Ratziu V, Bedossa P, et al. RESOLVE-IT Phase 3 of Elafibranor in NASH: Final Results of the Week 72 Interim Surrogate Efficacy Analysis. Hepatology 2020;72. 37. Francque SM, Bedossa P, Ratziu V, et al. A randomized controlled trial of the panPPAR agonist lanifibranor in NASH. N Engl J Med 2021. 38. Mantovani A, Dalbeni A. Treatments for NAFLD: state of art. Int J Mol Sci. 2021;22:1–27. 39. Han MAT, Altayar O, Hamdeh S, et al. Rates of and factors associated with placebo response in trials of pharmacotherapies for nonalcoholic steatohepatitis: systematic review and metaanalysis. Clin Gastroenterol Hepatol. 2019;17:616-629.e26. 40. Roskilly A, Hicks A, Taylor EJ, et al. Fibrosis progression rate in a systematic review of placebo-treated nonalcoholic steatohepatitis. Liver Int. 2021;41:982–95. 41. Arrese M, Arab JP, Barrera F, et al. Insights into nonalcoholic fatty-liver disease heterogeneity. Semin Liver Dis. Published Online First: 7 July 2021. 42. Majzoub AM, Nayfeh T, Barnard A, et al. Systematic review with network meta-analysis: comparative efficacy of pharmacologic therapies for fibrosis improvement and resolution of NASH. Aliment Pharmacol Ther. 2021;54:880–9. 43. Brunt EM, Kleiner DE, Wilson LA, et al. Improvements in histologic features and diagnosis associated with improvement in fibrosis in nonalcoholic steatohepatitis: results from the nonalcoholic steatohepatitis clinical research network treatment trials. Hepatology. 2019;70:522–31. 44. Vilar-Gomez E, Yasells-Garcia A, Martinez-Perez Y, et al. Development and validation of a noninvasive prediction model for nonalcoholic steatohepatitis resolution after lifestyle intervention. Hepatology 2016;63. 45. Ampuero J, Aller R, Gallego-Dura ´n R, et al. Definite and indeterminate nonalcoholic steatohepatitis share similar clinical features and prognosis: a longitudinal study of 1893 biopsy-proven nonalcoholic fatty liver disease subjects. Liver Int. 2021;41(9):2076–86. 46. Rockey D, Friedman S. Fibrosis regression after eradication of hepatitis c virus: from bench to bedside. Gastroenterology. 2021;160:1502–20. 47. Dezanet LNC, Miailhes P, Lascoux-Combe C, et al. Profiles of liver fibrosis evolution during long-term tenofovir treatment in HIV-positive patients coinfected with hepatitis B. Liver Int 2021. liv.15019. 48. Bardou-Jacquet E, Morandeau E, Anderson GJ, et al. Regression of fibrosis stage with treatment reduces long-term risk of liver cancer in patients with hemochromatosis caused by mutation in HFE. Clin Gastroenterol Hepatol. 2020;18:1851–7. 49. Chandan S, Mohan BP, Kumar A, et al. Safety and efficacy of biological therapy in chronic antibiotic refractory pouchitis: a systematic review with meta-analysis. J Clin Gastroenterol. 2021;55:481–91. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 123 J Gastroenterol (2022) 57:357–371 371 1 Vol.:(0123456789) Scientific Reports | (2022) 12:3418 | https://doi.org/10.1038/s41598-022-06614-9 www.nature.com/scientificreports Liver injury in non‑alcoholic fatty liver disease is associated with urea cycle enzyme dysregulation Rocío Gallego‑Durán1,2,13,12,13*, Javier Ampuero1,2,3,12, Helena Pastor‑Ramírez1,2, Leticia Álvarez‑Amor4,5, Jose Antonio del Campo6, Douglas Maya‑Miles1,2, Rocío Montero‑Vallejo1,2, Antonio Cárdenas‑García4,5, Mª Jesús Pareja7, Sheila Gato‑Zambrano1,2, Raquel Millán1,2, María del Carmen Rico1,2, Amparo Luque‑Sierra4,5, Antonio Gil‑Gómez1,2, Ángela Rojas1,2, Rocío Muñoz‑Hernández1,2, María García‑Lozano1,2, Rocío Aller8, Raúl J. Andrade2,9, Carmelo García‑Monzón2,10, Fausto Andreola11, Francisco Martín4,5, Rajiv Jalan11 & Manuel Romero‑Gómez1,2,3,13,13* The main aim was to evaluate changes in urea cycle enzymes in NAFLD patients and in two preclinical animal models mimicking this entity. Seventeen liver specimens from NAFLD patients were included for immunohistochemistry and gene expression analyses. Three‑hundred‑and‑eighty‑two biopsy‑ proven NAFLD patients were genotyped for rs1047891, a functional variant located in carbamoyl phosphate synthetase‑1 (CPS1) gene. Two preclinical models were employed to analyse CPS1 by immunohistochemistry, a choline deficient high‑fat diet model (CDA‑HFD) and a high fat diet LDLr knockout model (LDLr −/−). A significant downregulation in mRNA was observed in CPS1 and ornithine transcarbamylase (OTC1) in simple steatosis and NASH‑fibrosis patients versus controls. Further, age, obesity (BMI > 30 kg/m2), diabetes mellitus and ALT werefound to be risk factors whereas A‑allele from CPS1 was a protective factor from liver fibrosis. CPS1 hepatic expression was diminished in parallel with the increase of fibrosis, and its levels reverted up to normality after changing diet in CDA‑HFD mice. In conclusion, liver fibrosis and steatosis were associated with a reduction in both gene and protein expression patterns of mitochondrial urea cycle enzymes. A‑allele from a variant on CPS1 may protect from fibrosis development. CPS1 expression is restored in a preclinical model when the main trigger of the liver damage disappears. Abbreviations NAFLD Non-alcoholic fatty liver disease NASH Non-alcoholic steatohepatitis CPS1 Carbamoyl phosphate synthetase-1 CDA-HFD Choline deficient high-fat diet model LDLr −/− LDLr knockout model OTC1 Ornithine transcarbamylase UCEs Urea cycle enzymes OPEN 1SeLiver Group, Instituto de Biomedicina de Sevilla/Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Avda. Manuel Siurot s/n, 41013 Sevilla, Spain. 2Hepatic and Digestive Diseases Networking Biomedical Research Centre (CIBERehd), Madrid, Spain. 3UCM Digestive Diseases Unit, Hospital Universitario Virgen del Rocío, Avda. Manuel Siurot s/n, 41013 Sevilla, Spain. 4Andalusian Center of Molecular Biology and Regenerative Medicine-CABIMER, University Pablo Olavide-University of Seville-CSIC, Sevilla, Spain. 5Biomedical Research Network On Diabetes and Related Metabolic Diseases-CIBERDEM, Instituto de Salud Carlos III, Madrid, Spain. 6Digestive Diseases Unit, Hospital Virgen de Valme, Sevilla, Spain. 7Pathology Unit, Hospital Virgen de Valme, Sevilla, Spain. 8Digestive Diseases Unit, Hospital de Valladolid, Valladolid, Spain. 9Unit for the Clinical Management Gastroenterology, Instituto de Investigación Biomédica de Málaga-IBIMA, Hospital Universitario Virgen de La Victoria, Universidad de Málaga, Málaga, Spain. 10Liver Research Unit, Hospital Universitario Santa Cristina, Instituto de Investigación Sanitaria Princesa, Madrid, Spain. 11Liver Failure Group, Institute for Liver and Digestive Health, Royal Free Hospital, London, UK. 12These authors contributed equally: Rocío Gallego-Durán and Javier Ampuero. 13These authors jointly supervised this work: Rocío Gallego-Durán and Manuel Romero-Gómez. *email: [email protected]; [email protected] 2 Vol:.(1234567890) Scientific Reports | (2022) 12:3418 | https://doi.org/10.1038/s41598-022-06614-9 www.nature.com/scientificreports/ GS Glutamine synthetase BMI Body mass index rpm Revolutions per minute HDL-c High-density lipoprotein cholesterol LDL-c Low-density lipoprotein cholesterol NAS Score NAFLD Activity Score IHC Immunohistochemistry EVOO Extra virgin olive oil RIN RNA Integrity number ASS Argininosuccinate synthetase ARG Arginase-1 Non-alcoholic fatty liver disease (NAFLD) is currently considered the hepatic manifestation of the metabolic syndrome1. Non-alcoholic steatohepatitis (NASH) is a condition histologically characterized by inflammation, ballooning degeneration and may be or not accompanied by liver fibrosis, together with the typical fatty hepatocytes, in absence of relevant alcohol consumption. This entity has surfaced as a major health problem worldwide, due to its high global prevalence, reaching 25%, and not negligible hepatocellular carcinoma progression and mortality rates2. Liver fibrosis is the strongest predictor of mortality in NAFLD patients3, and it has been recently reported that significant fibrosis forecasts new onset diabetes mellitus and arterial hypertension in NASH patients4. The global outbreak of obesity fuels metabolic disarrangements and predicts a higher socioeconomic burden of this disease2. Nitrogen is essential for life-maintenance, and urea cycle is critical for the clearance of the excess of nitrogen that otherwise could be primarily accumulated as ammonia. Thus, urea cycle enzymes (UCEs) have been matter of study for a long time. Previous studies in different NAFLD animal models and clinical settings have demonstrated that messenger RNA from UCEs seems to be compromised, even resulting in a functional reduction in ureagenesis capacity5. Further, a recent study has shown that inhibition of urea-cycle flux generating carbamoyl phosphate synthetase 1 (CPS1) correlated with a loss of functionality for ureagenesis in NASH, whereas glutamine synthetase (GS) was increased6. Indeed, the use of ammonia scavenger ornithine phenylacetate prevented from hepatocyte death and diminished invitro and invivo liver fibrosis development7. NASH might produce a reversible reduction in UCEs expression and function, resulting in an impaired ammonia clearance and might generate hyperammonemia through the activation of hepatic stellate cells8. Moreover, rs1047891 variant located in CPS1 gene has been previously associated with functional consequences of the downstream availability of L-ornithine, precursor of nitric acid9. Therefore, it has been linked to several clinical conditions, such as differential serum lipid profiles in certain ethnic groups10 and A-allele carriers seem to be more prone to develop hyperammonemia associated to valproic acid intake for the treatment of epilepsy11. The lack of effective therapies to treat NAFLD and the high rate failure in randomized clinical trials, lead us to think that there is still unravelled pathogenic mechanisms, such as ammonia, that could influence disease progression. In this scenario, considering that NAFLD molecular mechanisms are not properly unravelled yet, there is a shortage of effective therapeutic treatments apart from lifestyle interventions12. Therefore, the aim of this study was to evaluate changes in the urea cycle enzymes in patients with NAFLD and in two preclinical animal models mimicking this disease. Results UCEs gene and protein expression results. Snap frozen and paraffin-embedded liver tissue was available for both gene expression and immunohistochemistry (IHC) purposes from 17 NAFLD biopsy-proven patients. To avoid biases, patients were selected as bland steatosis (n = 10), defined as steatosis presence, nor steatohepatitis or fibrosis, and NASH with fibrosis (n = 7). Seven healthy controls were also included for gene expression analyses. Main features of the cohort included in both IHC and gene expression analyses are shown in Table1. Briefly, mean age was 44.1 ± 13.1years, 47% (8/17) were male, and 41.2% (7/17) displayed significant fibrosis in liver biopsy. Regarding the gene expression, out of the five enzymes analysed, the two mitochondrial-located enzymes showed a significant downregulation in bland steatosis in comparison with healthy controls: CPS1 was 0.32 fold in steatosis [CI95% 0.02–0.63] vs. healthy controls (onefold, [CI95% 0.75–1.32]) (p = 0.0003), but not statistically significant in patients with NASH-fibrosis when compared to bland steatosis (p = ns) (Fig.1A). Additionally, there was a downregulation in OTC-1 in bland steatosis (0.28 fold [CI95% 0.08–0.48]) vs. healthy controls (onefold [CI95% 0.84–1.18]) (p < 0.0001), as well as in NASH-fibrosis patients when also compared to healthy individuals (0.50 fold [CI95% 0.06–0.94]) (p = 0.01) (Fig.1B). No statistical differences were found between both bland steatosis and NASH-fibrosis patients (p = ns). CPS1 expression by RNA-seq was found to be diminished in both bland steatosis (0.81 fold; FDR: 0.156) and NASH-fibrosis (0.78 fold; FDR: 0.052) when compared to healthy controls, but only reached statistical significance in the latter (Supplementary Fig.1A). Further, OTC1 was found to be increased in NASH-fibrosis versus healthy controls (0.31 fold, FDR: 0.004) (Supplementary Fig.1B) Further, in the second subset of patients, CPS1 expression by RNA-seq was found to be diminished in NASH patients without fibrosis when compared to bland steatosis (0.44 fold; FDR: 0.0005) and healthy controls (0.44 fold; FDR: 0.0003) (Supplementary Fig.1C). Finally, no differences in OTC1 expression were found when comparing these patients (Supplementary Fig.1D). On the other hand, in the immunohistochemistry staining, CPS1 was also found to be downregulated in NASH-fibrosis patients versus bland steatosis (6.6 × 105 ± 7.7 × 105 vs. 1.8 × 106 ± 2.2 × 105 stained pixels, p = 0.003) (Fig.2A). Further, CPS1 was found to be increased in bland steatosis patients versus healthy controls 3 Vol.:(0123456789) Scientific Reports | (2022) 12:3418 | https://doi.org/10.1038/s41598-022-06614-9 www.nature.com/scientificreports/ (1.8 × 106 ± 2.2 × 105 vs. 8.8 × 105 ± 3.8 × 105 stained pixels, p = 0.002). The qualitative scale showed that 60% (3/5) of patients with NASH-fibrosis were low positive for CPS1 and none of the simple steatosis patients were found to be low positive (p = 0.01). Further, OTC1 levels were not found to be statistically significant in patients with NASH versus steatosis simple patients or versus healthy controls (1.9 × 105 ± 1.4 × 104 vs. 1.6 × 105 ± 1.7 × 105 or vs. 2.8 × 105 ± 1.1 × 105 positive pixels, p = ns) (Fig.2C). Finally, glutamine synthetase (GS) was found to be upregulated in simple steatosis versus healthy controls (1.7 × 106 ± 6.9 × 105 vs. 4.1 × 105 ± 3.8 × 105 positive pixels; p = 0.002). These levels were found to be downregulated when compared NASH-fibrosis patients versus steatosis simple (1.0 × 106 ± 4.9 × 105; p = 0.05) (Fig.2B). The qualitative scale showed that 90% (9/10) of patients with simple steatosis were categorized as highly positive staining and 50% (3/6) of the NASH patients (p = ns), and none of the healthy controls were found to be highly positive (p = 0.001 when compared with simple steatosis patients). All the immunohistochemistry results are shown in Fig.2D. Findings in rs10478941 variant from CPS1. Main features of this cohort, composed by 382 biopsyproven NAFLD patients according to genotypes distribution are provided in Supplementary Table1. Liver fibrosis was distributed as follows: F0 58.4% (223/382), F1 22.3% (85/382), F2 7.9% (30/382), F3 8.6% (33/382) and cirrhosis 2.9% (11/382). In this cohort, 8.6% (33/382) of patients bore AA genotype from CPS1 rs10478941 SNP, 37.2% (142/382) AC genotype and 54.2% (207/382) CC genotype. None of the patients exhibiting AA genotype suffered from liver cirrhosis, and just 12.1% (4/33) showed significant fibrosis in liver biopsy (Supplementary Fig.1). Further, 64.2% (113/175) of patients bearing A-allele from CPS1 variant did not display hepatic fibrosis Table 1. Clinical and analytical characteristics of the study cohort employed for immunohistochemistry and gene expression analyses. BMI body mass index, T2DM type 2 diabetes mellitus, AST aspartate aminotransferase, ALT alanine aminotransferase, GGT gamma glutamil transferase. Variable Overall cohort (n = 17) Age, years 44.1 ± 13.1 Sex distribution (male), % 47.1% (8/17) BMI (kg/m2)29.04 ± 5.02 T2DM, % 29.4% (5/17) Arterial hypertension, % 29.4% (5/17) Dyslipidemia, % 52.9% (9/17) AST (IU/mL) 41 ± 23 ALT (IU/mL) 55 ± 34 GGT (IU/mL) 95 ± 105 Glucose (mg/dL) 100 ± 41 Insulin (microUI/mL) 13 ± 9 Triglycerides (mg/dL) 119 ± 87 Total cholesterol (mg/dL) 190 ± 46 Albumin (mg/dL) 4252 ± 310 Platelet count, ^109207 ± 56 Significant fibrosis (F2–F4), % 41.2% (7/17) Figure1. Relative expression results from CPS1 (A) and OTC1 (B) in liver biopsies from healthy controls, bland steatosis and NASH-fibrosis patients. CPS1: carbamoyl phosphate synthetase-1; OTC1: ornithine transcarbamylase. 4 Vol:.(1234567890) Scientific Reports | (2022) 12:3418 | https://doi.org/10.1038/s41598-022-06614-9 www.nature.com/scientificreports/ in liver biopsy vs. 53.4% (110/207) of patients without A-allele and without fibrosis (p = 0.03). Nevertheless, there was no association between A-allele and NASH when compared to simple steatosis patients (57.3% (71/124) vs. 52.2% (117/224), p = ns). The following variables were found to be associated with liver fibrosis presence in univariate analysis: age (p = 0.005), BMI higher than 30kg/m2 (p = 0.001), type 2 diabetes mellitus (p ≤ 0.001), AST (p ≤ 0.001), ALT (p ≤ 0.001), GGT (p = 0.023), triglycerides (p = 0.008) and A-allele from rs1047891 SNP (p = 0.036). In multivariate analysis, age (O.R. 1.02; 95% CI 1.02–1.05; p = 0.003), BMI > 30kg/m2 (O.R. 2.11; 95% CI 1.31–3.41; p = 0.002), T2DM (O.R. 2.66; 95% CI 1.39–5.09; p = 0.003), ALT (O.R. 1.01; 95%CI 1.01–1.03, p ≤ 0.001) and A-allele from CPS1 (O.R. 0.62 (95% CI 0.39–0.99; p = 0.047) were independently associated with presence of liver fibrosis (Table2). Figure2. CPS1 (A), GS (B) and OTC1 (C) staining according to different NAFLD stages. (D) represents immunohistochemistry findings in human liver biopsies. CPS1: carbamoyl phosphate synthetase-1; GS: glutamine synthetase; OTC1: ornithine transcarbamylase. 5 Vol.:(0123456789) Scientific Reports | (2022) 12:3418 | https://doi.org/10.1038/s41598-022-06614-9 www.nature.com/scientificreports/ Finally, ammonia was stained in liver sections of twenty-five NAFLD biopsy-proven patients. Main features of this cohort are described in Supplementary Table1. Regarding ammonia stain, AA-carriers from rs1047891 SNP showed a significantly reduced levels of hepatic ammonia than AC or CC-carriers (6.7 ± 2.5 vs. 25.4 ± 7.10 vs. 20.8 ± 5.5 percentage of area stained; p = 0.002 and p = 0.027 respectively). No significant differences between AC and CC-carriers were found (p = ns) (Supplementary Fig.1). UCEs expression in animal models. CDA‑HFD model (Choline‑deficient high‑fat amino‑acid controlled diet with 0.1% added methionine). First, liver fibrosis was found to be significantly higher in CDA-HFD groups when compared with controls, at both 16w (19.4 ± 1.76% vs. 0.71 ± 0.05% area threshold, p = 0.005) and 28w (14.70 ± 2.11% vs. 0.66 ± 0.45% area threshold, p = 0.023). On the other hand, after 4weeks with reversion diet, liver fibrosis was found to be diminished but still significant when compared to the CDA-HFD 28w group (5.40 ± 0.52 vs. 14.70 ± 2.11 area threshold; p = 0.003) (Fig.3). A significant decrease in CPS1 expression evaluated by IHC was observed after 28weeks of treatment when compared to control group at 16weeks (2.5 × 106 ± 2.6 × 105 vs. 5.59 × 106 ± 9.2 × 105 positive pixels, p = 0.004). Further, CPS1 protein expression was found to be decreased when compare both CDA-HFD, 28 and 16weeks (2.5 × 106 ± 2.6 × 105 vs. 4.24 × 106 ± 5.7 × 105 positive pixels; p = 0.008). Finally, in the reversion group, it was observed that just after 4weeks of administration of control diet, CPS1 expression was found to be reverted up to control levels (5.3 × 106 ± 4.1 × 105 positive pixels, p = ns) (Fig.4). LDLr knockout animals (knockout for LDL receptor animals). Hepatic fibrosis was found to be similar in all groups analysed (0.94 ± 0.05% area threshold for Lard and 1.91 ± 0.28% area threshold for EVOO group) when compared with controls (1.06 ± 0.17% area threshold) (p = ns) (Supplementary Fig.4). After 24weeks of diet, it was observed a decrease in CPS1 expression when compared to control group, regardless of the type of HFD, lard (8.41 × 106 ± 8.91 × 104 vs. 7.64 × 106 ± 2.6 × 105 stained pixels; p = 0.026) or extra virgin olive oil (7.51 × 106 ± 3.45 × 105 stained pixels, p = 0.026). There were no statistical differences among both groups (p = ns) (Fig.5). Discussion This study has demonstrated that both liver fibrosis and steatosis are associated with a reduction in gene and protein expression patterns of mitochondrial urea cycle enzymes. Our results show that CPS1, the first step limiting enzyme of the cycle, diminishes in parallel with the increase of the liver injury after a deleterious diet, such as CDA-HFD, and this CPS1 impairment could be reversed when the main trigger disappears. We have also confirmed an upregulation in GS protein expression, an ammonia scavenger, which might exert a compensatory effect to avoid ammonia gathering when urea cycle enzymes are impaired. Moreover, we have revealed the potential role of a functional variant previously linked with abnormal ammonia rates in hepatic fibrosis development, leading to a more aggressive phenotype and increased hepatic ammonia levels. This study might open novel hypotheses positioning CPS1 as a novel therapeutic target. The liver is the main organ responsible for removing nitrogen from blood to synthetize urea in order to be excreted from the body. When ammonia is not successfully cleared from the organism, hyperammonemia could promote different outcomes, from the liver to the brain. Ammonia seems to promote NAFLD progression mainly through the activation of hepatic stellate cells7. NASH is considered the progressive entity of NAFLD. Among the intra-organ factors that trigger NASH development it has been described lipotoxicity, innate immune response, apoptosis and cell death, endoplasmic reticulum stress and mitochondrial dysfunction13,14. Moreover, ammonia might generate an increased yield of reactive oxygen species that could eventually foster liver injury by mediating inflammatory processes, cell death and fibrosis, hallmarks of NASH15. Our findings continue along the same path, due to the main enzymes that found to be dysregulated are both located in this cellular compartment. Table 2. Multivariate analysis of factors independently associated with liver fibrosis. BMI body mass index, T2DM type 2 diabetes mellitus, AST aspartate aminotransferase, ALT alanine aminotransferase, GGT gamma glutamil transferase, CPS1 carbamoyl phosphate synthetase-1. Significant values are in Italics. Variable Univariate analysis Multivariate analysis Non-fibrotic (n = 224) Fibrotic (n = 159) p-value OR [95%CI] p-value Sex distribution (Males, %) 43.8% 39.6% 0.413 BMI > 30kg/m231.7% 48.5% 0.001 2.11[1.31–3.41] 0.002 Age (years) 44.8 ± 12.9 48.7 ± 13.1 0.005 1.02[1.01–1.05] 0.003 T2DM (yes) 9% 28.3% ≤ 0.001 2.66[1.39–5.09] 0.003 AST (IU/mL) 29.7 ± 21.7 39.3 ± 24.1 ≤ 0.001 ALT (IU/mL) 41 + 33.1 63.1 ± 44 ≤ 0.001 1.01[1.01–1.03] ≤ 0.001 GGT (IU/mL) 70.5 ± 79.4 90.5 ± 85.3 0.023 Triglycerides (mg/dL) 123.4 ± 63 157.2 ± 92.5 0.008 Total cholesterol (mg/dL) 190.8 ± 45.6 191.4 ± 43.1 0.941 Albumin (mg/dL) 4279 ± 329.4 4424.1 ± 463.5 0.076 CPS1 A-allele (AA/AC vs CC) 50.4% (113/224) 39.6% (63/159) 0.036 0.62 [0.39–0.99] 0.047 Citation: Gil-Gómez, A.; Rojas, Á.; García-Lozano, M.R.; Muñoz- Hernández, R.; Gallego-Durán, R.; Maya-Miles, D.; Montero-Vallejo, R.; Gato, S.; Gallego, J.; Francés, R.; et al. Impact of a Loss-of-Function Variant in HSD17B13 on Hepatic Decompensation and Mortality in Cirrhotic Patients. Int. J. Mol. Sci. 2022,23, 11840. https://doi.org/ 10.3390/ijms231911840 Academic Editors: Stefania Grimaudo and Rosaria Maria Pipitone Received: 7 September 2022 Accepted: 4 October 2022 Published: 6 October 2022 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). International Journal of Molecular Sciences Article Impact of a Loss-of-Function Variant in HSD17B13 on Hepatic Decompensation and Mortality in Cirrhotic Patients Antonio Gil-Gómez 1,2 ,Ángela Rojas 1,2 , María R. García-Lozano 1,2 , Rocío Muñoz-Hernández 1,2 , Rocío Gallego-Durán1,2, Douglas Maya-Miles 1,2, Rocío Montero-Vallejo 1,2, Sheila Gato 1,2 , Javier Gallego 1, Rubén Francés2,3,4 , Germán Soriano 2,5 , Javier Ampuero 1,2,6 and Manuel Romero-Gómez 1,2,6,* 1SeLiver Group, Institute of Biomedicine of Seville (IBiS), Virgen del Rocio University Hospital, CSIC, University of Seville, 41013 Seville, Spain 2CIBERehd, Instituto de Salud Carlos III, 28029 Madrid, Spain 3Clinical Medicine Department and IDIBE, Miguel Hernández University, 03550 San Juan de Alicante, Spain 4IIS ISABIAL, Hospital General Universitario de Alicante, 03010 Alicante, Spain 5Department of Gastroenterology, Hospital de la Santa Creu i Sant Pau, 08025 Barcelona, Spain 6UCM Digestive Diseases, Virgen del Rocío University Hospital, 41013 Seville, Spain *Correspondence: mromer[email protected] Abstract: A common splice variant in HSD17B13 (rs72613567:TA) was recently found to be associated with a reduced risk of developing chronic liver disease in NAFLD patients and a reduced risk of progression to advanced fibrosis and cirrhosis. In this study, we aimed to evaluate the prognosis of cirrhotic patients harboring this variant. We performed a retrospective analysis on 483 prospectively recruited patients from four different hospitals in Spain, followed-up for at least 5 years. We collected clinical, demographic, and biochemical data, and we performed a genotyping analysis for common variants previously associated with liver disease risk (HSD17B13 rs72613567:TA and PNPLA3 rs738409). Patients homozygous for the TA allele showed a higher MELD score (p= 0.047), Child– Turcotte–Pugh score (p= 0.014), and INR levels ( p= 0.046 ), as well as decreased albumin ( p= 0.004 ) at baseline. After multivariate analysis, patients with the “protective” variant indeed had an increased risk of hepatic decompensation [aHR 2.37 ( 1.09–5.06 ); p= 0.029 ] and liver-related mortality [aHR 2.32 (1.20–4.46); p= 0.012]. Specifically, these patients had an increased risk of developing ascites (Log-R 11.6; p< 0.001), hepatic encephalopathy (Log-R 10.2; p< 0.01 ), and higher mortality (Log-R 14.1; p< 0.001 ) at 5 years of follow-up. Interactions with the etiology of the cirrhosis and with the variant rs738409 in PNPLA3 are also described. These findings suggest that the variant rs72613567:TA in HSD17B13 has no protective effect, but indeed increases the risk of decompensation and death in patients with advanced chronic liver disease. Keywords: cirrhosis; polymorphism; PNPLA3;HSD17B13; NAFLD; fibrosis; ascites; hepatic encephalopathy; SNP; hepatic decompensation 1. Introduction Chronic liver disease progresses from fibrosis to cirrhosis, and ultimately to hepatic failure and hepatocellular carcinoma (HCC), being prominent causes of illness and death globally. The most common etiologies of cirrhosis are alcohol consumption, chronic hepatitis C or B infection, and non-alcoholic fatty liver disease (NAFLD), whose prevalence is increasing in parallel with that of metabolic syndrome [ 1 , 2 ]. Cirrhosis progression results in portal hypertension and hepatic decompensation, both of which substantially increase the risk of mortality, accounting for approximately 2 million deaths worldwide [ 3 ]. Notably, liver disease progression shows considerable interindividual variability; it has been estimated that host genetic factors could explain approximately half of this variability, contributing to the progression of cirrhosis and predisposing patients to the development of liver-related events [4,5]. Int. J. Mol. Sci. 2022,23, 11840. https://doi.org/10.3390/ijms231911840 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2022,23, 11840 2 of 11 Genome-wide association studies have identified sequence variations with robust effects on multiple stages of the disorder. The most validated so far is the variant p.I148M in 1-acylglycerol-3-phosphate O-acyltransferase (PNPLA3), associated with higher triglyceride levels, as well as with an increased risk of non-alcoholic steatohepatitis, cirrhosis decompensation, HCC development, and mortality [ 6 – 9 ]. Recently, an indel variant ( T > TA variant, rs72613567) in the 17-beta-hydroxysteroid dehydrogenase 13 (HSD17B13) gene, leading to the production of a truncated protein with reduced activity, has been associated with decreased aminotransferase levels and a reduction of liver injury in the setting of a fatty liver, mitigating the risk conferred by the carriage of the PNPLA3 variant [ 10 , 11 ]. Additionally, this variant has been associated with lower odds of developing cirrhosis and HCC among alcohol misusers [ 12 , 13 ]. However, the physiological role of this protein and the mechanisms underlying these associations have yet to be elucidated, since much of the genetic risk of chronic liver disease remains unexplained. The goal of this study was to evaluate the clinical impact of the genetic variant rs72613567:TA in HSD17B13 on the prognosis of cirrhotic patients and to explore possible risk interactions with the underlying etiology or the genotype of PNPLA3 rs738409. 2. Results 2.1. Baseline Characteristics and Genotyping The clinical and demographic characteristics of the patients are shown in Table 1. The majority of the patients were male (68.7%), with a mean age of 59.4 ± 11.9 years old. The most frequent etiology was alcohol (48.7%), followed by viral hepatitis (35.8% C and 6.2% B), with the other 9.3% being NAFLD, primary biliary cirrhosis, or autoimmunerelated cirrhosis. At the time of inclusion, liver function was CTP score A, with a Model for End-stage Liver Disease (MELD) score of 12.2 ±5.1, in 48.9% of patients. Table 1. Comparison of clinical and demographic characteristics of the patients according to the genotype of HSD17B13 rs72613567. Variable Total (n= 483) HSD17B13 T/T or T/TA (n= 427) TA/TA (n= 56) p-Value Age ±SD, years 59.4 ±11.9 59.2 ±11.5 60.8 ±14.8 0.448 Male sex 68.7% (332/483) 68.1% (291/427) 73.2% (41/56) 0.442 CTP 7.0 ±1.9 6.9 ±1.9 7.6 ±1.8 0.014 Etiology 0.800 Alcohol 48.7% (235/483) 49.2% (210/427) 44.6% (25/56) Viral 42.0% (203/483) 41.7% (178/427) 44.6% (25/56) Other 9.3% (45/483) 9.1% (39/427) 10.7% (6/56) MELD ±SD 12.2 ±5.1 12.0 ±5.1 13.4 ±5.0 0.047 Albumin ±SD (g/L) 35.0 ±7.9 35.4 ±7.8 32.2 ±7.6 0.004 Bilirubin ±SD (mg/dL) 2.14 ±2.17 2.13 ±2.14 2.24 ±2.46 0.727 Platelets ±SD (×109/L) 121.3 ±70.0 120.8 ±64.9 133.1 ±100.9 0.340 Creatinine ±SD (mg/dL) 0.98 ±0.45 0.98 ±0.46 0.96 ±0.36 0.762 AST ±SD (IU/L) 60.5 ±57.8 61.5 ±60.4 53.3 ±31.9 0.315 ALT ±SD (IU/L) 46.0 ±60.5 47.3 ±63.9 36.0 ±20.6 0.188 INR ±SD 1.32 ±0.32 1.32 ±0.32 1.40 ±0.32 0.046 Leukocytes ± SD (cells/ × 10 9 /L) 6.96 ±5.99 7.00 ±6.19 6.63 ±4.12 0.681 For the genotyping of rs72613567, we determined that 59.8% of patients were homozygous for the T allele, with 11.6% harboring the TA/TA variant (Figure 1A), resulting in a MAF (TA allele) of 0.255, comparable with that of previous studies (Abul-Husn et al., MAF = 0.242 [ 10 ]; Scheiner et al., MAF = 0.219) [ 14 ]. Nonetheless, we observed a deviation from the Hardy–Weinberg equilibrium (p< 0.05) for the variant in this cohort, which could indicate that the risk of disease per se is influenced by the polymorphism. Int. J. Mol. Sci. 2022,23, 11840 3 of 11 Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 3 of 11 For the genotyping of rs72613567, we determined that 59.8% of patients were homozygous for the T allele, with 11.6% harboring the TA/TA variant (Figure 1A), resulting in a MAF (TA allele) of 0.255, comparable with that of previous studies (Abul-Husn et al., MAF = 0.242 [10]; Scheiner et al., MAF = 0.219) [14]. Nonetheless, we observed a deviation from the Hardy–Weinberg equilibrium (p < 0.05) for the variant in this cohort, which could indicate that the risk of disease per se is influenced by the polymorphism. Figure 1. (A) Genotype frequency for rs72613567. (B) Comparison of minor allele frequency (MAF) across studies and in regards to liver function. (C) Alanine aminotransferase levels in patients with different genotypes (ANOVA). (D) Distribution of genotypes according to liver function measured by CTP score. Compared to a control European population (publicly available from https://gnomAD.broadinsitute.org (accessed on 22 September 2022); n = 55,616), as in [13], the frequency of TA allele carriers was lower in cirrhotic patients (40.2% vs. 46.2%; p = 0.008), as shown in Figure 1B. This suggests, in agreement with previous reports, a protective effect of the variant against cirrhosis compared to healthy subjects. This effect was maintained for ALD (OR: 0.68; CI 95% 0.53–0.86; p = 0.001), but not for viral etiology (OR: 0.95; 95% CI, 0.72-1.25; p = 0.725). In addition, we observed the expected trend towards decreased ALT levels in patients with the TA allele (Figure 1C). Curiously, the frequency of patients with worse liver function (CTP score B/C) was higher among those carrying the TA/TA genotype (Figure 1D). Furthermore, we also found that MELD scores (12.0 ± 5.1 vs. 13.4 ± 5.0; p < 0.05) and International Normalized Ratio (INR) levels (1.32 ± 0.32 vs. 1.40 ± 0.32; p < 0.05) were increased and albumin levels were decreased (35.4 ± 7.8 vs. 32.3 ± 7.6; p < 0.05) in rs72613567 TA/TA patients (Table 1). 2.2. Impact of rs72613567 Genotype on Mortality and Hepatic Decompensation The patients were followed for a median of 36.8 ± 23.9 months. During that period, 47 patients (9.7%) underwent liver transplantation. Of the remaining patients, 111 (25.5%) died within five years, 92 of which (82.3% of the deaths) were related to their underlying liver disease. Surprisingly, carriers of the TA/TA genotype not only were not protected, but indeed had an increased risk of liver transplantation [HR 1.56 (95% CI 0.70–3.50); p = Child-Pugh A Child-Pugh B/C 0 20 40 60 80 % of patients T/T or T/TA TA/TA p = 0.001 HSD17B13 genotype T/T T/TA TA/TA 30 40 50 60 U/L ALT T/T T/TA TA/TA HSD17B13 genotype p = 0.335 Total=483 T/T (59.83%) T/TA (28.57%) TA/TA (11.59%) AB CD Cirrhotic patients European population 0 20 40 60 80 Frequency T/T T/TA or TA/TA (n=483) (gnomAD n=55.616) p = 0.008 Figure 1. ( A ) Genotype frequency for rs72613567. ( B ) Comparison of minor allele frequency (MAF) across studies and in regards to liver function. ( C ) Alanine aminotransferase levels in patients with different genotypes (ANOVA). ( D ) Distribution of genotypes according to liver function measured by CTP score. Compared to a control European population (publicly available from https://gnomAD. broadinsitute.org (accessed on 22 September 2022); n = 55,616), as in [ 13 ], the frequency of TA allele carriers was lower in cirrhotic patients (40.2% vs. 46.2%; p = 0.008), as shown in Figure 1B. This suggests, in agreement with previous reports, a protective effect of the variant against cirrhosis compared to healthy subjects. This effect was maintained for ALD (OR: 0.68; CI 95% 0.53–0.86; p = 0.001), but not for viral etiology (OR: 0.95; 95% CI, 0.72–1.25; p = 0.725). In addition, we observed the expected trend towards decreased ALT levels in patients with the TA allele (Figure 1C). Curiously, the frequency of patients with worse liver function (CTP score B/C) was higher among those carrying the TA/TA genotype (Figure 1D). Furthermore, we also found that MELD scores (12.0 ± 5.1 vs. 13.4 ± 5.0; p < 0.05) and International Normalized Ratio (INR) levels (1.32 ±0.32 vs. 1.40 ±0.32; p < 0.05) were increased and albumin levels were decreased (35.4 ±7.8 vs. 32.3 ±7.6; p < 0.05) in rs72613567 TA/TA patients (Table 1). 2.2. Impact of rs72613567 Genotype on Mortality and Hepatic Decompensation The patients were followed for a median of 36.8 ± 23.9 months. During that period, 47 patients (9.7%) underwent liver transplantation. Of the remaining patients, 111 (25.5%) died within five years, 92 of which (82.3% of the deaths) were related to their underlying liver disease. Surprisingly, carriers of the TA/TA genotype not only were not protected, but indeed had an increased risk of liver transplantation [HR 1.56 (95% CI 0.70–3.50); p= 0.275 ] and death [1.96 (1.22–3.15); p = 0.005], especially if taking into account liverrelated deaths [2.52 (1.53–4.14); p = 0.001]. The cumulative incidence of liver-related death or transplantation is shown in Figure 2A. The resulting regression models, after adjusting for age, sex, liver function (MELD score), and factors influencing the progression of cirrhosis (i.e., alcohol intake or active viral infection), confirmed this trend (Table 2). Int. J. Mol. Sci. 2022,23, 11840 4 of 11 Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 4 of 11 0.275] and death [1.96 (1.22–3.15); p = 0.005], especially if taking into account liver-related deaths [2.52 (1.53–4.14); p = 0.001]. The cumulative incidence of liver-related death or transplantation is shown in Figure 2A. The resulting regression models, after adjusting for age, sex, liver function (MELD score), and factors influencing the progression of cirrhosis (i.e., alcohol intake or active viral infection), confirmed this trend (Table 2). Figure 2. (A) Mortality with liver transplantation as a competing event. (B) Frequency of patients with an increasing number of cirrhosis complications. Cumulative incidence of (C) ascites and (D) hepatic encephalopathy development, according to the HSD17B13 rs72613567 genotype. Table 2. Multivariate Cox regression analyses on the impact of the genotype of HSD17B13 rs72613567 on liver transplantation, mortality, and hepatic decompensation. Liver Transplantation Mortality Liver-Related Death Any (Further)$$$$$$Decompensation aHR$$$$$$(95 % CI) p-Value aHR$$$$$$(95 % CI) p-Value aHR$$$$$$(95 % CI) p-Value aHR$$$$$$(95 % CI) p-Value Sex 1.96$$$$$$(0.8 7–4.32) 0.097 1.17$$$$$$(0.7 1–1.93) 0.531 1.27$$$$$$(0.7 2–2.21) 0.404 1.04$$$$$$(0.6 4–1.70) 0.880 Age 0.98$$$$$$(0.9 5–1.00) 0.089 1.05$$$$$$(1.0 3–1.07) 0.001 1.05$$$$$$(1.0 3–1.08) 0.001 1.04$$$$$$(1.0 2–1.06) 0.001 MELD 1.09$$$$$$(1.0 4–1.15) 0.001 1.09$$$$$$(1.0 5–1.14) 0.001 1.11$$$$$$(1.0 7–1.16) 0.001 1.29$$$$$$(1.2 1–1.38) 0.001 Etiological factors 1.52$$$$$$(0.8 1–2.84) 0.197 1.52$$$$$$(0.9 7–2.39) 0.069 1.39$$$$$$(0.8 4–2.30) 0.204 1.99$$$$$$(1.2 9–3.07) 0.002 rs72613567 TA/TA genotype 1.21$$$$$$(0.4 9–2.93) 0.673 1.67$$$$$$(0.9 1–3.18) 0.089 2.32$$$$$$(1.2 0–4.46) 0.012 2.37$$$$$$(1.0 9–5.06) 0.029 MELD: Model for End-Stage Liver Disease; etiological factors: relapse in alcohol intake or active viral infection. 012 24 36 48 60 0 50 100 Time to liver transplantation or liver-related death (months) Probability Liver transplantation rs72613567 T/T-T/TA Liver transplantation rs72613567 TA/TA Liver-related death rs72613567 T/T-T/TA Liver-related death rs72613567 TA/TAlog-Rank=14.1 T/T or T/TA TA/TA 0 50 100 0 1 2 3 4 5 % of patients N° of complications p = 0.079 20.3 9.3 24.3 20.4 26.7 35.2 17.9 29.6 012 24 36 48 60 0 50 100 Time (months) Hepatic encephalopathy development (%) HE rs72613567 T/T-T/TA HE rs72613567 TA/TAlog-Rank=10.2 012 24 36 48 60 0 50 100 Time (months) Ascites development (%) Ascites rs72613567 T/T-T/TA Ascites rs72613567 TA/TAlog-Rank=11.6 A B C D Figure 2. ( A ) Mortality with liver transplantation as a competing event. ( B ) Frequency of patients with an increasing number of cirrhosis complications. Cumulative incidence of ( C ) ascites and (D) hepatic encephalopathy development, according to the HSD17B13 rs72613567 genotype. Table 2. Multivariate Cox regression analyses on the impact of the genotype of HSD17B13 rs72613567 on liver transplantation, mortality, and hepatic decompensation. Liver Transplantation Mortality Liver-Related Death Any (Further) Decompensation aHR (95% CI) p-Value aHR (95% CI) p-Value aHR (95% CI) p-Value aHR (95% CI) p-Value Sex 1.96 (0.87–4.32) 0.097 1.17 (0.71–1.93) 0.531 1.27 (0.72–2.21) 0.404 1.04 (0.64–1.70) 0.880 Age 0.98 (0.95–1.00) 0.089 1.05 (1.03–1.07) 0.001 1.05 (1.03–1.08) 0.001 1.04 (1.02–1.06) 0.001 MELD 1.09 (1.04–1.15) 0.001 1.09 (1.05–1.14) 0.001 1.11 (1.07–1.16) 0.001 1.29 (1.21–1.38) 0.001 Etiological factors 1.52 (0.81–2.84) 0.197 1.52 (0.97–2.39) 0.069 1.39 (0.84–2.30) 0.204 1.99 (1.29–3.07) 0.002 rs72613567 TA/TA genotype 1.21 (0.49–2.93) 0.673 1.67 (0.91–3.18) 0.089 2.32 (1.20–4.46) 0.012 2.37 (1.09–5.06) 0.029 MELD: Model for End-Stage Liver Disease; etiological factors: relapse in alcohol intake or active viral infection. In terms of decompensation, patients with the variant in HSD17B13 had an increased cumulative number of complications of cirrhosis (Figure 2B). Specifically, these patients had a higher risk of developing ascites (log-Rank 11.6; p < 0.001) and hepatic encephalopathy (log-Rank 10.2; p < 0.001) (Figure 2C,D). The Kaplan–Meier curves for the cumulative incidence of other complications, such as SBP (p = 0.059), EVB (p = 0.427), or hepatorenal syndrome (p = 0.236), did not differ significantly, probably due to the reduced number of cases in the cohort. Furthermore, and in contrast to recent studies, carriers of the TA/TA genotype were not protected from the 5-year development of HCC (log-Rank 0.1; p = 0.801). 2.3. Effect of Etiology and Interactions with PNPLA3 Genotype Previous studies have adjudged HSD17B13 to have a role in lipid droplet metabolism, and, so far, most ofthe literature showsassociationswith non-viral liverdisease (ALD/NAFLD Int. J. Mol. Sci. 2022,23, 11840 5 of 11 patients). We wondered if the underlying etiology of cirrhosis could differentially affect the prognosis of patients harboring the variant rs72613567. In order to find out, we compared patients with ALD/NAFLD to those with viral etiology. Of those with steatohepatitis, only 8.2% (21/256) had NAFLD; viral etiology was comprised 148 (72.9%) patients with hepatitis C virus (HCV) and 55 (27.1%) with hepatitis B virus (HBV). As shown in Table 3, patients with viral hepatitis carrying the variant had worse liver function, measured by either Child–Turcotte–Pugh (CTP) (7.8 ± 1.7 vs. 6.9 ± 2.0; p = 0.038) or MELD (13.4 ± 4.3 vs. 11.4 ± 4.8; p = 0.065), at baseline than those with the major allele (T-). In addition, INR levels (p = 0.009) were higher, and albumin levels were decreased (p = 0.001). Although the trend was similar, there were no significant changes in baseline clinical and biochemical parameters among ALD/NAFLD patients, suggesting that the genotype may influence differently depending on the etiology. Table 3. Comparison of patients’ characteristics according to the genotype of HSD17B13, depending on the etiology of the cirrhosis. Variable ALD/NAFLD (n= 256) Viral Hepatitis (n= 203) HSD17B13 p-Value HSD17B13 p-Value T/T or T/TA TA/TA T/T or T/TA TA/TA Age ±SD, years 59.1 ±10.1 58.8 ±13.2 0.853 59.3 ±13.2 63.2 ±16.4 0.259 Male sex 76.9% (176/229) 81.4% (22/27) 0.316 59.0% (105/178) 64.0% (16/25) 0.442 CTP 6.8 ±1.9 7.4 ±1.9 0.167 6.9 ±2.0 7.8 ±1.7 0.038 MELD ±SD 12.3 ±5.2 13.8 ±5.6 0.236 11.4 ±4.8 13.2 ±4.3 0.065 Albumin ±SD (g/L) 35.8 ±7.7 34.1 ±8.4 0.309 35.1 ±8.1 29.7 ±6.2 0.001 Bilirubin ±SD (mg/dL) 2.2 ±2.37 2.18 ±1.90 0.816 1.94 ±1.73 2.33 ±3.05 0.346 Platelets ±SD (×109/L) 125.3 ±65.4 141.1 ±69.1 0.279 112.1 ±63.5 125.9 ±131.5 0.610 Creatinine ±SD (mg/dL) 0.98 ±0.49 0.98 ±0.42 0.739 0.95 ±0.40 0.95 ±0.25 0.938 AST ±SD (IU/L) 54.7 ±63.9 48.5 ±29.4 0.620 71.3 ±56.3 59.5 ±34.5 0.121 ALT ±SD (IU/L) 36.1 ±42.5 31.8 ±19.1 0.585 63.3 ±83.5 41.2 ±21.9 0.191 INR ±SD 1.34 ±0.34 1.39 ±0.31 0.555 1.27 ±0.27 1.44 ±0.32 0.009 Leukocytes ±SD (cells/×109/L) 7.11 ±4.58 8.19 ±4.87 0.328 6.93 ±8.18 5.10 ±2.64 0.293 The etiology seemed to influence the cumulative incidence of cirrhosis complications. Patients with viral hepatitis carrying the mutation had an increased risk of developing any complication at five years (87.0% vs. 57.4%; log-R 9.8; p = 0.002) than those with the major allele, but we did not observe a significant difference among patients with ALD/NAFLD, neither at one year (38.1% vs. 48.4%; p = 0.249) nor at five ( 64.5% vs. 57.0% ; p= 0.363 ) years. However, carriers of the variant had reduced 5-year transplant-free survival, independently of the etiology (63.4% vs. 80.7%; log-R 5.3; p = 0.021 for ALD/NAFLD patients) (38.2% vs. 74.7%; log-R 9.0; p = 0.003 for viral hepatitis). Alternatively, we analyzed the interactions with the variant rs738409 (PNPLA3) in a subcohort of 294 patients. This polymorphism was in agreement with the Hardy–Weinberg equilibrium (p> 0.05). The result showed that the negative effect of the rs72613567-TA variant was present among carriers of the rs738409-GG genotype, who had a greater risk of liver-related mortality [aHR:8.550 (95% CI 1.490–21.864); p < 0.001], ascites [aHR:11.048 (3.83–32.129); p < 0.001], and HCC [aHR:3.641 (1.019–13.160); p = 0.048]. By contrast, this effect was diluted in non-carriers of the rs738409-GG genotype, and none of the outcomes were found to be significantly changed, although the rs72613567-TA/TA was still not protective. The cumulative incidences of these events are shown in Figure 3A–C. Int. J. Mol. Sci. 2022,23, 11840 6 of 11 Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 6 of 11 survival, independently of the etiology (63.4% vs. 80.7%; log-R 5.3; p = 0.021 for ALD/NAFLD patients) (38.2% vs. 74.7%; log-R 9.0; p = 0.003 for viral hepatitis). Alternatively, we analyzed the interactions with the variant rs738409 (PNPLA3) in a subcohort of 294 patients. This polymorphism was in agreement with the Hardy–Weinberg equilibrium (p > 0.05). The result showed that the negative effect of the rs72613567- TA variant was present among carriers of the rs738409-GG genotype, who had a greater risk of liver-related mortality [aHR:8.550 (95% CI 1.490–21.864); p < 0.001], ascites [aHR:11.048 (3.83–-32.129); p < 0.001], and HCC [aHR:3.641 (1.019–13.160); p = 0.048]. By contrast, this effect was diluted in non-carriers of the rs738409-GG genotype, and none of the outcomes were found to be significantly changed, although the rs72613567-TA/TA was still not protective. The cumulative incidences of these events are shown in Figure 3A–C. 012 24 36 48 60 0 50 100 Time (months) Probability rs72613567 T/T-T/TA rs72613567 TA/TA log-Rank=5.2; p = 0.023 aHR = 8.5; p < 0.001 012 24 36 48 60 0 50 100 Time (months) Probability rs72613567 T/T-T/TA rs72613567 TA/TA ns PNPLA3 GG PNPLA3 allele C Liver transplantation or liver-related death 012 24 36 48 60 0 50 100 Time (months) Probability rs72613567 T/T-T/TA rs72613567 TA/TA log-Rank = 16.5; p < 0.001 aHR = 11.1; p < 0.001 012 24 36 48 60 0 50 100 Time (months) Probability rs72613567 T/T-T/TA rs72613567 TA/TA ns Ascites 012 24 36 48 60 0 50 100 Time (months) Probability rs72613567 T/T-T/TA rs72613567 TA/TA log-Rank = 4.7; p = 0.031 aHR = 3.6; p = 0.048 012 24 36 48 60 0 50 100 Time (months) Probability rs72613567 T/T-T/TA rs72613567 TA/TA ns HCC A B C T/T-T/TA 116 101 89 72 59 44 TA/TA 7 5 4 3 3 1 T/T-T/TA 159 155 132 105 82 46 TA/TA 12 12 10 7 7 4 T/T-T/TA 116 98 81 67 53 37 TA/TA 7 4 2 2 2 1 T/T-T/TA 159 128 101 82 65 43 TA/TA 12 10 6 6 5 3 T/T-T/TA 116 102 88 73 63 51 TA/TA 7 5 5 4 3 1 T/T-T/TA 159 133 1 10 88 73 50 TA/TA 12 10 7 7 6 4 Figure 3. Effect of the PNPLA3 rs738409 GG genotype on the risk of ( A ) liver-related mortality, ( B ) ascites, and ( C ) hepatocellular carcinoma, according to the HSD17B13 rs72613567 genotype. aHR: adjusted hazard ratio; ns: not significant p-value. 3. Discussion Several studies have revealed an association of epidemiological factors (geographic location, race, and environment), underlying different genetic backgrounds, with the prevalence of chronic liver diseases. Some of these genetic factors correlate with disease severity, progression to HCC, and liver-related outcomes. Usually, these mutations lead to structural, functional, or expression changes in the protein they code for, influencing a metabolic pathway that is differentially activated in disease progression. However, the liver, the most metabolically active and ductile organ, can go through different metabolic stages throughout liver disease progression, for the maintenance of hepatic regeneration and homeostasis [ 15 ]. Moreover, the physiopathology usually expands beyond the liver, Int. J. Mol. Sci. 2022,23, 11840 7 of 11 and the consequences of these variants in peripheric tissues may also account for some of the risks, although they have not been properly evaluated yet. The HSD17B13 gene encodes for 17-beta-hydroxysteroid dehydrogenase 13, a retinol dehydrogenase highly expressed in the liver. Although its function is not completely clear, it is associated with lipid droplets, and it appears to be involved in the metabolism of lipids, as well as of steroid hormones, prostaglandins, xenobiotics, and retinoids [ 16 ]. The PNPLA3 protein also localizes to lipid droplets’ surfaces and plays a role in retinol metabolism; the I148M variant leads to increased fat accumulation in hepatic stellate cells, the main cellular site of retinol in the body. Retinol contributes to the regulation of gene transcription by binding nuclear hormone receptors (RAR/RXR), and the impairment induced by the variant in PNPLA3 is associated with a proinflammatory and profibrogenic phenotype [ 17 ]. Contrariwise, although the variant in HSD17B13 was linked with decreased levels of the spliced variant, it has been associated with protection against alcoholic and non-alcoholic fatty liver disease and against the progression towards fibrosis and HCC [18]. However, the findings of this work suggest that the variant rs72613567:TA in HSD17B13 has no protective effect, but indeed increases the risk of decompensation and death in patients with advanced chronic liver disease. Previously, Ting et al. discussed the possibility of losing the protective effect given by the mutation once advanced fibrosis develops [ 19 ]; Scheiner et al. [ 14 ] showed, in agreement with our results, that the variant was not protective against hepatic decompensation and mortality among patients with portal hypertension. Numerous studies reported associations between genetics and hepatic decompensations [ 20 , 21 ]; nevertheless, for the first time, as far as we are concerned, we report here that the effect of the variant might change according to the stage of the disease. Of note, most studies have been conducted on patients with ALD/NAFLD. Our cohort included few NAFLD patients (4.3% of total), and 42% were viral hepatitis patients, a prevalence that is consistent with epidemiological studies from the same period as this study [ 22 ], in which consecutive patients from outpatient clinics were included. In agreement with other studies, we showed that the underlying etiology of cirrhosis could impact the effect of the mutation on the prognosis, with a higher risk in patients with viral hepatitis. Certain differences in the pathophysiology could possibly explain this, such as the HCV lifecycle being closely linked to liver lipid metabolism [23]. In HBV patients, the mutation in PNPLA3 has been associated with treatment response [ 24 ]. Further studies examining the interactions between 17-beta-HSD13 and the steps of HBV and HCV infection are required. In any case, the variant in HSD17B13 did not show significant protection against liver-related outcomes in ALD/NAFLD patients. Interestingly, the negative impact of the HSD17B13 T>TA variant was diluted among non-carriers of the GG genotype of rs738409, suggesting an interaction between both polymorphisms, as previously reported [ 25 , 26 ]. Previous data showed a reduced effect of these variants on circulating retinol levels in patients with advanced chronic liver disease, in comparison to earlier stages [ 14 ]. Another study also observed that the impact of the variant in PNPLA3 was different depending on the stage of the disease and the etiology [ 8 ]. These findings highlight the role of lipid droplets and retinol metabolites, as well as the fact that the metabolic plasticity of the liver is needed to maintain homeostasis through different stages of liver disease, and lead us to hypothesize that genetic variants disrupting these pathways may have a singular impact in the context of end-stage liver disease. In this context, the recent literature revealed some of the metabolic abnormalities taking place in liver cells of compensated and decompensated cirrhotic patients, such as impaired oxidative phosphorylation and fatty acid hypometabolism due to the inhibition of beta-oxidation [ 27 ]. Genetics is now considered a promising approach for the development of precision medicine, with great clinical applications for improving liver disease risk stratification, especially in NAFLD/ALD patients. There is also interest in exploring the pharmaceutical inhibition of 17-beta-HSD13 in order to diminish its steatogenic effect and prevent liver damage, just as the rs72613567:TA variant does. However, future studies analyzing the mechanistic effect of the genotype of HSD17B13 (and other variants of interest) in different Int. J. Mol. Sci. 2022,23, 11840 8 of 11 stages of the disease are warranted before being considered therapeutic targets with wide clinical applications. This study has some limitations, mainly due to its retrospective nature. Nevertheless, we included a well-characterized, homogeneous cohort of cirrhotic patients that allowed us to assess the effect of the mutations in a specific setting. The majority of studies so far were population-based and limited in adjusting the results for liver function. Therefore, our sample size was modest compared to other studies, although a fair number of cases were included for some specific endpoints. Moreover, we only included individuals from our country; thus, the findings may not necessarily be extrapolated to other ethnicities, although the allele frequency observed here was similar to other studies using broader populations. Nevertheless, we acknowledge a deviation in HWE for HSD17B13; despite the possibility that it could be linked to methodological errors or non-random mating or sampling, this was taken into account, and it could indeed reflect an association with the disease. In conclusion, the results of this study suggest that the protective effect conferred by the carriage of the HSD17B13 rs72613567:TA genotype for liver disease may change in the context of cirrhosis, where it increased the risk of decompensation and death during the follow-up. A deeper understanding of the metabolic changes throughout the stages of advanced liver disease, focusing on the underlying effects of the HSD17B13 genotype, is indispensable. 4. Materials and Methods 4.1. Experimental Design Study population: A total of 483 patients suffering from liver cirrhosis of different etiologies were enrolled from 2007–2016 at the outpatient clinics of four different hospitals: Virgen del Rocio University Hospital, Valme University Hospital (Seville, Spain), Santa Creu i Sant Pau Hospital (Barcelona, Spain), and General University Hospital of Alicante (Alicante, Spain). The exclusion criteria were as follows: age <18 years; absence of signed informed consent; active alcohol consumption; previous liver transplant or transjugular intrahepatic portosystemic shunt procedure; hepatocellular carcinoma diagnosed within the first 6 months of follow-up; and the presence of comorbidities with an estimated survival of less than 6 months at the time of enrollment. Cirrhosis was diagnosed according to the clinical practice guidelines [28]. The study was performed by following ethical guidelines expressed in the Declaration of Helsinki, and in the E18 (a guideline on genomic sampling and the management of genomic data) and E6 (a guideline for good clinical practice) from the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use. Demographic, clinical, and epidemiological data were recorded at the first visit, and all data were coded in the original clinical database. Human blood samples were collected after obtaining a signed informed consent form, as approved by the ethical committee of each hospital. Definition of hepatic decompensation and transplant-free survival: Patients’ medical records were reviewed for the following events that defined (further) hepatic decompensation at baseline and during follow-up: large-volume paracentesis, spontaneous bacterial peritonitis (SBP), esophageal variceal bleeding (EVB), overt HE (West Haven grade > 1), hepatorenal syndrome, and liver-related death. Any (further) hepatic decompensation was defined by having at least one of these events occurring during follow-up. HCC was diagnosed without biopsy in the majority of the cases, because current clinical diagnostic approaches, including ultrasonography, computed tomography, or magnetic resonance imaging, were sufficient to diagnose HCC. The patients were followed-up for up to five years, being censored on the day of death, or the day of the surgery if they were transplanted, for the calculation of transplantfree survival. Liver- and non-liver-related death and liver transplants were considered competing events for hepatic decompensation and HCC development. Int. J. Mol. Sci. 2022,23, 11840 9 of 11 4.2. DNA Genotyping Genomic DNA was extracted from each sample of peripheral blood by using a MagNA Pure Compact Nucleic Acid Isolation Kit I. The concentration and purity of the DNA were evaluated by NanoDrop (ThermoFisher, Waltham, MA, USA). The genotyping of the singlenucleotide polymorphisms (SNP) of interest, PNPLA3 rs738409 (assay ID: C_7241_10) and HSD17B13 rs72613567 (custom TaqMan assay design), was performed using TaqMan SNP genotyping assays and chemistries (ThermoFisher, USA). PCRs were carried out, following the amplification protocol of FastStart Essential DNA Probes Master (Roche, Basel, Switzerland), in a QuantStudio 1 Real-Time PCR (ThermoFisher, USA). Genotype distribution was assessed for Hardy–Weinberg equilibrium by using a goodness-of-fit chi-square test; p> 0.05 was indicative of agreement with the equilibrium. 4.3. Statistical Analysis Categorical variables are presented as percentages, while continuous variables are presented as means ± standard deviation. The survival analysis was assessed by means of Cox regression for continuous variables, and by the Kaplan–Meier method, comparing subgroups by using the Mantel–Cox statistic, for categorical data. Adjusted multivariate analyses for the different outcomes were assessed by Cox regression, including age, sex, and variables with a known impact on prognosis, such as liver function measured by MELD score and etiology-dependent factors. These factors included the achievement of a sustained virological response for viral hepatitis, or the moderate relapse in alcohol intake for alcoholic hepatitis (any severe relapse leading to hospitalization was retrospectively excluded from the analysis). The strength of association between the variables and the outcomes under each model was represented by hazard ratios with 95% CIs. Missing data that were considered as missing completely at random biologically implausible values were set to missing values. Variables with <5% of missing data were imputed by multiple lineal regression, and variables with >5% of missing data were dropped. A two-sided p-value of <0.05 was set as the threshold for statistical significance. SPSS (v.25.0; SPSS, Inc., Chicago, IL, USA) and STATA (v.16, StataCorp LLC, College Station, TX, USA) statistical packages were used for analyses, and graphics were performed by using a GraphPad PRISM (v.8.0, GraphPad Software, Inc., San Diego, CA, USA). Author Contributions: A.G.-G. led the formal analysis and writing—original draft; Á.R. contributed to the data curation and supported the formal analysis; M.R.G.-L. contributed to the data curation and acquisition; R.M.-H. supported the formal analysis and methodology; R.G.-D. equally contributed to the data curation, and led the resources; D.M.-M. contributed to the data curation and methodology; R.M.-V., S.G. and J.G., equally contributed to the resources and methodology; R.F. and G.S. equally contributed to the validation; J.A. supported the conceptualization, led the supervision, and contributed to funding acquisition; M.R.-G. led the conceptualization, contributed to funding acquisition, and led the writing—review and editing. All authors have read and agreed to the published version of the manuscript. Funding: The research leading to these results received funding from Consejería de Salud de la Junta de Andalucía under grant agreements PE-0451-2018 and P20_01075 del programa PAIDI 2020 and from Instituto de Salud Carlos III, Ministerio de Ciencia y Competitividad under grant agreements PI19/01404 and PI19/00589. This research was also funded by European project LITMUS 777377 from Horizon 2020; a contract with fellowships from Consejería de Salud y Familias, Junta de Andalucía, supporting Antonio Gil-Gómez (RH-122-2020) and Angela Rojas (RH-002-2021); grant agreements PFIS FI20/00201 and IFI18/00041 from Instituto de Salud Carlos III to support Sheila Gato and Rocio Montero-Vallejo; and by support from Talento Doctores (PID Junta Andalucía, DOC_00866) to Rocío Muñoz-Hernández. Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee “CEI Virgen Macarena-Virgen del Rocío” (protocol C330020). Int. J. Mol. Sci. 2022,23, 11840 10 of 11 Informed Consent Statement: Written informed consent was obtained from all subjects involved in the study. Data Availability Statement: The data presented in this study are available upon request from the corresponding author. The data are not publicly available due to privacy reasons. Conflicts of Interest: The authors declare no conflict of interest. References 1. Cheemerla, S.; Balakrishnan, M. Global Epidemiology of Chronic Liver Disease. Clin. Liver Dis. 2021 ,17, 365–370. [CrossRef] [PubMed] 2. Sepanlou, S.G.; Safiri, S.; Bisignano, C.; Ikuta, K.S.; Merat, S.; Saberifiroozi, M.; Poustchi, H.; Tsoi, D.; Colombara, D.V.; Abdoli, A.; et al. The global, regional, and national burden of cirrhosis by cause in 195 countries and territories, 1990–2017: A systematic analysis for the Global Burden of Disease Study 2017. Lancet Gastroenterol. Hepatol. 2020,5, 245–266. [CrossRef] 3. Asrani, S.K.; Devarbhavi, H.; Eaton, J.; Kamath, P.S. Burden of liver diseases in the world. J. Hepatol. 2019 ,70, 151–171. [CrossRef] [PubMed] 4. Luukkonen, P.K.; Qadri, S.; Ahlholm, N.; Porthan, K.; Männistö, V.; Sammalkorpi, H.; Penttilä, A.K.; Hakkarainen, A.; Lehtimäki, T.E.; Gaggini, M.; et al. Distinct contributions of metabolic dysfunction and genetic risk factors in the pathogenesis of non-alcoholic fatty liver disease. J. Hepatol. 2022,76, 526–535. [CrossRef] [PubMed] 5. Loomba, R.; Schork, N.; Chen, C.H.; Bettencourt, R.; Bhatt, A.; Ang, B.; Nguyen, P.; Hernandez, C.; Richards, L.; Salotti, J.; et al. Heritability of Hepatic Fibrosis and Steatosis Based on a Prospective Twin Study. Gastroenterology 2015 ,149, 1784–1793. [CrossRef] [PubMed] 6. Buch, S.; Stickel, F.; Trépo, E.; Way, M.; Herrmann, A.; Nischalke, H.D.; Brosch, M.; Rosendahl, J.; Berg, T.; Ridinger, M.; et al. A genome-wide association study confirms PNPLA3 and identifies TM6SF2 and MBOAT7 as risk loci for alcohol-related cirrhosis. Nat. Genet. 2015,47, 1443–1448. [CrossRef] 7. Stickel, F.; Buch, S.; Nischalke, H.D.; Weiss, K.H.; Gotthardt, D.; Fischer, J.; Rosendahl, J.; Marot, A.; Elamly, M.; Casper, M.; et al. Genetic variants in PNPLA3 and TM6SF2 predispose to the development of hepatocellular carcinoma in individuals with alcohol-related cirrhosis. Am. J. Gastroenterol. 2018,113, 1475–1483. [CrossRef] 8. Mandorfer, M.; Scheiner, B.; Stättermayer, A.F.; Schwabl, P.; Paternostro, R.; Bauer, D.; Schaefer, B.; Zoller, H.; Peck- Radosavljevic, M.; Trauner, M.; et al. Impact of patatin-like phospholipase domain containing 3 rs738409 G/G genotype on hepatic decompensation and mortality in patients with portal hypertension. Aliment. Pharmacol. Ther. 2018 ,48, 451–459. [CrossRef] 9. Trépo, E.; Caruso, S.; Yang, J.; Imbeaud, S.; Couchy, G.; Bayard, Q.; Letouzé, E.; Ganne-Carrié, N.; Moreno, C.; Oussalah, A.; et al. Common genetic variation in alcohol-related hepatocellular carcinoma: A case-control genome-wide association study. Lancet Oncol. 2022,23, 161–171. [CrossRef] 10. Abul-Husn, N.S.; Cheng, X.; Li, A.H.; Xin, Y.; Schurmann, C.; Stevis, P.; Liu, Y.; Kozlitina, J.; Stender, S.; Wood, G.C.; et al. A Protein-Truncating HSD17B13 Variant and Protection from Chronic Liver Disease. N. Engl. J. Med. 2018 ,378, 1096–1106. [CrossRef] 11. Tang, S.; Zhang, J.; Mei, T.T.; Zhang, W.Y.; Zheng, S.J.; Yu, H.B. Association of HSD17B13 rs72613567: TA allelic variant with liver disease: Review and meta-analysis. BMC Gastroenterol. 2021,21, 490. [CrossRef] [PubMed] 12. Stickel, F.; Lutz, P.; Buch, S.; Nischalke, H.D.; Silva, I.; Rausch, V.; Fischer, J.; Weiss, K.H.; Gotthardt, D.; Rosendahl, J.; et al. Genetic Variation in HSD17B13 Reduces the Risk of Developing Cirrhosis and Hepatocellular Carcinoma in Alcohol Misusers. Hepatology 2020,72, 88–102. [CrossRef] [PubMed] 13. Yang, J.; Trépo, E.; Nahon, P.; Cao, Q.; Moreno, C.; Letouzé, E.; Imbeaud, S.; Bayard, Q.; Gustot, T.; Deviere, J.; et al. A 17-Beta- Hydroxysteroid Dehydrogenase 13 Variant Protects From Hepatocellular Carcinoma Development in Alcoholic Liver Disease. Hepatology 2019,70, 231–240. [CrossRef] 14. Scheiner, B.; Stättermayer, A.F.; Schwabl, P.; Bucsics, T.; Paternostro, R.; Bauer, D.; Simbrunner, B.; Schmidt, R.; Marculescu, R.; Ferlitsch, A.; et al. Impact of HSD17B13 rs72613567 genotype on hepatic decompensation and mortality in patients with portal hypertension. Liver Int. 2020,40, 393–404. [CrossRef] [PubMed] 15. Michalopoulos, G.K. Hepatostat: Liver regeneration and normal liver tissue maintenance. Hepatology 2017 ,65, 1384–1392. [CrossRef] [PubMed] 16. Ma, Y.; Belyaeva, O.V.; Brown, P.M.; Fujita, K.; Valles, K.; Karki, S.; de Boer, Y.S.; Koh, C.; Chen, Y.; Du, X.; et al. 17-Beta Hydroxysteroid Dehydrogenase 13 Is a Hepatic Retinol Dehydrogenase Associated with Histological Features of Nonalcoholic Fatty Liver Disease. Hepatology 2019,69, 1504–1519. [CrossRef] [PubMed] 17. Mondul, A.; Mancina, R.M.; Merlo, A.; Dongiovanni, P.; Rametta, R.; Montalcini, T.; Valenti, L.; Albanes, D.; Romeo, S. PNPLA3 I148M Variant Influences Circulating Retinol in Adults with Nonalcoholic Fatty Liver Disease or Obesity. J. Nutr. 2015,145, 1687–1691. [CrossRef] [PubMed] 18. Pirola, C.J.; Garaycoechea, M.; Flichman, D.; Arrese, M.; San Martino, J.; Gazzi, C.; Castaño, G.O.; Sookoian, S. Splice variant rs72613567 prevents worst histologic outcomes in patients with nonalcoholic fatty liver disease. J. Lipid Res. 2019 ,60, 176–185. [CrossRef] Page 6 of 10 Marañónetal. Biomarker Research (2022) 10:35 NPV = 0.909) and a SAN ≥ 0.6 to rule it in (SP = 0.96; PPV = 0.86). Discussion Non-alcoholic fatty liver disease (NAFLD), is the major cause of chronic liver disease worldwide, with a diverse histopathological spectrum ranging from simple steatosis without significant inflammation, also termed fatty liver (NAFL), to steatohepatitis (NASH) with varying stages of fibrosis and, ultimately, cirrhosis and hepatocellular carcinoma [24]. Given that NAFLD is usually asymptomatic at the time of diagnosis, the determination of the lipid content within the liver is an important clinical challenge in terms of early identification to allow appropriate therapeutic interventions in order to prevent the progression of liver disease to more advanced fibrotic states or, even, to the end-stage liver disease. In the last decade, there has been an explosion in the discovery of biomarkers for NAFLD diagnosis and/or prognosis. Considering the wide range of their potential applications in clinical practice, the growing interest of the academic and industrial community in biomarkers is not surprising. Indeed, lack of awareness about risk factors for NAFLD and its progression, combined with insufficient or unreliable screening and surveillance modalities, may contribute to a delay in diagnosis and may explain why many patients are diagnosed in the later stages of the disease [7]. A novel finding of the present study is that BMP2 expression is increased in the liver from NAFLD Fig. 2 Serum levels of BMP2 are increased in NAFLD patients. A and B Serum levels of BMP2 determined by ELISA. Data are expressed as pg/ml and presented as mean ± SD. C. Correlation in the study population of matched serum BMP2 levels with the steatosis grade. D Correlation in the study population of matched serum BMP2 levels with the NAFLD activity score. E Correlation in the study population of matched serum BMP2 levels with the BMI. F Correlation in the study population of matched serum BMP2 levels with the HOMA‑IR. Study population: Normal liver (NL) individuals (n = 75), NAFLD patients (n = 115: 56 NAFL and 59 NASH). **p < 0.01, NAFLD, or NASH vs. NL Page 7 of 10 Marañónetal. Biomarker Research (2022) 10:35 patients as well as that serum BMP2 levels were also higher in these patients than in subjects with histologically normal liver, positioning this protein as a new molecular target linked to NAFLD. One limitation of this study is the limited number of individuals taking into account that women were predominant in all the histological groups studied, especially within patients with histological NL, and NAFLD patients were significantly older than NL subjects. In fact, BMP2 expression might be affected by age and sex. However, no correlation was found between circulating BMP2 levels and age (r = − 0.05484 p =0.4524), and there were no differences when separating the population by sex (p =0.2035). In addition, we have also shown herein that human hepatocytes are able to produce and secrete BMP2 upon stimuli with PA mimicking steatotic conditions [23], indicating that BMP2 might be involved in hepatosteatosis development, not only in the fibrogenic process as some studies have previously reported [25, 26]. Interestingly, there are only two studies so far showing that the expression of some BMPs, specifically BMP6 and BMP8B, is increased in the liver of NAFLD patients [13, 27]. These studies, along with our findings, indicate that BMPs might play a role in NAFLD pathophysiology. In this regard, it has been demonstrated that the Fig. 3 Palmitic acid overload upregulates mRNA levels of BMP2 in Huh7 hepatocytes. A Representative images of Oil Red O (ORO) staining and its quantification. B Hepatic BMP2 mRNA levels were determined by RT‑qPCR and normalized to 36B4 gene expression. Data are expressed as fold increase and presented as mean ± SEM relative to control condition [1]. C BMP2 levels in the cell supernatant determined by ELISA. Data are expressed as pg/ml and presented as mean ± SEM. Experimental conditions: Huh7 cells treated with 750 μM palmitic acid (PA750) for 16 h (n = 3 independent experiments performed by duplicate). **p < 0.01 and ***p < 0.005, PA750 vs. Control (C) Table 2 Univariate and multivariate analysis of the independent variables associated with NASH in the study population OR Odds ratio, CI Confidence interval, BMI Body mass index, GGT Gamma-glutamyltransferase Independent variables Univariate analysis Multivariate analysis OR 95% CI p value OR 95% CI p value BMP2 (pg/ml) Cut‑off = 49.21 3.5 [1.47–8.37] 0.005 5.07 [1.55–16.6] 0.007 Gender (female/male) 2.77 [1.43–5.36] 0.002 2.96 [1.22–7.14] 0.016 Age (years) 1.04 [1.01–1.06] 0.004 1.03 [0.99–1.06] 0.06 BMI (kg/m2) 1.22 [1.14–1.31] < 0.001 1.28 [1.16–1.41] < 0.001 Glucose (mg/dL) 1.03 [1.01–1.04] < 0.001 1.02 [1–1.04] 0.018 GGT (IU/L) 1.01 [1.01–1.02] 0.001 1.01 [1–1.02] 0.016 Page 8 of 10 Marañónetal. Biomarker Research (2022) 10:35 pharmacological inhibition of BMP signaling reduced diet-induced hepatic steatosis in obese mice by targeting the production of triglycerides in the liver [18]; but also, this study has also shown that this effect is specific for the inhibition of BMP type I receptor ALK3, proposing a role of BMP2 and BMP4 in the establishment of NAFLD [18]. Going further, our study is the first which reports circulating BMP2 concentrations in NAFLD patients. Other authors have previously reported that serum levels of distinct BMPs varied widely in the context of other metabolic disorders. For instance, elevated serum BMP4 levels have been correlated with carotid atherosclerosis in T2D patients [28]. Conversely, various studies have associated BMP9 with glucose and lipid metabolism because a decrease in the circulating concentrations of BMP9 have been observed in patients who displayed features of metabolic syndrome such as insulin resistance, T2D or arterial hypertension [29–31]. BMP9 has also been related to other diseases, such as portopulmonary hypertension, with lowered levels of this protein acting as a potential risk factor [32]. Regarding liver diseases, while BMP7 has been proposed as a marker for hepatic carcinogenesis because its serum levels were found to be decreased in patients with liver cancer [33], other authors have reported that circulating BMP7 levels are augmented in patients with viral chronic liver diseases [34]. Considering BMP2 reports, some studies have associated increased serum levels of this BMP with coronary artery disease in diabetic patients [35]. Moreover, an elevation of serum BMP2 concentrations was observed in overweight and obese middle-aged and elderly women [36], whereas in the present study, a correlation between circulating levels of BMP2 and BMI was not observed in agreement with results recently reported by Lopes etal. [37]. Nevertheless, further investigations are warranted to elucidate the role of BMP2 regulating glucose and lipid homeostasis and its impact in the pathogenesis of metabolic diseases such as obesity, diabetes and NAFLD. Fig. 4 Receiver operating characteristic (ROC) curve showing diagnostic accuracy of SAN algorithm for detecting NASH. AUROC, area under the ROC curve; BMI, body mass index; GGT, gamma‑glutamyltransferase Table 3 Diagnostic accuracy of SAN SAN Screening Algorithm for NASH, SN Sensitivity, SP Specificity, PPV Positive predictive value, NPV Negative predictive value, LR+, positive likelihood ratio; LR-, negative likelihood ratio SAN cut point Accuracy % SN SP PPV NPV LR+LR- ≥ 0.10 0.66 60.2 0.93 0.54 0.46 0.95 2.02 0.13 ≥ 0.20 0.71 48.9 0.84 0.66 0.52 0.90 2.47 0.24 ≥ 0.30 0.80 36.6 0.77 0.81 0.63 0.89 4.05 0.28 ≥ 0.40 0.81 32.8 0.73 0.85 0.67 0.88 4.87 0.32 ≥ 0.50 0.83 24.2 0.62 0.92 0.78 0.85 7.75 0.41 ≥ 0.60 0.83 18.8 0.54 0.96 0.86 0.83 13.5 0.48 ≥ 0.70 0.84 15.1 0.48 0.99 0.96 0.82 48.0 0.53 ≥ 0.80 0.82 12.4 0.41 1.00 1.00 0.80 N/A 0.59 ≥ 0.90 0.78 8.1 0.27 1.00 1.00 0.76 N/A 0.73 Page 9 of 10 Marañónetal. Biomarker Research (2022) 10:35 One of the most striking findings of the present study is that serum BMP2 levels were significantly associated with biopsy-proven NASH in multivariate logistic regression model, suggesting that BMP2 might have utility as biomarker to screen NASH in different clinical settings. Using serum BMP2 levels combined with age, gender, BMI, glucose and GGT, an algorithm termed SAN (Screening Algorithm for NASH) was derived from a logistic regression model which discriminated NASH patients from subjects with histologically normal liver or NAFL. This is a breakthrough because there are no validated non-invasive tools for predicting NASH so far, the most clinically relevant form of NAFLD which can only be diagnosed by liver biopsy. In our population, a value of SAN ≥ 0.6 indicated a high risk for biopsy-proven NASH with a good accuracy (AUROC, 0.86; SP, 0.96) showing that SAN may be useful to screen populations at risk for NAFLD. Conclusion In conclusion, this proof-of-concept study provides the first scientific evidence that hepatic and serum levels of BMP2 are abnormally elevated in NAFLD patients, indicating that BMP2 is a new molecular target linked to human NAFLD. In addition, we describe a simple and efficient algorithm termed SAN which was able to discriminate NASH with a good accuracy in our study population. Further studies, however, in distinct and larger cohorts of NAFLD patients are needed to validate the potential utility of SAN as a non-invasive algorithm for NAFLD evaluation. Acknowledgements We thankfully acknowledge Elvira del Pozo for helpful clinical assistance. Authors’ contributions AGR conceived and supervised the study. CGM, JRdC, JA and MRG carried out and analyzed the clinical parameters. PM, ER, RGD and RMV were involved in data generation. PM, CEFG, CGM and AGR analyzed and discussed data. PM, AGR and CGM prepared the manuscript. All authors critically revised and approved the manuscript. Funding This work was supported by grants PI20/00837 from Instituto de Salud Carlos III (ISCIII, Spain) and Fondo Europeo para el Desarrollo Regional (FEDER), GLD18/00151 from Gilead Science and CIBEREHD (ISCIII) to CGM, and contracts CP14/00181 and CP19/00032, and grants PI16/00853 and PI19/00123 from ISCIII/FEDERFEDER and CIBERDEM (ISCIII) to AGR. Availability of data and materials Authors declared that all and the other data supporting the findings of this study are available within the paper. The raw data that support the findings of this study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate This study was performed in agreement with the Declaration of Helsinki, and with local and national laws. The Human Ethics Committee of the Hospital Universitario Santa Cristina (reference, PI‑688A) and Hospital Universitario Virgen del Rocío (reference, C.I. 0359‑N‑15) approved the study procedures, and all participants signed an informed written consent before inclusion in the study. Consent for publication Not applicable. Competing interests All authors have declared that no competing interest exists. Author details 1 Metabolic Syndrome and Vascular Risk Laboratory, Hospital Universitario Santa Cristina, Instituto de Investigación Sanitaria del Hospital Universitario de La Princesa, C/Maestro Vives 2, 28009 Madrid, Spain. 2 SeLiver Group, Instituto de Biomedicina de Sevilla/CSIC/Hospital Virgen del Rocío, Sevilla, Spain. 3 Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD), Madrid, Spain. 4 Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM), Madrid, Spain. 5 Present address: Instituto de Investigaciones Biomédicas Alberto Sols (Centro Mixto CSIC‑UAM), 28029 Madrid, Spain. Received: 13 April 2022 Accepted: 10 May 2022 References 1. Buzzetti E, Pinzani M, Tsochatzis EA. The multiple‑hit pathogenesis of non‑alcoholic fatty liver disease (NAFLD). Metabolism. 2016;65(8):1038– 48 PubMed PMID: 26823198. 2. Goh GB, McCullough AJ. Natural history of nonalcoholic fatty liver dis‑ ease. Dig Dis Sci. 2016;61(5):1226–33 PubMed PMID: 27003142. 3. Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, Henry L, Wymer M. Global epidemiology of nonalcoholic fatty liver disease‑Meta‑analytic assessment of prevalence, incidence, and outcomes. Hepatology. 2016;64(1):73–84 PubMed PMID: 26707365. 4. Younossi ZM, Stepanova M, Younossi Y, Golabi P, Mishra A, Rafiq N, et al. Epidemiology of chronic liver diseases in the USA in the past three dec‑ ades. Gut. 2020;69(3):564–8 PubMed PMID: 31366455. 5. Svegliati‑Baroni G, Ridolfi F, Di Sario A, Casini A, Marucci L, Gaggiotti G, et al. Insulin and insulin‑like growth factor‑1 stimulate proliferation and type I collagen accumulation by human hepatic stellate cells: differential effects on signal transduction pathways. Hepatology. 1999;29(6):1743–51 PubMed PMID: 10347117. Epub 1999/05/29. eng. 6. Sokol RJ, McKim JM Jr, Goff MC, Ruyle SZ, Devereaux MW, Han D, et al. Vitamin E reduces oxidant injury to mitochondria and the hepato‑ toxicity of taurochenodeoxycholic acid in the rat. Gastroenterology. 1998;114(1):164–74 PubMed PMID: 9428230. 7. Lindenmeyer CC, McCullough AJ. The natural history of nonalcoholic fatty liver disease‑an evolving view. Clin Liver Dis. 2018;22(1):11–21 PubMed PMID: 29128051. Pubmed Central PMCID: PMC6130315. 8. Castera L, Friedrich‑Rust M, Loomba R. Noninvasive assessment of liver disease in patients with nonalcoholic fatty liver disease. Gastroenterol‑ ogy. 2019;156(5):1264–81 e4 PubMed PMID: 30660725. Pubmed Central PMCID: 7505052. 9. Arthur MJ, Fibrogenesis II. Metalloproteinases and their inhibitors in liver fibrosis. Am J Physiol Gastrointest Liver Physiol. 2000;279(2):G245–9 PubMed PMID: 10915630. 10. Carreira AC, Alves GG, Zambuzzi WF, Sogayar MC, Granjeiro JM. Bone morphogenetic proteins: structure, biological function and therapeutic applications. Arch Biochem Biophys. 2014;561:64–73 PubMed PMID: 25043976. 11. Miyazono K, Kamiya Y, Morikawa M. Bone morphogenetic protein recep‑ tors and signal transduction. J Biochem. 2010;147(1):35–51 PubMed PMID: 19762341. 12. Herrera B, Sanchez A, Fabregat I. BMPS and liver: more questions than answers. Curr Pharm Des. 2012;18(27):4114–25 PubMed PMID: 22630083. 13. Arndt S, Wacker E, Dorn C, Koch A, Saugspier M, Thasler WE, et al. Enhanced expression of BMP6 inhibits hepatic fibrosis in non‑alcoholic fatty liver disease. Gut. 2015;64(6):973–81 PubMed PMID: 25011936. Page 10 of 10 Marañónetal. Biomarker Research (2022) 10:35 • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: 14. Mahli A, Seitz T, Beckroge T, Freese K, Thasler WE, Benkert M, et al. Bone morphogenetic protein‑8B expression is induced in Steatotic hepatocytes and promotes hepatic steatosis and inflammation in vitro. Cells. 2019;8(5):457. PubMed PMID: 31096638. Pubmed Central PMCID: 6562647. 15. Peng Q, Chen B, Wang H, Zhu Y, Wu J, Luo Y, et al. Bone morphogenetic protein 4 (BMP4) alleviates hepatic steatosis by increasing hepatic lipid turnover and inhibiting the mTORC1 signaling axis in hepatocytes. Aging (Albany NY). 2019;11(23):11520–40 PubMed PMID: 31831718. Pubmed Central PMCID: 6932923. 16. Platko K, Lebeau PF, Byun JH, Poon SV, Day EA, MacDonald ME, et al. GDF10 blocks hepatic PPARgamma activation to protect against diet‑ induced liver injury. Mol Metab. 2019;27:62–74 PubMed PMID: 31288993. Pubmed Central PMCID: 6717799. 17. Yang M, Liang Z, Yang M, Jia Y, Yang G, He Y, et al. Role of bone morpho‑ genetic protein‑9 in the regulation of glucose and lipid metabolism. FASEB J. 2019;33(9):10077–88 PubMed PMID: 31237775. 18. Thayer TE, Lino Cardenas CL, Martyn T, Nicholson CJ, Traeger L, Wunderer F, et al. The role of bone morphogenetic protein signaling in non‑ alcoholic fatty liver disease. Sci Rep. 2020;10(1):9831 PubMed PMID: 32561790. Pubmed Central PMCID: 7305229. 19. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta‑cell function from fasting plasma glucose and insulin concentrations in man. Diabeto‑ logia. 1985;28(7):412–9 PubMed PMID: 3899825. eng. 20. Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander‑Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol. 1999;94(9):2467–74 PubMed PMID: 10484010. Epub 1999/09/14. eng. 21. Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalco‑ holic fatty liver disease. Hepatology. 2005;41(6):1313–21 PubMed PMID: 15915461. Epub 2005/05/26. eng. 22. Miquilena‑Colina ME, Lima‑Cabello E, Sanchez‑Campos S, Garcia‑Medi‑ avilla MV, Fernandez‑Bermejo M, Lozano‑Rodriguez T, et al. Hepatic fatty acid translocase CD36 upregulation is associated with insulin resistance, hyperinsulinaemia and increased steatosis in non‑alcoholic steatohepa‑ titis and chronic hepatitis C. Gut. 2011;60(10):1394–402 PubMed PMID: 21270117. Epub 2011/01/29. 23. Gonzalez‑Rodriguez A, Mayoral R, Agra N, Valdecantos MP, Pardo V, Miquilena‑Colina ME, et al. Impaired autophagic flux is associated with increased endoplasmic reticulum stress during the development of NAFLD. Cell Death Dis. 2014;5:e1179 PubMed PMID: 24743734. Pubmed Central PMCID: 4001315. Epub 2014/04/20. eng. 24. Brunt EM, Kleiner DE, Carpenter DH, Rinella M, Harrison SA, Loomba R, et al. NAFLD: reporting histologic findings in clinical practice. Hepatology. 2021;73(5):2028–38 PubMed PMID: 33111374. 25. Chung YH, Huang YH, Chu TH, Chen CL, Lin PR, Huang SC, et al. BMP‑2 restoration aids in recovery from liver fibrosis by attenuating TGF‑beta1 signaling. Lab Investig. 2018;98(8):999–1013 PubMed PMID: 29789683. 26. Shen H, Huang GJ, Gong YW. Effect of transforming growth factor beta and bone morphogenetic proteins on rat hepatic stellate cell prolifera‑ tion and trans‑differentiation. World J Gastroenterol. 2003;9(4):784–7 PubMed PMID: 12679932. Pubmed Central PMCID: PMC4611450. Epub 2003/04/08. 27. Sanz S, Pucilowska JB, Liu S, Rodriguez‑Ortigosa CM, Lund PK, Brenner DA, et al. Expression of insulin‑like growth factor I by activated hepatic stellate cells reduces fibrogenesis and enhances regeneration after liver injury. Gut. 2005;54(1):134–41 PubMed PMID: 15591519. Pubmed Central PMCID: 1774353. 28. Son JW, Jang EH, Kim MK, Baek KH, Song KH, Yoon KH, et al. Serum BMP‑4 levels in relation to arterial stiffness and carotid atherosclerosis in patients with type 2 diabetes. Biomark Med. 2011;5(6):827–35 PubMed PMID: 22103619. 29. Huang H, Wang W, Yang G, Zhang Y, Li X, Liu H, et al. Circulating bone morphogenetic protein‑9 levels are associated with hypertension and insulin resistance in humans. J Am Soc Hypertens. 2018;12(5):372–80 PubMed PMID: 29550458. 30. Luo Y, Li L, Xu X, Wu T, Yang M, Zhang C, et al. Decreased circulating BMP‑9 levels in patients with type 2 diabetes is a signature of insulin resistance. Clin Sci. 2017;131(3):239–46 PubMed PMID: 27940998. 31. Xu X, Li X, Yang G, Li L, Hu W, Zhang L, et al. Circulating bone morphoge‑ netic protein‑9 in relation to metabolic syndrome and insulin resistance. Sci Rep. 2017;7(1):17529 PubMed PMID: 29235531. Pubmed Central PMCID: 5727514. 32. Nikolic I, Yung LM, Yang P, Malhotra R, Paskin‑Flerlage SD, Dinter T, et al. Bone morphogenetic protein 9 is a mechanistic biomarker of Portopul‑ monary hypertension. Am J Respir Crit Care Med. 2019;199(7):891–902 PubMed PMID: 30312106. Pubmed Central PMCID: 6444661. 33. Schmierer B, Hill CS. TGFbeta‑SMAD signal transduction: molecular speci‑ ficity and functional flexibility. Nat Rev Mol Cell Biol. 2007;8(12):970–82 PubMed PMID: 18000526. 34. Tacke F, Gabele E, Bataille F, Schwabe RF, Hellerbrand C, Klebl F, et al. Bone morphogenetic protein 7 is elevated in patients with chronic liver disease and exerts fibrogenic effects on human hepatic stellate cells. Dig Dis Sci. 2007;52(12):3404–15 PubMed PMID: 17415633. 35. Zhang M, Sara JD, Wang FL, Liu LP, Su LX, Zhe J, et al. Increased plasma BMP‑2 levels are associated with atherosclerosis burden and coronary calcification in type 2 diabetic patients. Cardiovasc Diabetol. 2015;14:64 PubMed PMID: 26003174. Pubmed Central PMCID: 4450848. 36. Ribeiro S, Lopes LR, Paula Costa G, Figueiredo VP, Shrestha D, Batista AP, et al. CXCL‑16, IL‑17, and bone morphogenetic protein 2 (BMP‑2) are associated with overweight and obesity conditions in middle‑aged and elderly women. Immun Ageing. 2017;14:6 PubMed PMID: 28293269. Pubmed Central PMCID: PMC5346187. Epub 2017/03/16. 37. Lopes LR, Ribeiro S, Figueiredo VP, Leite ALJ, Nicolato RLC, Gomes JAE, et al. The overweight increases circulating inflammatory mediators commonly associated with obesity in young individuals. Cytokine. 2018;110:169–73 PubMed PMID: 29763838. Epub 2018/05/16. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Marañónetal. Biomarker Research (2023) 11:46 https://doi.org/10.1186/s40364-023-00489-2 RESEARCH Open Access © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Biomarker Research Circulating bone morphogenetic protein 8A isanovel biomarker topredict advanced liver fibrosis Patricia Marañón1†, Stephania C. Isaza1†, Carlos Ernesto Fernández‑García1, Esther Rey1, Rocío Gallego‑Durán2,3, Rocío Montero‑Vallejo2,3, Javier Rodríguez de Cía1, Javier Ampuero2,3, Ángela M. Valverde4,5, Manuel Romero‑Gómez2,3, Carmelo García‑Monzón1,3 and Águeda González‑Rodríguez4,5* Abstract Background & Aims Advanced hepatic fibrosis is the main risk factor of liver‑related morbidity and mortality in patients with chronic liver disease. In this study, we assessed the potential role of bone morphogenetic protein 8A (BMP8A) as a novel target involved in liver fibrosis progression. Methods Histological assessment and BMP8A expression were determined in different murine models of hepatic fibrosis. Furthermore, serum BMP8A was measured in mice with bile duct ligation (BDL), in 36 subjects with histo‑ logically normal liver (NL) and in 85 patients with biopsy‑proven non‑alcoholic steatohepatitis (NASH): 52 with non‑ or mild fibrosis (F0‑F2) and 33 with advanced fibrosis (F3‑F4). BMP8A expression and secretion was also determined in cultured human hepatocyte‑derived (Huh7) and human hepatic stellate (LX2) cells stimulated with transforming growth factor ꞵ (TGFꞵ). Results Bmp8a mRNA levels were significantly upregulated in livers from fibrotic mice compared to control animals. Notably, serum BMP8A levels were also elevated in BDL mice. In addition, in vitro experiments showed increased expression and secretion to the culture supernatant of BMP8A in both Huh7 and LX2 cells treated with TGFꞵ. Noteworthy, we found that serum BMP8A levels were significantly higher in NASH patients with advanced fibrosis than in those with non‑ or mild fibrosis. In fact, the AUROC of circulating BMP8A concentrations to identify patients with advanced fibrosis (F3‑F4) was 0.74 (p˂0.0001). Moreover, we developed an algorithm based on serum BMP8A levels that showed an AUROC of 0.818 (p˂0.0001) to predict advanced fibrosis in NASH patients. Conclusion This study provides experimental and clinical evidence indicating that BMP8A is a novel molecular target linked to liver fibrosis and introduces an efficient algorithm based on serum BMP8A levels to screen patients at risk for advanced hepatic fibrosis. Keywords Liver fibrosis, Bone morphogenetic proteins, BMP8A, Non‑invasive diagnosis, Hepatic stellate cells †Patricia Marañón and Stephania C. Isaza these authors contributed equally to this work. *Correspondence: Águeda González‑Rodríguez aguedagr[email protected] Full list of author information is available at the end of the article Page 2 of 14 Marañónetal. Biomarker Research (2023) 11:46 Introduction Globally, chronic liver diseases (CLD) represent nearly 3.5% of the all-cause mortality and hepatocellular carcinoma (HCC) is the sixth leading cause of cancer [1], emphasizing the urgent need to implement multilevel interventions to tackle the CLD burden of health systems worldwide. The underlying etiologies of CLD are multiple and include virus-induced chronic hepatitis, autoimmune biliary tract diseases, alcohol-related liver disease (ALD) and non-alcoholic fatty liver disease (NAFLD), largely its more progressive form named non-alcoholic steatohepatitis (NASH) which can occur in more than 40–50% of NAFLD-affected individuals [2]. CLD of any etiology are commonly featured by progressive hepatic fibrosis which, if untreated, can eventually lead to cirrhosis and/or HCC [3, 4]. Liver fibrosis is a dynamic process consisting in an increased production and deposition within the liver of extracellular matrix proteins (ECM), such as collagens among others, as a wound-healing response to chronic liver injury. Hepatic stellate cells (HSCs) are the main contributors to liver fibrosis when activated after a chronic damage and transdifferentiate from a quiescent phenotype into highly proliferative and profibrotic myofibroblasts which are the main ECM-producing cells [5]. It is well known that the stage of hepatic fibrosis is the major prognostic factor of liver-related morbidity and mortality in patients with CLD [6, 7] and, therefore, is often taken as a key data to choose the appropriate therapy and to establish the prognosis of a patient with CLD. Consequently, an accurate assessment of liver fibrosis is needed for successful clinical management of CLD patients. Liver biopsy is still considered the gold standard to differentiate hepatic fibrosis stages [8] but, given that it is an invasive and costly method with potential significant complications and sampling errors, noninvasive methods to assess hepatic fibrosis stages with accuracy are urgently needed. In this regard, transient elastography (TE), which uses pulsed-echo ultrasound as a surrogate marker of fibrosis, is now widely available and used in clinical practice, though it has some limitations derived from central obesity [9]. Therefore, the identification of non-invasive biomarkers of liver fibrosis has been gaining interest since they can offer a faster and costeffective alternative to other diagnostic methods [10]. For instance, blood-based algorithms to predict liver fibrosis are mathematical models calculated from clinical or analytical data that have been validated for the quantification of the probability of advanced fibrosis state. Some of them, such as Fibrosis 4 score (FIB-4) and aminotransferase to platelet ratio index (APRI), have good sensitivity and specificity to rule out and rule in advanced fibrosis in CLD of varied etiologies [11, 12], but novel biomarkers or algorithms with higher accuracy to predict advanced fibrosis still remain as an unmet need. Interestingly, a number of clinical and translational studies have shown that serum levels of different bone morphogenetic proteins (BMPs) are modulated in patients with CLD, particularly BMP2 [13], BMP7 [14– 16], BMP9 and BMP10 [17]. As a member of BMP family, BMP8A has been described as developmental factor [18, 19]. In addition, circulating BMP8A has already been put forward as a novel biomarker for early diagnosis of thyroid cancer in humans [20], but little is known about its potential significance in CLD. Taking this background into account, the aim of the present study was to determine the expression pattern of BMP8A in the liver of different preclinical models of hepatic fibrosis and to assess whether serum BMP8A levels could be a good predictor of liver fibrosis stage in patients with NASH, the commonest cause of CLD worldwide. Material andmethods Fibrosis experimental models C57BL6J mice were maintained at room temperature (22°C) and controlled humidity, with light/dark cycles (12h light/ 12h dark), fed adlibitum and had free access to drinking water at the animal facilities of the Universidad Autónoma de Madrid (UAM). All animal experimentation was carried out following both Spanish and European legislations. Animals were submitted to different experimental models in order to induce fibrosis. First, for the bile duct ligation (BDL) model, anaesthetized mice underwent BDL operation, which consisted in the double ligation with 4–0 silk ligatures and transection of the bile duct, as previously described [21]. Control animal group underwent a sham operation, where the common bile duct was exposed but not ligated (N = 7 animals per group). Mice were sacrificed 28days after operation, On the other hand, for the carbon tetrachloride (CCl4) treatment, mice were injected CCl4 (1:4 in olive oil, 1.6 mg/kg body weight) intraperitoneally (i.p.) twiceweekly while control mice were injected vehicle (olive oil) (N = 6 animals per group) and after 6weeks of treatment mice were sacrificed. Finally, for the non-alcoholic fatty liver disease (NAFLD) model, mice were fed either a high fat diet (D12492, Rodent Diet With 60kcal% Fat Research Diets, Inc. New Brunswick, USA) or chow diet (CHD, SAFE A04-10 Panlab, Barcelona, Spain) for 16weeks (N = 5 animals per group). At the end of each experiment, liver and serum samples were collected for further analysis. Page 3 of 14 Marañónetal. Biomarker Research (2023) 11:46 Study population This study included 85 patients with biopsy-proven NASH, of them 52 showed non- or mild fibrosis (F0- F2) and 33 displayed advanced fibrosis stages (F3-F4). In addition, 36 subjects with histologically normal liver (NL) were also studied and considered as a control group. All patients met the inclusion criteria that included: age between 18 and 75years old, less than 20 gr per day of alcohol consumption, absence of treatment with any potentially hepatotoxic drug, no analytical evidence of iron overload and seronegativity for HIV and hepatitis B and C infection. This study was performed in agreement with the Declaration of Helsinki, and with local and national laws. The Human Ethics Committee of the Hospital Universitario Santa Cristina (report reference, PI-688A) and Hospital Universitario Virgen del Rocío (report reference, 0359-N- 15) approved the study procedures, and all participants signed an informed written consent before inclusion in the study. Clinical and biochemical characteristics of each patient were routinely determined by central laboratories. Blood-based fibrosis scores such as FIB-4 and APRI were calculated in the study population as previously described [11, 12]. Characteristics of all patients included in this study based on the histological diagnosis are shown in Table1. Histopathology assessment Paraffin embedded liver sections of 5μm were stained with Haematoxylin and Eosin (H&E) and Sirius Red. Liver biopsy sections were examined by a single-blinded pathologist and evaluated according to the NAFLD activity score (NAS Score) system regarding presence of steatosis, inflammation and hepatocyte ballooning [22], as well as determination of the presence of fibrosis in four stages, being 0 absence of fibrosis and 4 cirrhosis [23]. Representative images were taken using an optical microscope Nikon Eclipse E400 (Nikon, Tokyo, Japan). Cell culture andtreatment Human hepatic stellate cells (HSC) LX2 (ATCC, Manassas, VA, USA) were maintained in a medium prepared with 50% Dulbecco´s modified Eagle´s medium (Ref. Table 1 Characteristics of the study population Data are shown as mean ± SD or as number of cases (%). BMI, body mass index; AST, aspartate aminotransferase ALT Alanine aminotransferase, GGT Gamma-glutamyltransferase ** p < 0.01 *** p < 0.005 **** p < 0.0001, mild (F0-F2) or advanced fibrosis (F3-F4) vs. NL # p < 0.05 ## p < 0.01, advanced fibrosis (F3-F4) vs. mild fibrosis (F0-F2) Normal liver (NL) (n = 36) Mild fibrosis (F0-F2) (n = 52) Advanced fibrosis (F3-F4) (n = 33) Age (years) 56.82 ± 12.48 52.75 ± 11.89 60.64 ± 9.37## Gender (female/male, %) 58.33 / 41.67 59.62 / 40.38 48.48 / 51.52 BMI (kg/m2)27.72 ± 5.93 34.92 ± 7.99**** 32.6 ± 4.63*** Glucose (mg/dL) 94.17 ± 12.76 125.9 ± 55.49** 120 ± 46.13** Diabetes (%) 11.1 53.8**** 75.8****# Triglycerides (mg/dL) 107.2 ± 53.23 181 ± 133.7**** 159.7 ± 74.75*** Total cholesterol (mg/dL) 186 ± 41.1 197.5 ± 46.83 194.4 ± 35.55 AST (IU/L) 19.5 ± 5.79 61 ± 121.6**** 48.15 ± 29.74**** ALT (IU/L) 20.58 ± 12.03 67.33 ± 86.66**** 54.97 ± 35.9**** GGT (IU/L) 40.39 ± 29.79 96.52 ± 158** 136.5 ± 189.1****# Platelets (109/L) 233.3 ± 65.5 234.1 ± 64.38 186.6 ± 68.56**### Albumin (g/L) 44.85 ± 3.61 45.3 ± 3.64 44.71 ± 3.16 Fibrosis (%) Stage 0 100% 32.7% Stage 1 40.38% Stage 2 26.92% Stage 3 66.67% Stage 4 33.33% Page 4 of 14 Marañónetal. Biomarker Research (2023) 11:46 SH30022.01 DMEM, Cytiva, USA) containing high glucose and 50% DMEM containing high glucose without glutamine (Ref. 10,938–025 Gibco, Thermo Fisher Scientific, Inc.) and supplemented with antibiotics and 2% fetal bovine serum (FBS). LX2 cells were serum starved for 2h prior to the treatment with TGFβ (Ref. 100-21C, Peprotech Inc., NJ, USA) (10ng/ml) for 24h. Human hepatoma cell line Huh7 (ATCC, Manassas, VA, USA) was maintained in DMEM containing high glucose, antibiotics and supplemented with 10% FBS. Huh7 cells were treated with TGFβ (10ng/ml) for 24h. All cells were cultured at 37°C with 5% CO2 and relative humidity of 95%. Gene expression analysis Total RNA was extracted from liver tissue samples or cell lysate using TRIzol reagent (Vitro, Sevilla, Spain) and was reverse transcribed with ImProm-II™ Reverse transcription kit (Promega Inc., Madison, WI, USA) in a T100TM Thermal Cycler (BioRad Inc., Madrid, Spain). Quantitative Real Time PCR (RTqPCR) was performed with StepOnePlusTM Real Time PCR System sequence detector (Thermo Fisher Scientific, Inc.) using SYBR Green method and quantified with ΔΔCt method. All samples were run in duplicate and normalized by 36B4. Primer sequences used were: m-Bmp8a, 5’ AAC CAT GCC ATC TTG CAG TCT 3’ (forward) and 5’ CAG AGG TGG CAC TCA GTT TGG 3’ (reverse); m-Col1a1, 5’ TAG GCC ATT GTG TAT GCA GC 3’ (forward) and 5’ ACA TGT TCA GCT TTG TGG ACC 3’ (reverse); m-Acta2, 5’ CCC AGA CAT CAG GGA GTA ATGG 3’ (forward) and 5’ TCT ATC GGA TAC TTC AGC GTCA 3’ (reverse); m-Serpin1, 5’ ATG ACT GGG TGG AAA GGC ATAC 3’ (forward) and 5’ CAG GCG TGT CAG CTC GTC TA 3’ (reverse); m-36b4, 5’ AGA TGC AGC AGA TCC GCA T 3’ (forward) and 5’ GTT CTT GCC ATC AGC ACC 3’ (reverse); h-BMP8A, 5’ CAC CCT TCT CAT CTG GAT CG 3’ (forward) and 5’ CAG GAA GTA GGC ACC GAG AG 3’ (reverse); h-COL1A1, 5’ ACT GGT GAG ACC TGC GTG TA 3’ (forward) and 5’ GAA TCC ATC GGT CAT GCT CT 3’ (reverse); h-TGFβ1, 5’ AGG ACT GCG GAT CTC TGT GT 3’ (forward) and 5’ GGG CAA AGG AAT AGT GCA GA 3’ (reverse); h-36B4, 5’ CAG GCG TCC TCG TGG AAG TGAC 3’ (forward) and 5’ CCA GGT CGC CCT GTC TTC CCT 3’ (reverse). Determination ofserum BMP8A levels BMP8A concentration in serum samples was determined using the human BMP8A ELISA kit (CSB-EL002745HU, Cusabio Technology LLC, Hubei, China) for human samples and mouse BMP8A ELISA kit (MBS7205421, MyBioSource Inc., San Diego, CA, USA) for samples from mice, following manufacturer’s indications. Absorbance from samples was interpolated to a standard curve using a four-parameter logistic (4-PL) equation. Protein content analysis bywestern blot After treatment with TGFβ, LX2 and Huh7 cells were scraped in RIPA buffer (50mM Tris HCl, pH 7.4, 1% Triton X-100, 0.2% SDS, 1mM EDTA, 1mM PMSF and 5μg/ml leupeptin) to obtain total protein extracts. Also, protein content from LX2 and Huh7 culture supernatants were concentrated using ultrafiltration units (Vivaspin Turbo 4 Ultrafiltration Unit.VS04T21, Sartorius, Thermo Fisher Scientific Inc.). Protein extracts and concentrated supernatants were boiled in Laemmli sample buffer prior to electrophoresis in 10% SDS-PAGE. Proteins were then transferred to an immunoblot nitrocellulose membrane (Bio-Rad) that was blocked with 5% non-fat dry milk and exposed to primary BMP8A antibody (ab154373 Abcam plc, Cambridge, UK), αSMA (A-2547, Merck Life Science, Darmstadt, Germany) and COL1A1 (SC-8784, Santa Cruz Biotechnology Inc., Heidelberg, Germany) antibodies overnight at 4°C. After the incubation with the corresponding secondary antibody (Santa Cruz Biotechnology Inc.), immunoreactive bands were visualized using the ECL Western blotting protocol (Bio-Rad). Densitometric analysis of the bands was performed using Image J software. Ponceau staining was used as loading control. Statistical analysis Kolmogorov–Smirnov test was applied to evaluate if the variables were adjusted or not to a normal distribution. Qualitative variables are presented as relative frequencies, and data between groups compared with Pearson’s Chi-squared or Fisher’s exact test as appropriated. Quantitative variables are expressed as measures of central tendency (mean) and dispersion (standard deviation -SD- for patients’ data or standard error of mean -SEM- for (See figure on next page.) Fig. 1 Increased hepatic Bmp8a expression in BDL mice. A Representative images of H&E and Sirius Red staining. Evaluation of the fibrosis stage (upper panel) and quantification of Sirius Red expressed as % of stained area (lower panel). B Hepatic mRNA levels of Col1a1, Acta2 and Serpin1 determined by RT‑qPCR and normalized to 36b4 gene expression. C Hepatic mRNA levels of Bmp8a determined by RT‑qPCR and normalized to 36b4 gene expression. D, E, F and G Correlation of matched Bmp8a mRNA expression with fibrosis stage, Col1a1, Acta2 and Serpin1 mRNA expression, respectively. Experimental conditions: mice with either BDL or sham operation and sacrificed 28 days after operation (N = 7 animals per group). BDL, bile duct ligation. Data are expressed as fold increase relative to control condition (sham group, 1) and presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.005 and ****p < 0.0001, BDL vs. Sham Page 5 of 14 Marañónetal. Biomarker Research (2023) 11:46 Fig. 1 (See legend on previous page.) Page 12 of 14 Marañónetal. Biomarker Research (2023) 11:46 variables and, based on the formula from multivariate logistic regression model, we developed an algorithm called BMP8A fibrosis score (BFS) combining circulating BMP8A (OR = 2.15 for each 100 pg/mL, p = 0.002) with age (OR = 1.07, p = 0.02) and platelet count (OR = 0.99, p = 0.07) which had an excellent performance to predict advanced fibrosis in our NASH patients (AUROC = 0.818, p ˂ 0.0001). It is important to mention that the novel blood-based algorithm described herein and named BFS performed better than both APRI and FIB-4 index, two of the most widespread scoring systems to evaluate liver fibrosis in patients with CLD of different etiologies [11, 12], to predict advanced hepatic fibrosis in the study population. Currently, most of the fibrosis scores use two different cut-off points to rule in and to rule out advanced fibrosis, consequently establishing a grey zone with a significant Table 2 Univariate and multivariate analysis of the independent variables associated with advanced fibrosis in the study population OR Odds ratio, CI Confidence interval, BMI Body mass index, AST Aspartate aminotransferase, ALT Alanine aminotransferase, GGT Gamma-glutamyltransferase Univariate analysis Multivariate analysis Independent variables OR 95% CI p value OR 95% CI p value BMP8A (100 pg/ml) 2.01 [1.3–3.1] 0.002 2.15 [1.33–3.48] 0.002 Sex (female/male) 1.56 [0.65–3.78] 0.316 Age (years) 1.07 [1.02–1.13] 0.004 1.07 [1.01–1.13] 0.02 BMI (kg/m2) 0.95 [0.88–1.02] 0.308 ALT (IU/L) 0.99 [0.98–1] 0.464 AST (IU/L) 0.99 [0.99–1] 0.574 GGT (IU/L) 1 [0.99–1] 0.309 Platelets (109/L) 0.99 [0.98–1] 0.004 0.99 [0.98–1] 0.07 Albumin (g/L) 0.95 [0.82–1.09] 0.484 Table 3 Diagnostic accuracy of BFS BFS BMP8A fibrosis score, % number of patients with BFS ≥ cut-off point, SN Sensitivity, SP Specificity, PPV Positive predictive value, NPV Negative predictive value, LR + positive likelihood ratio, LR- Negative likelihood ratio BFS cutoff point Accuracy % SN SP PPV NPV LR + LR- ≥ 0.10 54.1 82.4 97.0 26.9 45.7 93.9 1.3 0.1 ≥ 0.20 62.4 71.8 93.9 42.3 50.8 91.7 1.6 0.1 ≥ 0.30 70.6 51.8 78.8 65.4 59.1 82.9 2.3 0.3 ≥ 0.40 76.5 43.5 75.8 76.9 67.6 83.3 3.3 0.3 ≥ 0.46 78.8 36.5 69.7 84.6 74.2 81.5 4.5 0.4 ≥ 0.50 77.6 35.3 66.7 84.6 73.3 80.0 4.3 0.4 ≥ 0.60 72.9 25.9 48.5 88.5 72.7 73.0 4.2 0.6 ≥ 0.70 70.6 14.1 30.3 96.2 83.3 68.5 8.0 0.7 ≥ 0.80 65.9 9.4 18.2 96.2 75.0 64.9 4.8 0.9 ≥ 0.90 62.3 1.2 3.0 100 100 61.9 ‑ 1.0 Table 4 Comparison of the operating characteristics of BFS and other widely used fibrosis scores to detect high risk of advanced fibrosis (F3‑F4) BFS BMP8A fibrosis score, % number of patients with BFS ≥ cut-point, SN Sensitivity, SP Specificity, PPV Positive predictive value, NPV Negative predictive value, LR + Positive likelihood ratio, LR- Negative likelihood ratio BFS APRI FIB-4 Operating char‑ acteristics ≥ 0.46 < 0.5 ≥ 1.5 < 1.30 ≥ 2.67 Accuracy 78.8 62.3 60.0 64.7 65.9 % 36.5 54.1 10.6 42.4 16.5 SN 69.7 60.6 12.1 78.8 27.3 SP 84.6 63.5 90.4 55.8 90.4 PPV 74.2 51.3 44.4 53.1 64.3 NPV 81.5 71.7 61.8 80.6 66.2 LR + 4.5 1.7 1.3 1.8 2.8 LR‑ 0.4 0.6 1.0 0.4 0.8 Page 13 of 14 Marañónetal. Biomarker Research (2023) 11:46 proportion of unclassified patients. According to APRI and FIB-4 categorization, the grey zone represented 35.3% and 41.1% of patients in our study population, respectively. The strength of BFS relies on its capacity to accurately distinguish mild from advanced fibrosis using a single cut-off point, thus avoiding unclassified patients, and most importantly, without heavily sacrificing sensitivity or specificity. The main limitation of our study is that we have included only NASH patients with liver fibrosis and, therefore, the results might be different in patients with CLD of different etiologies. However, our preclinical studies showed that BMP8A expression was upregulated in different experimental models of liver fibrosis. Further studies in distinct and larger cohorts of patients with CLD of varied etiologies are needed to validate the potential utility in clinical practice of BFS as a non-inva- sive algorithm to identify/screen those CLD patients at high-risk of advanced fibrosis. In conclusion, the results of the present study provide the first scientific evidence that hepatic and serum levels of BMP8A are abnormally elevated in experimental models of liver fibrosis and in the serum of NASH patients with advanced hepatic fibrosis, indicating that BMP8A is a new molecular target linked to liver fibrogenesis. In addition, we have defined a simple and efficient algorithm termed BFS which was able to discriminate advanced hepatic fibrosis with a good accuracy in NASH patients. Validation of these results in larger and independent cohorts of CLD patients are warranted to propose BFS as a clinically useful tool to screen CLD patients at primary care settings, in order to identify those patients at higher risk of advanced fibrosis to whom an in-depth evaluation in specialized hepatology units must be recommended. Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s40364‑ 023‑ 00489‑2. Additional file 1: Supplementary Figure 1. Increased hepatic Bmp8a expression in CCl4 mice. A. Hepatic mRNA levels of Col1a1 and Serpin1 determined by RT‑qPCR and normalized to 36b4 gene expression. B. Hepatic mRNA levels of Bmp8a determined by RT‑qPCR and normalized to 36b4 gene expression. C and D. Correlation of matched Bmp8a mRNA expression with Col1a1 and Serpin1 mRNA expression respectively. Experi‑ mental conditions: mice treated with CCl4 or vehiclei.p. injection twice weekly for 6 weeks.. Data are expressed as fold increase and presented as mean ± SEM relative to control group. *p<0.05, **p<0.01 and ***p<0.005, CCl4 vs. Oil. Authors’ contributions AGR conceived and supervised the study. CGM, JRdC, JA and MRG carried out and analyzed the clinical parameters. PM, SCI, ER, AMV, RGD and RMV were involved in data generation. PM, SCI, CEFG, CGM and AGR analyzed and dis‑ cussed data. PM, AGR and CGM prepared the manuscript. All authors critically revised and approved the manuscript. Funding This work was supported by grants PI16/00853 and PI19/00123 from Instituto de Salud Carlos III (ISCIII, Spain) and Fondo Europeo para el Desarrollo Regional (FEDER), CIBERDEM (ISCIII, Spain) and grant GLD18/00151 from Gilead Science. PM was supported by a research contract (PEJ‑2020‑AI/BMD‑19491) from Comunidad de Madrid (Spain). SCI was supported by a predoctoral contract (FI20‑00296) and CEGF was supported by a Sara‑Borrell contract (CD20/00199) from ISCIII/FEDER (Spain). Availability of data and materials Authors declared that all and the other data supporting the findings of this study are available within the paper. The raw data that support the findings of this study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate This study was performed in agreement with the Declaration of Helsinki, and with local and national laws. The Human Ethics Committee of the Hospital Universitario Santa Cristina (reference, PI‑688A) and Hospital Universitario Virgen del Rocío (reference, C.I. 0359‑N‑15) approved the study procedures, and all participants signed an informed written consent before inclusion in the study. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Author details 1 Metabolic Syndrome and Vascular Risk Laboratory, Hospital Universitario Santa Cristina, Instituto de Investigación Sanitaria del Hospital Universitario de La Princesa, Madrid, Spain. 2 SeLiver Group, Instituto de Biomedicina de Sevilla/CSIC/Hospital Virgen del Rocío, Seville, Spain. 3 Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD), Madrid, Spain. 4 Instituto de Investigaciones Biomédicas Alberto Sols (Centro Mixto CSIC‑UAM), Madrid, Spain. 5 Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM), Instituto de Investigaciones Biomédicas Alberto Sols (Centro Mixto CSIC‑UAM), Madrid, Spain. Received: 3 January 2023 Accepted: 18 April 2023 References 1. Asrani SK, Devarbhavi H, Eaton J, Kamath PS. Burden of liver diseases in the world. J Hepatol. 2019;70(1):151–71. 2. Cheemerla S, Balakrishnan M. Global epidemiology of chronic liver disease. Clin Liver Dis (Hoboken). 2021;17(5):365–70. 3. Parola M, Pinzani M. Liver fibrosis: Pathophysiology, pathogenetic targets and clinical issues. Mol Aspects Med. 2019;65:37–55. 4. Roehlen N, Crouchet E, Baumert TF. Liver fibrosis: mechanistic concepts and therapeutic perspectives. Cells. 2020;9(4):875. 5. Geerts A. History, heterogeneity, developmental biology, and functions of quiescent hepatic stellate cells. Semin Liver Dis. 2001;21(3):311–35. 6. Hardy T, Oakley F, Anstee QM, Day CP. Nonalcoholic fatty liver disease: pathogenesis and disease spectrum. Annu Rev Pathol. 2016;11:451–96. 7. Ekstedt M, Hagström H, Nasr P, Fredrikson M, Stål 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(5):1547–54. 8. Buzzetti E, Pinzani M, Tsochatzis EA. The multiple‑hit pathogenesis of non‑ alcoholic fatty liver disease (NAFLD). Metabolism. 2016;65(8):1038–48. 9. Jennison E, Patel J, Scorletti E, Byrne CD. Diagnosis and management of non‑alcoholic fatty liver disease. Postgrad Med J. 2019;95(1124):314–22. 10. Baranova A, Lal P, Birerdinc A, Younossi ZM. Non‑invasive markers for hepatic fibrosis. BMC Gastroenterol. 2011;11:91. Page 14 of 14 Marañónetal. Biomarker Research (2023) 11:46 • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: 11. McPherson S, Stewart SF, Henderson E, Burt AD, Day CP. Simple non‑ invasive fibrosis scoring systems can reliably exclude advanced fibrosis in patients with non‑alcoholic fatty liver disease. Gut. 2010;59(9):1265–9. 12. Lin ZH, Xin YN, Dong QJ, Wang Q, Jiang XJ, Zhan SH, et al. Performance of the aspartate aminotransferase‑to‑platelet ratio index for the staging of hepatitis C‑related fibrosis: an updated meta‑analysis. Hepatology. 2011;53(3):726–36. 13. Patricia M, Ernesto F‑GC, Stephania CI, Esther R, Rocío G‑D, Rocío M‑V, et al. Bone morphogenetic protein 2 is a new molecular target linked to nonalcoholic fatty liver disease with potential value as non‑invasive screening tool. MedRxiv. 2022. 14. AktugDemir N, Kolgelier S, Inkaya AC, Sumer S, Demir LS, Pehlivan FS, et al. Are bone morphogenetic protein‑7 (BMP‑7) serum levels correlated with development of hepatic fibrosis? J Infect Dev Ctries. 2014;8(5):605–10. 15. Tacke F, Gäbele E, Bataille F, Schwabe RF, Hellerbrand C, Klebl F, et al. Bone morphogenetic protein 7 is elevated in patients with chronic liver disease and exerts fibrogenic effects on human hepatic stellate cells. Dig Dis Sci. 2007;52(12):3404–15. 16. Chayanupatkul M, Honsawek S, Vejchapipat P, Chongsrisawat V, Poovorawan Y. Elevated serum bone morphogenetic protein 7 levels and clinical outcome in children with biliary atresia. Eur J Pediatr Surg. 2009;19(4):246–50. 17. Owen NE, Alexander GJ, Sen S, Bunclark K, Polwarth G, Pepke‑Zaba J, et al. Reduced circulating BMP10 and BMP9 and elevated endoglin are associ‑ ated with disease severity, decompensation and pulmonary vascular syndromes in patients with cirrhosis. EBioMedicine. 2020;56:102794. 18 Wu FJ, Lin TY, Sung LY, Chang WF, Wu PC, Luo CW. BMP8A sustains sper‑ matogenesis by activating both SMAD1/5/8 and SMAD2/3 in spermato‑ gonia. Sci Signal. 2017;10(477):eaal1910. 19. Chung D, Gao F, Jegga AG, Das SK. Estrogen mediated epithelial prolif‑ eration in the uterus is directed by stromal Fgf10 and Bmp8a. Mol Cell Endocrinol. 2015;400:48–60. 20. Liu K, Gao M, Qin D, Wang H, Lu Q. Serous BMP8A has clinical significance in the ultrasonic diagnosis of thyroid cancer and promotes thyroid cancer cell progression. Endocr Metab Immune Disord Drug Targets. 2020;20(4):591–8. 21. Villar‑Lorenzo A, Rada P, Rey E, Marañón P, Arroba AI, Santamaría B, et al. Insulin receptor substrate 2 (IRS2) deficiency delays liver fibrosis associ‑ ated with cholestatic injury. Dis Model Mech. 2019;12(7):dmm038810. 22. Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalco‑ holic fatty liver disease. Hepatology. 2005;41(6):1313–21. 23. Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander‑Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol. 1999;94(9):2467–74. 24. Puche JE, Saiman Y, Friedman SL. Hepatic stellate cells and liver fibrosis. Compr Physiol. 2013;3(4):1473–92. 25. Zhang L, Xu LM, Chen YW, Ni QW, Zhou M, Qu CY, et al. Antifibrotic effect of N‑acetyl‑seryl‑aspartyl‑lysyl‑proline on bile duct ligation induced liver fibrosis in rats. World J Gastroenterol. 2012;18(37):5283–8. 26. Zou GL, Zuo S, Lu S, Hu RH, Lu YY, Yang J, et al. Bone morphogenetic protein‑7 represses hepatic stellate cell activation and liver fibrosis. World J Gastroenterol. 2019;25(30):4222–34. 27. Arndt S, Wacker E, Dorn C, Koch A, Saugspier M, Thasler WE, et al. Enhanced expression of BMP6 inhibits hepatic fibrosis in non‑alcoholic fatty liver disease. Gut. 2015;64(6):973–81. 28. Mahli A, Seitz T, Beckröge T, Freese K, Thasler WE, Benkert M, et al. Bone morphogenetic protein‑8B expression is induced in steatotic hepato‑ cytes and promotes hepatic steatosis and inflammation in vitro. Cells. 2019;8(5):457. 29. Vacca M, Leslie J, Virtue S, Lam BYH, Govaere O, Tiniakos D, et al. Bone morphogenetic protein 8B promotes the progression of non‑alcoholic steatohepatitis. Nat Metab. 2020;2(6):514–31. 30. Addante A, Roncero C, Almalé L, Lazcanoiturburu N, García‑Álvaro M, Fernández M, et al. Bone morphogenetic protein 9 as a key regulator of liver progenitor cells in DDC‑induced cholestatic liver injury. Liver Int. 2018;38(9):1664–75. 31. Zoheiry MM, Hasan SA, El‑Ahwany E, Nagy FM, Taleb HA, Nosseir M, et al. Serum Markers of Epithelial Mesenchymal Transition as Predictors of HCV‑induced Liver Fibrosis. Cirr Hepatocellular Carc Elec Physician. 2015;7(8):1626–37. 32. Hao J, Wang Y, Huo L, Sun T, Zhen Y, Gao Z, et al. Circulating bone mor‑ phogenetic protein‑9 is decreased in patients with type 2 diabetes and non‑alcoholic fatty liver disease. Int J Gen Med. 2022;15:8539–46. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Liver International. 2023;43:1909–1919. | 1909wileyonlinelibrary.com/journal/liv Received: 17 January 2023 | Revised: 30 March 2023 | Accepted: 23 April 2023 DOI: 10.1111/liv.15604 ORIGINAL ARTICLE Role of EpCAM+ CD133+ extracellular vesicles in steatosis to steatohepatitis transition in NAFLD Rocío Muñoz- Hernández1,2,3 | Sheila Gato1,2 | Antonio Gil- Gómez1,2 | Rocío Aller4,5 | Angela Rojas1,2 | Laura Morán6,7 | Javier Gallego1 | Elena Blázquez- López2,6,8 | Rocío Gallego- Durán1,2 | Rocío Montero- Vallejo1,2 | Vanessa García- Fernández1,3 | Douglas Maya- Miles1,2 | María del C Rico1,2,9 | Francisco J Cubero2,6,7 | Javier Vaquero2,6,8 | Javier Ampuero1,2,3,9 | Rafael Bañares2,6,8 | Manuel Romero- Gómez1,2,3,9 1SeLiver Group, Instituto de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Sevilla, Spain 2Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD), Madrid, Spain 3Departamento de Medicina, Facultad de Medicina, Universidad de Sevilla, Seville, Spain 4Department of Medicine, Dermatology and Toxicology, Universidad de Valladolid / Gastroenterology Unit, Hospital Clínico Universitario de Valladolid / BioCritic, Group for Biomedical Research in Critical Care Medicine, Valladolid, Spain 5Centro de Investigación Biomédica en Red de Enfermedades Infecciosas (CIBERINFEC), Instituto de Salud Carlos III, Madrid, Spain 6Instituto de Investigación Sanitaria Gregorio Marañón (IiSGM), Madrid, Spain 7Department of Immunology, Ophthalmology and ENT, Complutense University School of Medicine, Madrid, Spain 8Servicio de Aparato Digestivo, Hospital General Universitario Gregorio Marañón, Madrid, Spain 9UCM Digestive diseases, Hospital Universitario Virgen del Rocío, Sevilla, Spain This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2023 The Authors. Liver International published by John Wiley & Sons Ltd. Abbreviations: ALT, alanine aminotransferase; ASGPR1, asialoglycoprotein receptor 1; AST, aspartate aminotransferase; BMI, body mass index; CAP, controlled attenuated parameters; CD133, prominin- 1; DM, diabetes mellitus; EpCAM, epithelial cell adhesion molecule; EVs, extracellular vesicles; GGT, gamma glutamyl transferase; HDL, highdensity lipoprotein; HFHCC, high fat, high carbohydrate and cholesterol; HTA, hypertension; LDL, lowdensity lipoprotein; MVs, microvesicles; NAFLD, nonalcoholic fatty liver disease; NASH, steatohepatitis; ROC, receiver operating characteristic curve; TC, total cholesterol. Correspondence Rocío Muñoz- Hernández and Manuel Romero- Gómez, SeLiver Group, Instituto de Biomedicina de Sevilla (IBiS), Hospital Universitario Virgen del Rocío/CSIC/ Universidad de Sevilla, Sevilla 41013, Spain. Email: rociomun[email protected] and [email protected] Funding information Consejería de Economía, Conocimiento, Empresas y Universidad, Junta de Andalucía; consejería de salud y Familia; Instituto de Salud Carlos III; Junta de Andalucía; Sociedad Andaluza de patología Digestiva Handling Editor: Luca Valenti Abstract Background and Aims: Extracellular vesicles (EVs) have emerged as a potential source of circulating biomarkers in liver disease. We evaluated circulating AV+ EpCAM+ CD133+ EVs as a potential biomarker of the transition from simple steatosis to steatohepatitis. Methods: EpCAM and CD133 liver proteins and EpCAM+ CD133+ EVs levels were analysed in 31 C57BL/6J mice fed with a chow or high fat, high cholesterol and carbohydrates diet (HFHCC) for 52 weeks. The hepatic origin of MVs was addressed using AlbCrexmT/mG mice fed a Western (WD) or Dual diet for 23 weeks. Besides, we assessed plasma MVs in 130 biopsyproven NAFLD patients. Results: Hepatic expression of EpCAM and CD133 and EpCAM+ CD133+ EVs increased during disease progression in HFHCC mice. GFP+ MVs were higher in AlbCrexmT/mG mice fed a WD (5.2% vs 12.1%) or a Dual diet (0.5% vs 7.3%). Most GFP+ MVs were also positive for EpCAM and CD133 (98.3% and 92.9% respectively), suggesting their hepatic Rocío Muñoz- Hernández and Manuel Romero- Gómez contributed equally as senior authors. 1910 | MUÑOZ-HERNÁNDEZ et al. 1 | INTRODUCTION Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease worldwide, affecting about a quarter of the world's adult population.1 The spectrum of liver damage in NAFLD ranges from simple steatosis (SS), an asymptomatic state to more severe forms of the disease, characterized by the appearance of steatohepatitis (NASH) and liver fibrosis. Liver biopsy remains the gold standard diagnostic tool for patients with NAFLD despite its significant disadvantages (low acceptance, invasiveness and heterogeneity). Indeed, the diagnosis of NASH requires a histological examination to confirm the presence of lobular inflammation and hepatocyte ballooning degeneration, with or without fibrosis.2 Thus, developing novel noninvasive biomarkers of NASH is strongly needed. The role of extracellular vesicles (EVs) in cellular communication within tissues and organs has gained significant attention during the last few years, particularly in acute and chronic diseases.3,4 EVs are small membrane vesicles released by different cell types in resting or under activating conditions, into several biological fluids such as blood, urine, or saliva. EVs contain proteins, lipids and coding/noncoding RNAs of the cell of origin.5 Microvesicles (MVs) are membranederived EVs in the size range of 100– 1000 nm that are generated by blebbing from the plasma membrane of nearly all cells, also called large EVs.6,7 Large EVs express cell surface markers from their parent cell, which enable the characterization of their cellular origin through flow cytometry approaches.8,9 Previous studies have shown that parenchymal and nonparenchymal liver cells release EVs in biological fluids indicating liver disease progression.10,11 In this context, several studies demonstrated that changes in hepatocytederived EVs under lipotoxic conditions promote inflammation and fibrosis in the liver.12 A twofold increase of circulating ASGPR1+ EVs in cirrhotic patients compared to healthy controls has been previously reported,13 and Haertel et al. showed that the combination of Annexin V, ASGPR1, EpCAM and CD133 MVs could help to identify patients with liver tumours.14 Hepatocytedriven liver regeneration involves the proliferation of preexisting hepatocytes which is the primary generation mode,15 when the regenerative capacity of hepatocytes and/or cholangiocytes is exhausted hepatic progenitor cells (HPCs) are thought to play an important role in this process.16 HPC may exist in or near the canal of Hering,17 and current studies have demonstrated HPCs activation and proliferation in patients with severe and chronic liver disease such as nonalcoholic fatty liver disease18,19 and have been identified and isolated from liver tissue from all age donors.20,21 Previous studies have shown that these cell populations are positive for endodermal progenitor markers such as epithelial cell adhesion molecule (EpCAM) and Prominin- 1 (CD133),22– 24 and they can express high levels of albumin.25– 27 Despite EpCAM+ CD133+ cells are associated with tumours, in an early stage of the disease, where the liver is submitted to chronic injury, EpCAM+ CD133+ cells are activated and are thought to play a role in liver regeneration without giving rise to hepatocellular carcinoma28 and could also be releasing specific EVs into circulation. Given the asymptomatic nature of the early stages of NAFLD, the development of noninvasive biomarkers capable of detecting the transition from steatosis to NASH is strongly needed. This study aimed to evaluate the potential of EpCAM+ CD133+ EVs as a novel biomarker to detect the transition to more severe forms of the disease. 2 | METHODS 2.1 | Animal procedures and ethics statements 2.1.1 | High fat, high carbohydrate and cholesterol animal model Thirtyone C57BL/6J mice were purchased from Charles River Laboratories (Cedex, France) at 5 weeks of age. Mice were raised in origin. In 71 biopsyproven NAFLD patients, EpCAM+ CD133+ EVs were significantly higher in those with steatohepatitis compare to those with simple steatosis (286.4 ± 61.9 vs 758.4 ± 82.3; p < 0.001). Patients with ballooning 367 ± 40.6 vs 532.0 ± 45.1; p = 0.01 and lobular inflammation (321.1 ± 74.1 vs 721.4 ± 80.1; p = 0.001), showed higher levels of these EVs. These findings were replicated in an independent cohort. Conclusions: Circulating levels of EpCAM+ CD133+ MVs in clinical and experimental NAFLD were increased in the presence of steatohepatitis, showing high potential as a noninvasive biomarker for the evaluation and management of these patients. KEYWORDS CD133, EpCAM, extracellular vesicles, NAFLD Key points Hepatic expression of EpCAM and CD133 increases with NAFLD progression and EpCAM+ CD133+ extracellular vesicles are produced by the liver after chronic liver injury. Our study showed that AV+ EpCAM+ CD133+ EVs represent a useful biomarker for the difficult task of separating patients with steatohepatitis from those with simple steatosis. 14783231, 2023, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.15604 by MANUEL ROMERO-GÓMEZ - Universidad De Sevilla , Wiley Online Library on [25/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 1911 MUÑOZ-HERNÁNDEZ et al. a temperaturecontrolled room of 23°C with constant humidity and maintained on a 12/12- h light/dark cycle with ad libitum access to diet and water at the Institute of Biomedicine of Seville (IBiS). After 1 week of acclimatization, the mice were assigned to the following study groups: (a) Control diet (n = 12), Teklad ENVIGO++++; (b) High fat, high carbohydrates and cholesterol (HFHCC) diet composed by 39.8% kcal from fat, 15.7% kcal from protein and 45.4% kcal from carbohydrates (5TJT TestDiet®) plus 42 g/L (55% fructose and 45% glucose) carbohydrates in drinking water for up to 52 weeks. All animal care and experimental protocols followed the guidelines for the care and use of laboratory animals of our Institution and the European Community Policy for Experimental Animal Studies. (19/02/2016/023). 2.1.2 | AlbCrexmT/mG mice— Liver damage mouse model The mT/mG mice mouse strain (Stock 07676, Jackson) was crossbred with mice expressing Crerecombinase under the control of the Alb promoter. In the litters, the EGFP protein is expressed in the plasma membrane (mG) in the cells expressing Crerecombinase, whereas the dTomato protein is expressed in the absence of Cre (mT) (AlbCrexmT/mG mice, EGFP in hepatocytes, dTomato in the rest of the cells), as described.29 Male and female 10- weekold C57BL6/J mice were treated with WD (D09100301; Research Diets, Inc., New Brunswick, NJ) for 23 weeks. Controls were fed with a chow diet (Altomin, Lage, Germany). In addition, male and 10- weekold female C57BL6/J mice were treated with a Dual diet consisting of Western diet and 10% vol/vol EtOH absolute in sweetened drinking water containing 6.75% D- glucose for 23 weeks, as previously described.30 All procedures were carried out according to Spanish legal requirements and animal protection law approved by Comunidad de Madrid (PROEX125.1/20). Animals were maintained in the Animal Facility in the Faculty of Biology, Complutense University of Madrid under a 2- h light/dark ad libitum according to the guidelines of the Federation for Laboratory Animal Science Associations (FELASA). 2.1.3 | Collection of samples and determination of EpCAM and CD133 MVs Retroorbital blood was extracted and centrifuged at 2000g for 20 min at 4°C in heparincoated microtubes. Samples were collected and pooled for a total of 100 μL of plasma from five mice per group to determine EpCAM and CD133 large EVs levels. In brief, 30 μL of sample were incubated with 5 μL of GFP antibody Alexa Fluor 488 (ThermoFisher scientific, MA, USA,), 5 μL of anti- EpCAM APC (Invitrogen, MA, USA) and 5 μL of CD133 PeCY7 (Invitrogen, MA, USA) for 15 min at RT. Next, we distinguished phosphatidylserine positive MVs using 5 μL of Annexin V- Brilliant Violet with 95 μL of Annexin buffer incubating for 20 min at RT (BD, Bioscience). MVs were analysed using a LSRFortessa (BD Biosciences, San Diego, CA). MVs size was determined using calibration beads Megamixplus side angle light scatters (SSC) (0.16, 0.20, 0.24 and 0.5 μm) (Biocytex, Marseille, France). Data represent the mean of two independent experiments. MVs levels of HFHCC- fed mice were determined in 5 μL of heparinplasma sample following the protocol above described with EpCAM FITC (Invitrogen, MA, USA), CD133 (PeCy7) and Annexin V- PE (BD, Bioscience, NJ, USA). 2.2 | Histological evaluation and immunohistochemistry Liver tissues from HFHCC- fed mice and Dual diet mice were fixed in 4% buffered formalin solution and embedded in paraffin. Successive 4- μmsections were obtained and prepared for immunohistochemistry. Liver tissue sections were deparaffinized with citrate buffer and rehydrated with decreasing concentrations of ethanol. EDTA- antigen retrieval was subsequently performed, and 3% hydrogen peroxide was used to quench peroxidase activity for 20 min. Sections were then incubated with CD133 (Cell signalling, 1:1000) and EpCAM (Abcam, 1:300) overnight at 4°C and the nuclei were counterstained with haematoxylin. For HFHCC- fed mice, EpCAM and CD133 expression in oval cells in liver tissue, were evaluated by an experienced pathologist from the pathology department of Virgen del Rocio University Hospital. In dual mice liver tissue, Images were captured by Olympus X61 at 20× magnification. Photomicrographs of EpCAM and CD133- positive areas for WD and Dual mice were taken at 20× magnification and analysed using the opensource software ImageJ. 2.3 | Immunofluorescence Successive 4- μm hepatic sections were obtained and prepared for immunofluorescence. Paraffin sections were deparaffinized in xylol and rehydrated in graded alcohol series. Sections were then washed in distilled water and heated in a pressure cooker for epitope retrieval (in 10 mM citrate buffer [pH 6.0], 5 min). Slides were blocked with 2.5% horse serum and then incubated (overnight at 4°C) with primary antibodies against CD133 (1:100) and EPCAM (1:50). Sections were then washed with PBS. For fluorescence visualization of antibody reactions, primary antibodies were detected using secondary antibodies labelled with the fluorochromes Alexa Fluor (Alexa antimouse A488 and Alexa antirabbit A568, Invitrogen, MA, USA). To detect cell nuclei, sections tissues were incubated with 4′, 6- diamidino- 2- phenylindole (DAPI) and then mounted on Vectashield mounting medium for fluorescence (Vector Laboratories, Peterborough, UK). Negative controls omitting the primary antibody step were run in parallel. Images were obtained using confocal microscopy (Leica Stellaris 8). 2.4 | Study population Patients with biopsyproven NAFLD were enrolled at the outpatient clinics of Virgen del Rocío University Hospital (Seville, Spain) or the 14783231, 2023, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.15604 by MANUEL ROMERO-GÓMEZ - Universidad De Sevilla , Wiley Online Library on [25/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 1912 | MUÑOZ-HERNÁNDEZ et al. Clinical Valladolid University Hospital. Exclusion criteria were as follows: age < 18 years, significant alcohol intake (≥30 g/day in men and ≥20 g/day in women), evidence of viral autoimmune hepatitis, HIV, recreational drugs abuse or druginduced fatty liver or other metabolic liver diseases (such as haemochromatosis or Wilson's disease). The clinical assessment included structured questionnaires to determine risk factors (hypertension, T2DM, dyslipidaemia, alcohol consumption and smoking status), monitor physical activity and control current medication along with other anthropometric measures (age, sex) and routine and liver function biochemical panels. Percutaneous liver biopsies were obtained in all patients under local anaesthesia and ultrasound guidance. Biopsy samples were fixed in formalin and embedded in paraffin blocks. Specimens were stained with haematoxylin– eosin and Masson's trichrome. An experienced pathologist assessed the samples to determine the grade of steatosis, ballooning and inflammation, in order to classify patients by SAF score.31 Blood samples were collected at inclusion. Fasting (overnight) blood samples were drawn from participants into tubes containing clot activator and separator gel (SST), K3- EDTA and heparin- Vacutainer TM. The index of insulin resistance (Homeostatic Model Assessment for Insulin Resistance, HOMA- IR) was calculated using the following formula: glucose (mg/dL) × insulin (mU/mL)/405. Finally, biochemical parameters were determined in routine laboratory at both hospitals. The study was conducted in accordance with the ethical guidelines described in the Declaration of Helsinki and the international conference on harmonization guidelines for good clinical practice. Both centres received the approval of its Ethics Committee and ensured that participants understood the study and provided written informed consent. 2.5 | Determination of EPCAM CD133 MVs in NAFLD patients Samples of peripheral blood were collected using a 21- gauge needle. Heparinbuffered blood samples were obtained and processed less than 1 h after venipuncture at 3000g for 10 min. For flow cytometry, 250 μL of plasma was centrifuged for 20 min at 3500g at room temperature to obtain plateletfree plasma sample; after that, samples were centrifuged for 30 min at 17 000g at RT twice to obtain large EVs pellet. All samples were frozen and stored at −80°C until assayed. The cellular origin of plasma MVs was determined by immunostaining using selective fluorochromelabelled antibodies. Fifty microlitre of processed plasma was incubated with 3 μL of a monoclonal antibody antihuman CD326 (EpCAM) coupled to Alexa fluor 488 (Invitrogen, MA, USA) and 3 μL of antihuman CD133 (Prominin- 1) coupled to PeCY7 (Invitrogen, MA, USA). After that, MVs were then incubated with 5 μL of Annexin V- PE (BD Biosciences, CA) for 20 min at RT to label phosphatidylserine. The negative control (zero value) was obtained using the isotype antibodies. Large EVs were identified as events with a 0.1– 1 μm diameter on forward light scatter (FSC) and sideangle light scatters (SSC) intensity dot plot representation, by comparison to flow cytometry calibration beads (Megamixplus side angle light scatters SSC [0.16, 0.20, 0.24 and 0.5 μm], Biocytex, Marseille, France), indicating that our EVs isolation strategy mainly detected large vesicles or microvesicles. Finally, 50 μL of Flow Count Beads (Beckman Coulter, Marseille, France) were added to obtain MVs/μL. Data represent the mean of two independent experiments. Microvesicles were analysed by flow cytometry (BD LSR Fortessa; BD Biosciences). 2.6 | Statistical analysis Categorical variables were expressed as absolute frequencies and percentages (n, %), and noncategorical variables as the mean ± SEM or SD. The Kolmogorov– Smirnov test was used to assess whether variables followed a normal distribution. Comparisons between groups were made using T- tests to compare variables with normal distribution, Mann– Whitney U tests for those with nonnormal distribution, and χ2 tests for qualitative variables. Bonferroni correction was applied to confirm differences between groups. All variables reaching statistical significance after Bonferroni correction (p < 0.05) in the univariate regression model were included in the multivariate regression model along with wellestablished prognostic indicators (age, sex and BMI). Possible associations were analysed using Spearman rankorder correlation. The predictive cutoff for EpCAM CD133 MVs was identified using the receiver operating characteristic (ROC) curve method and using the Youden index. Comparisons between AUCs were performed using STATA (v. 16, StataCorp LLC, TX). Data were compared and analysed using SPSS software (version 22.0; SPSS, Chicago, IL), and GraphPad Prism (version 7.0; GraphPad Software, La Jolla, CA) was used for graphics. For all analyses, p < 0.05 was considered statistically significant. 3 | RESULTS 3.1 | Hepatic EpCAM and CD133 expression increase with the progression of NAFLD and coexpresses in human liver tissue Mice fed with the HFHCC diet showed liver steatosis from week 13 and clear signs of steatohepatitis at 52 weeks of diet, showing: (a) higher body weight, macro– microvesicular steatosis (>90%), inflammatory foci at lobular, portal and periductal areas, hepatocyte degeneration (ballooning), moderate fibrosis and high prevalence of oval cells; (b) higher expression of EpCAM and CD133 proteins in liver tissue compared with mice with simple steatosis (Figure 1A,B). The reactivity of EpCAM and CD133 in oval cells appears from week 39 of diet onwards, 100% of animals had mild expression of CD133 in week 39 and 64% moderate levels of EpCAM at the same timepoint, reaching severe expression of both markers only when liver tumour appears (p < 0.001 for both EpCAM and CD133). 14783231, 2023, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.15604 by MANUEL ROMERO-GÓMEZ - Universidad De Sevilla , Wiley Online Library on [25/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 1913 MUÑOZ-HERNÁNDEZ et al. Next, we analysed the expression of EpCAM and CD133 in liver tissue of AlbCrexmT/mG mice fed a Dual diet (Figure 1C). Compared with control mice, the liver parenchyma of mice fed a Dual diet showed a significantly increased percentage of positive EpCAM area per view field and a clear tendency towards increased CD133 staining (Figure 1D). As shown in Figure 1, importantly, the expression of EpCAM and CD133 was also confirmed in liver tissue from patients with NAFLD (Figure 1E). As shown in Figure 1F, these cells frequently coexpressed both EpCAM and CD133, thus explaining the double positivity observed in the cellderived microvesicles. 3.2 | The liver is able to produce AV+ EpCAM+ CD133+ circulating MVs Mice fed a HFHCC diet for 52 weeks showed significantly higher circulating concentrations of EpCAM CD133 EVs compared to control mice fed with a chow diet (Figure 2A, p = 0.02). Besides, we found a positive and significative correlation between levels of this kind of EVs compare to EpCAM (r = 0.47; p = 0.005) and CD133 (r = 0.30; p = 0.036) liver tissue expression. To confirm if EpCAM CD133 EVs were released from hepatocytes, we used AlbCrexmT/mG mice fed a WD or a Dual diet (WD + alcohol). In AlbCrexmT/mG mice, albuminexpressing cells (hepatocytes and oval cells) present GFP in their plasma membrane (mG), whereas all other cells express the Tomatodye protein instead (mT) (Figure 2B). Using flow cytometry and amplification of signal with a fluorochromeconjugated anti- GFP antibody, we showed that AlbCrexmT/mG mice fed a WD for 23 weeks presented increased proportions of GFP- positive Annexin V EVs compared with controls (12.1% vs 5.2%), supporting hepatocytes as their cell of origin (Figure 2C). The proportion of AV+ dTomato + EVs was not increased by the WD (58.9% in control vs 44.5% after WD). Importantly, 92.8% and 98.3% of the AV+ GFP+ EVs population in the control and WD groups, respectively, were also EpCAM+ CD133+, indicating that most hepatocytederived (GFP+) EVs were also positive for these proteins, while within the dtomato+ population, 19.1% and 19.3% were also positive for EpCAM and CD133 in the control and the WD group respectively. FIGURE 1 Expression of EpCAM CD133 liver tissue. (A) Representative pictures of EpCAM and CD133 oval cell staining according to different stages of the disease in HFHCC- fed mice. (B) Percentage of animals with several degrees of EpCAM and CD133 liver tissue expression. (C) EpCAM and CD133 IHC staining in AlbCrexmT/mG mice with Dual diet after 23 weeks of Dual feeding. (D) Quantification of % positive area of EpCAM and CD133, respectively, using ImageJ software (n = 5– 6). (E) EpCAM and CD133 IHC in human liver biopsies (10× magnification). (F) Immunofluorescence double staining of hepatic tissue (40× magnification). Data represent the mean ± SEM. *p < 0.05 *** p < 0.001. 14783231, 2023, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.15604 by MANUEL ROMERO-GÓMEZ - Universidad De Sevilla , Wiley Online Library on [25/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 1914 | MUÑOZ-HERNÁNDEZ et al. Similar results were obtained in AlbCrexmT/mG mice fed a Dual diet for 23 weeks (Figure 2E,F). Compared with control mice, mice fed the Dual diet showed a significative elevation of circulating AV+ GFP+ EVs (0.5% vs 7.3%) while levels of dTomato+ EVs remain stable (27.8% vs 23.9%). Again, the proportion of MVs positive for EpCAM and CD133 among the total AV + GFP+ EVs was 87.9% in control mice and 92.9% in Dual mice. In summary, we observed an increase in GFP+ EVs in both NALFD models, and approximately 90% of these GFP+ EVs were EpCAM+ and CD133 + . 3.3 | Circulating levels of AV+ EPCAM+ CD133 EVs were associated with steatohepatitis in patients with NAFLD Among 130 patients with biopsyproven NAFLD included in this study, 71 of them composed the discovery cohort. Patients were classified by SAF score as simple steatosis (n = 20) or steatohepatitis (n = 51). Among their baseline demographic and clinical characteristics (Table 1), patients with steatohepatitis had more fibrosis (p = 0.046), higher levels of AST (p = 0.001) and ALT (p = 0.005), and a higher prevalence of arterial hypertension (p = 0.002) than patients with simple steatosis. Patients with steatohepatitis showed significantly higher plasma levels of EpCAM+ CD133+ EVs than patients with simple steatosis (Table 1, and Figure 3A). Regarding histopathological features of NAFLD, patients with lobular inflammation (mild or severe) had higher levels of EPCAM+ CD133+ EVs (321.1 ± 305.7 vs 721.4 ± 588.3, p = 0.001) than those without lobular inflammation (Figure 3B), but no difference was found between patients with mild and severe inflammation (data not shown). Patients with hepatocyte ballooning also showed higher levels of this kind of EVs compared to patients without ballooning (Figure 3C). In contrast, the circulating concentration of EpCAM+ CD133+ EVs was associated neither with the degree of steatosis (S1 (n = 37): 633.3 ± 99.1 U/μL vs S2 (n = 25): 672.3 ± 117.9 U/μL vs S3 (n = 9): 552.7 ± 141.9 U/μL, p = 0.907, Figure 3D) nor with the presence of fibrosis (No fibrosis: 611.5 ± 77.5 vs Fibrosis: 649.9 ± 123 U/ μL, p = 0.735) or its degree (Figure 3E). Multivariate regression showed that age, ALT and circulating EpCAM+ CD133+ EVs were independent predictors of steatohepatitis (rankordered by Wald, Table 1). FIGURE 2 (A) Circulating concentration of EpCAM+ CD133+ EVs in mice fed an HFHCC diet (MASH) or a chow diet (Control) for 31 weeks. (B) AlbCrexmT/mG mice. (C, D) GFP+/ – dTomato−+ EpCAM+ CD133+ EVs levels in AlbCrexmT/mG mice after 23 weeks with WD or Dual treatment (E, F). Samples were collected and pooled for a total of 100 μL of plasma from five mice per group in AlbCrexmT/mG mice. Data are shown as mean ± SD in the HFHCC- mouse model. *p < 0.05. 14783231, 2023, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.15604 by MANUEL ROMERO-GÓMEZ - Universidad De Sevilla , Wiley Online Library on [25/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 1915 MUÑOZ-HERNÁNDEZ et al. TABLE 1 Baseline demographic and clinical characteristics of patients in the discovery cohort. Variables Simple steatosis (n = 20) NASH (n = 51) Univariate pvalue Multivariate OR (CI 95%; pvalue) Age, years ± SD 50.3 ± 12.8 57.7 ± 10.6 0.016 1.153 (1.052– 1.264); p = 0.003 Male, sex, n (%) 14 (70) 24 (47) 0.081 BMI (kg/m2)31.8 ± 5.2 35.6 ± 7.8 0.047 Fibrosis (Kpa) 10.8 ± 6.8 14.8 ± 7.7 0.046 CAP (dB/m) 314.9 ± 64.6 332.8 ± 41.6 0.192 Bilirubin ± SD (mg/dL) 0.92 ± 1.20 0.56 ± 0.32 0.222 Platelets ± SD (×109/L) 198.8 ± 61.1 222.1 ± 66.6 0.181 Creatinine ± SD (mg/dL) 0.78 ± 0.19 0.88 ± 0.51 0.561 AST ± SD (IU/L) 30.80 ± 15.1 53.2 ± 36.7 0.001 ALT ± SD (IU/L) 36.3 ± 17.8 62.02 ± 45.2 0.005 1.055 (1.006– 1.107); p = 0.031 GGT ± SD (IU/L) 59.1 ± 25.8 101.8 ± 129 0.061 TC ± SD (mg/dL) 179.2 ± 54.2 181.4 ± 34.9 0.833 LDL ± SD (mg/dL) 106.1 ± 42.1 104.8 ± 32.9 0.364 HDL ± SD (mg/dL) 43.7 ± 12.3 46.4 ± 10.8 0.393 HTA, n (%) 8 (5.6) 40 (78.4) 0.002 DM, n (%) 9 (6.3) 11 (21.6) 0.292 EpCAM+ CD133+ (U/μL) ± SD 286.4 ± 61.9 758.4 ± 82.3 <0.001 1.006 (1.002– 1.009); p = 0.009 Note: Categorical variables are represented by absolute frequencies and percentages (n, %). Noncategorical variables are represented by mean ± SEM. pvalues are from the comparison of SS and NASH groups. Abbreviations: ALT, alanine aminotransferase, AST, aspartate aminotransferase; BMI, body mass index; CAP, controlled attenuated parameters; CI 95%, 95% confidence interval; DM, diabetes mellitus; GGT, gamma glutamyl transferase; HDL, highdensity lipoprotein; HTA, hypertension; LDL, lowdensity lipoprotein; TC, total cholesterol. FIGURE 3 EpCAM AND CD133 EVs levels in NAFLD patients. (A) AV+ EpCAM+ CD133+ EVs population in SS versus NASH. (B) EVs levels regarding inflammation, (C) Steatosis grade (D) Fibrosis stage and (E) Ballooning. Data are represented as mean ± SEM ns, no significant, *p < 0.05, **p < 0.005 and ***p < 0.001. 14783231, 2023, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.15604 by MANUEL ROMERO-GÓMEZ - Universidad De Sevilla , Wiley Online Library on [25/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 3 NAFL: non-alcoholic fatty liver NHPS: NASH Human Proximity Score GLP-1: glucagon-like peptide-1 foz/foz: fat aussie mice Alms1: Alström syndrome 1 HFD: high fat diet MC4R: melanocortin 4 receptor Mc4r−/−: melanocortin 4 receptor– deficient mice DPP-4: dipeptidyl peptidase-4 ASCVD: atherosclerotic cardiovascular disease CVD: cardiovascular disease Ldlr−/−: LDL receptor knockout Apoe−/−: apolipoprotein E knockout CETP: cholesteryl ester transfer LDL: low-density lipoprotein ApoB: apolipoprotein B ApoE: apolipoprotein E IR: insulin resistance OCA: obeticholic acid PNPLA3: patatin-like phospholipase domain-containing protein 3 LPAAT: lysophosphatidic acid acyltransferase TG: triglyceride ER: endoplasmic reticulum VLDL: very-low-density lipoprotein MBOAT7: membrane-bound o-acyltransferase domain-containing 7 4 LPIAT: lysophosphatidylinositol acyltransferase SREBP: sterol regulatory element binding protein HSD17B13: hydroxysteroid 17β-dehydrogenase Pten-/-: pten knockout PTEN: phosphatase and tensin homolog WDF: western diet supplemented with 5% fructose FLS: fatty liver Shionogi HFCS: high-fructose corn syrup HFHFD: high-fat high-fructose diet HFHSD: high-fat, high-sucrose diet CASH: cholesterol-associated steatohepatitis HC: high-cholesterol HFHC: high-fat, high-cholesterol ALIOS: American lifestyle-induced obesity syndrome DIAMOND: diet-induced animal model of non-alcoholic fatty liver disease CDAA: choline-deficient l-amino acid-defined WT: wild type CDAHFD: l-amino acid-defined, high-fat diet ALT: alanine aminotransferase AST: aspartate aminotransferase STAM: stelic animal model STAM-HFD: stelic animal model, high-fat diet CCL4: carbon Tetrachloride DEN: diethylnitrosamine model DEN-HFD: diethylnitrosamine high-fat diet 5 TAA: thioacetamide model TIRF: transgene-free Il2rg-/-/Rag2-/-/Fah-/- FFC: high-fat, high-fructose and cholesterol CCL4-HFD: carbon tetrachloride, high-fat diet AAs: amino acids Data availability statement: Not applicable. Conflict of interest: All authors affirm that there is no conflict of interest to declare. Ethics approvals: Not applicable. Patients consent: Not applicable. Funding: The research leading to these results has received funding from the Consejería de Salud de la Junta de Andalucía under grant agreement PC-0148-2016-0148 and PE- 0451-2018, and from the Instituto de Salud Carlos III under grant agreements CD21/00095, PI19/01404, PI19/00589; PI22/01342, PI22/01345 and IFI18/00041 and Ministerio de Economía y Competitividad under grant agreement AGL-2017-86927-R and PID2020-116731RB-C21. Rocío Montero-Vallejo has received the Juan Manuel Herrerías from Sociedad Andaluza de Patología Digestiva, CIBERehd grants to support mobility and improve mice handling. Rocío Gallego-Durán has received the Andrew K Burroughs Fellowship from European Association for the Study of the Liver (EASL), Aprendizaje de Nuevas Tecnologías fellowship from Asociación Española para el Estudio del Hígado (AEEH) and CIBERehd Grant to support researcher’s mobility. Lay summary: - MASLD pathophysiology is complex with various contributing factors from different routes leading to several clinical phenotypes. - There is no ideal animal model for MASLD, thus the selection of the preclinical model depends on endpoints desired. 6 - Animal welfare, ethical considerations, the cost-efficiency, length of the diet, reproducibility and expected outcomes need to considered prior the selection. ABSTRACT Metabolic associated steatotic liver disease (MASLD) encompasses a wide spectrum of metabolic conditions associated with an excess of fat accumulation in the liver, ranging from simple hepatic steatosis to cirrhosis and hepatocellular carcinoma. Finding appropriate tools to study its development and progression is essential to address essential unmet therapeutic and staging needs. This review discusses advantages and shortcomings of different dietary, chemical and genetic factors that can be used to mimic this disease and its progression in mice from a hepatic and metabolic point of view. Also, this review will highlight some additional factors and considerations that could have a strong impact in the outcomes of our model to end up providing recommendations and a checklist to facilitate the selection of the appropriate MASLD preclinical model based on clinical aims. Keywords: MASLD, liver fibrosis; steatohepatitis; animal models. 7 INTRODUCTION Metabolic associated steatotic liver disease (MASLD) encompasses a wide spectrum of conditions associated with excess fat accumulation in the hepatocyte, ranging from simple liver steatosis to metabolic associated steatohepatitis (MASH), fibrosis, cirrhosis and eventually, hepatocellular carcinoma (HCC)1. The prevalence of MASLD, which is closely related to metabolic syndrome, has manifested an upper trend over the past years and has been globally estimated to be around 25%2. Several recent studies point MASH as the fastest growing indication for both, liver transplantation and HCC, especially in Western countries 3,4,5. Selecting preclinical models is a difficult task in which the ability of models to recapitulate clinically relevant endpoints needs to be appropriately balanced to other issues like ethics (e.g. number of experimental animals required to achieve statistical significance) or budget (diet duration, mouse colony maintenance). In fact, the lack of good preclinical models in MASLD could be one of the explanations for the translational failure from preclinical studies to human clinical trials. Animal models should only be used with the purpose of answering relevant questions or clarifying mechanisms that cannot be addressed by using in vitro models. Besides, researchers have to comply the Principle of the 3R’s. Each R stands for a principle for the ethical use of animals in animal experimentation: Reduction consists in using as few animals as possible to achieve the objective of the study. Refinement is the application of methods that avoid animal suffering and Replacement is the most important concept, stands for finding a replacement for animal models in other systems such as cell cultures, organoids or similar (3R)6. 8 Moreover, it is accepted that preclinical models aiming to reproduce MASLD to a certain extent should be able to induce hepatic steatosis and although inflammation, ballooning and liver fibrosis might not be mandatory to study certain processes, there are all highly desirable. This will better reflect the staging of the disease in those patients that are more likely to require pharmacological therapy. Up to now, many animal models mainly focus on particular stages or traits of the disease. Thus, additionally, these models should ideally exhibit features associated to metabolic syndrome like glucose intolerance, insulin resistance and obesity. Other features that can be commonly observed in MASLD such as adipose tissue inflammation, in order to better mimic the full spectra of comorbidities that can act as both, cause or consequence, during disease establishment and progression are desirable 7. Overall, a wide variety of animal models are used in translational research, and although rats (Rattus norvegicus) are widely used, this review will focus on mice because of their convenient features. Mus musculus has dominated the field for many years based on its main advantages, such as: small size (therefore less drug dose is needed, among others), more cost-effectiveness, the availability of a wide spectrum of potential recombinant models ready-to-use and finally, their handling advantages 8. This review attempts to summarize the main advantages and disadvantages of different animal models available to study MASLD and to provide recommendations according to the desired clinical endpoints. 9 1. Genetic modelling of MASLD Genetics has a major impact in the progression and outcomes of MASLD, where heritability in humans is estimated to lie between 20-70%9,10. Genetic models constitute an invaluable tool in preclinical research that can be used to build and shape our animal models to better fit and reproduce those symptoms required to mimic desired endpoints, as well as to study genetic risk factors that linked to MASLD progression. 1.1. Hyperphagic genetic models 1.1.1. Leptin-deficient mice (ob/ob) & leptin receptor-deficient mice (db/db) Leptin is a satiety hormone and a master regulator of energy balance expressed mainly by white cells in adipose tissue. Lepob/Lepob mutant mice in the C57BL/6J background produces a deficit of leptin that in combination with ad libitum diet is known to induce overweight, increased body fat deposition, hyperglycaemia, and hyperinsulinemia due to hyperphagia and reduced energy expenditure 11,12. These mice also develop MASLD but they require modified diets to promote the development of steatohepatitis and fibrosis like Amylin Liver NASH (AMLN) or Gubra Amylin NASH (GAN) diet. Both diets are quite similar but GAN diet is more obesogenic and will further exacerbate hyperglycaemia 13,14. Leptin receptor-deficient mice (db/db) are metabolically alike the ob/ob but tend to accumulate less fat in the liver and develop stronger insulin resistance15. db/db mice are hyperleptinemic and develop obesity and severe type-2 diabetes partly due to a functional defect in the long-form leptin receptor (Ob-Rb), which plays a significant role in the regulation of food intake and the control of body weight. A recent comparative study has highlighted additional important differences in fat distribution (being db/db more subcutaneous and ob/ob mice more epididymal) and liver and adipose tissue inflammation (being the first more pronounced in ob/ob and the second 10 in the db/db background). ob/ob mice also show higher levels of circulating lipopolysaccharides (LPS), and alterations in the hepatic bile acid and short chain fatty acid profiles, which could be related to changes in gut microbiota composition16. ob/ob and db/db experimental models also show important differences in the appearance of fibrosis when fed with methionine choline deficient diet (MCD). While db/db mice are able to exacerbate histological damage including inflammation and marked pericellular fibrosis after 4 weeks of diet compared to controls, ob/ob mice treated for the same period of time develop much less inflammation and no obvious significant fibrosis17. It has been recently proven that ob/ob mice exposed to 23 weeks of GAN diet showed a good ability to reproduce the metabolic disturbances observed in MASLD patients. This also triggered significant accumulations of hepatic fibrosis and steatohepatitis according to a recent meta-analysis performed within the preclinical unit of the LITMUS initiative. This was based on an unbiased ranking system (NASH Human Proximity Score, NHPS) that incorporated (i) metabolic phenotype; (ii) centralized liver histology; (iii) liver transcriptome benchmarked against human MASH18. Leptin deficiency models are therefore able to trigger NAFL and most of the main metabolic risk factors. This can be useful to shorten significantly the time required to induce MASH and fibrosis after a nutritional intervention19. db/db mice might be more useful in studies aiming to assess the interaction between NAFL and diabetes or in studies testing therapies in which glucose homeostasis is the main target, like the glucagon-like peptide-1 (GLP-1) receptor agonist liraglutide. Liu et al. used this model quite recently to demonstrate that liraglutide could attenuate NAFL in db/db mice improving hepatic fat deposition and triggering changes on the composition of bacteria related to glucolipid metabolism and intestinal inflammation20. 11 Colony maintenance, a common handicap for all genetic models, can be an important drawback for these two models given the strong fertility problems caused by these mutations, especially in homozygosis, although there are several ways to overcome these issues (e.g. leptin treatment)21,22. Backcrossing to genetic backgrounds other than C57BL/6J can impact the extent of some of the metabolic disturbances developed and needs to be considered as well 23,24. 1.1.2. Fat aussie mice (foz/foz) Appetite-defective foz/foz mice constitute an overnutrition model caused by a truncating mutation in Alms1(Alström syndrome 1), a gene encoding a ubiquitous protein essential for proper primary cilium function. Mice carrying this mutation exhibit age-dependent loss of primary cilia, in hypothalamic neurons, implicated in regulating satiety responses, and causes Alström syndrome development in humans. These mice develop obesity, diabetes, hepatic steatosis, hypercholesterolemia and insulin resistance in the absence of additional stimuli25,26. Fat aussie mice fed high fat diet (HFD) exacerbate metabolic abnormalities and are able to trigger MASH after only 4 weeks of diet, advanced fibrosis after 12 weeks, and HCC after only 24 weeks according to a recent study27,28. These authors also provide evidence of an early disruption of the gut barrier, microbial dysbiosis, lipopolysaccharide leakage, and the intestinal inflammation that precedes MASH. They were able to demonstrate a certain extent of MASH regression when animals return to a normal chow diet. Old foz/foz mice (≈ 1 year) also display clear signs of hypertension and cardiac hypertrophy previously associated to their metabolic derangements29. The higher mortality caused by excessive fat consumption owe to cardiac failure stands in favour of this hypothesis 25. This interpretation, however, needs to be taken with caution as ALMS1 mutations in the same exon have been linked to rare neonatal cardiomyopathies in patients that seem to be linked to a direct role of this protein 12 in myocardiocyte development before the onset of metabolic symptoms 30-31.Severity of MASH seems to be strain related, being more severe in C57BL6/J mice32. Inconveniences for this model include fertility problems in females after the onset of obesity and metabolic syndrome together with progressive germ cell loss and sterility in males25. These mice also develop audiological and visual limitations that might compromise their utility in studies aiming to perform behavioural studies, although these symptoms appear late during disease progression (deafness > 150 days, retinal degeneration ≥ 1 year33). Wree et al. have used this model to prove that NLRP3 is a crucial contributor to liver damage and liver fibrosis and demonstrated that MCC950, an NLRP3 selective inhibitor, can improve MASLD pathology and fibrosis in obese diabetic mice34,35. 1.1.3. Melanocortin 4 receptor– deficient mice (Mc4r−/−) Melanocortin 4 receptor (Mc4r) is expressed in the hypothalamic nuclei and is involved in the regulation of food intake and body weight. Mutations occurring in the melanocortin type 4 receptor have been previously associated with severe familial obesity and reported as the most common monogenic cause of childhood obesity36. Mc4r-/- mice develop obesity associated with hyperphagia, hyperinsulinemia, and hyperglycemia37. Mc4r-/- mice develop steatosis together with marked MASH with fibrosis within 20 weeks of HFD. Interestingly, further progressing to well-differentiated HCC can be observed in all animals (100%) exposed to a high fat diet for at least 1 year according to two different reports 38. The apparent ability to recapitulate different stages of MASLD progression and the generation of well-differentiated HCC in most animals derived from metabolic disturbances without the addition of toxic compounds makes this model an interesting option to study processes related to MASH-HCC21. Kawakubo et al. have used this model to prove the ability of inhibitors of dipeptidyl peptidase-4 (DPP-4) to stop the progression 19 1.4.2. FATZO/Pco mice / MS-NASH/ PcoJ mice MS-NASH or commonly known as FATZO mice has been obtained by crossing both C57BL/6 and AKR/J strains. The name of this model was recently changed from FATZO to MS-NASH in conjunction with a new distribution agreement with The Jackson Laboratory88.These mice are prone to develop a metabolic pattern of obesity, dyslipidaemia and insulin resistance when fed a normal chow diet 89-90. Researchers proved that the addition of fat to the MS-NASH diet accelerated weight gain in parallel with disease progression91. Sun et al. fed 8-week-old male MS-NASH mice during 20 weeks with high fat Western diet supplemented with 5% fructose (WDF) and showed a development of MASLD/MASH similar to humans89. 1.4.3. Fatty liver Shionogi (FLS) FLS is a useful animal model resulted of inbreeding of dd Shionogi mice which developed hereditary hepatic steatosis without obesity92.Interestingly, they are able to show spontaneous hepatocellular tumors and HCC at 1year of age, and the incidence is much higher in male than female as in human settings93,94. A variation of this model is the named FLS- ob/ob mice. The Lepob gene of the C57BL/ 6JWakShi (B6)-Lepob/Lepob mouse was transferred into the genome of the FLS mouse by backcross mating, and mice showed obesity, hyperlipidemia and diabetes together with steatosis, steatohepatitis and hepatic tumors95,96. Considering this, FLS-ob/ob mice could be more attractive for researchers than FLS alone, nevertheless, more studies are still needed. 1.4.4. Which is the most suitable genetic model? The most appropriate genetic model will be the one that shows a phenotype that is closest to the disease in humans ideally, it should represent the full spectrum of the disease. This is what happens in foz/foz mice (with dietary requirements) (Please refer to Figure 1a). Foz/Foz mice exhibit obesity, insulin resistance and hypercholesterolaemia, among [Document text truncated for crawler view.]