The study included 240 consecutive patients who underwent liver transplantation (LT) for end-stage liver disease due to hepatitis B (n = 37, 15.4%), hepatitis C (n = 100, 41.7%) and alcoholic liver disease (n = 103, 42.9%). The main demographic and clinical characteristics are reported in Table 1. Two-hundred thirty-six healthy community blood donors served as controls. They were 164 males (69.5%) and 72 females (30.5%); the median age was 48 years with a range of 18-77 years. The male gender distribution in the patients and controls (74% vs 70%) was similar, whereas these two groups differed in age: 54 ± 8 years vs 46 ± 13 years (mean ± SD, P < 0.0001). Control subjects did not have any clinical and/or laboratory evidence of liver disease or other major pathological conditions, such as diabetes mellitus. All patients and controls were Caucasian. Informed consent to participate in the study was obtained from each subject in accordance with the Declaration of Helsinki and following the guidelines of our ethical committee. All study participants approved the storage of their frozen DNA specimens for research purposes in our laboratory.

All total hepatectomy specimens were sectioned at intervals of approximately one cm to search for suspicious focal hepatic lesions. Standard histological staining techniques were applied to confirm the presence of HCC and to evaluate the characteristics of the identified tumors, such as pathologic tumor grade (Edmondson grade) and macro- or micro-vascular invasion.

In 113 patients (47.1%), serum samples, which were collected the day before the transplant operation and were separated and stored at -80°C until analysis, were available to assay the pre-LT serum vitamin D concentration. Circulating 25-hydroxyvitamin D levels were measured using a chemo-luminescent immunoassay implemented on a Liaison automatic analyzer (DiaSorin Inc, Stillwater, MN, USA). Data were expressed in ng/mL. Reference values of serum vitamin D adopted in this study were in accordance with those proposed by the Scientific Advisory Committee of Nutrition[17], which considers serum vitamin D levels < 10-15 ng/mL to be inadequate for bone and overall health in healthy individuals.

Genomic DNA was isolated from whole blood using the QIAamp DNA Blood Mini Kit (Qiagen, Milan, Italy) according to the manufacturer’s instructions. Four diallelic polymorphisms of the VDR were genotyped: FokI C>T (rs10735810) and TaqI T>C (rs10735810) polymorphic sites on the coding sequence, BsmI A>G (rs1544410) and ApaI G>T (rs7975232) on the last intron. For the detection of the VDR polymorphisms, the polymerase chain reaction (PCR) technique was applied and followed by restriction fragment length polymorphism assays. The PCR amplifications were carried out in a total volume of 10 μL containing 10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, 0.01% Tween-20, 0.2 mmol/L deoxyribonucleotides, 2-4 pmol of each primer, 2.0 mmol/L MgCl₂ and 0.5 U hot-start Taq DNA polymerase (RighTaq, Euroclone, Milan, Italy). The sequences of primers used for FokI were (f) 5'-TGCAGCCTTCACAGGTCATA-3', (r) 5'-GGCCTGCTTGCTGTTCTTAC-3'; for TaqI and ApaI were (f) 5'-ACGTCTGCAGTGTGTTGGAC-3', (r) 5'-TCACCGGTCAGCAGTCATAG-3'; for BsmI were (f) 5'-CAGTTCACGCAAGAGCAGAG-3', (r) 5'-ACCTGAAGGGAGACGTAGCA-3'. All the primers were newly designed with the aid of the NCBI Primer-Blast Tool (http://www.ncbi.nlm.nih.gov/tools/primer-blast/). The cycling conditions for all the VDR polymorphisms were set as 40 cycles at 95°C for 30 s, 61°C for 30 s and 72°C for 1 min. In a total volume of 20 μL, amplified DNA (10 μL) was digested overnight with 2 U of restriction endonucleases using the buffers and temperatures recommended by the manufacturers. The presence of restriction sites for the FokI, TaqI, BsmI and ApaI enzymes were coded as ‘f’, ‘t’, ‘b’ and ‘a’ and the absence of restriction sites as ‘F’, ‘T’, ‘B’ and ‘A’, respectively. The FokI C>T (F/f) polymorphism was analyzed by digestion of a 157-bp PCR product with FokI (New England Biolabs, Hitchin, UK), which resulted in two fragments of 121 and 36 bp in the presence of the ‘f’ allele and in an uncut fragment in the presence of the ‘F’ allele. The TaqI T>C (T/t) and ApaI T>G (A/a) polymorphisms were analyzed by digestion of a 211-bp PCR product with TaqI (New England Biolabs, Hitchin, UK), which resulted in two fragments of 172 and 39 bp in the presence of the ‘t’ allele. These same polymorphisms were also analyzed by digestion with ApaI (New England Biolabs, Hitchin, UK), which resulted in two fragments of 121 and 90 bp for the ‘a’ allele. The BsmI A>G (B/b) polymorphism was analyzed by digestion of a 236-bp PCR product with BsmI (New England Biolabs, Hitchin, UK), which resulted in two fragments of 197 and 39 bp in the presence of the ‘b’ allele. All PCR reactions were carried out in a Techne TC-412 thermal cycle, and PCR products were sized by electrophoresis on a 3% agarose gel stained with ethidium bromide.

The statistical analysis of data was performed using the BMDP Dynamic Statistical Software Package 7.0 (Statistical Solutions, Cork, Ireland). Continuous variables were presented as median (range) or mean ± SD, whereas categorical variables were expressed as frequencies (%). Differences between continuous variables were assessed by the Student’s t-test, whereas differences between categorical variables were evaluated using the Pearson χ² test. The χ²G test “Goodness of Fit” was employed to verify whether the proportions of the four polymorphisms were distributed in controls and in patients in accordance with the Hardy-Weinberg equation. Differences in the allelic and genotypic frequencies between different groups were assessed by means of the Pearson χ² test and calculation of odds ratios with 95% confidence intervals (CI). Haplotype reconstruction from population genotype data and inferred phased diplotype calculation for each control subject or patient with liver cirrhosis were performed by means of the ARLEQUIN integrated software package for population genetics, version 3.1[18]. Analysis of molecular variance (AMOVA) with a global and a pair-wise approach was performed to assess whether haplotype allelic content differed among groups. Locus-by-locus AMOVA was utilized to assess the statistical contribution of each polymorphism. Pair-wise differences in haplotype frequencies were assessed using the exact test for sample differentiation. Linkage disequilibrium between the four analyzed VDR polymorphisms in the studied population was determined by means of the Haploview software[19]. Stepwise logistic regression analysis with a forward approach was used to verify whether the absence or presence of specific VDR haplotypes was an independent predictor of HCC.

HCC foci were detected in the native livers from 80 (33.3%) of the patients with end-stage liver disease; 33 patients (41.2%) were HCV positive, 17 (21.2%) HBV positive, and 30 (29.1%) had alcoholic liver disease. Table 1 illustrates the distribution of the values of the main demographic and clinical variables of cirrhotic patients according to the presence of HCC. Sixteen patients (20.0%) had an Edmondson grade of one, forty-eight patients (60.0%) had a grade of two, fifteen patients (18.8%) had a grade of three and one patient (1.2%) had a grade of four. Macro- and micro-vascular invasion was only observed in five cases (6.2%).

FokI C>T (F/f), BsmI A>G (B/b), ApaI T>G (A/a) and TaqI T>C (T/t) allele and genotype frequencies are reported in Table 2. No departure from the Hardy-Weinberg equilibrium equation was observed for each polymorphism in patients or controls. No significant difference was detected between patients and controls in allele or genotype frequencies. A strong linkage disequilibrium was detected between BsmI and TaqI (R² = 0.92); linkage disequilibrium was also detected between ApaI and TaqI (R² = 0.46) and between BsmI and ApaI (R² = 0.48) (Figure 1).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Schematic representation of linkage disequilibrium in the studied population (n = 476) between the four VDR polymorphisms: FokI C>T (F/f), BsmI A>G (B/b), ApaI T>G (A/a) and TaqI T>C (T/t). R² for linkage disequilibrium between each marker is reported. Shades of gray are proportional to the R² value, expressing the strength of the linkage disequilibrium.

AMOVA was performed by grouping the liver cirrhosis patients according to the etiology of their liver disease: patients with cirrhosis of viral origin (n = 137) and patients with cirrhosis of alcoholic origin (n = 103); a third group consisted of control subjects (n = 236). A significant difference was detected among these three populations (P < 0.05) by global AMOVA; the comparison of pairs of population samples demonstrated a significant difference between patients with cirrhosis of viral and alcoholic origin (P < 0.01). Locus-by-locus AMOVA showed the following significant differences for the single nucleotide gene polymorphisms tested: FokI C>T (F/f) P = 0.720, BsmI A>G (B/b) P = 0.047, ApaI T>G (A/a) P = 0.052 and TaqI T>C (T/t) P = 0.065. Estimated VDR haplotype frequencies of FokI C>T, BsmI A>G, ApaI T>G and TaqI T>C polymorphisms are reported in Table 3. A significant difference in haplotype frequencies was detected between patients with cirrhosis of viral and alcoholic origin.

Table 4 illustrates the genotype frequencies of FokI C>T (F/f), BsmI A>G (B/b), ApaI T>G (A/a) and TaqI T>C (T/t) polymorphisms in patients with liver cirrhosis grouped according to the presence (n = 80) or absence (n = 160) of HCC. Patients with HCC were more likely to carry the b/b genotype compared with the B/B + B/b genotypes of the BsmI A>G (B/b) polymorphism and the T/T genotype compared with the T/t + t/t genotypes of the TaqI T>C (T/t) polymorphism. Considering the BsmI A>G (B/b), ApaI T>G (A/a) and TaqI T>C (T/t) (BAT) polymorphisms, carriage of the A-T-C haplotype was associated with the absence of HCC and carriage of the G-T-T haplotype with the presence of HCC. We then grouped the patients as follows: group (a) carriers of the BAT A-T-C haplotype (n = 126), group (b) carriers of both BAT A-T-C and G-T-T haplotypes or of none of the two haplotypes (n = 75) and group (c) carriers of the BAT G-T-T haplotype (n = 39). A significantly linear trend for increasing frequencies of HCC was detected starting with group (a) (34/126, 27.0%) to group (b) (29/75, 38.7%) to group (c) (17/39, 43.6%, P < 0.05). Stepwise logistic regression analysis was performed to verify whether carriage of the BAT A-T-C and G-T-T haplotypes was a predictor of HCC independent of gender, age (≤/> 50 years), body mass index (</≥ 25 kg/m²), Child-Pugh score (≤/> 8) at the transplant operation, viral etiology of liver disease and presence of diabetes mellitus. The analysis confirmed that carriage of the BAT A-T-C and G-T-T haplotypes was a predictor of HCC occurrence (improvement of χ²P < 0.05, OR, 1.95, 95% CI: 1.06-3.57) independent of age > 50 years (improvement of χ²P < 0.001, OR 3.81, 95% CI: 1.75-8.29), male gender (improvement of χ²P = 0.001, OR 4.01, 95% CI: 1.80-8.96), viral etiology of liver disease (improvement of χ²P < 0.05, OR 2.19, 95% CI: 1.18-4.05) and body mass index ≥ 25 kg/m² (improvement of χ²P < 0.05, OR 1.82, 95% CI: 0.99-3.35).

AMOVA was performed by grouping the patients with liver cirrhosis according to the etiology of their liver disease (viral origin n = 137 and alcoholic origin n = 103) and the presence (n = 80) or absence (n = 160) of HCC. A significant difference was detected among these four populations (P < 0.002); the comparison of pairs of population samples demonstrated significant differences between patients with liver cirrhosis of alcoholic origin with HCC vs (a) patients with liver cirrhosis of viral origin with HCC (P < 0.0001), (b) patients with liver cirrhosis of viral origin without HCC (P < 0.0001) and (c) patients with liver cirrhosis of alcoholic origin without HCC (P < 0.01). Locus-by-locus AMOVA showed the following significant differences for the single polymorphisms: FokI C>T (F/f) P = 0.923, BsmI A>G (B/b) P = 0.000, ApaI T>G (A/a) P = 0.033 and TaqI T>C (T/t) P = 0.000. Estimated VDR haplotype frequencies of FokI C>T, BsmI A>G, ApaI T>G and TaqI T>C polymorphisms are reported in Table 5. A significant difference in haplotype frequencies was detected between patients with alcoholic liver cirrhosis with HCC vs patients with viral liver cirrhosis without HCC (P < 0.01) and patients with viral liver cirrhosis with HCC (P < 0.05). Although no relationship was detected between carriage of the BAT A-T-C and G-T-T haplotypes and the presence of HCC in patients with viral liver disease, a strong association was observed between carriage of the BAT A-T-C and G-T-T haplotypes and HCC in alcoholic liver disease (27/80 vs 17/39 vs 6/18, P = NS) (7/46 vs 12/36 vs 11/21, P < 0.002).

A significant linear trend was detected when patients were grouped according to gender and vitamin D serum levels (cut-off level 15 ng/mL) in relation to HCC occurrence. Group A and B comprised female patients with vitamin D serum levels ≤ 15 ng/mL and > 15 ng/mL, respectively; group C and D comprised male patients with serum levels of vitamin D ≤ 15 ng/mL and > 15 ng/mL, respectively. HCC was detected with increasing frequency starting with group A (0/12, 0.0%) to group B (2/12, 16.7%) to group C (16/38, 42.1%) to group D (26/51, 51.0%, P < 0.0005). A synergistic effect was found between vitamin D serum levels > 15 ng/mL and carriage of the BAT ATC haplotype; in these patients, HCC occurred in 5/26 (19.2%) cases compared with 39/87 (44.8%) of the remaining cases (P < 0.02).

Hepatocellular carcinoma (HCC) is the fifth most common cancer and the third leading cause of cancer-related death worldwide. Liver cirrhosis due to alcohol consumption or to chronic infection by hepatitis C virus (HCV) and hepatitis B virus (HBV) are considered the main etiologic agents for the development of HCC. The differences in the incidence rates and the strong gender distribution of HCC are probably not entirely due to differences in exposure to the three causative agents mentioned above. Recently, several genetic factors regarding gene polymorphisms of inflammatory cytokines and growth factors have been considered.

The vitamin D receptor (VDR) is a member of the nuclear receptor superfamily of ligand- inducible transcription factors which are involved in cell growth and differentiation. Several single nucleotide restriction length polymorphisms have been described in the VDR gene in association with cancers of the breast, prostate, colon, bladder and kidney. There are no data in the literature concerning the prevalence of VDR polymorphisms in HCC of different etiologies.

This is the first study to examine the potential role of VDR genetic polymorphisms in the occurrence of HCC in humans. In this study, the authors investigate the VDR single nucleotide gene polymorphisms FokI C>T (F/f), BsmI A>G (B/b), ApaI T>G (A/a) and TaqI T>C (T/t) (BAT) in a large series of patients who underwent liver transplantation due to liver cirrhosis with or without HCC. They demonstrate that carriage of the b/b genotype of BsmI and the T/T genotype of TaqI was significantly associated with HCC, whereas carriage of the BAT A-T-C and G-T-T haplotypes was significantly more prevalent in patients with alcoholic liver disease and HCC.

The characterization of VDR genetic polymorphisms in patients with liver cirrhosis could help to identify those who are at high risk of developing HCC. This observation could be used to modify the strategy of periodical ultrasound surveillance in this category of patients.

In this paper, the relationships between VDR gene polymorphisms and the presence of HCC have been evaluated in patients with viral or alcoholic cirrhosis. The most important result of this study is the observation that some VDR genetic polymorphisms are associated with the occurrence of HCC in patients with liver cirrhosis. The results are convincing, well presented and of some interest.