The present study analyzed the expression of distinct ZHX factors in cancers through the Oncomine database (15). When the transcriptional expression of ZHX factors in tumor tissues were compared to those in noncancerous tissues, P<0.01 with a fold-change = 2 was considered as statistically significant. Paired Student's t-test was used.

TIMER web server is an integrated online database for comprehensive analysis of immune infiltrates through multiple types of cancer (16). In the current study, the gene expression profile of ZHX factors in multiple types of cancer were evaluated via TIMER database analysis (https://cistrome.shinyapps.io/timer/).

The mRNA expression levels of specific ZHX factors in diverse types of cancer cell lines were determined using the CCLE database (http://portals.broadinstitute.org/ccle/), as described previously (3,4,17).

The prognostic impacts of ZHX mRNA levels were analyzed using the Kaplan-Meier Plotter online database, which includes the information of 54,675 genes on survival using 10,461 clinical cancer samples, including 364 from patients with liver cancer for outcome prediction analysis (18,19). Data sources contain those from the Gene Expression Omnibus (GEO), the European Genome-phenome Archive (EGA) and the Cancer Genome Atlas (TCGA). To investigate the overall survival (OS) and relapse-free survival (RFS) rates, patients were separated into high and low-expression groups according to the median mRNA expression levels so that survival analyses were conducted to produce Kaplan-Meier plots. Hazard ratio with 95% confidence interval and log-rank P-values were calculated.

The effects of genomic alterations of ZHX genes containing mutations and copy-number variance on OS and disease-free survival (DFS) rates in patients with liver cancer were analyzed using the cBioPortal online database (20,21). The raw data used prior to bioinformatic analysis are derived from GEO and TCGA. In the present study, OncoPrint in cBioPortal were employed to demonstrate the proportion and distribution of samples with genetic alterations in ZHX genes.

The immunohistochemical staining for ZHX3 protein expression was performed using a standard EnVision complex method previously described (3,4,22,23). One tissue microarray chip containing 94 primary HCC tissues and 86 adjacent noncancerous tissues was purchased from Outdo Biotech Co., Ltd. Following deparaffinization, rehydration and antigen retrieval, 4-µm sections of tissue samples were incubated with a rabbit polyclonal anti-ZHX3 antibody (catalog no. ab84677; dilution, 1:500; Abcam) overnight at 4°C. ZHX3 protein staining was visualized using an EnVision antibody complex (anti-Mouse/Rabbit) method with an Envision Detection kit (OriGene Technologies, Inc.) and 3,3′-diaminobenzidine as the chromogen substrate. Nuclei were counterstained with 0.5% hematoxylin for 2 min at room temperature.

A total of 10 random microscopic fields per slide (magnification, ×400) were evaluated by two independent observers who were unaware of the clinical information. Immunostaining was graded semi-quantitatively by multiplication of staining intensity and percentage of positive cells. The mean percentage of positively stained cells was scored as follows: 0–5% (0); 5–25% (1); 26–50% (2); 51–75% (3); and 76–100% (4). The staining intensity was categorized as follows: Absent (0); weak (1); moderate (2); and strong (3). Tumor samples exhibiting a final staining score of <2 were defined as low ZHX3 expression and those with scores ≥2 as high ZHX3 expression.

Statistical analyses were conducted using the SPSS 17.0 statistical software package (SPSS Inc.). Associations between the expression levels of ZHX factors and clinicopathological variables were assessed using the Pearson's χ² test or Fisher's exact test. Survival curves were produced using the Kaplan-Meier method and compared with the log-rank test. The prognostic significance of the clinicopathological variables was determined using a univariate Cox regression analysis. A Cox proportional hazards regression model for multivariate analysis was employed for factors that achieved significance in the univariate analysis. P<0.05 was considered to indicate a statistically significant difference.

Hitherto, three ZHX factors were characterized in a variety of types of human cancer. Our previous study revealed that the Oncomine database provided a total of 308, 434 and 416 unique analyses for ZHX1, ZHX2 and ZHX3, respectively (3). However, the mRNA levels of ZHX factors were not found in liver cancer datasets. The present study thus examined the mRNA expression of ZHX factors in multiple types of cancer using the TIMER online database. The expression of all three ZHX factors was significantly higher in liver hepatocellular carcinoma (LIHC) tissues than in normal tissues (Fig. 1). Additionally, analyses from the CCLE database revealed that the mRNA levels of ZHX1, ZHX2 and ZHX3 in liver cancer cells ranked the 27th, 11th and 23rd highest across all types of cancer, respectively (Fig. 2).

The present study next identified the prognostic impacts of ZHX family members on patient outcome via Kaplan-Meier plotter survival analysis. ZHX1 mRNA level was not significantly associated with OS in patients with liver cancer (Fig. 3A). Subgroup analyses showed no significant association between ZHX1 mRNA expression and male patients or female patients (Fig. 3B and C). Similarly, ZHX1 mRNA expression was not correlated with RFS in patients with liver cancer. Low expression of ZHX1 predicted a longer RFS rate in male patients, but not in female patients (Fig. 3E and F). Increased ZHX1 expression also displayed a longer RFS rate in patients without hepatitis virus infection (Fig. 3G and H).

No significant association was observed between ZHX2 mRNA levels and OS in patients with liver cancer (Fig. 4A). Subgroup analyses suggested that decreased ZHX2 expression indicated a longer OS rate in patients with stage III/IV tumors but not in patients with stage I/II tumors (Fig. 4B and C). Decreased ZHX2 mRNA level was associated with an improved OS in patients without hepatitis virus infection (Fig. 4D), whereas increased ZHX2 expression was associated a favorable OS in patients with hepatitis virus infection (Fig. 4E). Similarly, ZHX2 expression was not significantly associated with RFS in patients with liver cancer (Fig. 4F). High expression of ZHX2 implied longer RFS times in patients with micro vascular invasion, but not in those without vascular invasion (Fig. 3G and H).

Regarding ZHX3, its upregulation was found to be associated with a prolonged OS rate in patients with liver cancer (Fig. 5A). Subgroup analyses showed that no significant correlation between ZHX2 expression and OS either in male patients or in female patients (Fig. 5B and C). Increased ZHX3 expression exhibited longer OS times in patients with Stage I/II tumors and Stage III/IV tumors (Fig. 5D and E). High ZHX3 mRNA level represented an improved OS rate in patients with micro vascular invasion, but not in those without micro vascular invasion (Fig. 5F and G). In addition, elevated ZHX3 expression illustrated a longer OS in patients without hepatitis virus infection (Fig. 5H), but not in those with hepatitis virus infection (Fig. 5I).

The prognostic association between genetic alterations of ZHX factors and outcomes in patients with liver cancer was further characterized using the CbioPORTAL online database. The genetic alteration rates for ZHX1, ZHX 2 and ZHX3 were 10, 10 and 0.6%, respectively (Fig. 3). The genetic alteration of ZHX2 was found to be associated with OS in patient with liver cancer (Fig. 6C). Nevertheless, no other significant relationship was observed between genetic alterations of ZHX factors and patient survival, as regarding either OS or DFS (Fig. 6B and D-G).

To support the above results, the expression status of ZHX3 protein was thus examined using one tissue microarray chip containing total 94 primary HCC specimens. A high level of ZHX3 protein expression primarily in the cytoplasm of cancer cells in 48.9% (46/94) of the HCC specimens tested was observed (Fig. 7). Low ZHX3 expression was found to be associated with larger tumor size, advanced TNM staging and T stage, positive thrombus status and TP53 expression (Table I). Kaplan-Meier survival analyses demonstrated that patients with high ZHX3 expression had an improved OS compared with those with low ZHX3 expression (Fig. 8A). Subgroup analyses showed that high ZHX3 expression indicated an improved OS in patients both with T1/T2 tumors and T3/T4 tumors (Fig. 8B and C). ZHX3 overexpression also exhibited a longer OS in patients both with Stage I/II and Stage III/IV tumors (Fig. 8D and E). In addition, elevated ZHX3 suggested an improved OS in patients with histological grade I/II tumors (Fig. 8F). In the univariate analysis, larger tumor size, advanced TNM stage, higher histological grade, positive thrombus status and ZHX3 expression were determined to be associated with an unfavorable OS (Table II). After correcting the prognostic variables obtained in the univariate analysis, only histological grade and ZHX3 expression kept the independent implication in the multivariate analysis (Table II).