Serum pentraxin 3 as a biomarker of hepatocellular carcinoma in chronic hepatitis B virus infection

Biomarkers for early diagnosis of hepatocellular carcinoma (HCC) are needed in chronic hepatitis B virus (HBV) infection, a leading cause of HCC. We evaluated whether measurement of serum pentraxin 3 (PTX3) could improve diagnosis of HCC in chronic HBV infection. Data from patients with HBV-related chronic hepatitis (n = 159), cirrhosis (n = 99) and HCC (n = 107), and healthy controls (n = 151) were analyzed. Serum PTX3 concentration was measured by immunoassay. Area under the receiver operating characteristic curve (AUC) was applied to assess diagnostic accuracy. PTX3 levels were significantly higher in HBV patients than in healthy controls (P < 0.001) and in HCC than in chronic hepatitis (P < 0.001) or cirrhosis patients (P < 0.001). PTX3 was an independent risk factor of HCC [odds ratio (OR) 1.617, P < 0.001] and could distinguish HCC in chronic HBV infection [cutoff 9.231 ng/mL, AUC 0.929 with 95% confidence interval (CI) of 0.898–0.953], including α-fetoprotein (AFP) negative [cutoff 8.985 ng/mL, AUC (95%CI) 0.947 (0.908–0.973)] and early-stage HCC [cutoff 9.359 ng/mL, AUC (95%CI) 0.920 (0.885–0.947)]. Combination of PTX3 with AFP improved the discrimination of early HCC from chronic HBV infection [AUC (95%CI) 0.948 (0.918–0.970)]. In short, PTX3 measurement could identify HCC, including AFP-negative and early-stage HCC, in chronic HBV infection.


Scientific Reports
| (2020) 10:20276 | https://doi.org/10.1038/s41598-020-77332-3 www.nature.com/scientificreports/ With regard to liver diseases, plasma PTX3 was indicated to be associated with nonalcoholic steatohepatitis (NASH) 30 and could differentiate NASH from non-NASH 31 . Elevated levels of PTX3 were also documented in acute liver injury caused by paracetamol and linked with adverse consequences in paracetamol overdose 32 . Hepatic and plasma PTX3 was raised in alcoholic hepatitis patients and PTX3 expression was related to the disease severity and short-term mortality 33 . Serum PTX3 was clinically shown to have valid diagnostic accuracy as a marker of fibrosis in chronic hepatitis C virus (HCV) infection 34 and increased PTX3 plasma level was a risk factor for HCC occurrence in chronic HCV infection 35 . Furthermore, PTX3 was shown to be able to promote HCC progression and high PTX3 expression in tumor tissues was related to unfavorable prognosis in HCC patients 36 . Despite these studies, the potential role of circulating PTX3 levels in chronic HBV infection, especially HBV-related HCC, remains to be further examined. This study, therefore, was designed to determine the serum PTX3 levels in patients with various cross-sectional diseases including chronic hepatitis, cirrhosis and HCC during chronic HBV infection and to assess the potential diagnostic value in distinguishing HCC, including AFP-negative and early HCC, from other disease conditions in chronic HBV infection.

Methods
Study subjects. A total of 516 participants were included in this study. Among them, 365 were patients with chronic HBV infection (159 chronic hepatitis, 99 cirrhosis and 107 HBV-related HCC) and 151 were age-and sex-matched healthy controls. Chronic HBV infection and liver diseases were diagnosed in accordance with Clinical Practice Guidelines on the management of HBV infection 37 . Measurement of AFP level and ultrasound sonography were routinely performed in all the patients with chronic HBV infection. Chronic HBV infection was defined based on the presence of HBsAg for more than 6 months and detectable serum HBV DNA levels. Chronic hepatitis was diagnosed in view of abnormal biochemistry of liver function such as increased serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels in chronic HBV infection. Cirrhosis was determined based on imaging evidence of liver cirrhosis and existence of portal hypertension, ascites, varices and splenomegaly and/or changes of histopathology of liver biopsy samples with abnormal biochemistry of liver function such as increased ALT and AST levels and/or hypoalbuminia. HCC was determined on the base of characteristics of ultrasound, computed tomography and/or magnetic resonance and/or histopathology 38 . According to the commonly used upper limit of normal AFP level (20 ng/mL) 38 , the HCC was divided into AFP negative (≤ 20 ng/mL) or positive (AFP > 20 ng/mL) HCC. HCC stage was defined in line with the Barcelona Clinic Liver Cancer (BCLC) staging system. The diseases were classified as early-stage (stage 0 and stage A) 39 and late-stage (stage B, stage C and stage D) HCC. Patients with other liver disease such as nonalcoholic fatty liver disease, primary biliary cirrhosis, alcoholic liver disease and drug-induced liver injury were excluded. Patients with autoimmune diseases such as autoimmune hepatitis and systemic lupus erythematosus and severe diseases of other systems were also excluded. Patients aged under 18 years were also excluded. Blood samples were obtained from each participant. Serum was separated, aliquoted and stored at-80 °C until use.
Determination of laboratory parameters. Blood tests and liver functions were assayed at the central laboratory of the hospital. Serum HBV DNA was quantified by HBV quantitative polymerase chain reaction. HBsAg, anti-HBs, HBeAg, anti-HBe, and anti-HBc were determined by enzyme-linked immunosorbent assay. Serum AFP (ng/mL) was quantified by automated Eleceyes (Hoffman-La Roche Ltd., Basel, Switzerland).

Statistical analysis.
Statistical analysis was conducted using SPSS 24.0 software and MedCalc12.0 software.
Quantitative variables were presented as mean and standard deviation (SD) or median and interquartile range. Categorical variables were expressed as absolute or relative frequencies. Continuous variables were compared using the analysis of variance. Categorical variables were compared by Chi-Square test. Risk factors for HCC were estimated using multivariable logistic regression analysis. The predictive values of relevant parameters were analyzed by receiver operating characteristic (ROC) curve to calculate the area under the curve (AUC), 95% confidence interval (CI) and optimal cut-off value and its sensitivity, specificity, Yoden index, positive (PPV) and negative (NPV) predictive value, and positive and negative likelihood ratio (LR). Analysis of statistical differences between AUCs was performed using Z-test. To minimize the influence of data bias and confounding variables we also performed analysis of the data for patients with similar propensity scores using a 1:1 ratio matching after propensity score matching (PSM). Significance was defined as a two-tailed P value less than 0.05.
Among the patients with chronic HBV infection, 87 patients received antiviral treatment with nucleos(t)ide analogues for more than six months and 278 did not receive antiviral treatment (Table S3) Table S4).

Discussion
This study demonstrated significantly elevated serum PTX3 levels in patients with chronic HBV infection compared to healthy controls and in patients with HBV-related HCC compared to patients with chronic hepatitis or cirrhosis. High PTX3 level was an independent risk factor of HCC. The PTX3 levels in HCC patients appeared to be related to HCC differentiation, with low differentiation HCC having higher PTX3 levels in comparison to high differentiation HCC. Importantly, serum PTX3 levels could highly discriminate HCC from chronic hepatitis, cirrhosis and chronic HBV infection without HCC. The discrimination of HCC by PTX3 levels was not influenced by the underlying disease (chronic hepatitis or cirrhosis) or the usage of antiviral treatment with nucleos(t)ide analogues. In particular, PTX3 levels were highly discriminative of AFP-negative (≤ 20 ng/mL) and early-stage HCC from chronic hepatitis, cirrhosis and chronic HBV infection without HCC. These findings suggest that PTX3 may act as a major contributing factor to the progress of HBV disease and the development of HCC. The determination of PTX3 levels may have a significant diagnostic value for HBV-related HCC including AFP-negative and early-stage HCC in chronic HBV infection. PTX3 has been revealed to be an extrinsic oncosuppressor in preclinical models and certain tumors through regulating complement-driven macrophage-mediated tumor progression and tuning cancer-related Table 3. Diagnostic performance of serum PTX3, AFP and the combination of PTX3 with AFP levels for HCC in chronic hepatitis, cirrhosis and chronic HBV infection without HCC (CH + LC). PTX3 pentraxin 3, AFP alpha-fetoprotein, HBV hepatitis B virus, HCC hepatocellular carcinoma, CH chronic hepatitis, LC liver cirrhosis, AUC area under receiver operating characteristic (ROC) curve, PPV positive predictive value, NPV negative predictive value, LR likelihood ratio. www.nature.com/scientificreports/ inflammation 22,25,40,41 . However, it remains unknown whether PTX3 plays a protective or promotive role in cancer in that it may impose an important influence on various aspects of cancer such as tumor initiation, angiogenesis, metastasis and immune-regulation 42 . Elevated PTX3 expression has been shown to be associated with poor prognosis in certain cancers, such as breast cancer 25 , gastric cancer 26 , lung cancer 27 , pancreatic cancer 28 and prostate cancer 29 . Elevated PTX3 plasma level was a risk factor for HCC occurrence in chronic HCV infection 35 and higher PTX3 expression in tumor tissues was also related to poor prognosis in HCC patients 36 . Consistently, this study showed that elevated circulating PTX3 levels were related to HCC development in chronic HBV infection. The mechanisms by which PTX3 could be related to the occurrence and development of HCC are largely unknown. There are several potential explanations. First, PTX3 plays a complicated regulatory role in cancerrelated inflammation 22,24 . In liver diseases, PTX3 levels were markedly higher in NASH than in simple steatosis non-NASH patients 30,31 . PTX3 levels were also related to the severity of liver fibrosis in NASH 30 and HCV infection 34 . It is suggested that PTX3 is involved in pathogenesis of hepatic inflammation and fibrosis which are common milieu at the origin of HCC 43 . Second, the involvement of PTX3 in immune response may also play a role in its association with HCC. Expression of PTX3 is inducible by tumor necrosis factor (TNF)-α and interleukin (IL)-1 44 . Differentialy expressed TNF-α from HBV-specific CD8 + T cells was associated with the outcomes of chronic HBV infection including asymptomatic status, active chronic hepatitis and HBV-related HCC 45 . IL-1β levels were significantly and progressively increased with the disease advancement to HCC in chronic HBV infection 46 . Additionally, IL-10 could stimulate B cells in adaptive immunity through enhancing PTX3 production 20 . Elevated IL-10 was suggested to mediate disease progress, from inactive state to cirrhosis and HCC in HBV infection 47 . HCC patients also exhibited significantly higher frequencies of IL-10-expressing B cells that could suppress cytotoxic CD4+ T cell function related to poor survival and high recurrence rate of HCC 48 . Third, although PTX3 is believed not to be produced by hepatocytes 49 , it was shown to be able to enhance HCC cell proliferation and to induce epithelial-mesenchymal transition (EMT) 36 , a biologic process closely related to tumor cell invasion and metastasis.
An important factor for improving long-term prognosis of HCC patients is to detect and treat the tumor at its early stage 7,9 . AFP is the most commonly used biomarker but it is not suitable to detect early stage 11 and www.nature.com/scientificreports/ AFP-negative HCC 12 . Therefore, identification of novel biomarkers for HCC remains an urgent need. In the present study, PTX3 was highly discriminative of HCC in chronic HBV infection. In particular, PTX3 was highly and accurately discriminative of AFP-negative and early-stage HCC and the diagnostic performance of PTX3 was superior to AFP. These findings suggest the potentail of PTX3 to be used as a more sensitive biomarker for HCC including early HCC and as a supplemental biomarker for AFP-negative or low AFP-displaying HCC in chronic HBV infection. For diagnostic accuracy, PTX3 exhibited higher AUC, sensitivity, and specificity than did AFP in HCC patients in relation to chronic hepatitis, cirrhosis and chronic HBV infection without HCC. Detecting of both PTX3 and AFP improved the diagnostic accuracy for HCC in comparison with either detection alone. Although the addition of AFP to PTX3 did not significantly increase the diagnostic ability of PTX3 alone for AFP-negative HCC, the combination of PTX3 with AFP improved diagnostic accuracy when only early-stage HCC was evaluated.
Addition of AFP to ultrasound was shown to significantly increase sensitivity of early HCC detection 50 . PTX3 exhibited good diagnostic performance for the detection of early HCC from HBV chronically infected populations. Whether addition of PTX3 to ultrasound can improve the sensitivity of HCC detection deserves further investigation. Moreover, this study showed that the addition of PTX3 to AFP may allow further identification of HCC in the HBV-related liver diseases. Whether addition of both PTX3 and AFP to ultrasound can further improve the ability and accuracy of HCC detection also deserves investigation. Moreover, serum des-gammacarboxy prothrombin (DCP) was revealed to be an ideal marker for the diagnosis of HBV-related HCC 51 . We did not compare PTX3 and DCP in the study because measurement of DCP was not carried out in this study. The comparison of PTX3 and DCP is an interesting issue to be addressed in future studies.
We recognized some limitations of the study. First, this is a retrospective study in a relatively small number of patient population with cross-sectional diseases of chronic HBV infection and did not examine the changes of PTX3 levels before and after HCC development. Second, the lack of a validation population may affect the strength of the study. Third, this study only included patients with chronic HBV infection and patients with other etiologies were not included for investigation and comparison. This limits the generalization of the findings to disease conditions of other etiologies such as chronic HCV infection and NASH. Nevertheless, this study demonstrated consistent findings of the diagnostic accuracy of PTX3 for HCC in various disease conditions of chronic HBV infection, and both multivariable analysis and the analysis of data after PSM exhibited the independent role and diagnostic potential of PTX3 in HBV-related HCC, providing evidence for further studies of enhancing and extending the diagnostic potential of PTX3 in HCC detection and optimization of HCC surveillance.

Conclusion
This study demonstrated the differences of circulating PTX3 levels in various diseases in chronic HBV infection and identified serum PTX3 as an independent factor associated with HCC. Particularly, PTX3 can be used as a new indicator for HCC diagnosis. The performance of PTX3 for differentiating HCC from chronic HBV infection is superior to AFP. Furthermore, PTX3 is highly and accurately discriminative of AFP-negative and earlystage HCC. The combination of PTX3 with AFP may allow further identification of HCC in the HBV-related liver disease population. Therefore, PTX3 may be used as a simple and potentially readily available test that can accurately discriminate HCC in chronic HBV infection. Prospective studies with large patient populations of various etiologies are warranted to validate these results and to directly compare PTX3 with other noninvasive markers before establishing its diagnostic value. Ethics approval. The study was conducted under approval of Ethics Committee of the First Affiliated Hospital of Xi'an Jiaotong University. All experiments were conducted in accordance with the Declaration of Helsinki. All the participants provided written informed consent to participate in the study.

Data availability
The data generated and analysed during this study are available from the corresponding author on reasonable request.