Telomeres and Telomerase in the Development of Liver Cancer

Liver cancer is one of the most common cancer types worldwide and the fourth leading cause of cancer-related death. Liver carcinoma is distinguished by a high heterogeneity in pathogenesis, histopathology and biological behavior. Dysregulated signaling pathways and various gene mutations are frequent in hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA), which represent the two most common types of liver tumors. Both tumor types are characterized by telomere shortening and reactivation of telomerase during carcinogenesis. Continuous cell proliferation, e.g., by oncogenic mutations, can cause extensive telomere shortening in the absence of sufficient telomerase activity, leading to dysfunctional telomeres and genome instability by breakage–fusion–bridge cycles, which induce senescence or apoptosis as a tumor suppressor mechanism. Telomerase reactivation is required to stabilize telomere functionality and for tumor cell survival, representing a genetic risk factor for the development of liver cirrhosis and liver carcinoma. Therefore, telomeres and telomerase could be useful targets in hepatocarcinogenesis. Here, we review similarities and differences between HCC and iCCA in telomere biology.


Introduction
Liver cancer is predicted to be the sixth most common tumor disease worldwide and the fourth leading cause of cancer-related death [1]. Liver cancer presents with a high heterogeneity in pathogenesis, histopathology and biological behavior. The heterogeneous disease in terms of etiologies reflects the poor prognosis of patients with liver cancer. The two most common types of liver cancer are hepatocellular carcinoma (HCC) (75-85% of cases) and intrahepatic cholangiocarcinoma (iCCA) (10-15% of cases) [1]. Most liver carcinomas are diagnosed at advanced stages despite the surveillance program of patients with liver cirrhosis to diagnose early liver tumors. To date, therapy options are limited to the multikinase inhibitors sorafenib and lenvatinib as first-line treatment options and regorafenib and cabozantinib as second-line treatment options for liver cancer patients [2][3][4]. A recent clinical trial revealed significantly longer overall and progression-free survival in patients with unresectable hepatocellular carcinoma, who received atezolizumab, a programmed death ligand 1 (PD-L1) inhibitor, combined with bevacizumab, a monoclonal antibody targeting the vascular endothelial growth factor (VEGF), in comparison to sorafenib only [5]. These findings point to new treatment options in patients with unresectable hepatocellular carcinoma and support the development of new therapy options.
The major risk factors in hepatocarcinogenesis are chronic viral infections with hepatitis B virus (HBV) and hepatitis C virus (HCV), heavy alcohol consumption, obesity, type 2 diabetes, smoking and

Telomere Shortening in Liver Cirrhosis and Hepatocellular Carcinoma
Chronic liver disease is associated with chronic liver inflammation, which can lead to cell death and compensatory cell regeneration. The liver is characterized by a high regenerative reserve [23], which decreases in the context of chronic liver disease, consequently leading to telomere shortening and limiting the regenerative reserve of the liver. The frequent appearance of senescent hepatocytes in liver cirrhosis is the result of the loss of telomeric repeats and the extensive proliferation [21,[24][25][26][27]. These senescent hepatocytes exhibit markers like p16 INK4a and p21 WAF1/Cip1 and are positive for senescence-associated β-galactosidase staining. A disruption of the p53-signaling pathway overcomes the senescence checkpoint and leads to further cell division of hepatocytes with already-shortened telomeres until the telomeres become critically short. At this point, the cells enter the crisis checkpoint, which is characterized by massive cell death [25,27,28].
In the liver, telomerase activity is downregulated during early embryonic development, and telomerase activity is absent in the adult liver. The healthy liver is a slowly proliferating organ, and most hepatocytes are in a quiescent stage (only one out of 20,000 cells (0.005%) is in the cell cycle [29]). Thus, telomerase activity seems not to be essential for hepatocyte function in a healthy liver. However, telomere shortening occurs in the absence of sufficient telomerase activity in hepatocytes under conditions of chronic liver diseases or upon injury [21,25]. Importantly, low levels of telomerase activity were observed under regenerative conditions, indicating the potential physiological activation of telomerase in adult hepatocytes [30,31]. So far, two distinct mechanisms were made responsible for telomere shortening upon increased proliferative signals, either due to oncogene activation (e.g., Ras mutations or c-Myc amplification) or expression of viral oncogenic proteins: (i) in most of the cases, absence of sufficient telomerase reactivation [32,33] and (ii) in some cases, germline mutations within the coding region of telomerase, which impair the enzymatic activity of telomerase in proliferating hepatocytes (see Section 5) [34,35]. Reactivation of telomerase activity has been shown in more than 80% of HCCs, which suggests that telomerase activation is a rate-limiting process for liver cancer formation [21,36,37]. The reactivation of telomerase correlates with the upregulation of both essential components TERT and TERC, respectively. It has been reported that the re-expression of TERT and activation of telomerase occurs at early premalignant stages in regenerative nodules and cirrhotic livers [31,38,39]. Importantly, telomerase activity was detected both in HCC and in iCCA to a similar extent (80-85%) [36,40]. Thus, it is important to emphasize at this point that the majority of both HCCs and iCCAs are similarly telomerase-positive, highlighting the necessity of telomerase activity for telomere functionality and tumor progression.
Irrespective of the mechanism, insufficient telomerase activity leads to accelerated telomere shortening in proliferating liver cells and as a result to genomic instability by breakage-fusion-bridge cycles (see review by Meena et al. [41]). In cells with intact DNA-damage response (DDR) checkpoints, telomere shortening leads to senescence/apoptosis and functions as a tumor suppressor mechanism. On the other hand, in cells lacking functional DDR, telomere shortening promotes genome instability and tumor formation.
Consequently, telomere shortening is an important risk factor for tumor initiation in liver carcinogenesis. The risk of tumor formation drastically increases at the cirrhosis stage, which is characterized by increased hepatocyte senescence, and upon further cell division at the crisis checkpoint by apoptosis of hepatocytes. Several studies have shown that telomere shortening is more pronounced in liver carcinoma compared to the surrounding liver tissue (see review by Satyanarayana et al. [42]). Furthermore, the progressive shortening of telomeres and the inactivation of cell cycle checkpoints in premalignant lesions led to the identification of a preneoplastic-sequence in human hepatocarcinogenesis, suggesting that small cell changes (SCC) are more advanced precursor lesions compared to large cell changes (LCC) [43]. In addition, telomere shortening was more pronounced in HCCs with a high degree of aneuploidy compared to diploid HCCs [44,45]. In fact, several studies provide evidence for a role of telomere shortening in the induction of chromosomal instability and increased risk for tumor formation [46][47][48][49]. The importance of telomere shortening and dysfunctional telomeres in HCC initiation was shown in transgenic mouse models (see Section 3).

Mouse Models of Telomere Dysfunction in Hepatocarcinogenesis
To understand the severe situation of dysfunctional telomeres and telomere shortening during chronic liver disease, transgenic mouse models were used to analyze the functions of telomeres. To this end, it is important to note that, firstly, there is a substantial difference in the regulation of telomerase between mouse and human liver. Telomerase activity is detectable in resting mouse liver but not in resting human liver [30,[50][51][52][53]. The limiting component, restricting telomerase activity in human tissues, is the catalytic subunit of the telomerase TERT [54]. Concordantly, there is a marked difference in TERT mRNA levels in human and mouse livers [52][53][54]. In fact, in vivo experimental evidence supports the idea that the species-specific differential regulation is based on different promoter organization [30,53,55]. Secondly, the average telomere length in laboratory mice is about five times longer than that of human telomeres, to some extent due to constitutive telomerase activity in mouse cells [56,57].
The telomerase knockout mouse (Terc−/−) lacking the RNA component of the telomerase enzyme was used to analyze telomere shortening in liver regeneration, chronic liver disease, and hepatocarcinogenesis [24,[58][59][60]. Mice are characterized by the existence of longer telomeres compared to humans [56]. Mice of different backgrounds differ a lot in telomere length [61]. For this reason, the Terc−/− mouse has to be crossed until the third to the sixth generation, depending on the used background strain to generate mice with critically short telomere lengths [58,60,62]. In an experimental model of liver regeneration involving the removal of two-thirds of the liver by partial hepatectomy of G3 Terc−/− and Terc+/+ mice, telomere shortening was observed to be a heterogeneous event at the cellular level, which led to the inhibition of a subpopulation of cells with critically short telomeres to enter the cell cycle and prevent those cells from participating in liver regeneration [58]. By comparing the mean telomere fluorescence intensities measured by FISH analysis, Satyanarayana and colleagues [58] observed no significant differences in TERC+/+ mice between BrdU-positive cells (952. 53 ± 144.19) and BrdU-negative cells (957.44 ± 130.57) but saw a dramatic reduction of the mean telomere fluorescence intensity of BrdU-negative cells (364. 94 ± 116.45) in comparison to BrdU-positive cells (509. 65 ± 101.30) in G3 TERC−/− mice. In a mouse model of experimentally induced acute liver damage in which Terc−/− mice were subjected to genetic, surgical and chemical impairment of the liver, dysfunctional telomeres were associated with defective liver regeneration and accelerated formation of liver cirrhosis, which could be partly rescued by adenoviral delivery of the telomerase RNA [24]. In an approach comprising three different cancer-prone model systems-(1) treatment with CCl 4 (carbon tetrachloride), (2) treatment with DEN (diethylnitrosamine) and (3) a genetic model (urokinase plasminogen activator transgenic mice) in Terc−/− mice-it could be shown that telomere dysfunction has a differential impact on tumor initiation and tumor progression. In all three model systems, dysfunctional telomeres were associated with higher amounts of tumor initiation and a decline in tumor progression [63]. In mouse models of chronic liver damage achieved by crossing HBsAg-expressing mice (the mice express the hepatitis B surface antigen under the liver-specific albumin promoter [64]) with Terc−/− mice, contrary effects of telomere shortening were shown between the beneficial effect on suppression of tumor growth and the negative effect on organismal survival [59]. In another mouse model of chronic liver disease, HBsAg mice were crossed with Terc−/− and Trp53 cKO mice. We generated mice with critically short telomeres by an intercross of Terc−/− and Terc+/− to generate siblings with loss of telomerase function in one group and telomerase expression in the other group. This study yielded the evidence for telomerase to be a critical component in the progression of Trp53-deficient hepatocellular carcinoma with short telomeres in the setting of chronic liver damage [60]. In addition, it was also shown that telomerase limits the accumulation of telomere dysfunction and the generation of aneuploidy by the activation of TRP53-independent checkpoints which suppress carcinogenesis [60]. Increased rates of chromosomal aberrations could be also shown in a DEN-induced liver cancer mouse model with dysfunctional telomeres. Telomerase knockout mice (Terc−/−) with chronic telomere dysfunction as well as a model of transient telomere dysfunction by inducing a dominant-negative variant of the TRF2 (telomeric repeat-binding factor 2) protein exhibited higher levels of chromosomal aberrations. In summary, the model of transient telomere dysfunction promotes chromosomal instability and liver carcinogenesis in telomerase-competent mice [65]. RAP1 (Ras-proximate-1 or Ras-related protein 1), like TRF2, is a component of the shelterin complex, which caps the telomere end for the protection of chromosome ends [66]. A recent publication suggested an important role of RAP1 in the protection of liver damage and liver carcinogenesis. DEN-induced Rap1−/− female mice were more prone to liver damage and hepatocellular carcinoma [67]. These models reflect the complexity and opposing roles of dysfunctional telomeres in hepatocarcinogenesis.

Telomere Shortening in Cholangiocarcinoma
Intrahepatic cholangiocarcinoma (iCCA) is the second most common malignant liver tumor which arises from the biliary tract and is characterized by a very poor prognosis with rising incidence and mortality in recent years [68,69]. The main risk factors described for HCC are also reported for iCCA. Additional risk factors are primary sclerosing cholangitis (PSC), hepatobiliary flukes, biliary duct cysts and hepatolithiasis [13]. A study by Verma and colleagues analyzed telomere shortening during aging in normal liver with no history of liver disease. Interestingly, they observed that the cholangiocytes exhibited the longest telomeres compared to all other analyzed intrahepatic lineages [70]. Similar to CD4 + and CD8 + lymphocytes, no significant telomere shortening was observed in cholangiocytes and hepatocytes of individuals without liver disease during aging. The authors only observed an age-related telomere shortening in Kupffer cells and stellate cells [70]. On the other hand, a consistent telomere shortening was reported during the development of biliary tract carcinoma, starting early in carcinogenesis in the inflamed biliary tract, metaplasia, dysplasia and carcinoma [71]. In contrast, the normal and the inflamed epithelium of the biliary tract showed a uniform telomere length [71].
Within cholangiocarcinoma, a frequent intratumoral heterogeneity of telomere length is reported [71]. As indicated above, telomere shortening in hepatocytes triggers cellular senescence in the context of intact DDR checkpoints. Similarly, an investigation of telomere shortening and senescence in the pathogenesis of primary biliary cirrhosis (PBC) showed telomere shortening. Moreover, DNA damage accumulation was detectable in biliary epithelial cells in the damaged small bile ducts and bile ductules in PBC in comparison to normal-looking bile ducts and bile ductules in PBC, chronic viral hepatitis and normal livers [72]. Of note, the accumulation of DNA damage foci correlated with increased expression of p16 INK4a and p21 WAF1/Cip1 , which characterize biliary cellular senescence [72].

Loss of Function Mutations in Telomerase Components
Germline loss-of-function mutations in the telomerase components were found in a variety of human diseases, including dyskeratosis congenita, aplastic anemia, familial idiopathic fibrosis and acute myeloid leukemia [73][74][75][76][77][78][79][80]. These mutations provoked an impaired tissue regeneration due to telomere dysfunction and stem/progenitor cell exhaustion. Similar mutations were also reported in a subset of liver cancer samples [34,35]. The authors analyzed TERT and TERC mutations in buccal mucosa tissue and peripheral blood of patients with liver cirrhosis and compared them with healthy non-cirrhotic controls. An increased number of telomerase mutations were found in the group with liver cirrhosis. The study by Calado et al. [34] reported nine patients with a mutation in the TERT gene and one patient with a mutation in the TERC gene among 134 patients with liver cirrhosis. Similarly, Hartmann et al. [35] reported mutations in the TERT and TERC genes in 16 out of 521 patients. The Calado study reported a significantly higher allele frequency for the gene variants in the TERT and TERC genes in patients with cirrhosis (allele frequency 0.037) compared to controls (0.008; p = 0.0011). A similar result was shown by the Hartmann study, which stated an increased incidence of telomerase mutations detected in cirrhosis patients (allele frequency 0.017) compared to non-cirrhotic controls (0.003, p = 0.0007). The mutations in telomerase components led to decreased telomerase activity in comparison to wildtype telomerase enzyme activity. Consequently, patients with these mutations showed shorter telomeres in peripheral white blood cells [34,35]. Rare TERT mutations were also reported in patients with nonalcoholic fatty liver disease (NAFLD). Here, an enrichment of TERT mutations could be found in NAFLD-associated HCC [81]. Functional evaluation of these mutations exposed reduced protein synthesis from some of the mutations compared to the TERT wild-type protein. It is speculated that these TERT mutations could also impair the DNA-binding function of TERT. In summary, these results indicate that TERT mutations result in impaired telomerase activity, accelerated telomere shortening and impaired regeneration in chronic liver disease. These findings are supported by the above-mentioned studies indicating low/absent telomerase activity in resting liver and telomerase activation in the regenerating liver [30,31]. Taken together, these studies show that telomerase activity acts as a protective mechanism in chronic diseases to prevent telomere shortening during accelerated cell proliferation, whereas TERT mutations result in telomere shortening and may promote hepatocarcinogenesis by dysfunctional telomeres.

Telomerase Reactivation during Hepatocarcinogenesis
It has been proposed for a long time that up to 90% of human tumors can reactivate telomerase [20,82]. Telomerase reactivation is associated with the alteration of transcriptional regulators of the TERT promoter in cancer, TERT promoter mutations or rearrangements and DNA copy number amplifications [30,[83][84][85][86][87][88]. Reactivation of the telomerase enzyme has been shown in more than 80% of HCCs, which suggests reactivation of telomerase as a rate-limiting process for liver cancer formation. The reactivation of telomerase correlates with the upregulation of both essential components, TERT and TERC. It has been reported that the re-expression of TERT and re-activation of telomerase occurs at early premalignant stages in regenerative nodules and cirrhotic livers [31,38,39]. Telomerase activation was also reported in iCCA (85%) [40]. Thus, it is important to note at this point that the majority of HCCs and iCCAs are telomerase-positive, highlighting the necessity of telomerase activity for telomere functionality and tumor progression in the two most common liver cancer entities.

TERT Promoter Mutations
TERT promoter mutations which result in increased TERT expression were first identified in melanoma and were subsequently reported in other cancers like bladder cancer, glioma, thyroid cancer and HCC [84,[89][90][91]. In the following, we present data that report on TERT promoter mutations in HCC, iCCA and tumors with a mixed-differentiation HCC/iCCA (Tables 1-3). These data were achieved by PubMed literature search, using the keywords "TERT promoter mutations", "liver carcinoma" and "biliary tract cancer". Additionally, the list includes studies that analyzed TERT promoter mutations in different tumor entities. Due to limited sample sizes in the reported studies, we excluded studies/data on rare entities, such as fibrolamellar carcinoma (FLC) and cholangiolocellular carcinoma (coCC), as well as studies/data based on liquid biopsies. In HCCs, TERT promoter mutations were identified with an overall prevalence of 20-82% as the most frequent somatic genetic alterations (Table 1) .   However, it should be noted that in 80% of cases which lacked TERT promoter, mutations showed an enhanced TERT expression. Moreover, telomerase activity was detectable, indicating the existence of alternative mechanisms of telomerase reactivation during liver carcinogenesis [118]. For example, it was shown that the insertion of HBV in the TERT promoter region can activate TERT gene expression and telomerase promoting hepatocarcinogenesis in HBV-related HCC [119,120]. Unlike HBV, HCV is not able to integrate into the host genome, but HCV can also induce chromosomal instability by direct effects of its proteins [121]. Interestingly, mutations in the TERT promoter were identified to be significantly more common in HCV-related HCC tumors compared with tumors without HCV infection [93]. It is also notable that the TERT promoter mutation frequency is higher in HCV-related HCC (64%) compared to HBV-related HCC tumors (37%) [98].  The occurrence of TERT promoter mutations in liver carcinogenesis is identified in premalignant lesions, and the prevalence of mutations gradually increased with the degree of dysplasia, indicating that TERT promoter mutations are highly associated with the stepwise transformation from premalignant dysplastic nodules to malignant HCC. TERT promoter mutations were identified in 6% of low-grade dysplastic nodules, 19% of high-grade dysplastic nodules and 61% of early HCCs. Interestingly, TERT promoter mutations were not detected in cirrhotic liver [94,117]. TERT promoter mutations are so far the earliest recurrent genetic events in cirrhotic preneoplastic lesions and belong to the most frequent alterations in hepatocellular carcinoma [94,122]. These results support the idea that telomerase reactivation is required for the malignant transformation of liver cells from cirrhosis to cancer.
In HCC, TERT promoter mutations at several positions were described. The two most frequent mutations occur at positions −124 (G > A) and −146 (G > A) upstream of the ATG translation start site (Figure 1) [98,106]. Interestingly, the TERT promoter mutation at the −124 bp hotspot appears more often compared with the TERT promoter mutation at the −146 bp hotspot [93,94,[96][97][98]100,101,106,108,111,112,114] was shown that the TERT promoter mutations generate a de novo consensus binding site for the E-twenty-six (ETS) transcription factor family [83,84], leading to an increase in TERT protein amounts, telomerase activity and telomere length [123]. Recently a new TERT promoter mutation was found in 7.5% of the analyzed HCCs at the position −297 (C > T) upstream of the ATG translation start site, which creates an AP2 consensus sequence [104]. consensus binding site for the E-twenty-six (ETS) transcription factor family [83,84], leading to an increase in TERT protein amounts, telomerase activity and telomere length [123]. Recently a new TERT promoter mutation was found in 7.5% of the analyzed HCCs at the position −297 (C > T) upstream of the ATG translation start site, which creates an AP2 consensus sequence [104]. Interestingly, the TERT promoter mutation frequency in HCC differs geographically. By comparing different studies, Pezzuto et al. [106] reported that TERT promoter mutations seem to be most common in Europe (56.6%). In Africa, TERT promoter mutations were identified in 53.3% and in Asia in 42.5% of HCCs. The overall TERT promoter mutation rate in America is around 40% [106].
In contrast to frequent mutations in the TERT promoter region in HCC, TERT promoter mutations are less frequently analyzed in cholangiocarcinoma [92,95,96,106,116,[124][125][126]. These studies identified TERT promoter mutations in 31-47% of the HCCs. Nakamura et al. analyzed the largest dataset of 145 iCCAs (named as ICCs in the study). A TERT promoter mutation was detected only in one iCCA sample. Furthermore, the same study also analyzed a set of 86 extrahepatic cholangiocarcinomas, with none of them exhibiting a TERT promoter mutation [125].
While no TERT promoter mutation was detected in iCCA samples in the majority of studies [92,95,96,116], in some studies, performed on lower sample numbers, a low frequency of TERT promoter mutations were observed in iCCA [106,124,126] (Table 2). Interestingly, the TERT promoter mutation frequency in HCC differs geographically. By comparing different studies, Pezzuto et al. [106] reported that TERT promoter mutations seem to be most common in Europe (56.6%). In Africa, TERT promoter mutations were identified in 53.3% and in Asia in 42.5% of HCCs. The overall TERT promoter mutation rate in America is around 40% [106].
In contrast to frequent mutations in the TERT promoter region in HCC, TERT promoter mutations are less frequently analyzed in cholangiocarcinoma [92,95,96,106,116,[124][125][126]. These studies identified TERT promoter mutations in 31-47% of the HCCs. Nakamura et al. analyzed the largest dataset of 145 iCCAs (named as ICCs in the study). A TERT promoter mutation was detected only in one iCCA sample. Furthermore, the same study also analyzed a set of 86 extrahepatic cholangiocarcinomas, with none of them exhibiting a TERT promoter mutation [125].
While no TERT promoter mutation was detected in iCCA samples in the majority of studies [92,95,96,116], in some studies, performed on lower sample numbers, a low frequency of TERT promoter mutations were observed in iCCA [106,124,126] (Table 2).
The high percentage of TERT promoter mutations in HCC/iCCA may reflect the presence of the HCC part consisting of the TERT promoter mutation or indicate that both tumor types may arise from the same cell type of origin (Table 3).
In addition, a clear difference between hepatitis-positive LCBs and hepatitis-negative LCBs was observed (20% vs. 6%) [124]. Taken together, the fact that TERT promoter mutations are detectable only in a small subset of cholangiocarcinoma indicates that the mechanism of telomerase activation is evidently different from telomerase activation in many HCCs (Figure 2). In addition, a clear difference between hepatitis-positive LCBs and hepatitis-negative LCBs was observed (20% vs. 6%) [124]. Taken together, the fact that TERT promoter mutations are detectable only in a small subset of cholangiocarcinoma indicates that the mechanism of telomerase activation is evidently different from telomerase activation in many HCCs (Figure 2).

Amplification and Genomic Rearrangements of TERT
Structural and numerical aberrations in the organization of the genome can lead to TERT promoter activation and eventually to telomerase reactivation. Frequent amplifications of the TERT gene have been shown in human cancer cell lines and in human tumors [85,128]. The amplified TERT

Amplification and Genomic Rearrangements of TERT
Structural and numerical aberrations in the organization of the genome can lead to TERT promoter activation and eventually to telomerase reactivation. Frequent amplifications of the TERT gene have been shown in human cancer cell lines and in human tumors [85,128]. The amplified TERT gene was found in around 22% of hepatocellular carcinoma and shows a higher incidence in poorly differentiated hepatocellular carcinoma [129]. Nevertheless, the TERT mRNA level did not correlate with the number of TERT gene copies, as was also reported in colorectal cancer [129,130]. Focal amplifications of 5p15.33 (TERT) were observed with a reduced overall survival, independently of other clinicopathological parameters in patients with hepatocellular carcinoma [131]. However, it should be noted that TERT amplifications do not always lead to increased TERT mRNA levels, which are by themselves not assignable to increased TERT translation levels.
Structural rearrangements in the TERT gene have been shown in several cancers like neuroblastoma, renal cell carcinoma, sarcoma and prostate carcinoma, and also in liver carcinoma [82,86,132,133]. A specific event that occurs in hepatocellular carcinoma is integration of viral genomes. Integration of the Hepatitis B virus into the host genome has been described in hepatocellular carcinoma, frequently in the TERT promoter region [98,99,120,[134][135][136]. The high frequency of HBV integration within the TERT promoter region might be a reason for a low rate of TERT promoter mutations in HBV-related HCC, which is known to induce telomerase transcription [94,119]. This could represent an alternative mechanism or two different mechanisms of telomerase reactivation.
AAV2 (adeno-associated virus type 2) is the second virus reported to integrate into HCC cell lines and hepatocellular carcinoma of human patients [137]. The insertion of AAV2 can take place in the TERT promoter region and thereby lead to overexpression of the TERT gene [137]. The AAV2 genome achieves a liver-specific enhancer-promoter activity in the 3 UTR and binding sites of hepatic transcription factors, which might be linked to TERT overexpression and telomerase reactivation [138].

Altered Transcriptional Regulation of TERT Gene Expression
Studies addressing telomerase activation in liver cells mainly focused on TERT gene regulation in hepatocytes and HCC. The RB/E2F gene regulatory circuit regulates TERT promoter activity during liver regeneration and cancer [30]. The role of the RB/E2F circuit in TERT regulation is supported by a recent report showing that the Human Krüppel related 3 (HKR3) is capable of repressing TERT gene expression in HCC cell lines, resulting in the subsequent activation of CDKN2A (encoding the p16 tumor suppressor factor) and cell cycle inhibition as a tumor suppressor mechanism [139]. In another study, the RNA-binding fox-1 homolog 3 (RBFOX3) protein cooperates with AP2β to activate TERT gene expression in HCC cell lines, indicating a role of telomerase activation in promotion of HCC [140]. Concurrently, impaired E2F1 binding to the TERT promoter has been shown to correlate with increased patient mortality [141].
On the other hand, the bromodomain PHD finger transcription factor (BPTF) seems to regulate TERT gene expression via nucleosome remodeling in HCC cell lines [142]. BPTF is required for c-MYC transcriptional activity in carcinogenesis, linking TERT gene activation to c-MYC, which was the first transcription factor shown to regulate TERT expression by direct promoter binding [143][144][145][146]. Of note, c-MYC amplification correlates with liver cancer progression, mainly HCC, and to a lesser extent, iCCA [147,148]. In a recent study using a genome-wide shRNA screening strategy in HepG2 cell line, Chr15orf55 (also known as NUTM1) and Chr7orf43, two regulatory factors with currently unknown functions, were found to activate TERT gene expression through SP1 or YAP1, respectively [149]. The exact nature of these factors and their role in HCC and iCCA remain yet to be clarified.

Epigenetic Mechanisms in the Regulation of TERT Gene Expression
There are a number of studies indicating that TERT promoter methylation and histone acetylation may be involved in the regulation of TERT transcription and hence telomerase activity. The mechanism of MAD1-mediated repression involves modification of the chromatin structure by histone deacetylation [146]. In fact, treatment of telomerase-negative human cell lines with the histone deacetylase inhibitor TSA (trichostatin A) results in the reactivation of TERT gene expression and telomerase activity [150], whereas the overexpression of HDAC1 causes the repression of the TERT gene and telomerase activity [151]. To what extent histone acetylation/deacetylation is involved in TERT regulation during carcinogenesis remains pending.
The overexpression of DNMTs is thought to be responsible for suppressing the expression of tumor suppressor proteins by methylating the promoters of their genes, which leads to the early switching off of these factors. DNMT3B in particular is strongly overexpressed in breast tumors and correlates with the degree of tumorigenesis [152]. It can therefore also be assumed that the epigenetic regulation of TERT gene expression takes place indirectly, namely via the promoter methylation of genes for transcription factors which are responsible for TERT gene expression. A recent report showed the cooperation between TERT and the transcription factor SP2 to stimulate DNMT3B transcription, while TERT depletion inhibited DNMT3B expression [153]. Higher levels of TERT and DNMT3B expression predicted shorter survival in HCC patients based on the TCGA database [153].
On the other hand, the methylation status of the TERT promoter was studied in some detail. Initial reports regarding the methylation state and the activity of the TERT promoter were contradictorily unequivocal: while Devereux et al. [154] and Dessain et al. [155] did not observe an impact of promoter methylation of TERT expression, Guilleret et al. [156] found a correlation between the methylation state and the activity of the TERT promoter. Moreover, studies on TERT promoter methylation in HCC revealed contradictory results. Zhang et al. [157] observed that aberrant methylation of TERT promoter in HCC patients of the Han Chinese population showed a nearly 56-fold increase of TERT expression from the hypermethylated promoter. In contrast to that study, an analysis of 106 patient tissues (64 with HCC and 42 without liver disorders) and hepatocarcinoma cell lines revealed that the TERT promoter was methylated in normal liver but was hypomethylated in most of the hepatocellular carcinomas [158].
Recent studies point out the importance of a comprehensive mapping of the methylation landscape within the TERT promoter using more advanced technologies, arguing that the methylation analysis of a small number of CpG sites may not represent the methylation landscape of the whole TERT promoter [159]. Using a next-generation sequencing technique, the authors performed a comprehensive methylation analysis and revealed a TERT hypermethylated oncological region (THOR) defining a 433 bp genomic region within the TERT promoter as a cancer-associated epigenetic mechanism of TERT upregulation. This region encompasses 52 CpG sites and is located upstream of the TERT promoter mutations (TPMs) [159]. Five CpG sites within THOR accurately presented the average THOR methylation and were used for a bigger screen of 1352 tumor samples and 80 normal samples. Of the tumor samples 91.4% exceeded the median THOR methylation level detected in normal samples. Low THOR methylation levels were detected in thyroid cancers, which was linked to a lower malignant potential and a better prognosis [159]. Tumors from skin and bone showed low and heterogeneous methylation levels, which indicates the use of other mechanisms of telomere maintenance like TERT promoter mutations or the alternative lengthening of telomeres (ALT). In line with these results, Esopi et al. [160] described hypermethylation in the region upstream of the recurrent C228T and C250T promoter mutations in immortalized and cancer cell lines (including hepatocarcinoma cell lines). In contrast, non-malignant primary cells were rather hypomethylated [160]. Interestingly, the authors could show on an allele level that the hypermethylation of TERT promoter sequences in cancer cells is associated with TERT repression, while the remaining unmethylated allele marked with an open chromatin is largely responsible for the TERT expression in cancer cells [160]. In cancers with TERT promoter mutations, the expressed allele is mutated, while the WT allele is silenced [160,161]. In summary, current data support the assumption that TERT promoter activity can be regulated by epigenetic mechanisms in a tissue-and cancer-type-specific manner.

Telomere and Telomerase-Based Cancer Therapy
The fact that telomerase is repressed in somatic tissues but active in cancer cells qualifies it as an ideal target for cancer therapy. On the other hand, impairment of telomere length and/or structure could be a target for tumor therapy as telomere length in cancer cells is usually shorter than in the corresponding normal tissue cells. Moreover, due to their high turnover, functional telomeres have to be rebuilt more often in tumor cells than in normal cells. Thus, deterioration of the intact telomere structure by compounds may represent an attractive target to cure cancer.
Several telomerase inhibitors have been used in preclinical studies and clinical trials, involving telomerase inhibitor molecules, such as the antisense oligonucleotide Imetelstat (GRN163L) or the small-molecule inhibitor BIBR1532, as well as the G-quadruplex stabilizers (BRACO, RHPS4, Telomestatin). There are also interesting alternative strategies such as telomerase vaccination (peptide-based: GV1001; mRNA-based: GRNVAC1) to induce anti-tumor responses or tumor cell lysis through TERT promoter-driven oncolytic adenovirus (Telomelysin (OBP-301)) or TERT promoter-driven pro-apoptotic protein (e.g., TRAIL) to take advantage of high telomerase activity in cancer cells for therapy options. For further reading, we refer to a recent excellent article, which discusses this topic in greater detail [162]. Furthermore, we refer to a recent review article, which summarizes clinical trials on telomerase-based cancer therapeutics [163].

Conclusions
The biology of telomere maintenance plays a major role in the process of cirrhosis formation as well as in initiation and progression in liver carcinoma. A shortening of telomeres occurs due to progressive chronic liver diseases and highly correlates with an increased tumor risk. Telomere shortening occurs in HCC as well as in iCCA. In addition, telomerase reactivation is detectable at a high frequency in both tumor entities. The mechanisms of telomerase activation seem, at least in part, to be different between the two liver cancer entities. TERT promoter mutations were more frequently observed in HCCs compared to iCCAs. The low frequency of TERT promoter mutations reported in iCCA might be due to a not-detected HCCs part (mixed differentiation) of the analyzed tumors. This links to the phenomenon that some tumor entities lacking specific mutations in the TERT promoter region and where other mechanisms leading to telomerase activation exist. As described above, the TERT promoter region contains binding sites for many transcription factors, i.e., c-MYC, E2Fs, and others, which contribute to the tissue-specific regulation of TERT gene expression [22]. It would be interesting to explore whether TERT expression is regulated differentially in HCC and iCCA. The identification of the regulatory mechanisms contributing to telomerase reactivation in HCC and iCCA could shed light on the differential tumor initiation and progression pathways and may provide alternative and specific therapy options.
The observation that TERT promoter mutations occur early during liver carcinogenesis highlights the importance of telomerase activity for tumor cell survival. Two possible scenarios are conceivable how telomerase contributes to tumorigenesis in liver cancer. On the one side, telomerase reactivation before entering the crisis checkpoint may stabilize critically short telomeres, providing growth advantage for cells with oncogenic mutations (Figure 3). In melanoma, a two-step mechanism was described, which showed that mutations in the TERT promoter region contribute to tumorigenesis [164]. This study showed that TERT promoter mutations are not sufficient to counteract telomere shortening, but they contribute to tumorigenesis by promoting the immortalization and chromosomal instability in two phases: (1) extend the cellular life span by healing the shortest telomeres without prevention of bulk telomere shortening, and (2) the existence of critically short telomeres conducts to genome instability and thereby, telomerase is further up-regulated to proceed cell proliferation.
On the other side, early reactivation of telomerase may be related to its non-canonical functions. There is strong experimental evidence indicating that oncogenic mutations or mutations which result in genome instability induce cellular senescence by inducing telomere replication stress [41,165].
This mechanism functions to limit continuous proliferation of cells carrying detrimental mutations, preventing tumorigenesis.
However, telomerase activity can alleviate telomere replication stress by a yet-unidentified mechanism and promote tumorigenesis [22,88,166,167].
highlights the importance of telomerase activity for tumor cell survival. Two possible scenarios are conceivable how telomerase contributes to tumorigenesis in liver cancer. On the one side, telomerase reactivation before entering the crisis checkpoint may stabilize critically short telomeres, providing growth advantage for cells with oncogenic mutations (Figure 3). In melanoma, a two-step mechanism was described, which showed that mutations in the TERT promoter region contribute to tumorigenesis [164]. This study showed that TERT promoter mutations are not sufficient to counteract telomere shortening, but they contribute to tumorigenesis by promoting the immortalization and chromosomal instability in two phases: (1) extend the cellular life span by healing the shortest telomeres without prevention of bulk telomere shortening, and (2) the existence of critically short telomeres conducts to genome instability and thereby, telomerase is further upregulated to proceed cell proliferation.  Whether telomere maintenance or even elongation of telomeres during chronic liver diseases or an inactivation of the telomerase could be beneficial for the treatment of liver carcinoma remains to be elucidated in future studies.

Conflicts of Interest:
The authors declare that they have no conflict of interest.