HPV and methylation
The realization that viral infections, by insertion of viral genes into host genomes, can trigger host defense mechanisms such methylation machinery activation has aroused interest in the study of epigenetic events occurring in both virus and host genomes [
99]. Human genomes harbor DNA sequences resembling retroviral long terminal repeats and the transposable elements, and indeed there are indications that under some situations inappropiate "activation" of these normally silenced sequences could play a role in the carcinogenic process [
100]. In addition, it is also established that some viruses can find ways to adapt different tactics to regulate expression of their genes through modulation of DNA methylation; thus, a virus may silence activation of its genes in a manner that favors establishment of persistent infection and evades the host immune defense [
101]. In addition to this, viral oncoproteins can possess the ability to modulate directly or indirectly the methylation machinery in order to silence cellular genes that could interfere with its tumor promoting actions. A very illustrative example of this is how the Epstein-Barr virus oncogene product, latent membrane protein 1, induces downregulation of
E-cadherin gene expression via activation of DNA methyltransferases [
102].
The role of HPV genome DNA hypermetylation has of late been the subject of study. One of the first indications of the importance of DNA methylation and viral gene expression came from studies of cell transfection with HPV-16
in-vitro methylated genomes, demonstrating that under these circumstances DNA is transcriptionally repressed [
103]. In SiHa and CasKy cell lines that harbor HPV-16 and have a couple of and multiple viral genome copies, respectively, a conserved profile of CpG hyper and hypomethylation was found by using scanning with the restriction enzyme McfBC. Hypermethylation was found in genomic segments overlying late genes, while the long control region and the E6 and E7 oncogenes were unmethylated in SiHa cells. Interestingly, evaluation of smears of normal, precursor, and invasive lesion smears of 81 patients showed that as lesion severity increases, there is progressive hypomethylation on these LCR and E6 gene regions; thus, hypermethylation was found in 52% of smears from asymptomatic women, in 21.7% of preinvasive lesions, and only in 6.1% of invasive-case smears. These findings lead the authors to postulate that neoplastic transformation can be suppressed by gene hypermethylation, whereas hypometylation accompanies or causes cancer progression [
104]. These findings however, were not totally coincident in another study that studied L1 and LCR regions by bisulfite modification in 115 clinical samples. First, high heterogeneity on methylation status was noted among patients and even in different samples of the same patient. As expected, methylation frequency was ca. 30% in the L1 region and lower in other positions, particularly at a CpG site located in the linker between two nucleosomes positioned over HPV-16 enhancer and promoter. However, methylation at most sites was consistently higher in carcinomas as compared with dysplasia, possibly related to the tandem repetition and chromosomal integration that occurs in invasive lesions [
105]. In another study performed in two HPV-18 cervical cancer cell lines, HeLa and C4-1, and clinical samples, there was also clonal heterogeneity in the methylation status of the different regions analyzed. When it came to clinical sample analysis, there was complete or partial HPV-enhancer methylation in three of six tumors and complete demethylation in eight smears from asymptomatic patients. Likewise, promoter methylation was found in three of six cancers and in four of six smears [
106]. The latter two studies suggest that methylation status of viral oncogenes in lesions perhaps is perhaps solely the result of their transcriptional activity level and not a causal event for neoplastic progression.
Further data on the influence of DNA methylation in the HPV life cycle comes from another study that focused on methylation of
E2, the early gene that contributes to multiple biological processes including viral transcription and viral DNA replication. It has been shown that the capacity of E2 protein to bind E2BS
in vitro is inhibited by methylation of these cytosines [
107]. Kim et al. performed a methylation analysis by bisulfite modification of
E2 binding site within LCR in DNA isolated from an immortalized epithelial cell line isolated from an HPV 16-infected patient and demonstrated that this region is selectively hypomethylated in the highly differentiated cell populations, whereas poorly differentiated basal-like cells were heavily methylated particularly in E2 binding sites. These observations may indicate that the methylation state of the viral genome, and particular that of E2BSs, may vary during the viral life cycle, providing a novel means for modulating E2 function as infection progresses [
108]. It will be of major interest to analyze human papillomavirus oncogene expression in cervical tumors before and after treatment of patients with DNA methylation inhibitors.
It is now established that failure of cells to undergo apoptosis is crucial for cancer development and progression, but most importantly this phenomenon participates in intrinsic or acquired resistance of cancer cells to chemotherapy and radiation. Identification of points in the apoptotic pathway at which dysregulation occurs would potentially open up new therapeutic opportunities in situations where conventional cancer treatments fail. One of the first indications of the role of methylation for inactivation of key apoptotic genes came from the study showing that
Apaf-1 was silenced in melanoma instead of being lost or mutated [
109].
Studies analyzing apoptosis-related genes that can be inactivated by methylation in cervical cancer are limited. One study has shown that decoy receptors DcR1 and DcR2 can be the target for abnormal methylation that leads to their silencing [
110]. These molecules are members of the tumor necrosis factor receptor superfamily, which includes TNFR1,
Fas, and the decoy receptors for TRAIL. Upon engagement by their respective ligands, TNFR1 and FAS recruit adaptor molecules and activate a cascade of caspases. Death-inducing decoy receptors
DR4 and
DR5 and
DcR1 and
DcR2 are structurally related; nonetheless,
DcR1 completely lacks the intra-cellular death domain and
DcR2 contains a truncated, nonfunctional death domain and appears unable to induce apoptosis. Hence,
DcR1 or
DcR2 have been postulated to function as oncogenes because of their postulated anti-apoptotic effects [
111,
112]. In cervical carcinoma, a study has found that all 50 cases analyzed had methylation of either DcR1 and/or DcR2 [
113], suggesting that cervical cancer cells, by downregulating decoy receptor expression, obtain a growth advantage.
Telomerase activation is a critical element in cellular immortalization and cancer. The end of the chromosome, the telomere, plays a critical role in chromosome structure and function. A certain length of the telomere is important for cell division, and the telomere may serve as a "mitotic clock" for cell proliferation. Normal human somatic cells express low or undetectable telomerase activity, whereas in immortal eukaryotic cells as well as in cancer cells telomerase activity increase is apparently necessary to ensure proliferation. Telomerase is a ribonucleoprotein comprising an RNA template, the telomerase-associated protein, and the catalytic subunit (
hTERT) [
114,
115]. Telomerase activity has been demonstrated in various types of gynecologic cancers [
116]. Data on
hTERT expression in cervical cancer has revealed that 0–33% of normal cervices exhibited
hTERT mRNA expression, whereas 80–100% of cervical cancers showed
hTERT expression [
117‐
120] The fact that the
hTERT gene promoter has a CpG island and high overall GC content suggests a possible role for methylation in regulation of
hTERT gene expression; however, the relationship between gene promoter and expression is unclear for this gene. Despite it is expected that hypermethylation decreases gene expression, a study has found a correlation between reduced expression and catalytic subunit activity with demethylation [
121]. This may explain what was found with regard to better prognosis of patients with cervical cancer whose tumors lack
hTERT methylation [
122].
The p53 pathway responds to stresses that can disrupt the fidelity of DNA replication and cell division, resulting in activation of the p53 protein as a transcription factor that initiates either growth arrest or apoptosis. This apoptosis pathway is disrupted in the majority of human cancers by downregulation or loss of p14
ARF, upregulation of MDM2, or mutation of p53 [
123,
124]. However, this pathway, by virtue of its multiple positive and negative feedback loops, can be the target of aberrant methylation in some of their components.
p73 is a member of the
p53 family that encodes two different proteins expressed under the control of two independent promoters and that have opposite activities: the transcriptionally active full-length
TAp73 and the NH
2-terminally truncated dominant-negative
Np73 [
125].
TAp73 has been reported as involved in cellular response to DNA damage induced by radiation and chemotherapeutic agents and when it is overexpressed in cells, it activates transcription of
p53-responsive genes such as
p21,
Bax,
Mdm2, and
GADD45 and inhibits cell growth in a
p53-like manner by inducing apoptosis [
126,
127]. It has been reported that
p73 transcription can be regulated by the promoter and exon 1, which is rich in CpG dinucleotides [
128], and its transcriptional silencing through methylation is a common event in some leukemias, lymphomas, and brain tumors, as well as in ovarian cell lines but not in breast, renal, and colon cancers. A recent study found that epigenetic modification of
p73 via CpG-island hypermethylation represents a critical alternative mechanism for inactivation of this gene in cervical cancer and high incidence of
p73 hypermethylation (38.8%) in cervical cancer but not in controls; in addition, its methylation was correlated with loss of its p73 expression. Importantly, radioresistant cancer samples had significantly higher methylation rate than radiosensitive cancer samples, and
in vitro demethylation successfully restored p73 expression in cervical cancer cell lines previously found to have methylated
p73 and lack of p73 mRNA and protein expression [
129].
It is well-established that cancer cells evolve in part by overriding normal cell-cycle regulation. Normal cell cycle progression relies on the cell's ability to translate extracellular signals such as those produced by growth factor receptor stimulation and extracellular matrices to efficiently replicate DNA and divide. Proper cell- cycle regulation is essential for cells and requires a number of players, among them cyclin-dependent kinases and their binding partners along with natural inhibitory molecules such as p16, Rb, and p15 that play an essential role. Within this class, the p16 gene has been one of the most studied in cervical cancer. Aberrant methylation of the
p16 gene occurs early within tumor cell populations in both
in situ and invasive tumors at frequencies that vary from 10 up to 100% [
130‐
136]. As the cell cycle is primarily deregulated by the HPV, the molecular contribution of p16 inactivation is unclear; however, the fact that this not only is a very early event in cervical carcinogenesis but is more frequently methylated in advanced tumors [
132] suggests that its reactivation could have therapeutic value. Despite the fact that Rb and p15 are known to be inactivated by methylation in other tumors, no reports exist on cervical tumors.
WNT pathway
The Wnt signaling pathway, named for its most upstream ligands, the Wnts, is involved in various differentiation events during embryonic development and leads to tumor formation when aberrantly activated. Within this pathway, there are a number of participating molecules, and the pathway is regulated by a multiprotein complex consisting of, among others, members of β-catenin, the key component, adenomatous polyposis coli APC, Axin, and GSK-3β. [
137]. In the absence of Wnt stimulation, β-catenin accumulates in cytosol to then be translocated to nucleus, leading to transcription of target genes. This pathway is also involved in calcium-dependent cell adhesion by virtue of the interaction between β-catenin and cadherin [
138]. There are mutations in APC, another key regulator of the pathway that promotes β-catenin proteolysis and reduces its transcriptional activity. PTEN is a lipid and protein phosphatase that is a negative regulator of phosphatidylinositol 3 (PI-3) kinase-dependent signaling and influences the WNT pathway by hindering activation of integrin-linked kinase (ILK), which inhibits GSK-3 β and thereby causes accumulation of β-catenin [
139]. The WNT signaling pathway is the most frequently altered pathwayin the majority of cancers; for instance, it has been demonstrated that nearly all colorectal cancers have at least one activating mutation in this pathway [
140]. As such, individual components of the pathway can be targeted by epigenetic inactivation. A recent study analyzing 310 colorectal carcinomas for eight members of the signaling cascade, including APC, β-catenin, AXIN2, TCF4, WISP3, E-cadherin, and PTEN. Hypermethylation on E-cadherin and APC were found at frequencies between 36 and 42% [
141].
Studies on cervical cancer have uncovered that hypermethylation of these genes is not uncommon. A study in 62 cases of squamous cell carcinomas showed that while PTEN mutations were absent, promoter methylation was found in 58% of cases. Interestingly, patients with persistent disease or patients who died of the disease had a significantly higher percentage of PTEN methylation than those without evidence of recurrence. Multivariate Cox regression models confirmed PTEN was an important significant predictor for both total and disease-free survival after controlling age, pathologic grade, and clinical stage [
142]. Inactivation of the cadherin-mediated cell adhesion system caused by aberrant methylation is a common finding in human cancers. Methylation frequency of
E-cadherin in cervical cancer varies from between 28 and 80.5% [
143‐
145] and appears to have prognostic significance, cases with no promoter methylation having a better outcome in univariate and multivariate analyses [
146]. Mutations are the main mechanism of inactivation for
APC, particularly for colon and other tumors from the gastrointestinal tract. However,
APC promoter hypermethylation occurs in other cancers. Frequencies of methylation in 208 primary human tumors of multiple origins were as follows: stomach (13 of 38, 34%); pancreas (6 of 18, 33%); liver (6 of 18, 33%), and esophagus (4 of 27, 15%; it was less common in tumors of bladder (2 of 19, 10%), kidney (1 of 12, 8%), or breast (1 of 19, 5%), or was not observed at all in brain (0 of 10), lung (0 of 17), head and neck (0 of 10), or ovary (0 of 20) [
147]. In endometrial cancer, hypermethylation occurs at an increased frequency, particularly in MSI+ endometrial tumors [
148] as well as in cervical cancer, with rates varying from 11 to 94% [
133,
135,
136].
DNA repair
Alkyating agents induce O6-alkylguanines that can lead to mutations and to cell death unless repaired. The major pathway of repair involves transfer of the alkyl group from DNA to a cysteine acceptor site in the protein O6-alkylguanine-DNA alkyltransferase. Alkyltransferase brings about this transfer without The need for cofactors and DNA is restored completely by the action of a single protein, but the cysteine acceptor site is not regenerated and the number of O6-alkylguanines that can be repaired is equal to the number of active alkyltransferase molecules. A significant fraction of human tumor cell lines do not express the alkyltransferase gene; thus, they are much more sensitive to mutagenesis and killing by alkylating agents [
149]. The
MGMT gene product removes mutagenic and cytotoxic adducts from O(6)-guanine in DNA, the preferred point of attack of many carcinogens (i.e., methylnitrosourea) and alkylating chemotherapeutic agents (i.e., BCNU, temozolamide, etc.). As a consequence, its lack of expression produces opposite effects for cancer development and progression: First, tumors acquire a mutator phenotype characterized by generation of transition point mutations in key genes such as
p53 and
K-ras, but at the same time lack of enzymatic activity renders tumors more sensitive to the killing effects of alkylating drugs [
150]. While these observations bear clinical and practical implications as predictive or prognostic markers for response in CNS tumors [
151], its silencing by hypermethylation can be associated with higher stages, worse survival, or mutations in secondary genes that adversely affect the prognosis of patients with tumors such as gastric, colorectal, head, and neck carcinomas, [
152‐
154] and even in low-grade astrocytomas [
155]. There is scarce information concerning the role of
MGMT gene in cervical cancer; a number of studies have analyzed the frequency of MGMT promoter hypermethylation, which varies from 5–81% [
133,
134,
136,
144]. Interestingly, the five cases with
MGMT or
BRCA1 methylation did not respond to chemoradiation [
133].
FA-BRCA pathway
Fanconi anemia (FA) is an autosomal recessive chromosomal instability syndrome characterized by hypersensitivity to DNA cross-linking agents and predisposition to cancer, especially leukemia [
156]. FA patients are also prone to various solid malignancies, including squamous cell carcinoma. FA is a genetically heterogeneous disease with genes for seven FA complementation (FANC) groups identified [
157].
FANC genes are essential in DNA repair pathways in normal cellular response to cisplatin and other DNA cross-linking agents. FANC proteins interact with
BRCA genes in a pathway that involves a number of other genes [
158]. Recently, it has been shown that promoter hypermethylation of
FANCF gene disrupts the FA-BRCA pathway, resulting in cisplatin resistance in ovarian cancer [
159].
FANCF promoter hypermethylation has also been identified in squamous cell carcinomas of lung and oral cavity [
160]. In cervical cancer, a study has shown methylation of BRCA1 in 6.1% [
133] of invasive tumors, whereas
FANCF hypermethylation rate was 30% [
161]. Interestingly, hypermetylation of these genes was mutually exlusive in the analyzed cases [
161], suggesting the important role of disruption of this pathway for cancer. This abnormality seems to be a late event in cervical carcinogenesis, as no hypermetylation was observed in any case of preinvasive disease [
161].
Mismatch repair
Cells with dysfunction of mismatch repair genes
hMLH1 and
hMSH2, as well as
hMSH3,
hMSH6, and
hPMS2, show mutation rates up to 1,000-fold greater than those observed in normal cells [
162]. The mutator phenotype, which can be measured by microsatellite instability analysis, has been detected in tumors of patients with hereditary nonpolyposis colorectal sporadic and other types of cancers [
163]. Mutations and loss of expression due to gene promoter hypermetylation are the main mechanisms of inactivation of members of this gene family [
164]. Hypermethylation and loss of hMLH1 protein expression has been associated with chemotherapy resistance in ovarian and other tumors [
165]. The relevancy of this phenomenon has been recently demonstrated by acquisition of hypermethylation of the gene in relapsed ovarian cancer after being treated with chemotherapy, which predicts poor overall survival [
166]. Existing data on cervical cancer with regard to hMHL1 expression status and methylation is limited. While some studies have found protein expression loss in invasive lesions [
167], others have found the opposite [
168], while presence of microsatellite instability appears to correlate with a worse prognosis [
169] but not with response to cisplatin in a neoadjuvant setting [
170]. Regarding gene promoter methylation, its presence is rare in cervical cancers [
134].