The systemic and hepatic alternative renin–angiotensin system is activated in liver cirrhosis, linked to endothelial dysfunction and inflammation

We aimed to assess the systemic and hepatic renin-angiotensin-system (RAS) fingerprint in advanced chronic liver disease (ACLD). This prospective study included 13 compensated (cACLD) and 12 decompensated ACLD (dACLD) patients undergoing hepatic venous pressure gradient (HVPG) measurement. Plasma components (all patients) and liver-local enzymes (n = 5) of the RAS were analyzed using liquid chromatography–tandem mass spectrometry. Patients with dACLD had significantly higher angiotensin (Ang) I, Ang II and aldosterone plasma levels. Ang 1–7, a major mediator of the alternative RAS, was almost exclusively detectable in dACLD (n = 12/13; vs. n = 1/13 in cACLD). Also, dACLD patients had higher Ang 1–5 (33.5 pmol/L versus cACLD: 6.6 pmol/L, p < 0.001) and numerically higher Ang III and Ang IV levels. Ang 1–7 correlated with HVPG (ρ = 0.655; p < 0.001), von Willebrand Factor (ρ = 0.681; p < 0.001), MELD (ρ = 0.593; p = 0.002) and interleukin-6 (ρ = 0.418; p = 0.047). Considerable activity of ACE, chymase, ACE2, and neprilysin was detectable in all liver biopsies, with highest chymase and ACE2 activity in cACLD patients. While liver-local classical and alternative RAS activity was already observed in cACLD, systemic activation of alternative RAS components occurred only in dACLD. Increased Ang 1–7 was linked to severe liver disease, portal hypertension, endothelial dysfunction and inflammation.

While advanced chronic liver disease (ACLD) is initially asymptomatic and therefore considered compensated (cACLD), it eventually progresses into decompensated disease (dACLD), displaying characteristic symptoms which arise from portal hypertension such as ascites and bleeding of oesophageal varices 1,2 . Aside from symptomdependent treatment, the current therapies of dACLD focus on lowering portal pressure and include pharmacological treatment using non-selective betablockers 3,4 as well as invasive procedures such as the implantation of a transjugular intrahepatic portosystemic shunt (TIPS) 5 .
A better understanding of the role of the renin-angiotensin-system (RAS) including its classical and alternative components in ACLD may reveal novel therapeutic targets. Through renin cleaving hepatic angiotensinogen, angiotensin I (Ang I) is created and further converted to angiotensin (Ang) II by angiotensin-converting enzyme (ACE). Ang II is also frequently converted by the locally active chymase 6 and mediates its effects through binding to members of the angiotensin-receptor family, which lead to vasoconstriction and aldosterone secretion facilitating sodium and water retention in the kidneys as well as mediating inflammation 7 . By acting not only systemically but also locally, the RAS is involved in growth and remodeling processes in the heart and blood vessels 8 . A permanently upregulated RAS is strongly associated with various health complications such as hypertension, diabetes and aging processes 9 . In liver disease, elevated Ang II is associated with both hepatic resistance as well as portal pressure 10 . Additionally, an increased expression of RAS components has even been observed in activated hepatic stellate cells (HSC) following liver injury, which enables local synthesis of Ang II mediating fibrosis of the liver 11,12 .
As opposed to the classical RAS-axis, the metalloprotease angiotensin-converting enzyme 2 (ACE2) has been identified as the key enzyme of the alternative RAS. It is able to use Ang II as substrate, converting it into Angiotensin 1-7 (Ang 1-7) 7 , which binds to the mas-receptor (masR) and has vasodilative and anti-inflammatory qualities 13 .
Ang 1-7 can also be formed by way of conversion through neprilysin (NEP) from Ang I. Upregulated Ang 1-7 is associated with lower incidence of steatosis 14 and even plays a role in the inhibition of Ang II-mediated pathways facilitating fibrosis in cultured HSC 15 . Hence it can be argued, that ACE2 and Ang 1-7 as part of the alternate RAS axis act as a compensating measure to the classical RAS in the liver.
There was no significant difference in plasma levels of Ang III and Ang IV between patients with cACLD and patients with dACLD ( Table 2).

Differences of RAS activity in the plasma and liver tissue of ACLD patients
However, interestingly, the enzymatic activity of the RAS was more complex in the liver tissue, as outlined in Fig. 3.

Discussion
In this study, we assessed the classical and alternative RAS fingerprint in well-characterized patients with ACLD, divided into two cohorts: cACLD versus dACLD. We found that patients with dACLD had pronounced systemic activation of the classical RAS, and activation of the alternative RAS, Ang 1-7 being detectable in 83.3% of patients with dACLD (and in only one patient with cACLD [7.7%]). Importantly, we also demonstrated that, compared to other RAS components, Ang 1-7 showed the strongest correlations with parameters of liver disease severity, portal hypertension, LSM and endothelial dysfunction. Moreover, Ang 1-7 was the only angiotensin that correlated with IL-6 as a marker for systemic inflammation. Finally, suggest that the local activity of the classical and non-classical RAS in the liver is different from systemic RAS activity. Of note, patients with cACLD and patients with clinically significant portal hypertension exhibited the highest hepatic activities of classical and alternative RAS enzymes.
A systemic activation of the classical RAS in patients with dACLD, particularly in patients with ascites is welldocumented [16][17][18][19] . In line with this, patients with dACLD in our cohort showed pronounced systemic elevation of classical RAS components including PRA-S, Ang I, Ang II and aldosterone. Higher classical RAS activity in dACLD aggravates hyperdynamic circulation, as it promotes splanchnic vasodilation and higher intrahepatic vascular resistance 19 . Moreover, ACE inhibition lowers portal pressure in cirrhotic patients [20][21][22] .
To date, data on alternative RAS activity in patients with ACLD are still scarce and evidence of alternative RAS activation in liver cirrhosis is mostly based on animal studies 19 . Our study shows that in the plasma of patients with dACLD, alternative RAS components including Ang 1-7 and Ang 1-5 were higher than in patients with cACLD. Similarly, activation of the alternative RAS was observed in diabetic patients with vascular disease and hypertension, in which ACE2 upregulation was interpreted as an attempt to counteract disease progression 23 . Whether this is the case in patients with dACLD or whether increased Ang 1-7 and Ang 1-5 are simply byproducts of overall RAS activation cannot be answered definitively with our data and requires further research.
At the same time, Ang III and Ang IV levels did not differ significantly between patients with cACLD and dACLD, but were numerically higher in patients who were decompensated. This result is in line with previous findings of a study displaying activation of the Ang 1-7/Mas receptor axis in 7 patients with liver cirrhosis undergoing liver transplantation 24 and another study that showed increased levels of Ang 1-7 in the plasma of 9 cirrhotic patients and 23 non-cirrhotic patients with hepatitis C compared to healthy controls 25 . Moreover, www.nature.com/scientificreports/ several animal studies demonstrated systemic elevation of alternative RAS components including Ang 1-7 and ACE2 in chronic liver disease 26,27 . Interestingly, previous studies suggest an activation of the systemic alternative RAS already in pre-cirrhotic stages of chronic liver disease, particularly in viral hepatitis 25 . In contrast, our cohort of patients who had strictly compensated ACLD of majorly viral etiology with mostly subclinical portal hypertension showed virtually no alternative RAS activity in their plasma, with Ang 1-7 being undetectable in 92.3% of patients in this cohort. This may be due to the fact that these were primarily patients after cure of chronic hepatitis C with correspondingly lower levels of (hepatic) inflammation.
Importantly, there were statistical correlations between Ang 1-7 and parameters of liver disease severity (MELD), LSM, portal hypertension (assessed by the gold standard HVPG) and endothelial dysfunction (assessed by its well-established surrogate biomarker VWF antigen [28][29][30][31], suggesting a link between activation of the alternative RAS and worse outcomes of patients with ACLD 32,33 . Moreover, unlike other components of the RAS, Ang 1-7 correlated with IL-6, linking alternative RAS activation in ACLD to systemic inflammation 1 . Of note, systemic Ang II did not correlate with MELD or HVPG, indicating a special status of particularly the systemic alternative RAS in ACLD and portal hypertension. Interestingly, HVPG was the only parameter independently associated with Ang 1-7 in linear regression analysis, linking portal hypertension to alternative RAS activation.
An important role of the local intrahepatic classical RAS has been reported to drive an increase in portal hypertension 19,34 . Moreover, ACE inhibitors were associated with a decreased risk of liver-related events in nonalcoholic fatty liver disease 35 . There has been evidence that ACE2 is upregulated in livers of humans and rats 24,26,36 with chronic liver disease. Furthermore, NEP was higher in livers of cirrhotic animals 37 . Our study, examining liver biopsies of 2 individuals with cACLD and 2 with dACLD, demonstrated that hepatic expression of enzymes of the classical and alternative RAS is complex and not limited to patients with dACLD. Despite exhibiting little classical and alternative RAS activity in their plasma, patients with cACLD had the highest hepatic activity of ACE, chymase, NEP and ACE2, suggesting a pronounced activation of intrahepatic classical and alternative RAS already in early stages of ACLD. This finding is of high relevance, as Ang II is known to induce and promote liver fibrosis via contraction, proliferation and activation of HSCs 11,38,39 and hepatic chymase has also been linked to liver fibrosis in animals and humans 40,41 . Moreover, induction of intrahepatic ACE2 activity has been suggested to exert antifibrotic effects 42 and NEP seems to be critically involved in the development of portal hypertension 37 and kidney function in cirrhotic animal models 43 . Modifications such as ACE2 activation or Ang 1-7 supplementation seemed to have beneficial effects in murine models 44,45 .Thus, targeted pharmacological interventions targeting RAS components in cACLD patients may attenuate disease progression.
Our study also has limitations. Firstly, we recruited strictly compensated patients with cACLD, as well as patients with dACLD with a history of at least two decompensation events. Accordingly, this study does not investigate RAS activation patterns in other stages of ACLD. Secondly, this is a pilot study, including only a small number of patients with cACLD and dACLD. This limitation applies for the systemic (plasma) RAS fingerprint results and even more so for the intrahepatic RAS characterization (assessed in only 5 patients). Thus, the limited sample size represents an important limitation of our study and further studies are required to examine the systemic RAS fingerprint throughout all stages of ACLD, as well as the intrahepatic RAS in cACLD and dACLD. Still we want to emphasize that these data are novel and of important value for the liver community. Thirdly, collinearity between dACLD and other parameters, which were correlated with Ang 1-7, cannot be excluded, since these parameters increase with ACLD severity 1,46 . To account for this, linear regression analysis considering dACLD as a potential confounding factor was conducted. Also, patients with dACLD in our study still had relatively high median MAP, which indicates that patients with hyperdynamic circulation may be underrepresented in our study. Moreover, the majority of patients with dACLD had intake of aldosterone antagonists. This represents a confounding factor regarding the aldosterone levels reported in this study. Finally, plasma aldosterone levels were not available for all patients, but our results are well in line with previous studies 18,47 .
In conclusion, in this comprehensive characterization of the RAS fingerprint in patients with cACLD and dACLD, we were able to show that in addition to the classical RAS, the alternative RAS is activated in the plasma of those with dACLD. Importantly, Ang 1-7 correlated with parameters of liver disease severity, portal hypertension, endothelial dysfunction and inflammation. Hepatic activation of classical and alternative RAS components is complex and ACE, chymase, ACE2 and NEP are already upregulated in liver tissue of patients with cACLD.

Methods
Study population and sampling. This prospective single center study included patients who underwent hepatic venous pressure gradient (HVPG) measurement at the Vienna Hepatic Hemodynamic Lab (Division of Gastroenterology and Hepatology, Medical University of Vienna) between 06/2017 and 12/2021. ACLD was defined as liver stiffness measurement (LSM) ≥ 10 kPa and/or hepatic venous pressure gradient > 5 mmHg and/ or histological fibrosis stage F3/F4. While exclusively strictly compensated ACLD patients in Child-stage A with without intake of diuretics or ACE inhibitors were evaluated for inclusion in the cACLD group, only clinically stable decompensated patients with ascites and another hepatic decompensation event (hepatic encephalopathy, refractory ascites, history of variceal bleeding, history of spontaneous bacterial peritonitis) were considered for the dACLD group. Patients with evidence of infection at the time of HVPG measurement, as well as patients with portal vein thrombosis, TIPS, hepatocellular carcinoma, or liver transplantation were excluded.
We aimed to measure components of both the classical and the alternate RAS from local and systemic sources. Liquid biopsies were collected as EDTA plasma through peripherally drawn venous blood samples. Liver tissue was collected by way of transjugular liver biopsy as previously described 48 . Both plasma and liver tissue samples were immediately stored at −80 °C. www.nature.com/scientificreports/ HVPG measurement. HVPG was measured according to a standardized procedure 48 . After local anesthesia, a catheter introducer sheath was placed in the right internal jugular vein. Using fluoroscopic guidance, a specifically designed angled-tip balloon catheter 49 and was advanced into a large hepatic vein via the inferior vena cava. HVPG was calculated by subtracting free from wedged hepatic venous pressure. For further analyses, the mean of 3 measurements was used. Hemodynamic parameters including mean arterial pressure (MAP) were measured at the beginning of HVPG measurement.
Assessment of liver stiffness measurement, liver function and endothelial dysfunction. Liver stiffness measurement (LSM) was measured via vibration-controlled transient elastography (VCTE) using Fibroscan© (Echosens, Paris, France). Severity of portal hypertension was evaluated by calculating the HVPG during liver vein catheterization. Serum von-Willebrand-factor was measured for evaluation of endothelial dysfunction and as an additional surrogate marker for portal pressure 33 . Child-Turcotte-Pugh (CTP) and Model for End stage Liver Disease score (MELD) of 2016 32 were used to assess severity of ACLD.
Laboratory analysis. Measurement of routine laboratory parameters such as interleukin-6 (IL-6) were performed by the facility's department of laboratory medicine. For assaying enzyme activities in tissue samples, liver tissue samples were homogenized under liquid nitrogen to enable stabilized peptide extraction and assayed as described previously 50 . Quantification of equilibrium RAS metabolites Ang I, Ang II, Ang III Ang IV as well as Ang 1-5 and Ang 1-7 in plasma was performed by liquid chromatography-tandem mass spectrometry (Attoquant Diagnostics, Vienna, Austria) using previously established protocols 51 . Plasma renin activity surrogate (PRA-S), a well-established biomarker for plasma renin activity was calculated as the sum of Ang I and Ang II 52,53 and systemic ACE surrogate (ACE-S) was calculated as Ang II divided by Ang I 53,54 . Plasma aldosterone was assessed by chemiluminsescence immunoassay (DiaSorin, Liaision XL, Saluggia, Italy). Of note, plasma aldosterone levels were not available in all patients, as they were only measured starting 04/2018.

Statistical analysis.
Categorical variables were reported as number (n) of patients and % of patients with the characteristic of interest. Continuous data was presented as median and interquartile range (IQR). The data sets were tested for normal distribution with D' Agostino & Pearson and Shapiro-Wilk normality test. Mann-Whitney U test was used for comparing continuous variables without normal distribution between two groups and group comparisons of categorical variables were conducted using Pearson's Chi-squared test. Correlations between two metric non-normally distributed variables were assessed with Spearman's Rho. Linear regression analysis was conducted to investigate factors associated with Ang 1-7. Parameters that were at least tendentially associated with Ang 1-7 in univariate analysis (p < 0.100) were included in the multivariate model. GraphPad Prism 8 (Graphpad Software, La Jolla, CA, USA) and IBM SPSS 25.0 statistic software (IBM, Armonk, NY, USA) were used for statistical analysis. Two-sided p-values equal to or below 0.05 were considered as statistically significant.
Ethics. All included patients were part of the Vienna Cirrhosis Study (VICIS; NCT: NCT03267615), a prospective observational trial. This study was performed in accordance with the current version of the Helsinki declaration and approved by the local Ethics Committee of the Medical University of Vienna (No. 1262/2017). All patients gave their informed written consent before study inclusion.

Data availability
The data is available upon reasonable request to the corresponding author. www.nature.com/scientificreports/