Introduction

Hepatocellular carcinoma (HCC) is one of the malignant cancer-related deaths worldwide and is the third most common malignancy and the second leading cause of cancer-associated death in Taiwan [1]. The prognosis of HCC is destitute and only rarity of HCC patients are cured because cancer cells inevitably survive, which results in HCC recurrence or metastasis, and the therapeutic effect for inhibition of cancer cell growth also remains in straitened circumstances [2–5]. Therefore, detection of candidates who are susceptible to HCC for preventing this malignance is strongly recommended.

Tissue inhibitor of metalloproteinase (TIMP) family, including-1, 2, 3, and 4, are associated with the inhibition of matrix metalloproteinase (MMP), which contributes to the degradation of extracellular matrix (ECM) for tissue remodeling, cancer cell invasion and metastasis [6–8]. Among TIMP family, TIMP-3 is suggested to induce apoptosis, cause cell cycle arrest, and inhibit MMP activity, angiogenesis, tumor growth, invasion, and migration of HCC cells [6, 7, 9–12]. The expression of TIMP-3 are markedly suppressed in human HCC cancer cells [6, 11, 13] and the liver of mice with diet-induced hepatocarcinogenesis [10]. It was demonstrated that repeated injection of plasmid encoding TIMP-3 gene can inhibit tumor growth of human HuHZ HCC cells xenografted into nude mice [9]. The expressions of TIMP-3 and TIMP-4 are significantly reduced in malignant prostate tissues compared to benign tissues [14] as well as significantly downregulated in lung tumors and lung cancer cell line, and suggested that hypermethylation in the promoter region of TIMP-4 could be involved in the transcriptional repression of TIMP-4 in lung cancer [15]. Overexpression of TIMP-4 by transfecting of a full-length TIMP-4 complementary DNA(cDNA) into MDA-MB-435 human breast cancer cells significantly inhibited MMP activity and cancer cell invasion as compared to non-transfected cells [16]. Besides, an animal study showed that TIMP-4 positive tumors grew more slowly and revealed a less metastasis than TIMP-4 negative tumors [16]. Also, both TIMP-3 and TIMP-4 were significantly increased in response to treatment with 20 μM and 40 μM of curcumin, a biphenyl compound in the herb Curcumal longa and possesses anticancer effect, in MDA-MB-231 human breast cancer cell line, and their overexpression correlated with downexpression of MMP-2 and MMP-9 in a concentration- and time-dependent manner [8]. We therefore hypothesized that both genes contribute to the regulation of HCC development and prognosis.

TIMP-3 and TIMP-4 genes are separately located on chromosome 22q12.1 and 3p25 [15]. The promoter polymorphisms of C allele at the −1296 T>C (rs9619311) position of TIMP-3 [17], and C allele at the −55 T>C (rs3755724) position of TIMP-4 were respectively identified [18]. We suggested both TIMP-3 -1296 T>C and TIMP-4 -55 T>C gene polymorphisms could alter the production or stabilization of functional proteins and influence the susceptibility of HCC [15]. However, the roles of these two gene polymorphisms on the risk of HCC are not investigated. The aim of this study was to estimate genetic impact of TIMP-3 -1296 T>C (rs9619311) and TIMP-4 -55 T>C (rs3755724) gene polymorphisms on the susceptibility and clinicopathological characteristics of hepatocellular carcinoma among Taiwanese.

Materials and methods

Subjects and specimen collection

This was a hospital-based case–control study. A total of 229 patients with hepatocellular carcinoma diagnosed at Chung Shan Medical University Hospital, Taichung, Taiwan, were recruited as a case group between April 2006 and August 2013. The diagnoses of HCC were according to the characteristic criteria of the national guidelines for HCC [19], such as liver injury diagnosed by either histology or cytology irrespective of a-fetoprotein (AFP) titer where imaging data showed either one of following three cases: (1) one or more liver masses more than or equal to 2 cm in diameter via both computed tomography (CT) and magnetic resonance imaging (MRI), (2) one imaging data with early enhancement and a high level of AFP more than or equal to 400 ng/mL, and (3) one imaging data with early arterial phase contrast enhancement plus early venous phase contrast washout regardless of AFP level. A total of 530 healthy controls, who visited the Department of Family Medicine, Chung Shan Medical University Hospital, Taiwan for health examination, were selected based on no risks related to hepatocellular carcinoma and matched on demographic data of race, ethnic groups, and residential area. The whole blood specimens, collected from healthy controls and HCC patients, were placed in tubes containing EDTA and immediately centrifuged and stored at −80 °C. Associated clinicopathological characteristics, such as HBsAg, anti-HCV, liver cirrhosis history, Child-Pugh grade, AFP, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and stage of HCC, were verified by chart review. Before commencing the study, approval was obtained from the Institutional Review Board of Chung Shan Medical University Hospital and informed written consent was obtained from each individual.

Genomic DNA extraction

Genomic DNA was extracted from whole blood samples collected from study subjects by QIAamp DNA blood mini kits (Qiagen, Valencia, USA) according to the manufacturer’s instructions [20]. DNA was dissolved in TE buffer [10 mM Tris (PH 7.8), 1 mM EDTA] and then quantitated by a measurement of OD260. Final preparation was stored at −20 °C and used as templates in polymerase chain reaction (PCR) [21].

Real-time PCR

Allelic discrimination of the TIMP-3 -1296 T>C (rs9619311) and TIMP-4 -55 T>C (rs3755724) gene polymorphisms was assessed with the ABI StepOne™ Real-Time PCR System (Applied Biosystems, Foster City , CA , USA) and analyzed using SDS vers. 3.0 software (Applied Biosystems), with the TaqMan assay. The FAM-primers used for analysis of TIMP-3 -1296 T>C (rs9619311) and TIMP-4 -55 T>C (rs3755724) genes polymorphisms were designed as FAM-5’- GAAGGGTGGAGCCCTGTC and FAM-5’- AGCTGCAGGAAGTGCTTTCAA, respectively. For each assay, appropriate controls (nontemplate and known genotype) were included in each typing run to monitor reagent contamination and as a quality control. The final volume for each reaction was 5 μL, containing 2.5 μL TaqMan Genotyping Master Mix, 0.125 μL TaqMan probe mix, and 10 ng genomic DNA. The real-time PCR included an initial denaturation step at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. To validate results from real-time PCR, around 5 % of assays were repeated and several cases of each genotype were confirmed by the DNA sequence analysis [22].

Statistical analysis

Hardy–Weinberg equilibrium was assessed using a goodness-of-fit χ 2 test for biallelic markers. The average is presented as the mean ± SE. A Fisher’s exact test was used to compare the differences of demographic characteristics distributions between healthy controls and patients with HCC, since the small sample size was present in some categorical variables. The one-way ANOVA test was used to detect the difference of continuous variables among three groups, and Scheffe correction was performed to check statistically significant difference between groups. The odds ratios (ORs) with their 95 % confidence intervals (CIs) of the association between genotype frequencies and hepatocellular carcinoma risk as well as clinical pathological characteristics were estimated by multiple logistic regression models after controlling for other covariates. A P value <0.05 was considered significant. The data were analyzed on SAS statistical software (Version 9.1, 2005; SAS Institute Inc., Cary, NC).

Results

In our recruited control group, the frequencies of genetic polymorphisms such as TIMP-3 -1296 T>C (p > .05; χ 2 value, 0.30) and TIMP-4 -55 T>C (p > .05; χ 2 value, 0.047) were in the Hardy–Weinberg equilibrium.

The demographical data and genetic distributions are shown in Table 1. There were significant differences of age and tobacco consumption between hepatocellular carcinoma patients and healthy controls. No association between gene polymorphisms and HCC was found. The adjusted odds ratios and 95 % confidence intervals were 0.83 (95%CI = 0.52–1.33; p = 0.48) and 0.29 (95%CI = 0.03–2.70; p = 0.28) to have HCC among individuals with TC and CC alleles compared to TT alleles of TIMP-3 -1296 T>C. People with TC and CC alleles of TIMP-4 -55 T>C genetic polymorphisms had 1.18 fold (95%CI = 0.83–1.68; p = 0.35) and 1.20 fold (95%CI = 0.70–2.07; p = 0.50) increased risk of developing HCC compared with those with TT homozygotes (Table 1).

Table 1 The distributions of demographical characteristics and gene polymorphisms in 530 healthy controls and 229 patients with hepatocellular carcinoma
Full size table

The study also determined whether there was an interaction effect of gene-to-related-risk-factors on HCC susceptibility. The adjusted odds ratios and 95 % confidence intervals of genotypic frequencies and HCC susceptibility were estimated among persons with exposure to HCC-related risk factors, respectively. There was no significant association between genetic polymorphisms of TIMP-3 -1296 T>C and TIMP-4 -55 T>C and HCC susceptibility among participants who had exposure to related environmental risk factors, including alcohol and tobacco consumption (Table 2). Also, there was no significant age-related association between genetic polymorphisms of TIMP-3 -1296 T>C and TIMP-4 -55 T>C and HCC susceptibility (Table 3). However, among women group, subjects with TC or CC alleles of TIMP-3 -1296 T>C protected against HCC (AOR = 0.35, 95%CI = 0.12–0.97; p = 0.04) compared to subjects with TT alleles, after adjusting for other confounders. Also, women with TC alleles and with TC or CC alleles of TIMP-4 -55 T>C had a 2.52-fold (95%CI = 1.23–5.13; p = 0.01) and 2.47-fold (95%CI = 1.26–4.87; p = 0.008) of developing HCC compared to women with TT homozygotes, after adjusting for other confounders (Table 4).

Table 2 Adjusted odds ratio (AOR) and 95 % confidence interval (CI) of interaction effect between genotypic frequencies and hepatocellular carcinoma related environmental risk factors
Full size table
Table 3 Adjusted odds ratio (AOR) and 95 % confidence intervals (CIs) of hepatocellular carcinoma associated with genotypic frequencies in different age groups
Full size table
Table 4 Adjusted odds ratio (AOR) and 95 % confidence intervals (CIs) of hepatocellular carcinoma associated with genotypic frequencies in different gender groups
Full size table

The relationship between gene polymorphisms of TIMP-3 -1296 T>C and TIMP-4 -55 T>C and clinical pathological characteristics are shown in Table 5 and Table 6, respectively. There was no significantly association between gene polymorphisms and clinicopathological status of hepatocellular carcinoma.

Table 5 Adjusted odds ratio (AOR) and 95 % confidence intervals (CI) of clinical statuses and TIMP-3 genotype frequencies in hepatocellular carcinoma patients (n = 229)
Full size table
Table 6 Adjusted odds ratio (AOR) and 95 % confidence intervals (CI) of clinical statuses and TIMP-4 genotype frequencies in hepatocellular carcinoma patients (n = 229)
Full size table

Furthermore, we estimated the relations between genotypic frequencies and serum levels of liver-related clinicopathological markers such as alpha-fetoprotein, aspartate aminotransferase (AST), alanine aminotransferase (ALT), as well as the ratio of AST to ALT among HCC patients. Also, there was no significantly association between liver-related clinicopathological markers and genotypic distribution of TIMP-3 -1296 T>C and TIMP-4 -55 T>C (Table 7).

Table 7 The expression of clinicopathological markers in different TIMP-3 and TIMP-4 genotype frequencies of hepatocellular carcinoma patients (n = 229)
Full size table

Discussion

In this study, we firstly provide novel information of the impacts of TIMP-3 -1296 T>C (rs9619311) and TIMP-4 -55 T>C (rs3755724) gene polymorphisms on the susceptibility of hepatocellular carcinoma among women.

TIMP-3 and TIMP-4 are associated with the inhibition of MMPs activity, tumor cell proliferation, invasion, and migration; and their expression are downregulated in several cancer cells; and therefore, their genetic polymorphisms could affect the process of carcinogenesis [6–16]. Only few studies investigated the effect of TIMP-3 -1296 T>C gene polymorphism on cancer susceptibility [23–25], but no study estimated the association between TIMP-4 -55 T>C genetic polymorphism and cancer risk. Lei et al. recruited 959 patients with breast cancer and 952 healthy control to investigate the association between polymorphisms and breast cancer susceptibility and found the C allele carriers of the TIMP-3 -1296 T>C SNP slightly increased the susceptibility to have breast cancer among Swedish (OR = 1.25, 95%CI = 1.05–1.50) [23]. However, Peterson et al. found no association between TIMP-3 -1296 T>C gene polymorphism and breast cancer risk and survival among enrolled 1,062 patients with breast cancer and 1,069 healthy control [24]. Also, there was no significantly different distribution of TIMP-3 -1296 T>C polymorphism between 241 patients with bladder cancer and 199 healthy controls [25].

To the best of our knowledge, both TIMP-3 -1296 T>C and TIMP-4 -55 T>C genetic polymorphisms have not been studied in HCC. Our current study was designed to estimate the relationship between both gene polymorphisms and HCC susceptibility, and we firstly found that C allele of TIMP-3 -1296 T>C and TIMP-4 -55 T>C genetic variants significantly associated with the susceptibility of HCC among women but not in male people. In female population, individuals with TC or CC alleles of TIMP-3 -1296 T>C gene polymorphism significantly protected against HCC (AOR = 0.35, 95%CI = 0.12–0.97; p = 0.04) compared to individuals with TT alleles, after adjusting for other confounders. Moreover, women with TC alleles and with TC or CC alleles of TIMP-4 -55 T>C polymorphisms had a 2.52-fold risk (95%CI = 1.23–5.13; p = 0.01) and 2.47-fold risk (95%CI = 1.26–4.87; p = 0.008) of developing HCC compared to individuals with TT alleles, after adjusting for other confounders. It has been reported that promoter hypomethylation at −699, −502, −880, and –928 bp upstream of the transcription start point of TIMP-3 significantly increased the gene expression of TIMP-3 [26]; however, hypermethylation of TIMP-3 resulted in decreased expression of TIMP-3 and was associated with the risk of glioblastoma and pancreatic endocrine tumors [27, 28]. It is believed that women have a lower incidence of HCC than men partly because of protective effect of estrogen [29, 30]. It was found that postmenopausal hormone replacement therapy could be a protective factor from HCC [29]. Estrogen receptor-α promoter hypermethylation and reduced gene expression were found in HCC samples [30, 31], and the decreased expression was significantly related to stimulate hepatoma cell proliferation [30], high liver damage score, pathological invasion of the intrahepatic portal vein as well as the size of tumor [31]. Celebiler Cavusoglu et al. collected 62 invasive breast tumors to estimate breast cancer-related gene expression, they found the level of TIMP-3 was positively correlated with that of estrogen receptor-α (p = 0.0001) [32]. Lei et al. found that breast cancer patients who were both TT homozygotes at position −1296 in TIMP-3 tended to have both estrogen and progesterone receptor negative tumors (OR = 1.81, 95 % CI = 1.03–3.21 and OR = 2.10, 95%CI = 1.18–3.86, respectively) [23]. Moreover, it was speculated that estrogen regulated the expression of TIMP-4 by binding to the TIMP-4 promoter region and downregulation of TIMP-4 mRNA followed downregulation of estrogen receptor-α [33, 34]. We suggested that women with TC or CC alleles of TIMP-3 -1296 T>C gene polymorphism could tend to promote the level of estrogen receptor or advantage of a high level expression of TIMP-3 that maybe modulated by estrogen receptor-α, and lead to their protective effect to against HCC. In addition, both estrogen and estrogen receptor-α gene expressions make a profit of high level gene expression of TIMP-4 among women, which benefit to women from HCC; unfortunately, female individuals with TC alleles and with TC or CC alleles of TIMP-4 -55 T>C polymorphisms could limit the expression of TIMP-4 which upregulated by estrogen and estrogen receptor-α, and subsequently increased their risk to have HCC. Our novel findings promote us to pay attention to different cancer-related mechanism and preventive strategy between female and male population and offer valuable information for further investigation.

It was reported that gene expressions of TIMP-3 and TIMP-4 were increased in heart tissue among alcohol-exposed rats [35]. Genetic polymorphisms, including CCL5-28, CCL5-403, and CCR5, and their synergic effects with alcohol and tobacco consumption have been found to increase HCC risk [36]. Moreover, phase II enzymes, including glutathione s-transferase theta (GSTT1) and glutathione s-transferases P1 (GSTP1), gene polymorphisms show increased age-related susceptibility to HCC [37, 38]. Also, the level of TIMP-3 protein in HBV-derived hepatomal cells was significantly lower than that in non-HBV-derived hepatoma cells and human normal liver hepatocyte cell line [6]. We therefore estimated interaction effect of both TIMP-3 -1296 T>C and TIMP-4 -55 T>C genetic polymorphisms and HCC-related environmental risk factor, such as alcohol and tobacco consumptions, as well as age-related susceptibility on HCC risk. Beside, the relationships between genetic polymorphisms and clinical statues and serum expression of clinicopathological markers among HCC patients were estimated. There was no synergistic effect between gene polymorphism and environmental risk factors and age-related susceptibility to HCC. Also, no association between these two genetic polymorphism and clinical statuses as well as clinicopathological markers was found. We hypothesized that these two genetic variants might not influence the progression of hepatocellular carcinoma and the contribution of these two gene polymorphisms on the susceptibility of HCC among women could be through the estrogen-related mechanism instead of affecting liver injury-related clinicopathological characteristics.

In conclusion, both TIMP-3 -1296 T>C (rs9619311) and TIMP-4 -55 T>C (rs3755724) gene polymorphisms are important factors for the susceptibility of hepatocellular carcinoma among Taiwan women but they might not influence the clinical pathological progression of HCC, and the contribution of these two gene polymorphisms on the susceptibility of HCC among women could be through estrogen-related mechanism.