Sera from 56 patients with AFP-negative HCC, 86 with AFP-positive HCC and 168 with CLD, and from 59 NHC samples were obtained from outpatients or inpatients between January 2015 and January 2016 at Beijing You’an Hospital, Capital Medical University. The AFP levels of all patients were measured with a commercially available electrochemiluminescence immunoassay kit (reagents from Roche Ltd, Indianapolis, IN, United States). Samples with AFP < 20 ng/mL were defined as AFP-negative. The diagnosis of AFP-negative and AFP-positive HCC patients was based on ultrasound, computed tomography (CT) or magnetic resonance imaging (MRI) characteristics and biochemistry (AFP serology and liver function enzymes), according to the Primary Liver Cancer Treatment Protocols (2011 edition). No HCC patients received any surgical treatment, such as resection, ablation or transarterial chemoembolization, chemotherapy, radiotherapy or multikinase inhibitor (sorafenib). Patients with CLD were followed up for at least 12 mo to exclude individuals with autoimmune liver diseases. This study was approved by the Institutional Review Board of Capital Medical University, Beijing, China. All enrolled patients gave written consent.

NPM1 construct GFP-NPM WT (plasmid ID: 17578), 14-3-3zeta construct GST-14-3-3 WT (plasmid ID: 1944) and MDM2 construct pGEX-4T MDM2 WT (plasmid ID: 16237) were purchased from Addgene (Cambridge, MA, United States), and the first two were subcloned into the pET28a vector. The recombinant protein NPM1 expressed in Escherichia coli BL21 (DE3) was purified using nickel column chromatography. The recombinant protein 14-3-3zeta expressed in ArcticExpress (DE3) RP was purified using nickel column chromatography. The recombinant protein MDM2 expressed in ArcticExpress (DE3) RP was purified using SP-sepharose and DEAE Sephocel.

The recombinant proteins were examined in SDS-PAGE and the expected molecular size of expression products were determined using Coomassie blue staining. In addition, western blot analysis was used to confirm that the bands seen in SDS-PAGE were reactive with corresponding antibodies.

Three recombinant proteins were respectively diluted in phosphate-buffered saline (PBS) to a final concentration of 0.5 μg/mL for coating a 96-well microtiter plate (No. 3590; Corning, Corning, NY, United States) overnight at 4 °C. The antigen-coated wells were blocked with 10% fetal bovine serum (FBS) at 37 °C for 1 h. Human serum diluted 1:200 was incubated in the antigen-coated wells for 60 min. Horseradish peroxidase (HRP)-conjugated goat anti-human IgG (Zhongshan Golden Bridge Biological Technology Co Ltd, Beijing, China) as a secondary antibody was diluted 1:10000 for coating (1 h) followed by washing with PBS containing 0.1% Tween 20 (PBST). The 3,3’,5,5’-Tetramethylbenzidine Liquid Substrate System (Solarbio Science & Technology Co Ltd, Beijing, China) was used as the detecting agent.

The optical density (OD) value of all wells was read at 450 nm, and the cut-off value for defining a positive reaction was designated as the mean OD value of the 59 normal sera plus three standard deviations (SDs). Each microtiter plate included 10 NHC samples representing a range of absorbance values above and below the mean of 59 NHC samples, and the average OD value of 10 NHC samples was used to normalize all OD values to the standard mean of the 59 NHC samples. Each sample was tested in triplicate.

The purified recombinant proteins of three TAAs were electrophoresed on 12% SDS-PAGE and subsequently transferred to a nitrocellulose membrane. After blocking in Tris-buffered saline with 5% nonfat milk and 0.1% Tween-20 for 1 h at ambient temperature, the membranes were cut into strips and incubated with patient sera diluted 1:200, polyclonal anti-NPM1 antibody diluted 1:1000, anti-14-3-3zeta antibody diluted 1:1000 or polyclonal anti-MDM2 antibody diluted 1:1000 separately, and finally incubated with HRP-conjugated goat anti-human IgG or HRP-conjugated goat anti-rabbit IgG diluted 1:10000 for 1 h. Positive signals were detected by the ECL kit (Thermo Scientific, Waltham, MA, United States).

The liver cancer tissue array slides with normal tissue controls (9 AFP-negative HCC tissues/10 normal tissues, including pathological diagnosis and clinical information) were purchased (Outdo Biotech Co Ltd, Shanghai, China) and used to detect the expression of the three antigen proteins. Tissue array slides were baked for 1 h and deparaffinized with xylene, and dehydrated with ethanol. Antigen retrieval was performed by microwave heating method in citrate antigen retrieval solution for 20 min. After incubation with acid methanol for 15 min, goat serum blocking solution was used to prevent nonspecific binding of antibodies. The tissue microarrays were incubated with polyclonal NPM1 antibody, polyclonal 14-3-3zeta antibody or polyclonal MDM2 antibody (1:100 dilution) for overnight at 4 °C. The HRP Detection System (HRP streptavidin label and polyvalent biotinylated link) and DAB Substrate Kit (Zhongshan Golden Bridge Biotechnology Co Ltd) were used as detecting reagents. The sections were counterstained with hematoxylin, dehydrated, and mounted. The slides were observed by light microscopy (Model BX51; Olympus, Tokyo, Japan).

A χ² test with Yates’ correction was used to determine whether the frequency of autoantibodies to three TAAs in each cohort of patient sera was significantly higher than that in sera from normal individuals. Two significant levels (0.05 and 0.01) were used. Methods for calculating the sensitivity, specificity and accuracy were based on the methodology provided in Introduction to Epidemiology (6^th edition, by Ray M. Merrill, published by Jones & Bartlett Learning Company, Burlington, 2012).

The baseline characteristics of patients in the HCC, CLD and NHC groups are summarized in Table 1. Most of the patients with AFP-negative HCC at an early stage had good liver function, lower Child-Pugh score and Barcelona Clinic Liver Cancer (BCLC) grade, and were without ascites, hepatic encephalopathy, vascular invasion or metastasis.

Three recombinant proteins were used as coating antigens in enzyme-linked immunosorbent assay (ELISA) to screen for autoantibodies against NPM1, 14-3-3zeta and MDM2 in sera from patients with HCC and CLD as well as NHCs. The prevalence of autoantibodies against NPM1, 14-3-3zeta and MDM2 was 21.4% (12/56), 19.6% (11/56) and 19.6% (11/56) in AFP-negative HCC, which was significantly higher than in CLD and NHCs (P < 0.01) and higher than in AFP-positive HCC, although not significantly (Table 2). These results were confirmed by western blot analysis. Representative HCC sera with a positive reaction to NPM1, 14-3-3zeta and MDM2 in ELISA also had strong reactivity in western blotting compared to CLD and normal human sera (Figure 1).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Western blot analysis of representative sera of three anti-tumor-associated antigens autoantibodies assessed by enzyme-linked immunosorbent assay. Lane 1: The polyclonal anti-NPM1 autoantibody and anti-14-3-3zeta autoantibody were used as positive control; Lanes 2 and 3: Two representative AFP-negative HCC serum samples which were positive in ELISA also had strong reactivity to 14-3-3zeta recombinant protein in western blot analysis; Lanes 4 and 5: Randomly selected chronic liver disease sera and normal human control, respectively, with negative reactivity to 14-3-3zeta recombinant protein. AFP: Alpha fetoprotein; ELISA: Enzyme-linked immunosorbent assay; HCC: Hepatocellular carcinoma.

The presence or absence of co-expression of autoantibodies to any combination of two of the three TAAs in AFP-negative HCC, AFP-positive HCC, CLD and NHCs is shown in Figure 2. The frequency of AFP-negative HCC sera co-expressing antibodies to NPM1 and MDM2 was highest among all the combinations and significantly higher than in AFP-positive HCC, CLD and NHCs. Most of the normal human sera showed a low level of co-expression of antibodies to any combination of two of the three TAAs. Similar results were observed in AFP-positive HCC and CLD.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Analysis to determine the presence or absence of co-expression of antibodies to any combination of two of the three tumor-associated antigens in α fetoprotein-negative hepatocellular carcinoma, α fetoprotein-positive hepatocellular carcinoma, chronic liver disease and normal human control. The height of the bar represents the percentage of sera with co-expression of two antibodies, e.g., NPM1 antibody with 14-3-3zeta antibody, and NPM1 antibody with MDM2 antibody. AFP: Alpha fetoprotein; CLD: Chronic liver disease; HCC: Hepatocellular carcinoma; NHC: Normal human control; TAAs: Tumor-associated antigens.

The sensitivity and specificity of diagnosis for AFP-negative HCC were 21.4% and 95.6% with NPM1 autoantibody, 19.6% and 95.2% with 14-3-3zeta autoantibody, and 19.6% and 93.0% with MDM2 autoantibody (Table 3). In a further analysis, with combined NPM1 with 14-3-3zeta autoantibodies, the sensitivity and specificity for immunodiagnosis of AFP-negative HCC reached 25% and 94.3%, respectively. When we combined 14-3-3zeta with MDM2 autoantibodies, the sensitivity and specificity reached 25% and 92.1%, respectively. When we combined NPM1 with MDM2 autoantibodies, the sensitivity and specificity reached 30.4% and 93.0%, respectively. Finally, when we combined the three antigens, the sensitivity was still 30.4% and specificity was maintained at 91.6%. This suggested that the three TAA autoantibodies had higher consistency in the diagnosis of AFP-negative HCC. The accuracy for this TAA array was 79.5%.

The expression profiles of three proteins in AFP-negative HCC tissues and normal liver tissues was examined by immunohistochemistry of tissue array slides. Tissue array slides were commercially available for this study, and included 9 AFP-negative HCC tissues and 10 normal hepatic tissues. The polyclonal NPM1 antibody, 14-3-3zeta antibody and MDM2 antibody were used as primary antibodies to detect the expression of the three proteins in liver cancer and normal hepatic tissues. The characteristics of patients and protein expression in AFP-negative HCC are shown in Table 4.

All 9 HCC tissues and 2/10 normal hepatic tissues were positively stained in the NPM1 autoantibody group. Five of the 9 HCC tissues and 1/10 of the normal hepatic tissues were positively stained in the 14-3-3zeta autoantibody group, and 3/9 HCC tissues and 2/10 normal hepatic tissues were positively stained in the MDM2 autoantibody group. Due to the small sample size of tissues in this study, it was difficult to perform a statistical analysis. The expression of the three proteins in AFP-negative HCC tissues and normal hepatic tissues is shown in Figure 3.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Expression of NPM1, 14-3-3zeta and MDM2 in α-fetoprotein-negative hepatocellular carcinoma tissues and normal hepatic tissues by immunohistochemistry. The three polyclonal anti-TAAs antibodies were used as a primary antibody to detect their expression in liver cancer and normal hepatic tissues. A and B: HCC tissue with positive staining and normal hepatic tissue with negative staining in anti-NPM1 antibody; C and D: HCC tissue with strong positive staining and normal hepatic tissue with negative staining in anti-14-3-3zeta antibody; E and F: HCC tissue with strong positive staining and normal hepatic tissue with negative staining in anti-MDM2 antibody. AFP: Alpha fetoprotein; HCC: Hepatocellular carcinoma; TAAs: Tumor-associated antigens.

Many autoantibodies to tumor-associated antigens (TAAs) have been reported in hepatocellular carcinoma (HCC), and have been suggested to be useful tools for immunodiagnosis of HCC. However, no previous study has specifically evaluated the diagnostic value of TAA autoantibodies in α-fetoprotein (AFP)-negative HCC.

No previous study has specifically evaluated the diagnostic value of TAA autoantibodies in AFP-negative HCC.

This is believed to be the first study to evaluate specifically the diagnostic value of TAA autoantibodies in AFP-negative HCC.

This study demonstrated that autoantibodies to nucleophosmin 1, 14-3-3zeta and mouse double minute 2 homolog may be useful biomarkers for immunodiagnosis of AFP-negative HCC.

In this study, the authors determined the prevalence and diagnostic value of autoantibodies in AFP-negative HCC. Partially positive sera were further evaluated by western blotting. Immunohistochemistry was used to detect the expression of three TAAs in AFP-negative HCC and normal control tissues. The frequency of autoantibodies to three TAAs in AFP-negative HCC sera was significantly higher than in chronic liver diseases and normal human controls as well as AFP-positive HCC.