From the Editor’s Desk...

Abstract

Tsai, Kuo et al. evaluated whether metformin reduces the hepatocellular carcinoma (HCC) risk among individuals with diabetes and chronic hepatitis C (CHC) after successful antiviral therapy. Of 7,249 individuals with CHC enrolled in the study, 781 had diabetes with 647 metformin users. During a median follow-up of 4.4 years, 227 patients developed new-onset HCC. The 5-year cumulative HCC incidence was 10.9% in non-metformin users and 2.6% in metformin users, compared to 3.0% in non-diabetic patients. Cirrhosis was the most important factor significantly associated with higher HCC risk followed by non-metformin use, older age, male sex, and obesity, whereas hyperlipidemia with statin use was associated with a lower HCC risk. Human hepatitis viruses, including HAV, are limited in their host species range to primates, a factor complicating the development of small animal models. This limited host range could be explained by species-specific viral strategies to disrupt innate immune responses, including viral protease cleavage of the innate immune adaptor protein MAVS. Sun, Feng et al. show that HAV 3ABC protease cleavage of MAVS enhances viral replication and lessens liver inflammation in mice lacking interferon receptors, but that it is insufficient by itself to overcome the cross-species barrier to infection in mice. The data suggest that HAV possesses additional mechanisms to disrupt cell intrinsic responses to early rounds of viral replication in the liver, and provide a specific perspective on the cleavage of MAVS by hepatitis viruses. Population-level uptake of direct-acting antiviral (DAA) treatment for HCV infection, including retreatment, can be estimated through administrative pharmaceutical dispensation data. J.M. Carson et al. developed a machine learning model to classify retreatments as reinfection or treatment failure at a national level. Nationally, 95,272 individuals initiated DAAs, with treatment uptake declining from 32,454 in 2016 to 6,566 in 2021. Of those treated, 6,980 (7%) were retreated. The model classified 51.8% as reinfection and 48.2% as treatment failure. Retreatment for reinfection increased steadily over the study period from 14 in 2016 to 1,092 in 2020, stabilizing in 2021. Retreatment for treatment failure increased from 73 in 2016 to 1,077 in 2019, then declined to 515 in 2021. Among individuals retreated for treatment failure, 50% had discontinued initial treatment. Sanyal et al. led an international effort meant to improve the performance of an elastography-based test (Fibroscan) for the detection of advanced fibrosis and cirrhosis. The objective was to design the test and the cut-offs in order to minimise the number of patients in the indeterminate zone (grey zone between the low and high cut-offs) and to maximise the positive predictive value in the rule-in zone. The authors used a large international cohort from 9 centres, North America, Europe and Asia (2,719 patients) and designed two tests, Agile 4 (for the detection of cirrhosis) based on liver stiffness, AST/ALT ratio, platelet count, sex and diabetes and Agile 3 (for detection of advanced fibrosis) based on the same parameters plus age. Both tests had an AUROC ranging between 0.89-0.91 in the different training and external validation sets and only around 17-18% of patients fell in the indeterminate zone. Overall, the diagnostic performance was better than that of elastography or that of FIB-4 alone. However, the training set had a high prevalence of the target condition thus contributing to the optimal positive predictive value (expected to be lower in less specialised, more general, settings). It is unclear if non-alcoholic steatohepatitis (NASH) improvement is mainly related to overall body weight changes or if changes in the partitioning of the adipose tissue may play a role. Visceral adipose tissue (VAT) is associated with liver fibrosis and inflammation while subcutaneous adipose tissue (SAT) comprises two functionally different layers, superficial SAT (sSAT) and deep SAT (dSAT), the latter functionally related to VAT. Shen et al. studied a subgroup (n = 76) of a previous phase IIb trial in NASH and performed volumetric assessment of the different fat compartments together with measurement of liver fat fraction and liver histology at the beginning and the end of the trial in order to determine if longitudinal reductions in abdominal adipose tissue depot volumes were independently associated with histologic improvement in NASH. They found that individuals with greater longitudinal reductions in dSAT and VAT volumes, but not sSAT, were more likely to have histologic improvement. Correlations with histologic improvement were independent of changes in hepatic fat content. Changes in overall body weight were not associated with histological improvement. The study is important because it indicates that future interventions should be designed to assess longitudinal reductions of dSAT and VAT volumes as outcome measures, rather than simply weight reduction. Nitrofurantoin (NTF) is widely used short-term for treating and long-term for preventing urinary tract infections. Chalasani et al. aim was to describe the clinical characteristics, outcomes, and HLA risk factors for NTF-liver injury (NTF-DILI) among individuals enrolled in the Drug Induced Liver Injury Network (DILIN). Seventy-eight individuals with definite, highly likely, or probable NTF-DILI were enrolled into DILIN studies between 2004-2020. HLA alleles were compared between NTF-DILI and 3 control groups: population, idiopathic autoimmune hepatitis, and non-NTF DILI. Liver injury was hepatocellular in 69%, and icteric in 55%. In long-term NTF-DILI, bridging fibrosis, nodularity or cirrhosis, or clinical and imaging evidence for cirrhosis were present in 38%, and massive or sub-massive necrosis was seen in 20%. No one in the short-exposure group died or underwent transplantation, whereas 7 (12%) patients from intermediate and long-term exposure groups combined died or underwent transplantation. After covariate adjustments, HLA-DRB1∗11:04 was significantly more frequent in NTF-DILI compared to population controls, idiopathic AIH, and non NTF DILI. Thus, NTF-DILI can result in parenchymal necrosis, bridging fibrosis, cirrhosis, and death or liver transplantation, especially with long-term exposure, and is associated with HLA-DRB1∗11:04. To mitigate against serious liver injury associated with NTF, the regulators should revise the prescribing information and consider other mitigation strategies. In individuals with compensated advanced chronic liver disease (cACLD), the severity of portal hypertension (PH) determines the risk of decompensation. In this study, Jiří Reiniš, Petrenko et al. evaluated the utility of machine learning models (MLMs) based on standard laboratory parameters in predicting the severity of PH in individuals with cACLD. A detailed laboratory workup of those with cACLD recruited from the VIENNA cohort was utilised with the aim of predicting clinically significant portal hypertension (CSPH) and severe PH. Among 1,232 individuals with cACLD, the prevalence of CSPH and severe PH in the VIENNA and validation cohorts were similar. The 5P-MLM performed robustly for CSPH (AUROCs: 0.813) and for severe PH (AUROCs: 0.887) in the VIENNA cohort, and better than liver stiffness measurement (AUROC: 0.808). However, the performance of the MLMs in external validation data sets was heterogeneous (AUROCs: 0.589-0.887). That said, the proposed 3P/5P online tool, based on widely available laboratory tests, could assist in identifying individuals with CSPH or severe PH. Although the effect of bacterial infection on cirrhosis has been well-described, the effect of non-hepatotropic virus (NHV) infection is unknown. Li, Hong et al. aimed at evaluating the genome fragments of circulating microorganisms using metagenomic next-generation sequencing (mNGS) in individuals with acute decompensation (AD) of cirrhosis, focusing on NHVs and relating the findings to clinical outcomes. Plasma mNGS was performed in a cohort of 129 individuals with AD (study cohort) and in 10 healthy volunteers and 20 individuals with stable cirrhosis, 39 with severe sepsis and 81 with haematological malignancies, enrolled as controls. Validation assays for human cytomegalovirus (CMV) reactivation in a validation cohort (n = 58) were performed and exploratory treatment instituted. In this study cohort, 188 microorganisms were detected in 74.4% patients, including viruses (58.0%), bacteria (34.1%), fungi (7.4%) and chlamydia (0.5%). NHV signatures, of which CMV was the most frequent, correlated with the clinical effect of empirical antibiotic treatment, progression to acute-on-chronic liver failure (ACLF), and 90-day mortality. Of the 14 cases with detectable CMV by mNGS, 9 were further validated by DNA RT-PCR or pp65 antigenemia testing. Three individuals with CMV reactivation experienced clinical resolutions with ganciclovir therapy. NHVs may have a pathogenic role in complicating the course of AD, but validation is needed to define whether their detection should be incorporated into the management of individuals with AD of cirrhosis. Hypercoagulability and hypo-fibrinolysis in individuals with AD of cirrhosis may be implicated in disease progression and haemostatic complications. In this prospective study, Zanetto et al. aimed to characterise haemostatic alterations in individuals with AD and to evaluate whether such alterations can predict ACLF and bleeding and/or thrombosis. One hundred and sixty-nine hospitalised individuals with AD were prospectively recruited and underwent an extensive haemostatic profiling. Inflammation severity was assessed by C-reactive protein. Haemostasis in individuals with AD was compared to that in controls (study part 1). Then individuals with AD were followed for 1 year and predictors of ACLF and bleeding and/or thrombosis were investigated (study part 2). Compared to controls, individuals with AD showed a more pronounced hypercoagulability, without differences in fibrinolysis activation. During the follow-up, 55 patients developed ACLF. CLIF-C AD, C-reactive protein, and Child-Pugh were independently associated with ACLF. A predictive model combining these variables (Padua model) accurately identified patients at higher risk of ACLF. The prognostic value of the Padua model was validated in an independent, bicentric European cohort. To conclude, inflammation severity, and not coagulopathy, seems to be the most important predictor of ACLF and bleeding in AD. After recovery from overt hepatic encephalopathy (HE), minimal HE (MHE), which impairs quality of life (QoL), can persist. Fagan et al. aim was to determine the impact of albumin vs. saline on MHE and QoL in individuals with prior overt HE already on standard of care in a double-blind, placebo-controlled, randomised clinical trial. Outpatients with cirrhosis and prior overt HE, MHE and hypoalbuminemia already on treatment for HE were included. Patients on regular IV albumin infusions were excluded. Participants were randomised 1:1 to receive either weekly infusions of 25% IV albumin 1.5 g/kg or saline over 5 weeks and then 1-week post-infusion. Forty-eight participants were randomised and were balanced at baseline, including for HE-medication use. Adverse events were similar, with MELD score and ammonia remaining stable between/within groups. Albumin levels increased and ischaemia-modified albumin decreased only in the albumin group at end of treatment and end of study vs. baseline. MHE improvement was greater in the albumin group. Sickness Impact Profile total and psychosocial, but not physical, domain improved in the albumin but not the placebo group while markers of inflammation and endothelial dysfunction were reduced in the albumin group. Thus, albumin infusions were associated with improved cognitive function and psychosocial QoL in outpatients with prior HE and current MHE. Liver resection for intrahepatic cholangiocarcinoma (iCCA) may or may not include lymphadenectomy when imaging procedures show no obvious signs of lymph node metastasis. To answer that question, Sposito et al. retrospectively studied more than 700 consecutive patients from 5 experienced tertiary care centres that were followed for a median of 33 months. 59% of them received adequate lymphadenectomy (≥6 retrieved lymph nodes) while 61% did not, and overall and recurrence-free survival of these two subgroups were compared using inverse probability of treatment weighting. Patients with nodal metastasis at pathology, had better overall survival (28 vs. 23 months) and disease-free survival (13 vs. 9 months) if they received adequate lymphadenectomy. Importantly, the benefit of adequate lymphadenectomy was shown in individuals without chronic liver disease, and in those with less advanced tumours (solitary, smaller than 5 cm and withCA 19.9 <200 U/ml). These results support the routine practice of adequate lymphadenectomy during liver resection in individuals with iCCA, even in the absence of suspected lymph-node metastasis at imaging. Heritability of biliary tract cancer (BTC) is uncertain and homologous recombination deficiency has not been investigated in BTC. Studying 27 cancer-predisposing genes in 1,292 Japanese individuals with BTC and 37,583 controls without a personal or family history of cancer, Nakagawa et al. identified 317 pathogenic germline variants and found that 5.5% of individuals with BTC and 1.4% of controls had at least one pathogenic germline variant. Pathogenic variants in BRCA1, BRCA2, APC, and MSH6 were enriched in BTC while variants of unknown significance were not. Carriers of pathogenic variants in these genes were younger and had increased personal and family histories of breast cancer. APC variants were predominantly found in ampullary cancers and the other genes had similar distributions across tumour locations. Whole-genome sequencing of 45 BTCs showed that three BTCs with pathogenic germline variants in BRCA2 and PALB2 accompanied with loss of heterozygosity displayed homologous recombination deficiency and had family histories of cancer. In contrast, seven BTCs with homologous recombination-related variants of unknown significance were predicted as homologous recombination proficient. Exploring the pathogenesis of CCA, Duwe, Munoz-Garrido et al. have focused on the dysregulation of the miRNome in CCA, including its impact on transcriptome homeostasis and cell behaviour. They performed small RNA sequencing on more than 200 samples of resected CCAs, peritumoral liver, and normal liver and found that 13% of miRs were differentially expressed between CCA and the surrounding liver. Unsupervised hierarchical clustering of tumour miRNomes identified three subgroups, with different prognosis. In parallel, they performed high-throughput miR mimic screens in human cholangiocyte cultures. Integration of patient transcriptomes and miRseq together with miR screening data identified an oncogenic miR (miR-27a-3p) which had consistently increased expression and activity in several cohorts. FoxO signalling was predominantly downregulated by miR-27a-3p in CCA while miR-27a knock-out increased FOXO1 levels and arrested tumour growth. Therefore, miR-27a-3p could be a therapeutic target in CCA. Claudin-1 (CLDN1) is a protein expressed in the membrane of hepatocytes, inside and outside of tight junctions. Roehlen et al. have investigated the role of non-junctional CLDN1 as a therapeutic target in hepatocellular carcinoma (HCC) using humanised monoclonal antibodies (mAbs) which target specifically the extracellular loop of human non-junctional CLDN1. CLDN1 was found overexpressed in HCC, particularly in high-grade tumours, and showed aberrant non-junctional localisation in HCC in contrast to primarily junctional localisation in the healthy liver. Highly CLDN1-expressing HCC cells exhibited a progenitor or stem cell-like phenotype and an immune-low or -excluded tumour microenvironment. In cell line-based models and in patient-derived 3D ex vivo models, targeting non-junctional CLDN1 showed a potent antitumor effect. This effect was further confirmed in several cell line- and patient-derived xenograft mouse models. Mechanistic studies suggested that CLDN1 regulates tumour stemness, metabolism, oncogenic signalling and perturbs the tumour immune microenvironment. These results open the door to the therapeutic development of CLDN1-directed mAbs. Immunotherapy with immune checkpoint inhibitors is the current standard of care for advanced HCC but not all patients respond to this therapy. To investigate how the immune HCC microenvironment is influenced by the genetic tumour background, Yuen, Chiu, Law et al. have established a collection of mouse HCCs using genome-editing systems, with genotypes that closely simulate the genetic composition found in human HCCs, including alterations in TP53, PTEN, TSC2, Keap1 and APC genes. Distinct CD8+ T-cell infiltration levels and chemokine expression profiles were observed in tumours with different genetic driver mutations. Tumours with different genotypes were considered cold or hot based on the level of CD8+ T-cell infiltration. PD-1 inhibition had no effect in cold tumours and was effective in hot tumours. Tumour-infiltrating CD8+ T cells from Keap1 knocked-out/MYC overexpressed tumours expressed higher levels of proinflammatory chemokines and exhibited enrichment of progenitor exhausted CD8+ T-cell phenotype compared to Trp53 knocked-out/MYC overexpressed tumours. Interestingly, sorafenib sensitised Trp53 knocked-out/MYC overexpressed tumours to anti-PD-1 treatment, providing a rationale to investigate the influence of the genetic background on responses to combination therapies with immune checkpoint inhibitors and tyrosine kinase inhibitors. Patrizia Burra∗ at Department of Surgery, Oncology and Gastroenterology, Padua University Hospital, Padua, Italy. Frank Tacke at Department of Hepatology and Gastroenterology, Charité Universitätsmedizin Berlin, Berlin, Germany. Vlad Ratziu at Insitute for Cardiometabolism and Nutrition, Sorbonne Université and Hospital Pitié Salpêtrière, Paris, France. Stefan Zeuzem at Department of Medicine I, Goethe University Hospital, Frankfurt, Germany. Bruno Sangro at Liver Unit, Clinica Universidad Navarra and CIBEREHD, Pamplona, Spain. Paolo Angeli at Unit of Internal Medicine and Hepatology, University of Padua, Padua, Italy.