Laboratory experiments were performed in collaboration with the Central Scientific Research Laboratory of the Institute of Medical Sciences (Ulaanbaatar, Mongolia) and the Hepato-Biliary-Pancreatic Surgical Department of the National Cancer Center of Mongolia (Ulaanbaatar, Mongolia). Liver tissue and serum samples from patients with HCC were collected between October 2022 and March 2023 at the National Cancer Center in Ulaanbaatar, Mongolia. Serum samples from the control and risk groups were collected during the same period at The Third Central Hospital in Ulaanbaatar, Mongolia.

Serum samples were collected from a total of 270 participants, comprising the HCC group (n=90), the risk group (RG) (n=90) and the healthy control group (n=90). The RG included patients with chronic hepatitis, toxic hepatitis, alcohol-associated liver disease and other liver disorders. The average age of the participants was 61.0±9.5 years. A total of 107 (39.7%) of the subjects were males and 163 (60.3%) were females.

A total of 64 people who were diagnosed with HCC and underwent surgical treatment were included in the tissue analysis. Following surgery, 14 participants were excluded from the study, as pathological examination of the cancer tissues extracted during surgery revealed that their diagnoses were other than HCC. GPC3 protein expression was evaluated in both the cancerous tissue and the surrounding tissue of 50 patients with HCC using an IHC staining method. The average age of the participants was 64.06±9.1 years. A total of 21 (42%) of the subjects were males and 29 (58%) were females.

Tumor staging was conducted based on the 8th tumor-node-metastasis (TNM) classification system for liver cancer [The National Comprehensive Cancer Network^® (NCCN^®), version 5.2020(27): AJCC, 8th edn, 2017(28)]. Tumors were classified according to the predominant histological subtype proposed by the 2017 American Joint Committee on Cancer classification system.

The present study was approved by the Ethics Committee of Mongolian National University of Medical Sciences (approval nos. 2022/3-05 and 2022.05.20; Ulaanbaatar, Mongolia), and the various experimental procedures were performed according to the Declaration of Helsinki (2013). All patients provided written informed consent.

GPC3 in serum of patients was detected using a Human Glypican 3 Quantikine^® QuicKit™ enzyme-linked immunosorbent assay (ELISA) Kit (cat. no. DGLY30) from R&D Systems, Inc. ELISA analysis was performed according to the manufacturer's protocol using a BIOBASE-EL10A ELISA microplate reader. The optical density was determined at 450 nm. All procedures were performed at room temperature and all samples were measured in triplicate.

The protein expression of GPC3 was assessed on extracted HCC tissues through IHC staining using an anti-GPC3 antibody (200:1; cat. no. SC-65443; Santa Cruz Biotechnology, Inc. https://www.scbt.com/home). Liver tissues were fixed in 10% neutral buffered formalin for 24 h at room temperature. Paraffin embedded tissue blocks were cut into 4-µm sections for IHC staining. The prepared tissues were deparaffinized in xylene at room temperature (18-22˚C) for 10 min, followed by rehydration in a 100, 100, 95, 80 and 70% ethanol series at room temperature (2 min each dilution of ethanol). Subsequently, the tissue slices were subjected to antigen retrieval at 120˚C for 10 min in 10 mmol/l citrate buffer (pH 6.0).

Subsequently, endogenous peroxidase activity was quenched by treatment with 3% hydrogen peroxide dissolved in methanol, followed by treatment with the anti-GPC3 primary antibody. The antibody was incubated with the tissues at room temperature for 1 h. Subsequently, a secondary antibody [broad-spectrum secondary antibody solution; cat. no. D01-110; Golden Bridge International (GBI) (Labs) Ltd. https://www.gbiinc.com/] was added, and incubated with the tissue at room temperature for 15 min according to the manufacturer's protocols. Streptavidin-HRP was added, and incubated at room temperature for a further 15 min. Finally, a drop of aminoethyl carbazole solution [cat. no. C01-12; GBI (Labs) Ltd. https://www.gbiinc.com/] was applied to the tissue, which was observed under a microscope (Olympus BX41; Olympus Corporation) until a red color developed.

The paraffin-embedded liver cancer tissue was observed under a light microscope (Olympus CHA; Olympus Corporation), and a suitable area was marked out. A 2-mm diameter punch was inserted into the cancer tissue to a depth of 5 mm, which was then placed into a newly prepared paraffin block. After inserting all the tissues, 1-µm thick tissue sections were cut from the prepared tissue microarray block using a microtome prior to IHC analysis as aforementioned.

When evaluating the results, the expression of GPC3 protein was categorized as negative, weakly positive or strongly positive based on the percentage of cells exhibiting red-brown staining, and the intensity of the staining within the cell membrane, cytoplasm and nucleus. The following criteria were used to assign scores, which were determined according to the average scores calculated by histopathologists and researchers: Score, 0: Percentage of stained cells <5%; score, 1: Percentage of stained cells 5-25%; score, 2: Percentage of stained cells 26-50%; and score, 3: Percentage of stained cells >50%. In terms of the color intensity, the following values were assigned: Score, 0: No staining; score, 1: Weak staining; score, 2: Moderate; and score, 3: Strong staining. To calculate the final score, the percentage of stained cells score was multiplied by the intensity of staining score (range 0-9) as follows: Negative, 0-1; weakly positive, 2-4; moderate staining, 5-7; strongly positive, 8-9.

Statistical analysis for the tables was performed using Chi-square (χ²) test or Fisher's exact test [to test the association between the sGPC3 level and the patient's clinical and histopathological characteristics] with SPSS software (version 24.0; IBM Corp.). One way ANOVA test was applied and followed by Tukey's test to analyze the differences in sGPC3 levels among the groups. P<0.05 was considered to indicate a statistically significant difference. Receiver operating characteristic (ROC) curves were performed to define the optimal cut-off values, and to assess sensitivity, specificity and respective areas under the curve (AUCs). Data are expressed as the mean ± SD, or n (%). All experiments were conducted in triplicate.

The sGPC3 levels in the 3 different experimental groups (HCC group, RG and control group) are shown in Fig. 1. The median sGPC3 level was found to be 327.25±98.22 pg/ml in the HCC group (n=90), whereas the median values ± range were 253.21±29.53 pg/ml in the RG (n=90) and 245.31±23.38 pg/ml in the control group (n=90).

Significant differences in the sGPC3 level were noted among these 3 groups (P=0.001); however, the sGPC3 levels in the RG were not significantly different from those in the control group (P>0.05; Fig. 1A).

The ROC curve of sGPC3 is presented in Fig. 1B. Based on the ROC curve, the cut-off value was set to 270 pg/ml. The sensitivity and specificity percentage values were 83.3 and 84.4% respectively, with AUC=0.892. The positive predictive value was 83%, whereas the negative predictive value was 91.02%.

The χ² test and Fisher's exact test were performed on the HCC group (n=90) to test the association between the sGPC3 level and the patients' clinical and histopathological characteristics (Table I). Patients were grouped into two groups with respect to the sGPC3 level, namely >270 pg/ml (n=17) or <270 pg/ml (n=73). A statistically significant association was observed between sGPC3 levels and cirrhosis, as well as between sGPC3 levels and hepatitis C virus (HCV) infection. However, no significant associations were observed between sGPC3 and the other clinical characteristics of the patients. Statistically significant values are shown in bold (Table I).

A total of 50 participants from the HCC group (21 men and 29 women aged 33-79 years, with an average age of 64.06±9.1 years) were evaluated for GPC3 expression with IHC analysis. Through IHC staining, GPC3 protein was found to be positively stained in the cytoplasm, membrane and canaliculi of cells in 38 out of 50 (76%) participants. Representative images of IHC staining are demonstrated in Fig. 2. Among these, 16 of 38 (42.1%) participants exhibited weak positive staining, whereas the remaining 22 (57.9%) displayed strong positive staining. TMA construction and representative images in different magnifications are shown in Fig. 3.

In early-stage cancers, GPC3 protein expression was found to be absent in 8 out of 32 cases (25%), weakly positive in 10 out of 32 cases (31.3%), and strongly positive in 14 out of 32 cases (43.8%). Similarly, the expression of GPC3 was absent in 4 out of 18 cases (22.2%), weakly positive in 6/18 cases (33.3%), and strongly positive in 8/18 cases (44.4%) (P>0.05) in late-stage cancers.

In Table II, the patients are grouped according to the tissue expression of GPC3. After having performed the χ² test and Fisher's exact test, clinical characteristics such as age, sex, hepatitis B virus (HBV) infection, HCV infection, cirrhosis, tumor number and tumor size did not show statistically significant association with GPC3 protein expression (P>0.05).

However, a statistically significant association was observed between GPC3 protein expression and serum AFP (sAFP) levels: In particular, when the sAFP level was either normal or <20 ng/ml, the GPC3 protein was positively stained in 13/21 (61.9%) of cases. On the other hand, when the sAFP level was >20 ng/ml, the GPC3 protein was positively stained in 25/29 (86.2%) of cases (P=0.047).

In addition, histopathological characteristics, including fibrosis, steatosis, histological grade, histological cell type, differentiation, TNM stage and vascular invasion, did not show statistically significant association with GPC3 protein expression (P>0.05) (Table II).