Decreased expression of zinc-alpha2-glycoprotein in hepatocellular carcinoma associates with poor prognosis

Paraffin-embedded specimens from 246 patients with HCC between 1997 and 2001 by immunohistochemistry, as well as clinical and pathological data, and 20 cases of resected fresh tissues with HCC between Jan and Jul 2011 for analysis of mRNA and protein expression, were collected from Sun Yat-Sen University Cancer Center, Guangzhou, China. The 246 cases, with histologically conformed were randomly selected from 400 patients with HCC who underwent initial surgical resection and did not receive preoperative radiotherapy or chemotherapy. Child-Pugh classification and the 15-min retention rate for indocyanine green (ICGR15) were essential to be considered in the preoperative assessment. The cases selected were based on availability of 10 years’ follow-up data and complete medical records. Patients whose cause of death remain unknown were excluded form our study. The 246 cases aged from 14 to 78 years (mean, 48 years) include 219 males (91.3%) and 27 females (8.7%). Clinicopathologic characteristics for these patients were described in Table 1. Tumor differentiation was assessed according to the criteria proposed by Edmonson and Steiner [20]. Tumor stage was defined according to American Joint Committee on Cancer/International Union Against Cancer tumor-node-metastasis (TNM) classification system [21]. The use of tissues for this study was approved by the Institute Research Medical Ethics Committee of Sun Yat-Sen University Cancer Center and written informed consent was obtained from each patient.

Total RNA was extracted from 20 pairs of HCC tissues and the adjacent non-tumorous liver tissues from 20 cases of fresh tissues with HCC using Trizol regent (BIOO Scientific Corp., Austin, TX,USA) and cDNA was synthesised using SuperScript RT kit (Promega, Madison, WI, USA) according to the manufacturer's instructions. The expression levels of AZGP1 and β-actin were measured using SYBR green-based real-time PCR performed on the Stratagene Mx3000P Real-Time PCR system. The primer sequences for AZGP1 were sense, ACGACAGTAACGGGTCTCAC and antisense, AGGCTGGGATTTCTTTGTTGAA, whereas those for β-actin were sense, TGGCACCCAGCACAATGAA and antisense CTAAGTCATAGTCCGCCTAGAAGCA. The optimised amplification protocol consisted of an initial denaturation step of 95°C for 10 min, followed by 40 amplification cycles at 95°C for 10 s, annealing at 59°C for 20 s and elongation at 72°C for 10 s. The fold-changes for AZGP1 expression levels were calculated using 2-ΔΔCt.

The 20 pairs of fresh samples including tumor tissues and the adjacent non-tumorous liver tissues were homogenated in a RIPA lysis buffer, and lysates were cleared by centrifugation (12,000 rpm) at 4°C for 30 min. The supernatant was collected and total protein (50 μg) was separated on 12% SDS-polyacrylamide gel electrophoresis (PAGE), and electrotransferred on a polyvinylidene difluoride (PVDF) membrane (Pall Corp., Port Washington, NY). After blocking non-specific binding sites for 1 h with 5% non-fat milk, the membranes were then incubated with primary mouse monoclonal antibodies against AZGP1 (Santa Cruz Biotechonlogy, CA, USA, at 1:1000 dilution), and GAPDH (Santa Cruz Biotechonlogy, CA, USA, at 1:4000 dilution) at 4°C overnight. After washing, the membrane was then incubated with the secondary anti-mouse antibody (Santa Cruz Biotechnology, CA, USA) for 1 h at room temperature. The immunoreactive signals were detected with enhanced chemiluminescence kit (Amersham Biosciences, Uppsala, Sweden). The procedures followed were conducted in accordance with the manufacturer’s instructions.

Tissue microarray was constructed according to the method described previously [22]. Briefly, all samples were fixed in formalin and embedded in paraffin. The corresponding histological hematoxylin and eosin (H&E)-stained sections were reviewed by two pathologists to determine and mark out representative areas of the paraffin blocks. Each case had four core biopsies, which contained two tumor tissues and two adjacent non-tumorous liver tissues. Each tissue core with a diameter of 0.6 mm was punched from the marked areas and re-embedded into a recipient paraffin arrayblock. Multiple sections (3-μm thick) were cut and mounted on microscope slides. After then, one section from the tissue array block was stained with H&E and confirmed that the punches had got the representative areas.

Immunohistochemical detection for AZGP1 was performed using a standard two-step method. The TMA sections were dried overnight at 37°C, deparaffinized in xylene and hydrated through a series of graded alcohol. Before antigen retrieval, endogenous peroxidase activity was inhibited with 3% hydrogen peroxide for 20 min. Then the slides were boiled in ethylene diamine tetraacetic acid (EDTA; 1 mmol/L; PH 8.0) by a pressure cooker for 5 min. After washing five times in phosphate buffered saline (PBS; 0.01 mol/L; pH = 7.4), the slides were preincubated with 10% normal goat serum for 30 min to reduce nonspecific antibody binding. Subsequently, they were incubated overnight at 4°C with the rabbit polyclonal antibody against AZGP1 (Protein Tech Group, Chicago, IL, USA; 1:800 dilution). After rinsing, the slides were incubated with a secondary antibody (Dako Corp., Glostrup, Denmark) for 1 h at room temperature, and stained with 3, 3-diaminobenzidine tetrahydrochloride (DAB) after washing in PBS again. Finally, they were counterstained with Mayer's hematoxylin, dehydrated, and mounted. Slides were immunoreacted with PBS as the negative controls. Known immunostaining-positive slides were used as positive controls.

Semi-quantitative immunohistochemical detection was used to determine the AZGP1 protein levels. Cytoplasm immunoreactivity for the AZGP1 protein was scored by evaluating the sum of positive tumor cells and the staining intensity over the total number of tumor cells. In brief, the percent of positive cells was scored as "0" (0%), "1" (1%-10%), "2" (11%-50%), "3" (51%-80%), "4" (81%-100%). Intensity was scored as "0" (no staining), "1" (weakly stained), "2" (moderately stained), and "3" (strongly stained). Both the scores were decided under double-blind conditions by three independent professionals. The final immunoreactive scores were determined by multiplying the intensity scores with the extent of positivity scores of stained cells [23, 24]. The scores were then classified as follow: "-" (score 0), " + " (score 1–4), "++" (score 5–8), and "+++"" (score 9-12). For the purpose of statistical analysis, the cohort was grouped into low expression group ("-"," + ") and high expression group ("++","+++") for either 246 tumor tissues or their adjacent non-tumorous liver tissues.

Statistical analyses were performed with the SPSS 13.0 software (SPSS, Chicago, IL, USA). The correlation between AZGP1 expression and various clinical and pathological characteristics were evaluated by Paired sample t-test. Kaplan-Meier analyses were used to analyze AZGP1 as univariate in prediction of patients’ survival. Comparisons of survival distributions were done with log-rank test. The multivariate survival analyses were performed by a stepwise Cox proportional hazard model using the Wald statistics. P values of < 0.05 were considered significant.

Results showed that in 16 out of 20 cases, AZGP1 mRNA expressions in HCC tissues were noticeably lower than those in the adjacent non-tumorous liver tissues, as determined by qRT-PCR. As depicted in Figure 1A, mRNA level of AZGP1 in HCC tissue was on average 15.5-fold lower than that in corresponding adjacent liver tissue. There was significant difference in the average level of AZGP1 between tumor tissues and non-tumorous tissues (P = 0.001). To examine the protein level of AZGP1 in HCC, western blot was performed (Figure 1B). The ratio of AZGP1 expression in non-tumorous tissues to that in tumorous tissues was calculated. Results showed that AZGP1 expression in tumorous tissues was on average 1.55-fold lower than that in the paired non-tumorous tissues.

The protein status of AZGP1 was further examined by TMA and IHC, using 246 pairs of HCC and adjacent non-tumorous liver tissues (Figure 2). Results indicated that reduced expression of AZGP1 presented in 80.1% of HCC tissues (Figure 2A; C). Kruskal-Wallis's analysis revealed that AZGP1 expression was significantly decreased by more than 2-fold in tumorous tissues, compared to that in adjacent non-tumorous liver tissues (P <0.001, t = -6.502) (Figure 2D).

To investigate the relationship between AZGP1 expression and clinicopathological parameters in 246 cases with HCC, these cases were first divided into two subgroups: “Low AZGP1 expression” and “High AZGP1 expression” as defined in the Immunohistochemistry section of Materials and Methods. Significant correlations were found between AZGP1 expression and three parameters including serum AFP levels (P = 0.013), cirrhosis (P = 0.002) and tumor differentiation (P = 0.017). Patients with low AZGP1 expression had a higher tendency to have high level of serum AFP, liver cirrhosis, and poor tumor differentiation. There were no statistical connections between AZGP1 expression and the rest clinicopathological parameters, such as patient age, gender, HBsAg, tumor size, tumor multiplicity, clinical stage and vascular invasion (P > 0.05, Table 1).

The association between AZGP1 expression in HCC and the survival time of selected patients was analyzed with Kaplan-Meier survival analysis (Figure 3). The median overall survival (OS) time of low AZGP1 expression group was 34.1 months, significantly shorter than that of high AZGP1 expression group (72 months) (P = 0.006, Figure 3A). Furthermore, the median disease-free survival (DFS) time of low AZGP1 expression group was remarkably shorter than that of high AZGP1 expression group (P = 0.025, Figure 3B). Statistically, the 5-year OS rate and DFS rate of low AZGP1 expression group were 31.0% and 22.5%, respectively, whereas they were 44.8% and 32.0%, respectively, in high AZGP1 expression group. Moreover, we stratifiedly analyzed the subtype of patient with well tumor differentiation (OS: P = 0.001, Figure 3C; DFS: P = 0.017, Figure 3D) and liver cirrhosis (OS: P = 0.014, Figure 3E; DFS: P = 0.042, Figure 3F). Results showed that HCC patients with high AZGP1 expression survived longer.

Univariate analysis was also performed on 11 clinialpathological parameters that may affect postoperative survival of HCC patients (Table 2). Results indicated that AZGP1 expression, as well as serum AFP level, tumor size, tumor multiplicity, tumor differentiation, clinical stage and vascular invasion, was one of the factors that are responsible for efficacy of surgical treatment in HCC patient, by showing that AZGP1 expression was significantly correlated to OS (P = 0.007) and DFS (P = 0.033) of HCC patients. Next, we subjected the statistically significant parameters to Cox proportional hazards regression model to evaluate the significance of AZGP1 expression in HCC prognosis. Results suggested AZGP1 expression was not an effective independent prognostic factor (Table 3).

We further assessed whether AZGP1 expression in HCC was related to the risk of recurrence and metastasis. Survival analysis indicated that AZGP1 expression was negatively associated with the risk of clinical recurrence (HR = 0.698, 95% CI = 0.488 to 0.999, P = 0.046; Figure 4A) and distant metastatic progression (HR = 0.679, 95% CI = 0.471 to 0.981, P = 0.036; Figure 4B). This may suggest AZGP1 may be involved in tumor recurrence and metastasis.