Author Affiliations: Institute of Infectious Diseases, Southwest Hospital, Third Military Medical University; Chongqing Key Laboratory for Research of Infectious Diseases, Chongqing 400038, China (Xu C, Zeng XH, Wang L, Tao SQ, Wu QX, Zhu P, Deng GH and Wang YM)

Corresponding Author: Yu-Ming Wang, Institute of Infectious Diseases, Southwest Hospital, Third Military Medical University, Chongqing 400038, China (Tel/Fax: +86-23-68754858; Email: wym417@163.com)

 

© 2015, Hepatobiliary Pancreat Dis Int. All rights reserved.

doi: 10.1016/S1499-3872(15)60352-6

Published online March 18, 2015.

 

 

Acknowledgment: We thank Drs. Xiao-Mei Xiang and Zhao-Xia Tan from Institute of Infectious Diseases, Southwest Hospital, for their help in collecting and measuring samples.

Contributors: XC and WYM proposed the study. XC and WYM performed research and wrote the first draft. XC, ZXH, WL, TSQ, WQX, ZP and DGH collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. WYM is the guarantor.

Funding: This study was supported by grants from the National Natural Science Foundation of China (81071694) and the State Key Project specialized for HBV-related severe hepatitis of China (2012ZX10002004).

Ethical approval: This study was approved by the Ethics Committee of Southwest Hospital.

Competing interest: No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

 

 

BACKGROUND: The current methods used for diagnosing hepatocellular carcinoma (HCC) are unsatisfactory. Here, we assessed the serum levels of secreted frizzled related protein 4 (sFRP-4) for diagnosing HCC in patients infected with chronic hepatitis B (CHB).

 

METHODS: In 272 patients with CHB enrolled, 142 were patients with HCC. Thirty-three healthy subjects were recruited as healthy controls. The CHB patients were assigned to a test group or a validation group based on the time of enrollment. Human antibody arrays were used to screen 15 patients (8 CHB-related HCC patients, 7 CHB patients) for serum markers. Four markers and one candidate marker were assessed in the test group and validation group, respectively.

 

RESULTS: Human antibody assays indicated that the serum levels of sFRP-4 in HCC patients were significantly higher than those in CHB patients (P<0.05). Additionally, serum sFRP-4 levels were significantly higher in the HCC patients than those in the non-HCC patients in both test group (79.7 vs 41.3 ng/mL; P<0.001) and validation group (89.0 vs 39.0 ng/mL; P<0.001). Areas under the Receiver Operating Characteristic curves (AUCs) for alpha-fetoprotein (AFP) and sFRP-4 were similar in both test group and validation group. In the test group, the combination of sFRP-4 (a sensitivity of 94.4%, a specificity of 60.5% at 46.4 ng/mL) and AFP (a sensitivity of 75.0%, a specificity of 87.2% at 11.3 ng/mL) showed better performance for diagnosing HCC (a sensitivity of 79.2% and a specificity of 95.3%). The AUC for combined sFRP-4 and AFP increased to 0.941 (95% CI: 0.908-0.975), and similar results were seen in the validation group.

 

CONCLUSION: sFRP-4 is a candidate serum marker for diagnosing HCC in CHB patients, and the combination of sFRP-4 with AFP may improve the diagnostic accuracy of HCC.

 

(Hepatobiliary Pancreat Dis Int 2015;14:164-170)

 

KEY WORDS: chronic hepatitis B; hepatocellular carcinoma; sFRP-4; serum marker

Hepatocellular carcinoma (HCC) is one of the most prevalent malignancies worldwide.^[1] In most cases, HCC develops from a pre-existing chronic liver disease, and HBV infection is a major risk factor for HCC. Patients with chronic hepatitis B (CHB) related HCC are often asymptomatic in the early stage; thus most such HCC cases are diagnosed at an advanced stage with a poor prognosis. The currently well recognized biomarker is AFP. However, AFP has a poor sensitivity and specificity for the diagnosis of HCC.^[2-4] Additionally, imaging techniques including CT, MRI, and abdominal ultrasound are not cost-effective, especially when used to detect lesions <2 cm in size. Since none of the current available modalities is satisfactory, new method which is convenient, sensitive and cost-effective is needed for detecting HCC.

 

Secreted frizzled related protein 4 (sFRP-4) is a member of the sFRP family of proteins; it binds Wnt and inhibits the signaling pathway.^[5] The Wnt pathway is involved in numerous cellular processes and has been suggested as a regulator of tumorigenesis.^[6] sFRP-4 is increased in some malignancies, such as colorectal carcinoma.^[7] The present study was to evaluate the diagnostic role of sFRP-4 in CHB related HCC.

 

 

An overview of the study design is presented in Fig. 1. Human antibody arrays were used to screen for candidate serum markers, which were later evaluated in specific tests. Markers which displayed good test results were further evaluated in a group of validation tests. Additionally, the diagnostic performance of selected markers was also evaluated in combination with AFP.

 

From May 2013 to October 2014, 287 patients infected with HBV and 33 healthy volunteers at the Institute of Infectious Diseases, Southwest Hospital were enrolled. The study protocol was approved by the Ethics Committee of Southwest Hospital, and all patients provided their verbal informed consent before enrollment. All individuals involved in data analysis were masked to the patients' identities.

 

Fifteen patients (48.6±12.3 years; 13 males and 2 females) enrolled in May 2013 were taken as a screening group; the patients and healthy volunteers enrolled between June 2013 and February 2014 served as the test group; and patients enrolled from March 2014 to October 2014 were designated as a validation group. Consecutive HCC patients were enrolled according to the following criteria: a) chronic HBV infection for more than 5 years; b) age >25 years; c) previously untreated HCC diagnosed by both an imaging examination (enhanced CT or ultrasound) and histopathological examination (<3 lesions, no evidence of lymphatic, portal, intra-hepatic or extra-hepatic metastasis). Cirrhosis was diagnosed by clinical manifestations, endoscopy combined with ultrasound or CT. In healthy controls, there was no evidence of liver disease, diabetes or malignancies.

 

The exclusion criteria were as follows: co-infection with HCV, HIV or other pathogens; autoimmune hepatitis; alcoholic liver diseases; drug induced hepatitis; other types of cirrhosis; and other types of malignancies.

 

A total of 15 patients (8 with CHB related HCC and 7 with CHB) were assessed using human antibody array-1 kits coated with 507 different cytokine antibodies (assay type: AAH-BLG-1, RayBiotech Inc., Norcross, GA, USA). Briefly, the protein chips were blocked at room temperature for 30 minutes, and then incubated with biotin-labeled samples (400 µL) overnight at 4 ��. Next, the protein chips were washed and incubated with Cy3-conjugated streptavidin (400 µL) at room temperature for 2 hours. After wash, the signals were detected at a frequency of 532 nm using a GenePix 4000B Microarray Scanner (Molecular Devices, LLC, Sunnyvale, CA, USA). Data were analyzed using RayBio® Analysis Tool software, and the results were displayed as densities relative to those of positive controls. Biotin-conjugated IgG and antibody capture diluent served as positive and negative controls, respectively. Candidate cytokines were selected if their relative densities were significantly different between HCC and CHB patients, and fold changes were higher than 1.5 or less than 0.66.

 

Serum samples were obtained from HCC patients before surgery or radiotherapy, and stored at -80 �� before analysis. The serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), total bilirubin (TBIL), albumin (ALB), and glucose (Glu) were analyzed using the kinetic method with a Hitachi 7600 chemistry analyzer (Hitachi, Tokyo, Japan). Serum concentrations of AFP were determined using an Architect I2000 luminescence analyzer (Abbott, North Chicago, IL, USA) and serum levels of human C-C chemokine receptor-7 (CCR-7), human neurotrophin-3 (NT-3), human fibroblast growth factor-23 (FGF-23) and sFRP-4 were determined using commercially available enzyme-linked immunosorbent assay (ELISA) kits (CUSABIO Biotech, Wuhan, China). The detection ranges for CCR-7, NT-3, FGF-23 and sFRP-4 were 18.75-1200 pg/mL, 1.25-80 ng/mL, 3.12-200 pg/mL and 1.56-100 ng/mL, respectively. The steps in these measurements were in accordance with the user manuals and previously described procedures according to the instructions of manufacturers, and samples were diluted ten times before the measurement of sFRP-4, while samples for other ELISA kits were undiluted.

 

All statistical data were analyzed using SPSS Statistics for Windows (Version 17.0. Chicago, IL, USA). Normally distributed values were shown as mean±standard deviation, and nonparametric values as the median with interquartile ranges. The Chi-square test was used to analyze categorical variables, and either Student's t test or the Mann-Whitney U test was used to analyze continuous variables as appropriate. A receiver operating characteristic (ROC) curve was plotted to assess diagnostic performance. Correlation analyses were performed based on the characteristics of variables. A P<0.05 was considered statistically significant. Bonferroni's correction method was used to reduce alpha error in multiple testing.

 

 

The 15 patients in the screening group were negative for both diabetes and co-infection with other viruses. Four cytokines (CCR-7, NT-3, FGF-23 and sFRP-4) showed significant difference in patients with HCC compared to those with CHB (Table 1). Each patient of HCC had a single lesion <2 cm in diameter and showed no evidence of metastasis. The serum levels of sFRP-4 in the 8 HCC patients were significantly higher than those in the 7 CHB patients (P=0.020). Four selected cytokines (CCR-7, NT-3, FGF-23 and sFRP-4) were evaluated in the test group, only sFRP-4 was significantly different between HCC and non-HCC patients (P<0.001, Table 1), and was thus selected as the candidate marker for validation.

 

Clinical characteristics of the 272 patients and 33 healthy volunteers enrolled in the study are shown in Table 2. In the test group, the serum levels of AFP and sFRP-4 in the HCC patients were significantly higher than those in the non-HCC patients [AFP (ng/mL): 61.8 (5.6-199.9) vs 3.6 (2.4-5.3), P<0.001; sFRP-4 (ng/mL): 79.7 (54.3-117.3) vs 41.3 (35.9-56.2), P<0.001], and similar results were found in the validation group [HCC vs non-HCC; AFP (ng/mL): 50.4 (7.6-404.1) vs 3.7 (2.5-5.2), P<0.001; sFRP-4 (ng/mL): 89.0 (63.6-131.1) vs 39.0 (28.9-54.5), P<0.001]. Additional comparisons between the HCC patients (cirrhosis and non-cirrhosis) and the non-HCC patients (cirrhosis, non-cirrhosis, and healthy controls) with or without diabetes are shown in Fig. 2.

 

The performance of sFRP-4 as a diagnostic marker was further assessed in the test group (Fig. 3A). The area under the ROC curve (AUC) for the ability of sFRP-4 to distinguish the HCC patients from the non-HCC patients was 0.847 (95% CI: 0.788-0.906), which was similar to the AUC for AFP (0.857, 95% CI: 0.796-0.918, P>0.05). The combination of serum AFP and sFRP-4 levels increased the AUC to 0.941 (95% CI: 0.908-0.975). When we excluded patients with diabetes, the AUC for sFRP-4 was still similar to that for AFP [0.847 (95% CI: 0.787-0.907) vs 0.860 (95% CI: 0.799-0.922), P>0.05; Fig. 3C].

 

The observed diagnostic characteristics of AFP and sFRP-4 are described in Table 3. The maximum diagnostic performance of sFRP-4 (a sensitivity of 94.4% and a specificity of 60.5%) was reached at a cut-off value of 46.4 ng/mL, while AFP reached its maximum diagnostic performance (a sensitivity of 75.0% and a specificity of 87.2%) at a cut-off value of 11.3 ng/mL. In comparison, the combination of AFP and sFRP-4 showed a sensitivity of 79.2% and a specificity of 95.3%. When patients with diabetes were excluded, the best cut-off values for sFRP-4 and AFP remained unchanged, and sFRP-4 still showed a sensitivity of 94.3% and a specificity of 60.7%, while AFP showed a sensitivity of 75.7% and a specificity of 85.7%.

 

We subsequently investigated the relative and combined performances of sFRP-4 and AFP in the validation group (Fig. 3B, 3D). There was no significant difference in the AUC values for AFP and sFRP-4 [0.850 (95% CI: 0.781-0.919) vs 0.891 (95% CI: 0.839, 0.942), P>0.05] when applied individually. The combination of sFRP-4 and AFP again had a high AUC value [0.932 (95% CI: 0.890-0.975)]. Similar results were obtained when we investigated the AUCs of these two markers in patients without diabetes [AFP vs sFRP-4, 0.868 (95% CI: 0.800-0.936) vs 0.889 (95% CI: 0.837-0.942); P>0.05]. The AUC value for combined sFRP-4 and AFP in non-diabetic patients was 0.933 (95% CI: 0.888-0.978).

 

The performance of sFRP-4 as a diagnostic marker remained encouraging (Table 3), as it showed a sensitivity of 78.6% and a specificity of 84.4% at the cut-off value of 61.4 ng/mL. When applying the cut-off value of 46.4 ng/mL obtained from the test group, sFRP-4 showed a sensitivity of 91.4% and a specificity of 64.9%. AFP demonstrated a sensitivity of 71.4% and a specificity of 97.4% at the cut-off value of 13.5 ng/mL, whereas the combination of AFP and sFRP-4 showed a sensitivity of 88.6% and a specificity of 89.6%. Additionally, sFRP-4 also showed excellent performance in patients without diabetes, with higher sensitivity than AFP (78.5% vs 73.8%) at the cut-off value of 61.4 ng/mL. At a cut-off point of 13.5 ng/mL, the combination of sFRP-4 and AFP displayed a sensitivity of 87.7% and a specificity of 92.1%.

 

Possible correlations between the diagnostic power of sFRP-4 and age, gender, ALT, AFP, glucose, cirrhosis, and the size of a single lesion were assessed in the HCC patients (Table 4). The diagnostic value of sFRP-4 was not correlated with age, gender, AFP, glucose or cirrhosis status. While the size of a single lesion showed a slight positive correlation with sFRP-4 (Spearman coefficient: 0.248, P=0.003); there were no significant differences in the degrees of correlation between sFRP-4 and single lesions of different sizes (<3 cm vs 3-5 cm vs >5 cm; P>0.05).

 

 

Patients with advanced stage HCC have a poor prognosis and a high rate of mortality. The current methods used for diagnosing HCC, especially early stage HCC, are unsatisfactory. Serum proteins and other components can serve as convenient and inexpensive biomarkers of diseases, and are expected to play important roles in the diagnosis of early stage HCC. Recent studies offered various serum markers (i.e. PIVKA-II, GPC-3, GP73 and OPN) for malignant liver disease,^[8-11] but subsequent meta-analyses of these markers did not show any satisfactory performance.^[12-14] Thus novel serum markers are needed for diagnosing HCC.

 

The present study evaluated a novel serum biomarker for HCC (sFRP-4) and showed that its performance was comparable to that of AFP, which is regarded as the gold standard serum marker for HCC in clinical practice. As there was no correlation between the serum levels of AFP and sFRP-4, their combined use can significantly improve the diagnosis of HCC. The size of a single lesion had no significant effect on the diagnostic performance of sFRP-4, and patients with early stage HCC in our study could be efficiently diagnosed by their serum sFRP-4 levels. The diagnosis of early stage HCC is important for patient prognosis and outcome. Moreover, measurements of serum sFRP-4 levels are convenient, cost-effective, and can be easily applied in clinical practice.

 

sFRP-4 is a secreted protein with a cysteine-rich domain homologous to the putative Wnt-binding site of frizzled proteins. Additionally, sFRP-4 is considered as a potential antagonist of the Wnt signaling pathway. While numerous studies have demonstrated methylation of the sFRP-4 gene promoter site in other malignant tumors,^[15-17] only sFRP-1, sFRP-2 and sFRP-5 have shown promoter methylation in HCC,^{[18, 19]} and the role of sFRP-4 in HCC remains unknown. In our study, we found higher serum sFRP-4 levels in HCC patients compared to non-HCC subjects, suggesting that development of HCC may facilitate sFRP-4 expression. If this finding is validated, it would be opposite to the effects of HCC on sFRP-1, sFRP-2 and sFRP-5 levels. Furthermore, the exact mechanism for the effects of HCC development on all sFRP isotypes remains unknown.

 

Our screening results were different from the markers reported by Wang et al,^[20] the size of populations for screening and diverse genetic background of two populations may contribute to the differences of the two studies. Although initial screening results suggested four candidates as HCC markers, only sFRP-4 proved to be useful for diagnosing HCC in the test group. The different results obtained using the human antibody array and ELISA kits may attribute to the different methodologies.

 

Our results may be biased by the limited number of patients and control subjects, and further investigation is needed to assess the diagnostic value of serum sFRP-4 levels in a larger population. Interestingly, in previous studies, sFRP-4 was reported to be overexpressed in type 2 diabetes, and was correlated with phosphate homeostasis.^[21-24] However, these studies focused on the sFGP-4 gene and mRNA levels. Christov et al^[25] suggested that sFRP-4 does not play a role in the long-term regulation of serum phosphate levels in mice. In our study, few patients had diabetes, and the effect of type 2 diabetes on the diagnostic value of sFRP-4 for HCC remains unclear. However, our results showed no relationship between levels of serum glucose and sFRP-4 which implies that diabetes mellitus does not affect the accuracy of sFRP-4 in HCC diagnosis.

 

In summary, our results suggest that sFRP-4 may serve as a novel serum marker for diagnosing CHB-related HCC. However, the underlying mechanisms for this relationship require further investigation. Finally, the combination of serum sFRP-4 and AFP improves the diagnostic accuracy in patients with HCC.

 

 

1 Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:69-90. PMID: 21296855

2 Hussein TD. Serological tumor markers of hepatocellular carcinoma: a meta-analysis. Int J Biol Markers 2015;30:e32-42. PMID: 25450646

3 Sang W, Zhang W, Cui W, Li X, Abulajiang G, Li Q. Arginase-1 is a more sensitive marker than HepPar-1 and AFP in differential diagnosis of hepatocellular carcinoma from nonhepatocellular carcinoma. Tumour Biol 2015 Jan 11 [Epub ahead of print]. PMID: 25577247

4 Jia X, Liu J, Gao Y, Huang Y, Du Z. Diagnosis accuracy of serum glypican-3 in patients with hepatocellular carcinoma: a systematic review with meta-analysis. Arch Med Res 2014;45:580-588. PMID: 25446613

5 Cruciat CM, Niehrs C. Secreted and transmembrane wnt inhibitors and activators. Cold Spring Harb Perspect Biol 2013;5:a015081. PMID: 23085770

6 Herr P, Hausmann G, Basler K. WNT secretion and signalling in human disease. Trends Mol Med 2012;18:483-493. PMID: 22796206

7 Han FQ, Zhao W, Bentel J, Shearwood AM, Zeps N, Joseph D, et al. Expression of sFRP-4 and beta-catenin in human colorectal carcinoma. Cancer Lett 2006;231:129-137. PMID: 16356838

8 Lamerz R, Runge M, Stieber P, Meissner E. Use of serum PIVKA-II (DCP) determination for differentiation between benign and malignant liver diseases. Anticancer Res 1999;19:2489-2493. PMID: 10470180

9 Nakatsura T, Yoshitake Y, Senju S, Monji M, Komori H, Motomura Y, et al. Glypican-3, overexpressed specifically in human hepatocellular carcinoma, is a novel tumor marker. Biochem Biophys Res Commun 2003;306:16-25. PMID: 12788060

10 Marrero JA, Romano PR, Nikolaeva O, Steel L, Mehta A, Fimmel CJ, et al. GP73, a resident Golgi glycoprotein, is a novel serum marker for hepatocellular carcinoma. J Hepatol 2005;43:1007-1012. PMID: 16137783

11 Cui BK, Zhang CQ, Zhang Y, Yuan YF, Zhang YQ, Shi M, et al. Osteopontin as a potential biomarker of metastasis and recurrence for hepatocellular carcinoma. Ai Zheng 2006;25:876-879. PMID: 16831281

12 Huang TS, Shyu YC, Turner R, Chen HY, Chen PJ. Diagnostic performance of alpha-fetoprotein, lens culinaris agglutinin-reactive alpha-fetoprotein, des-gamma carboxyprothrombin, and glypican-3 for the detection of hepatocellular carcinoma: a systematic review and meta-analysis protocol. Syst Rev 2013;2:37. PMID: 23738605

13 Xu C, Yan Z, Zhou L, Wang Y. A comparison of glypican-3 with alpha-fetoprotein as a serum marker for hepatocellular carcinoma: a meta-analysis. J Cancer Res Clin Oncol 2013;139:1417-1424. PMID: 23743582

14 Cheng J, Wang W, Sun C, Li M, Wang B, Lv Y. Meta-analysis of the prognostic and diagnostic significance of serum/plasma osteopontin in hepatocellular carcinoma. J Clin Gastroenterol 2014;48:806-814. PMID: 24247813

15 Kohno H, Amatya VJ, Takeshima Y, Kushitani K, Hattori N, Kohno N, et al. Aberrant promoter methylation of WIF-1 and SFRP1, 2, 4 genes in mesothelioma. Oncol Rep 2010;24:423-431. PMID: 20596629

16 Griffiths EA, Gore SD, Hooker C, McDevitt MA, Karp JE, Smith BD, et al. Acute myeloid leukemia is characterized by Wnt pathway inhibitor promoter hypermethylation. Leuk Lymphoma 2010;51:1711-1719. PMID: 20795789

17 Lin YW, Chung MT, Lai HC, De Yan M, Shih YL, Chang CC, et al. Methylation analysis of SFRP genes family in cervical adenocarcinoma. J Cancer Res Clin Oncol 2009;135:1665-1674. PMID: 19513747

18 Shih YL, Hsieh CB, Yan MD, Tsao CM, Hsieh TY, Liu CH, et al. Frequent concomitant epigenetic silencing of SOX1 and secreted frizzled-related proteins (SFRPs) in human hepatocellular carcinoma. J Gastroenterol Hepatol 2013;28:551-559. PMID: 23215838

19 Takagi H, Sasaki S, Suzuki H, Toyota M, Maruyama R, Nojima M, et al. Frequent epigenetic inactivation of SFRP genes in hepatocellular carcinoma. J Gastroenterol 2008;43:378-389. PMID: 18592156

20 Wang WW, Ang SF, Kumar R, Heah C, Utama A, Tania NP, et al. Identification of serum monocyte chemoattractant protein-1 and prolactin as potential tumor markers in hepatocellular carcinoma. PLoS One 2013;8:e68904. PMID: 23874805

21 Bukhari SA, Shamshari WA, Ur-Rahman M, Zia-Ul-Haq M, Jaafar HZ. Computer aided screening of secreted frizzled-related protein 4 (SFRP4): a potential control for diabetes mellitus. Molecules 2014;19:10129-10136. PMID: 25019556

22 Bergmann K, Sypniewska G. Secreted frizzled-related protein 4 (SFRP4) and fractalkine (CX3CL1) - Potential new biomarkers for β-cell dysfunction and diabetes. Clin Biochem 2014;47:529- 532. PMID: 24675103

23 Mahdi T, Hänzelmann S, Salehi A, Muhammed SJ, Reinbothe TM, Tang Y, et al. Secreted frizzled-related protein 4 reduces insulin secretion and is overexpressed in type 2 diabetes. Cell Metab 2012;16:625-633. PMID: 23140642

24 Kumar R. The phosphatonins and the regulation of phosphate homeostasis. Ann Endocrinol (Paris) 2006;67:142-146. PMID: 16639365

25 Christov M, Koren S, Yuan Q, Baron R, Lanske B. Genetic ablation of sfrp4 in mice does not affect serum phosphate homeostasis. Endocrinology 2011;152:2031-2036. PMID: 21427221