Introduction
Liver cirrhosis is the advanced stage of all chronic liver diseases [
1]. Common complications are ascites, esophageal varices, and renal impairment [
2]. Portal hypertension accounts for these severe consequences [
3]. Transjugular intrahepatic portosystemic shunt (TIPS) lowers portal pressure and is a valuable option for the treatment of patients with refractory ascites and bleeding esophageal varices [
4]. Adverse events of TIPS are hepatic encephalopathy [
4] and thrombosis [
5]. Liver tumors are considered a contraindication for TIPS, and this intervention was discussed to promote the onset of hepatocellular carcinoma (HCC) [
6].
Pentraxin 3 (PTX3) is an acute-phase protein, and thus its expression is strongly induced in inflamed and injured tissues. Unlike C-reactive protein (CRP), which is primarily synthesized in hepatocytes, PTX3 secretion was increased in hepatic stellate cells, neutrophils, and monocytes [
7]. Impaired hepatocyte function and accordingly CRP synthesis limit the diagnostic value of CRP in patients with decompensated liver cirrhosis [
8]. PTX3 is a reliable marker for the activation of immune cells and was supposed to be superior to CRP in these patients [
9].
PTX3 has an important function in tissue repair and wound healing. PTX3 deficiency was associated with increased clotting and fibrin deposition in line with a role of PTX3 in plasmin-mediated fibrinolysis [
10,
11]. Hepatic stellate cells are the main cells responsible for tissue repair in the liver [
12]. Accordingly, in experimental models of liver injury, PTX3 expression increased in these cells and contributed to enhanced synthesis of extracellular matrix proteins like collagens. Moreover, hepatic inflammation and injury improved [
11].
The prevalence of the metabolic syndrome, which is a risk factor for non-alcoholic fatty liver disease (NAFLD), is increasing. Plasma PTX3 was induced in patients with the metabolic syndrome. Of note, PTX3 was negatively correlated with high density lipoprotein [
13]. A further study described a positive association of PTX3 mRNA expression in white blood cells with low-density lipoprotein [
13,
14]. Accordingly, serum PTX3 was positively correlated with LDL-cholesterol in a cohort of type 2 diabetes patients [
15]. Notably, type 2 diabetes patients with NAFLD had serum PTX3 levels as high as patients with normal liver function illustrating an association of serum PTX3 with dyslipidemia rather than NAFLD [
13,
15]. Anyhow, a separate study described an association of plasma PTX3 with the stages of liver fibrosis in patients with NAFLD [
16]. Moreover, PTX3 was strongly induced in alcoholic liver disease and was related to the model for end-stage liver disease (MELD) score. Plasma PTX3 was positively correlated with hepatic PTX3 gene expression indicating that the liver contributed to higher systemic levels [
11]. In hepatitis C virus (HCV)-infected patients, plasma PTX3 did not differentiate mild from severe fibrosis [
17]. Accuracy of fibrosis assessment in HCV-infected patients was nevertheless improved by a score combining PTX3 levels, gamma-glutamyl transpeptidase/platelet count ratio, hyaluronic acid, and age [
18].
The prevailing opinion nowadays is that systemic PTX3 is induced in patients with chronic liver diseases [
11,
19] and has a positive predictive value for adverse clinical outcomes [
20,
21]. Liver cirrhosis is a risk factor for HCC, and PTX3 may also have a role herein [
17,
22].
Inflammation is a critical component of carcinogenesis and drugs which boost the anti-tumor immune response were successful in cancer therapy [
23]. Genetic diversity of killer-cell immunoglobulin-like receptor and human leukocyte antigen was related to the risk of HCV-related HCC further highlighting the role of the innate immune system in cancer [
24]. Indeed, PTX3 facilitated progression of different cancers, and PTX3 expression in hepatocellular carcinoma (HCC) tissues was positively associated with shorter survival time of the patients [
19]. Accordingly, plasma PTX3 was high in HCC compared to patients with severe liver fibrosis [
17]. Cancer-associated fibroblasts support tumor growth and metastatic disease. These cells were characterized by a high expression of profibrotic and proinflammatory genes and suppression of PTX3 [
25]. PTX3 deficiency was linked to cancer-related inflammation, angiogenesis, and mutations in models of skin cancer [
22]. Whether PTX3 functions as a tumor-suppressor or tumor-promoting factor in HCC requires future work.
In this study, we hypothesized that induced PTX3 synthesis in the cirrhotic liver results in higher PTX3 protein levels in the hepatic vein. Moreover, we postulated positive correlations of hepatic vein PTX3 with measures of liver function in patients with liver cirrhosis and patients with HCC.
Discussion
Main findings of the present study are that levels of PTX3 (1) were higher in the portal as compared to the hepatic vein (2) and were increased after TIPS insertion. This suggested that the liver eliminates PTX3 from the blood.
PTX3 did not correlate with resistin and galectin-3, which are produced by immune cells [
35]. Though circulating PTX3 in patients with liver cirrhosis was supposed as a marker of immune cell activation, high levels are also attributed to diminished hepatic excretion.
Negative correlations of PTX3 with prothrombin time were in accordance with a function of PTX3 in the extrinsic pathway of coagulation [
38,
39]. Tissue factor initiates extrinsic blood coagulation and PTX3 enhanced the expression of this protein in activated monocytes and endothelial cells [
38,
39]. Thus, shorter prothrombin time in cirrhotic patients with high PTX3 may be because of higher tissue factor expression.
PTX3 further negatively correlated with antithrombin 3, factor V, and fibrinogen. PTX3 indeed binds fibrinogen and enhances plasmin-induced fibrinolysis [
10]. Thus, PTX3 promotes fibrin deposition and fibrinolysis, and therefore, has a regulatory function in coagulation pathways [
10,
39]. Several pro- and anticoagulant factors are produced in the liver, and consequently, both pathways are impaired in patients with liver cirrhosis [
40]. PTX3 was higher in cirrhotic patients than controls, whereas prothrombin conversion as well as thrombin inactivation was impaired in these patients [
21,
40]. This suggests that raised PTX3 in liver cirrhosis may in part compensate for deficiencies in pro- as well as anticoagulatory pathways. Correlations of PTX3 with antithrombin 3 and factor V cannot provide information on a functional association of these proteins nor does it prove whether PTX3 directly regulates the hepatic synthesis or the excretion of these proteins.
Negative correlations of PTX3 with fibrinogen and antithrombin 3 were also described in patients with acute Puumala hantavirus infection, a disease mainly characterized by renal dysfunction [
41]. In contrast, such associations were not identified in patients with sickle cell painful crisis and women with preeclampsia [
42,
43]. Additionally, serum PTX3 did not correlate with prothrombin time in HCC patients. Associations of PTX3 with measures of coagulation were related to pathological condition, suggesting a complex regulatory network. PTX3 was correlated with inflammation, dyslipidemia, and renal function, illustrating that multiple influential factors are involved [
7,
13,
14,
44].
PTX3 was not associated with the severity of liver disease defined by the Child–Pugh or MELD score in the cohort analyzed herein. A separate study described modest positive associations of PTX3 with these measures. Mean MELD score in that study was nearly twofold higher than of the patients enrolled in the present investigation. Indeed, patients with a MELD score over 20 had higher systemic PTX3 than patients with a lower score [
21]. However, similar to our findings, PTX3 levels were not different in Child–Pugh class A, B, and C patients [
21]. Hence, PTX3 is not associated with the severity of liver disease. High levels are more likely related to severe complications such as acute-on-chronic liver failure or infections [
21].
This suggestion was supported as circulating PTX3 showed no association with histological liver steatosis, inflammation, or fibrosis in patients with HCC. Of note, serum PTX3 did not increase with the TNM stages in liver cancer. This shows that high plasma PTX3 in HCC, which was even considered as a risk factor for HCC development [
17], has no prognostic value.
In cirrhosis patients, PTX3 was neither related to ascites volume nor to variceal size. Moreover, PTX3 levels were not induced in patients presenting with ascites or grades III/IV encephalopathy [
21]. This excludes PTX3 as a marker of residual hepatic function and common complications of liver cirrhosis. PTX3 was, nevertheless, higher in patients with variceal bleeding compared to those with refractory ascites. These two cohorts had similar MELD score, prothrombin time, bilirubin, CRP, and resistin (data not shown). Related to the function of PTX3 in coagulation, tissue injury was associated with higher local and systemic PTX3 in a skin-wounding model [
10], and similar mechanisms may account for higher PTX3 in patients with variceal bleeding.
Moreover, pathological bacterial translocation is aggravated in patients with variceal bleeding [
45]. This boost of proinflammatory gut-derived mediators may further impact on PTX3 levels.
Interestingly, PTX3 was higher in the portal compared to the hepatic vein. PTX3 is mostly released by immune cells, and thus, the spleen and even bowel or omental adipose tissue localized cells may contribute to portal vein PTX3. Median ratio of portal to hepatic vein PTX3 was 117% in Child–Pugh A, 107% in Child–Pugh B, and 100% in Child–Pugh C patients. Though these values did not significantly differ, they tend to decline in patients with more impaired hepatic function. PTX3 may be eliminated by the liver and impaired liver function may thus contribute to higher systemic levels. In line with this assumption, PTX3 levels increased after TIPS. The more prominent rise of PTX3 in HVP compared to PVP shortly after TIPS further supports this hypothesis. A clinically interesting question is whether increased PTX3 shortly after this intervention contributes to higher thrombosis risk [
5].
Post-TIPS PTX3 in the portal vein still correlated with prothrombin time, antithrombin 3, and factor V. In HVP, associations with prothrombin time and factor V persisted after TIPS. This principally points to a relatively strong association of PTX3 with coagulation. SVP PTX3 before and after TIPS showed now associations with coagulation factors or prothrombin time. This illustrates that, despite the high correlation of PTX3 levels among each of the blood compartments, additional mechanisms exist that specifically modulate its systemic concentration. Whether coagulation factor levels vary in the distinct blood compartments needs further analysis.
There are limitations within our study. The cohort sizes were low, and effects were rather small. Plasma of HCC patients to analyze PTX3 was not available, and PTX3 levels of the cohorts could not be compared.Moreover, MELD score of HCC patients was not documented.
In summary, the present study provides evidence that the liver eliminates PTX3 from the blood. Impaired hepatic removal of PTX3 in liver cirrhosis may contribute to increased plasma levels. Anyhow, circulating PTX3 is not of prognostic value in liver cirrhosis or HCC.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.