Phillipp Hartmann ,Sonja Lang ,Rort Schirwagn ,Sain Klin ,Michal Praktiknjo ,Jonl Tricka ,f,Brnd Schnal

a Department of Pediatrics, University of California San Diego, La Jolla, CA, USA

b Division of Gastroenterology, Hepatology & Nutrition, Rady Children’s Hospital San Diego, San Diego, CA, USA

c Department of Medicine, University of California San Diego, La Jolla, CA, USA

d Department of Gastroenterology and Hepatology, Faculty of Medicine, University of Cologne, University Hospital Cologne, Cologne, Germany

e Department of Internal Medicine B, University of Münster, Münster, Germany

f European Foundation for the Study of Chronic Liver Failure, Barcelona, Spain

g Department of Medicine, VA San Diego Healthcare System, San Diego, CA, USA

Keywords: Liver disease Acute decompensation Acute-on-chronic liver failure Microbiome Model for end-stage liver disease

ABSTRACT Background: Cirrhosis with acute decompensation (AD) and acute-on-chronic liver failure (ACLF) are characterized by high morbidity and mortality. Cytolysin,a toxin from Enterococcus faecalis ( E. faecalis ),is associated with mortality in alcohol-associated hepatitis (AH). It is unclear whether cytolysin also contributes to disease severity in AD and ACLF.Methods: We studied the role of fecal cytolysin in 78 cirrhotic patients with AD/ACLF. Bacterial DNA from fecal samples was extracted and real-time quantitative polymerase chain reaction (PCR) was performed. The association between fecal cytolysin and liver disease severity in cirrhosis with AD or ACLF was analyzed.Results: Fecal cytolysin and E. faecalis abundance did not predict chronic liver failure (CLIF-C) AD and ACLF scores. Presence of fecal cytolysin was not associated with other liver disease markers,including Fibrosis-4 (FIB-4) index,‘Age,serum Bilirubin,INR,and serum Creatinine (ABIC)’ score,Child-Pugh score,model for end-stage liver disease (MELD) nor MELD-Na scores in AD or ACLF patients.Conclusions: Fecal cytolysin does not predict disease severity in AD and ACLF patients. The predictive value of fecal cytolysin positivity for mortality appears to be restricted to AH.

Introduction

Intestinal dysbiosis,or an imbalance of beneficial and potentially pathogenic microbes [ 1,2 ],is present in several hepatic diseases [ 3,4 ]. Besides fungal [5–8] and viral dysbiosis [ 9,10 ],bacterial dysbiosis plays a central role in the pathogenesis of liver disease [ 11,12 ]. Particularly liver cirrhosis is associated with dysbiosis and increased gut permeability [ 13,14 ]. We recently found that cytolysin,a toxin fromEnterococcusfaecalis(E.faecalis),directly lyses hepatocytes in alcohol-associated liver disease (ALD)models [15] . Cytolysin-positiveE.faecalisis a significant predictor for mortality in alcohol-associated hepatitis (AH) with 180-day mortality of 89% in cytolysin-positive AH patients compared with only 3.8% in cytolysin-negative AH patients [15] . It is unclear whether its predictive value is confined to ALD,or AH specifically. Cytolysin-positiveE.faecalisdoes not associate with disease severity in nonalcoholic fatty liver disease (NAFLD) [16] . The role of cytolysin-positiveE.faecalisis unknown in other liver diseases,including cirrhosis with acute decompensation (AD) or acute-onchronic liver failure (ACLF). The pathophysiology of ACLF is still incompletely understood and it is a difficult-to-treat condition with a very high transplant-free 90-day mortality up to 50% [17] . AH is one of the most common triggers for the development of ACLF and,together with bacterial infection,is present in almost all patients who develop ACLF [ 18,19 ].

The present study aimed to investigate the role and predictive value of cytolysin in cirrhosis with AD and ACLF.

Methods

Patients

Fecal samples were collected between September 1998 and August 2003 at the Department of Internal Medicine I,University of Bonn,Germany,when patients were evaluated for a transjugular intrahepatic portosystemic shunt (TIPS). Inclusion criteria were over 18 years of age with clinical signs of liver cirrhosis and an indication for TIPS insertion. Exclusion criteria were the presence of systemic infection,severe hepatic encephalopathy of unknown cause,severe hyperbilirubinemia,pulmonary hypertension or pregnancy as previously described [20] .

Bacterial DNA extraction and real-time quantitative polymerase chain reaction (PCR)

Human stool samples were stored at -80 °C,DNA from fecal samples was extracted and qPCR was performed as described [15] .Primer sequences,E.faecalisforward: 5 ′ -CGC TTC TTT CCT CCC GAG T-3 ′ ; reverse: 5 ′ -GCC ATG CGG CAT AAA CTG-3 ′ .E.faecaliscylLL forward: 5 ′ -CTG TTG CGG CGA CAG CT-3 ′ ; reverse: 5 ′ -CCA CCA ACC CAG CCA CAA-3 ′ .E.faecaliscylLS forward: 5 ′ -GTA AAA TAA GTA AAA TCA AGA AAA CTA TTA CTC-3 ′ ; reverse: 5 ′ -CAA AAG AAG GAC CAA CAA GTT CTA ATT-3 ′ [ 15,16 ].

Statistical analysis

Statistical analyses were performed using R statistical software(R version 1.3.1093,2020 the R Foundation for Statistical Computing). Results were expressed as median and interquartile range for each continuous outcome and as number and percentage for factor variables,if not stated otherwise. Area under the receiver operating characteristic (ROC) curve (AUC),best threshold to maximize the Youden Index,sensitivity,specificity,accuracy,precision,positive predictive value,negative predictive value,andPvalue between two AUCs per DeLong method were calculated using the pROC library in R. Kaplan-Meier curves to visualize survival were created employing the survival library in R. Mean decrease Gini score and mean decrease accuracy were calculated with the randomForest library,and odds ratios (ORs) with the question library in R. Continuous variables were compared using Mann-WhitneyUtest. Categorical variables were compared using Pearson’s Chi-square test. APvalue<0.05 was considered statistically significant.

Results

Study population with cirrhosis and AD or ACLF

The study population consisted of 63 AD patients and 15 ACLF patients ( Table 1 ). All subjects were enrolled when they received TIPS to treat complications of portal hypertension,most frequently for ascites (47.4%) and variceal bleeding (26.3%). Half of the subjects were male (52.6%),the median age was 56.5 years. Among them,57.7% had ALD. Liver cirrhosis was cryptogenic in 9.0%,due to hepatitis C virus (HCV) in 6.4%,and mixed in 6.4%. ACLF patients had a significantly higher model for end-stage liver disease(MELD) score (medians 16.20 vs. 10.20,P= 0.001),MELD-Na score(18.20 vs. 10.20,P<0.001),European Foundation for the study of chronic liver failure (CLIF-C) organ failure score (8.00 vs. 6.00,P<0.001),higher grade of hepatic encephalopathy,significantly more frequent renal failure (46.7% vs. 0%,P<0.001),circulatory failure (33.3% vs. 1.6%,P= 0.001),and respiratory failure (13.3%vs. 0%,P= 0.035) than AD patients. The median CLIF-C AD score was 48.08 in the AD group and the median CLIF-C ACLF score was 43.83 in the ACLF group. White blood count (WBC) and serum neutrophils were significantly higher in the ACLF group compared with the AD group (7.55 × 109/L vs. 5.39 × 109/L,P= 0.019 and 5.47 × 109/L vs. 3.67 × 109/L,P= 0.045,respectively). Both cohorts had a similar rate of antibiotic use but ACLF patients had a significantly higher use of rifaximin (40.0% vs. 12.7%,P= 0.023).However,the groups did not significantly differ regarding fecal cytolysin positivity (ACLF 6.7% vs. AD 15.9%,P= 0.681). Only one of the 78 patients died within 90 days after enrollment.

Table 1Baseline demographic and laboratory data of the study population.

Presence of fecal cytolysin does not predict liver disease severity in cirrhosis with AD or ACLF per CLIF-C AD and ACLF scores

Since fecal cytolysin positivity predicts mortality in alcoholassociated hepatitis (AH) patients [ 15,21 ],we aimed to examine its role in patients with cirrhosis and AD or ACLF. To determine disease severity in patients with acutely decompensated cirrhosis,the CLIF-C AD score was employed,which has a significantly better performance for predicting mortality in AD patients than Child-Pugh score (CPS),MELD,and MELD-Na scores [22] . Similarly,disease severity in ACLF patients was quantified using the CLIF-C ACLF score,which also has better predictive value in ACLF patients than CPS,MELD,and MELD-Na [ 23,24 ]. To create binary disease measures for the AD and ACLF groups,“High CLIF-C AD” and “Low CLIF-C AD” as well as “High CLIF-C ACLF” and “Low CLIF-C ACLF”were designated using the respective medians as cutoffs ( Table 1 ).In our cohort,fecal cytolysin had a worse performance than all other tested liver disease markers [including Fibrosis-4 index (FIB-4),‘Age,serum Bilirubin,INR,and serum Creatinine’ (ABIC) score,CPS,MELD,MELD-Na] in AD patients with a low AUC of 0.53 and low best Youden Index of 0.06 compared with the best performer CPS with an AUC of 0.75 and best Youden Index of 0.47 ( Fig. 1 A,Table 2 ). The performance of fecal cytolysin was similar to fecalE.faecalisabundance ( Fig. 1 A,Table 2 ). Likewise,its predictive performance was poor in the ACLF group with an AUC of 0.56 and best Youden Index of 0.12,whereas the ABIC score had a high AUC of 0.83 and best Youden Index of 0.58 ( Fig. 1 B,Table 3 ).

Fig. 1. Presence of fecal cytolysin does not predict liver disease severity in cirrhosis with acute decompensation or acute-on-chronic liver failure per CLIF-C AD and ACLF scores. A: ROC curves of liver disease markers for high CLIF-C AD score in AD patients (cytolysin n = 63,E. faecalis n = 63,FIB-4 n = 60,ABIC n = 63,Child-Pugh n = 59,MELD score n = 63,and MELD-Na score n = 63). B: ROC curves of liver disease markers for high CLIF-C ACLF score in ACLF patients (cytolysin n = 15,E. faecalis n = 15,FIB-4 n = 14,ABIC n = 14,Child-Pugh n = 15,MELD score n = 14,and MELD-Na score n = 14). ABIC: ‘Age,serum bilirubin,INR,and serum creatinine score’; ACLF: acuteon-chronic liver failure; AD: acute decompensation; CLIF-C: European Foundation for the study of chronic liver failure; FIB-4: Fibrosis-4 index; INR: international normalized ratio; MELD: model for end-stage liver disease; MELD-Na: sodium-adjusted model for end-stage liver disease; ROC: receiver operating characteristic.

Table 2High CLIF-C AD score predictors.

Table 3High CLIF-C ACLF score predictors.

Next,we used the best threshold that maximized the Youden Index in AD ( Table 2 ) and calculated the ORs for a high CLIF-C AD score for values exceeding the best threshold for the respective marker. All ORs for high liver disease markers – but not for cytolysin positivity or highE.faecalisabundance – were significant per simple logistic regression: cytolysin positivity [OR = 1.56,95% confidence interval (CI) 0.40-6.69;P= 0.53],highE.faecalis(OR = 2.06,95% CI: 0.66-6.95;P= 0.22),high FIB-4 (OR = 0.21,95% CI: 0.07-0.62;P= 0.006),high ABIC (OR = 9.92,95% CI: 2.39-68.18;P= 0.005),high CPS (OR = 8.36,95% CI: 2.63-30.85;P<0.001),high MELD (OR = 6.68,95% CI: 2.16-23.91;P= 0.002),and high MELD-Na (OR = 6.68,95% CI: 2.16-23.91;P= 0.002) ( Fig. 2 A).As high MELD and high MELD-Na had the same values for OR and significance,the MELD/MELD-Na score was only used once for multiple logistic regression. After multiple logistic regression using all markers,high FIB-4 (OR = 0.05,95% CI: 0.003-0.330;P= 0.009),high ABIC (OR = 37.73,95% CI: 3.98-1033.18;P= 0.007),and high MELD/MELD-Na (OR = 15.88,95% CI: 1.88-364.24;P= 0.026) retained significance,whereas cytolysin positivity (OR = 0.83,95% CI:0.10-6.40;P= 0.86),highE.faecalis(OR = 1.99,95% CI: 0.26-19.15;P= 0.52),and high CPS (OR = 3.19,95% CI: 0.58-20.22;P= 0.19)were not significant ( Fig. 2 B). ORs for high CLIF-C ACLF for the ACLF group using the aforementioned markers could not be calculated or were not significant (not shown).

Fig. 2. Cytolysin positivity does not confer higher odds ratio for high CLIF-C AD score. Odds ratios for high CLIF-C AD score per simple logistic regression (A) and multiple logistic regression (B) were calculated for cytolysin positivity and various liver disease markers for values that exceeded the best thresholds per Table 2 (95% confidence interval indicated by black brackets for each marker). Cytolysin n = 63,E. faecalis n = 63,FIB-4 n = 60,ABIC n = 63,Child-Pugh n = 59,MELD score n = 63,and MELD-Na score n = 63. ABIC: ‘Age,serum bilirubin,INR,and serum creatinine score’; AD: acute decompensation; CLIF-C: European Foundation for the study of chronic liver failure;FIB-4: Fibrosis-4 index; INR: international normalized ratio; MELD: model for end-stage liver disease; MELD-Na: sodium-adjusted model for end-stage liver disease. ?P <0.05,??P < 0.01,and ???P < 0.001.

Consistent with the findings above,presence of fecal cytolysin had a low feature importance for high CLIF-C AD and ACLF scores per random forest analysis with a very low mean decrease Gini score and mean decrease accuracy in relation to the other markers in AD ( Fig. 3 A,B) and ACLF ( Fig. 3 C,D).

Fig. 3. Fecal cytolysin is not an important feature for liver disease severity in cirrhosis with acute decompensation and acute-on-chronic liver failure per random forest analysis. Mean decrease Gini score ( A ) and mean decrease accuracy ( B ) by random forest analysis were quantitated for presence of cytolysin and multiple liver disease markers to determine their respective feature importance for high CLIF-C AD score (cytolysin n = 63,E. faecalis n = 63,FIB-4 n = 60,ABIC n = 63,Child-Pugh n = 59,MELD score n = 63,and MELD-Na score n = 63). Mean decrease Gini score ( C ) and mean decrease accuracy ( D ) for high CLIF-C ACLF score (cytolysin n = 15,E. faecalis n = 15,FIB-4 n = 14,ABIC n = 14,Child-Pugh n = 15,MELD score n = 14,and MELD-Na score n = 14). ABIC: ‘Age,serum bilirubin,INR,and serum creatinine score’;ACLF: acute-on-chronic liver failure; AD: acute decompensation; CLIF-C: European Foundation for the study of chronic liver failure; FIB-4: Fibrosis-4 index; INR: international normalized ratio; MELD: model for end-stage liver disease; MELD-Na: sodium-adjusted model for end-stage liver disease.

Other liver disease markers do not depend on fecal cytolysin positivity in cirrhosis with AD or ACLF

Next,we performed simple linear regression with fecal cytolysin as the independent variable and the liver disease markers FIB-4,ABIC,CPS,MELD,and MELD-Na as the dependent outcome variables. None of the liver disease markers were significantly associated with presence of fecal cytolysin in the combined AD and ACLF cohort ( Table 4 ) nor for the AD (Table S1) or ACLF cohorts separately (Table S2). As fecal cytolysin positivity predicts mortality in AH patients [ 15,21 ],we further assessed the predictive value of fecal cytolysin in patients with AD or ACLF of our cohort in whom liver disease was alcohol-associated. However,also in this subgroup,none of the liver disease markers were significantly associated with fecal cytolysin (Table S3). The cytolysin positivity in that subgroup was 13.3%,similar to 14.1% overall ( Table 1 ).Finally,various liver disease indicators [e.g. MELD score,CPS,laboratory markers alanine aminotransferase (ALT),aspartate aminotransferase (AST),gamma-glutamyltransferase (GGT),and alkaline phosphatase (AP)] were not significantly different between fecal cytolysin positive and fecal cytolysin negative subjects in the combined AD and ACLF cohort ( Fig. 4 ),the AD cohort only (Fig. S1),nor in the alcohol-associated combined AD and ACLF cohort (Fig. S2).Rifaximin use was significantly higher in the ACLF group compared with the AD group ( Table 1 ). To account for possible confounding by rifaximin,we excluded all patients that were given rifaximin and evaluated for an association between fecal cytolysin and the various liver disease markers,and again none of the disease measures depended on fecal cytolysin (not shown).

Fig. 4. Fecal cytolysin-positive patients with cirrhosis and acute decompensation or acute-on-chronic liver disease do not have more severe liver disease than cytolysinnegative patients. A: MELD score ( n = 77). B: MELD-Na score ( n = 77). C: FIB-4 ( n = 74). D: ABIC score ( n = 77). E: Child-Pugh score ( n = 74). F: ACLF grade ( n = 78). G:CLIF-C organ failure score ( n = 78). H: Hepatic encephalopathy ( n = 78). I: ALT ( n = 78). J: AST ( n = 75). K: GGT ( n = 77). L: AP ( n = 68). ABIC: ‘Age,serum bilirubin,INR,and serum creatinine score’; ACLF: acute-on-chronic liver failure; ALT: alanine aminotransferase; AP: alkaline phosphatase; AST: aspartate aminotransferase; CLIF-C: European Foundation for the study of chronic liver failure; FIB-4: Fibrosis-4 index; GGT: gamma-glutamyltransferase; HE: hepatic encephalopathy; INR: international normalized ratio;MELD: model for end-stage liver disease; MELD-Na: sodium-adjusted model for end-stage liver disease; OF: organ failure.

Table 4Simple linear regression of fecal cytolysin with liver disease markers for combined cohort

Discussion

We have shown previously that fecal cytolysin positivity predicts mortality in AH patients [ 15,21 ]. We therefore aimed to investigate its role in patients with cirrhosis and AD or ACLF. However,we found that fecal cytolysin did not predict disease severity in acutely decompensated cirrhosis or ACLF. We have also demonstrated previously that cytolysin positivity is not associated with disease severity in NAFLD [16],which indicates that cytolysin might only confer risk in ALD. However,even if we select subjects with alcohol-associated cirrhosis and AD or ACLF in the current cohort,no liver disease markers were associated with presence of fecal cytolysin. This further supports the idea that cytolysin positivity indicates a higher risk for worse outcome in AH specifically.Even though AH is a major trigger of AD and ACLF,the colonization withE.faecalisassociated with fecal cytolysin does not appear to be an important mechanism in AD and ACLF. This study shows once again that AD and ACLF are multifactorial,and besides AH as one of the triggers,other infections are also relevant inducers of AD and ACLF.

Strengths of this study include that this is a well-characterized AD and ACLF cohort. Moreover,the predictive value of fecal cytolysin for disease severity in this cohort was evaluated comprehensively from various angles and the results have been consistent throughout different methodologies. Limitations of the study include that the studied cohort had overall mild disease with only one dead subject in the first 90 days after enrollment,which appears low in the context of cirrhosis with AD or ACLF,compared with the transplant-free 90-day mortality up to 50% in other literature [17] . The low cytolysin positivity of only 14.1%of the overall study population (compared with up to 30% in AH) [15] likely did not contribute to the relatively low mortality,as cytolysin was not associated with any liver disease severity measures in our study cohort. It is possible that higher rifaximin use in the ACLF group might have contributed to a better survival in that group. Rifaximin is known to killE.faecalis[25] and reduce its translocation into the blood stream [26] . However,even after exclusion of rifaximin-exposed subjects,there was no association between fecal cytolysin and liver disease measures in our cohort.

In conclusion,fecal cytolysin positivity does not indicate higher disease severity in patients with AD and ACLF. Fecal cytolysin appears to exclusively predict worse outcome in AH.

Acknowledgments

None.

CRediT authorship contribution statement

PhillippHartmann: Formal analysis,Writing - original draft,Writing - review & editing.SonjaLang:Formal analysis,Writing- review & editing.RobertSchierwagen:Data curation,Writing- review & editing.SabineKlein:Data curation,Writing - review& editing.MichaelPraktiknjo:Data curation,Writing - review &editing.JonelTrebicka:Data curation,Writing - review & editing.BerndSchnabl: Supervision,Writing - review & editing.

Funding

This study was supported in part by National Institutes of Health (NIH) grant ( K12 HD85036 ); University of California San Diego Altman Clinical and Translational Research Institute (ACTRI)/NIH grant ( KL2TR001444 ); Pinnacle Research Award in Liver Diseases Grant ( PNC22-159963 ) from the American Association for the Study of Liver Diseases Foundation (to Hartmann P); Deutsche Forschungsgemeinschaft (DFG,German Research Foundation) fellowship ( LA 4286/1-1 ); the “Clinical and Translational Research Fellowship in Liver Disease” by the American Association for the Study of Liver Diseases (AASLD) Foundation (to Lang S),by National Institutes of Health grants ( R01 AA24726,R01 AA020703,U01 AA026939 ),and by Award Number BX004594 from the Biomedical Laboratory Research & Development Service of the VA Office of Research and Development ; Biocodex Microbiota Foundation Grant (to Schnabl B) and services provided by NIH centers ( P30 DK120515 and P50 AA011999 ). This study was also supported by the German Research Foundation (DFG) project ( 403224013 – SFB 1382 ) (to Trebicka J); the German Federal Ministry of Education and Research (BMBF) for the DEEP-HCC project (to Trebicka J); and the Hessian Ministry of Higher Education,Research and the Arts(HMWK) for the ENABLE and ACLF-I cluster projects (to Trebicka J). The MICROB-PREDICT (825694),DECISION (847949),GALAXY(668031),LIVERHOPE (731875),and IHMCSA (964590) projects (all to Trebicka J) have received funding from the European Union’s Horizon 2020 research and innovation program.

Ethical approval

The study protocol conformed to the ethical guidelines of the1975DeclarationofHelsinkiand was approved by the Human Research and Ethical Committee of the University of Bonn (121/14).Written informed consent was obtained from all patients.

Competing interest

Schnabl B has been consulting for Ferring Research Institute,HOST Therabiomics,Intercept Pharmaceuticals,Mabwell Therapeutics,Patara Pharmaceuticals and Takeda. Schnabl B’s institution UC San Diego has received grant support from Axial Biotherapeutics,BiomX,CymaBay Therapeutics,NGM Biopharmaceuticals,Prodigy Biotech and Synlogic Operating Company. Schnabl B is the founder of Nterica Bio. UC San Diego has filed several patents with Schnabl B and Lang S as inventors related to this work. The other authors declare no conflicts of interest.

Supplementary materials

Supplementary material associated with this article can be found,in the online version,at doi:10.1016/j.hbpd.2023.05.003 .