The medical ethics committee of the Hospices Civils de Lyon approved the study, and the informed consent was obtained from patients in accordance with the Declaration of Helsinki and with institutional guidelines. The study population consisted of paired tumor and non-tumor tissues from 40 patients suffering from HCC including 10 HBV-, 10 HCV-, 10 alcohol-related and 10 HCC without HCV viral or alcohol exposure (idiopathic HCC). Serological markers for HBV (HBsAg, HBeAg and anti-HBe), HCV (anti-HCV) were tested in all cases by commercial enzyme immunoassays. The clinical and histological data accompanying the samples analyzed are shown in Table 1. Tissue samples from patients resected for HCC were collected during surgery and divided into two parts: one was immediately cut into small pieces, snap-frozen in liquid nitrogen and stored in deep freezer; the other was fixed in 10% formalin and paraffin-embedded for histopathological examination and immunohistochemistry. Histological analysis and immunohistochemistry were performed at the Department of Pathology at the Hospices Civils de Lyon.

Each sample was homogenized and total cellular DNA was extracted using phenol chloroform. The average telomere length was measured in all samples using the TeloTAGGG Telomere length Assay (Roche). Briefly, purified genomic DNA (6–8 μg) was digested by specific restriction enzymes. The DNA fragments were separated by gel electrophoresis and transferred to a nylon membrane using Southern blotting. The blotted DNA fragments were hybridized to a digoxigenin-labeled probe specific to telomere repeats and incubated with a digoxigenin-specific antibody coupled to alkaline phosphate. Finally, the immobilized probe was visualized by a sensitive chemiluminescence substrate and the average TRF length was assessed by comparing the signals relative to a molecular weight standard.

The telomeric repeats amplification protocol (TRAP) was combined with real-time detection of amplification products to determine telomerase activity using a Quantitative Telomerase Detection kit (US Biomax) following the manufacturer’s recommendations. Total protein extracts (0.5 μg) were used for each reaction. The end products were resolved by PAGE on a 12.5% non-denaturing gel, stained with Sybr Green Nucleic Acid gel stain (Invitrogen) and visualized using the Bio-Rad Molecular Imager ChemiDoc System.

Each tissue sample was homogenized and total cellular RNA was extracted using the MasterPure Complete DNA and RNA Purification Kit (Epicentre) according to the manufacturer’s instructions. Before reverse transcription, RNA was treated with DNase (Invitrogen-Life technology) to prevent DNA contamination. First-strand complementary DNA (cDNA) was synthesized from 0.5 μg RNA using random primers (Promega) and Superscript II reverse transcriptase (Invitrogen). The RNA concentration and purity were determined using a NanoDrop instrument (Thermo Scientific). The primer sequences are available upon request. Primer sets used to quantify gene expression were first tested in PCR with a control cDNA to ensure specific amplification, as evidenced by the presence of a unique specific signal after agarose gel electrophoresis. PCR assays were performed on an ABI Prism 7000 sequence detection system (Applied Biosystems) using 5 μL of cDNA, 6 μL of SYBR Green Master Mix, 0.25 μL of ROX (Invitrogen) and 0.75 μL of primers at 10 μM. Thermal cycling consisted of a first cycle at 50°C for 2 min and 95°C for 10 min, followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 min. Finally at the end of each PCR run, temperature was raised up to 95°C in order to check the melting curve. The expression of each gene of interest was normalized against 2 housekeeping genes, Gus (NM_000181) and ERCC1 (NM_001983), which have been validated using the BestKeeper software tool 32 to adjust for variations in RNA amount and cDNA synthesis. The relative quantification was depicted as the fold-change in expression of each gene using the formula 2ΔΔCt, as previously described 33 . Each assay was performed in duplicate.

The antibodies to MRE11 were purchased from Santa Cruz Biotechnology (Sc-22,767), hTERT (ab-32,020) and POT1 (ab-124,784) were purchased from Abcam, TRF2 (05-521) was purchased from Millipore, and the antibodies to Ku80 (2753S) and beta-Actin (4967S) were purchased from Cell Signaling Technology. Tumor extracts were homogenized and then lysed. The protein concentration was determined using the Bio-rad D_c Protein Assay Kit (Bio-rad). Equal amounts of proteins were subjected to sodium dodecyl sulfate 10% polyacrylamide gel electrophoresis (MiniProtean TGX, BioRad) and fractionated proteins were transferred to PVDF membranes (Transblot Turbo, BioRad). These membranes were blocked in TBS containing 5% nonfat milk, 0.05% Tween 20, and then probed with the appropriate antibody followed by incubation with a secondary IgG HRP-linked antibody (Cell Signaling Technology). The blots were then developed using an enhanced chemiluminescence detection system (Clarity Western ECL, BioRad).

Ki67 was assessed using the anti-Ki67 MoAbs (clone MIB-1 Dako, on BenchMark Ventana XT). CC1 treatment (1/50) was performed before ultraview revelation. Ki67 immunohistochemistry was quantified by a pathologist. The percentage of labeled nuclear area over the total neoplastic and the non-neoplastic nuclear area in the section was quantified from 2000 cells in areas of highest nuclear labeling.

Statistical analysis was performed using the 2-tailed Student’s t test or the Mann–Whitney U rank sum test. P < 0.05 was considered statistically significant in all analyses. All data analyses were performed using SPSS statistical software version 20.

Expression of the proliferative marker Ki67 was not significantly different between the 8 HBV positive cirrhotic samples and the 12 non-cirrhotic liver samples deriving from patients with HCC. As illustrated in Figure 1A, the level of hTERT expression was significantly higher in the 8 HBV positive cirrhotic samples than in the 12 non-cirrhotic liver samples (p = 0.040, Mann–Whitney test). In contrast, there was no significant difference in the level of TA between the cirrhotic and non-cirrhotic sample categories. HBV-associated cirrhosis expressed significantly lower hTR levels when compared to histologically non-cirrhotic liver tissue: 0.0053 versus 0.3574 arbitrary units (p < 10^-4, Mann–Whitney test) (Figure 1A). The TRF length was longer in HBV positive cirrhotic samples than in non-cirrhotic samples (6.60 kbp versus 5.69 kbp) but the difference was not statistically significant. Comparative Western-blot analysis of hTERT expression in HBV positive cirrhotic samples versus non-cirrhotic liver samples confirmed the qRTPCR results for hTERT expression (Figure 2B). Table 2 and Figure 1B show that all shelterin and non-shelterin telomere factors except HMRE11A and RAD50 were significantly underexpressed in HBV positive peritumoral cirrhotic samples. Comparative Western-blot analysis confirmed that protection of telomeres 1 (POT1), telomere repeat factor 2 (TRF2), HMR11A/B, and Ku80 had lower expression levels in HBV positive cirrhotic samples than in non-cirrhotic liver samples (Figure 2C and D). These results suggest that at the telomere level, the development of HBV-associated cirrhosis includes strong hTERT overexpression and considerable repression of hTR, shelterin, and non-shelterin telomere factors. Similar results were obtained when the 8 HBV⁺ cirrhotic samples were compared with the 9 non-cirrhotic liver samples derived from patients with idiopathic HCC (data not shown).

Expression of the Ki67 proliferation marker was not significantly different between the 9 HCV positive cirrhotic samples and the 12 non-cirrhotic liver samples deriving from patients with HCC. There was no significant difference in the expression level of TA, hTERT and hTR between the two sample categories (Figure 1A). Western-blot analysis of hTERT expression confirmed the qRTPCR results for hTERT expression (Figure 2B). Shelterin, POT1 and repressor-activator protein 1 (RAP1) were demonstrated to be significantly overexpressed in HCV positive cirrhotic samples when compared with non-cirrhotic liver samples. The remaining factors displayed an identical (TRF2) or a non-significant reduced expression level (Table 2). In contrast to HBV, all telomere factors except Pinx1 non-shelterin were overexpressed in cirrhotic peritumoral HCV positive samples, as compared to non-cirrhotic liver samples (Figure 1C, Table 2). Indeed, the expression of Ku80 (p = 0.029) and RAD50 (p = 0.018) was approximately 3 times higher than that of the control samples. Western-blots confirmed that POT1, HMRE11A/B, and KU80 were more expressed in HCV positive cirrhotic samples than in non-cirrhotic liver samples (Figure 2D). These results suggested that at the telomere level, the main changes that accompany the development of HCV-associated cirrhosis predominately involve the overexpression of POT1, RAP1, Ku80, and RAD50 telomere factors.

Expression of the Ki67 proliferative marker was not significantly different between alcohol-associated cirrhotic and non-cirrhotic liver tissues deriving from patients with HCC. There was no significant difference in TRF length, TA, hTERT and hTR expression between the two sample categories (Figure 1A). Western-blot analysis of hTERT expression confirmed the qRTPCR results (Figure 2B). Shelterin, POT1 (p = 0.005) and RAP1 (p = 0.006) were demonstrated to be significantly overexpressed in alcohol-associated cirrhotic tissues, whereas other shelterins were found to be underexpressed, with TRF1-interacting nuclear protein 2 gene (TIN2) showing a significant difference (Table 2). All non-shelterin telomere factors, except TANK2 and Pinx1, contained a transcriptional pattern that resembled that in HCV cirrhotic samples. Accordingly, all telomere factors except the TANK2 non-shelterin were overexpressed in cirrhotic alcohol-exposed liver with significant differences demonstrated for HMRE11A, HMRE11B, Ku70, Ku80, RAD50, TANK1, and Pinx1 (Table 2, Figure 1C). Western-blot analyses confirmed the qRTPCR results for POT1, TRF2, HMR11A/B, and KU80 (Figure 2C and D). These results suggested that at the telomere level, the main changes accompanying the development of alcohol-associated cirrhosis and fibrosis predominantly involve the overexpression of POT1, RAP1, HMRE11A, HMRE11B, Ku70, Ku80, RAD50, TANK1, and Pinx1 telomere factors. Taken together, these results indicate that the development of HBV-, HCV-, and alcohol-related cirrhosis rely on clearly distinct telomere perturbations and suggests that these distinct carcinogens possess specific effects on telomere homeostasis. Consequently, 3 kinds of cirrhotic tissues displayed significant differences in the expression of telomere factors (Figure 1, Additional file 3: Table S3).

Having demonstrated the cause-specific changes in telomere factors’ expression between cirrhotic and non-cirrhotic livers, i.e. during early hepatocarcinogenesis, we next sought to investigate whether these differences persist at the late stages of HCC development. To this end we compared telomere deregulations between cirrhotic and tumoral samples deriving from patients with HCC. We first compared the 10 HBV-associated HCC samples with their 8 cirrhotic peritumoral samples. Expression of the Ki67 proliferative marker was significantly increased in HBV-associated HCC, as compared with HBV-associated cirrhosis (p = 0.002, Mann–Whitney test). The TRF length was significantly shorter in tumor samples than in cirrhotic samples (p = 0.05, Mann–Whitney test) whereas the levels of TA and hTERT expression were significantly higher in HBV positive HCC (p = 0.017 for hTERT and p = 0.002 for TA, Mann–Whitney test) without any significant difference in the level of hTR expression between the 2 tissues (Figure 1A). Western-blotting analyses confirmed the qRTPCR results for hTERT expression (Figure 2B). Table 3 shows that with the exception of Pinx1, where there was a trend for higher expression in HCC, all shelterin and non-shelterin genes remained underexpressed in HBV positive HCC without any significant difference between cirrhosis and HCC. Western-blot analysis of TRF2, HMRE11A/B, Ku80, and POT1 confirmed the qRTPCR results (Figure 2C and D). These results suggested that at the telomere level, augmented TA and hTERT expression represent the major significant telomere deregulation distinguishing HBV-associated HCC from HBV-associated cirrhosis. Accordingly, comparison of HBV-related HCC with non-cirrhotic liver samples demonstrated similar differences as the comparison of HBV-related cirrhosis with non-cirrhotic liver samples (Additional file 4: Table S4).

HCV-associated HCC expressed higher levels of the Ki67 proliferative marker (6% versus 1%) than peritumoral cirrhotic tissue samples but the difference was not statistically significant. When compared to their peritumoral cirrhotic tissue samples, HCV positive HCC expressed higher amounts of hTERT transcripts (p = 0.54) and hTR (p = 0.021) and they displayed increased TA (p = 0.036) when compared with HCV positive cirrhosis (Figure 1A). The TRF length was shorter in HCV-associated cirrhosis than in HCC but the difference was not statistically significant (5.1 kbp versus 6.6 kbp, p = 0.39) (Figure 1A). Table 3 shows that the pattern of shelterin and non-shelterin genes expression was not significantly different between HCV-associated HCC and HCV-associated cirrhosis. Western-blot analysis confirmed qRTPCR results (Figure 2B,C, and D). These results suggested that at the telomere level, increased TA and hTR expression represent the major significant telomere deregulation that distinguishes HCV-associated HCC from HCV-associated cirrhosis.

When compared to their peritumoral cirrhotic tissue samples, alcohol-associated HCC expressed higher levels of the Ki67 proliferative marker (8% versus 1%) but the difference was not statistically significant. Figure 1A shows that TA, hTERT and hTR expressions were augmented in alcohol-associated HCC but these differences were not statistically significant. Table 3 shows that the pattern of shelterin and non-shelterin genes expression was not significantly different between alcohol-associated HCC and alcohol-associated cirrhosis. Western-blot analysis confirmed the qRTPCR results (Figure 2C and D). These results suggested that at the telomere level, there is no significant deregulation that distinguishes alcohol-associated HCC from alcohol-associated cirrhosis.