Hypermethylation, risk factors, clinical characteristics, and survival in 235 patients with laryngeal and hypopharyngeal cancers
It has been established that promoter hypermethylation occurs in several genes during the pathogenesis of head and neck cancer. The authors investigated the role played by the hypermethylation of 4 cancer-related genes in the survival of patients who had laryngeal and hypopharyngeal cancer and in the occurrence of second primary tumors.
Archival paraffin-embedded tissue (PET) samples were available from patients who were enrolled in a multicentric European case-control study that was performed between 1979 and 1982 and was followed up to 2000. Genomic DNA extracted from 235 PET samples were analyzed for promoter methylation status of the p16, O6-methylguanine-DNA methyltransferase (MGMT), death-associated protein kinase (DAP-K), and E-cadherin genes by using a methylation-specific polymerase chain reaction assay.
Hypermethylation was present in 44% of samples for p16, in 27% of samples for MGMT, in 42% of samples for DAP-K, and in 43% of samples for E-cadherin. Hypermethylation of either individual genes or their combination was not associated with mortality from all causes, mortality from upper aerodigestive tract cancer, or the occurrence of second primary tumors.
The analysis of a large series of patients with laryngeal and hypopharyngeal cancer suggested that hypermethylation is a frequent event in laryngeal and hypopharyngeal cancer, but it is not a predictor of mortality or second primary cancer. Cancer 2007. © 2007 American Cancer Society.
Improvements in diagnosing and understanding the molecular mechanisms of head and neck cancer have taken place during the last decade. Despite these advancements, the 5-year survival rate for patients with head and neck cancer has remained unchanged at approximately 50% over the past 30 years.1 Previously, we described the role of clinical and lifestyle-related factors in the prognosis for patients with laryngeal and hypopharyngeal cancer.2, 3 Recent studies also have investigated the role played by genetic and epigenetic alterations in the progression to head and neck cancer4, 5 and, specifically, have analyzed the cytosine methylation of CpG islands in gene promoters. It has been suggested that aberrant promoter methylation is associated with loss of gene function that can provide a selective advantage for neoplastic cells, similar to what occurs in mutations.6 The factors that determine hypermethylation are considered to be gene-specific rather than chromosome-specific, and unique hypermethylation profiles in specific genes for given tumor types have been suggested.7 Indeed, higher frequencies of hypermethylation in the p16, O6-methylguanine-DNA methyltransferase (MGMT), death-associated protein kinase (DAP-K), and E-cadherin genes, but not in the early-onset breast cancer 1 (BRCA1), glutathione S-transferase π (GSTP1), p14, or human mutL homolog 1 (hMLH1) genes, have been reported in head and neck tumors.7 The inactivation of cancer-related genes through promoter hypermethylation also may affect tumor recurrence, the occurrence of second primary tumors (SPTs), and patient survival. Such effects have been suggested for head and neck cancer5, 8 and for other types of cancer.9–11
Therefore, we hypothesized that hypermethylation in the 4 genes p16, MGMT, DAP-K, and E-cadherin—likely resulting in their silencing—can affect the prognosis for patients with laryngeal and hypopharyngeal cancer. We tested this hypothesis in a series of 235 patients with laryngeal and hypopharyngeal cancer who had up to 21 years of follow-up available, including information on clinical and lifestyle-related factors. Our objectives were to measure cause-specific and overall survival and to measure the occurrence of SPTs according to hypermethylation status of the selected genes. Because epigenetic events potentially are reversible and, hence, can be targeted for therapeutic intervention, such information may have clinical implications in the management of laryngeal and hypopharyngeal cancer.
MATERIALS AND METHODS
A multicentric, population-based, case-control study on laryngeal and hypopharyngeal cancer was performed between 1972 and 1982 in 6 South European regions that were covered by cancer registries and allowed the collection of information on lifestyle, demographic, and clinical factors.12 The patients were followed until 2000 in 5 of 6 centers for the occurrence of SPTs and vital status. Follow-up was conducted by linkage with population, mortality, and cancer registry files. Results concerning patient survival and the occurrence of SPTs according to demographic and lifestyle factors recently have been published.3, 13 Surgical paraffin-embedded tissue (PET) sections were collected for the current study from 235 patients who had complete follow-up information available. The patients were comprised of 155 men and 12 women from the city of Turin, Italy, and 67 men and 1 woman from the province of Navarra, Spain.
DNA Extraction and Bisulfite Modification
Genomic DNA extracted in June 200514 from PET samples was used to study the methylation profile of 4 cancer-related genes: p16, MGMT, DAP-K, and E-cadherin. Genomic DNA was extracted from 3 to 5 (10 μm thick) sequential paraffin sections that were cut with disposable microtome blades to prevent tissue carry over. Tissue specimens were dewaxed in xylene and washed in 100% ethanol. All samples underwent the conventional overnight digestion at 55°C in lysis buffer (Tris 100 mM, ethylenediamine tetraacetic acid [EDTA] 5 mM, NaCl 200 mM, and 0.2% sodium dodecyl sulfate) and 1:10 proteinase K (20 mg/mL; Promega, Madison Wis). Then, DNA was purified by a double step with phenol, followed by precipitation with isopropanol, and washing in 70% ethanol, according to the standard protocols.15 Final elution was performed in 50 μL of 1× buffer containing Tris, HCl, and EDTA (TE buffer).
Bisulfite modification was performed on DNA samples based on the principle that sodium bisulfite treatment converts unmethylated cytosines to uracil, leaving methylated cytosines unaltered. Therefore, methylated and unmethylated DNA sequences become distinguishable after bisulfite conversion by sequence-specific primers. Genomic DNA samples underwent bisulfite modification using the CpGenome DNA modification kit (Intergen Company, New York, NY) according to the manufacturer's recommendations. Two microliters of a carrier DNA that were included in the kit (DNA Modification Reagent IV) were added to all samples.
Positive controls for methylated status (CpGenome universal methylated DNA; Intergen Company) and unmethylated status (normal human lymphocyte DNA) were included in each modification set.16, 17 The bisulfite-modified genomic DNA was resuspended in 35 μL of 1× TE buffer and either was used immediately for methylation-specific polymerase chain reaction (PCR) amplification or was stored at −80°C.
The modified DNA served as a template for PCR amplification using primers specific for either for the methylated or for the modified, unmethylated DNA. The specific primers sets and their relative annealing temperatures for the methylated and unmethylated forms of the MGMT, p16, DAP-K, and E-cadherin gene promoters were selected according to published sequences and are listed in Table 1.18, 19 For PCR amplification, 4 μL of bisulfite-modified DNA were added in a final volume of 25 μL PCR mix containing 1× PCR buffer (15 mM Tris [pH 8.0], 50 mM KCl, and 6.7 mM MgCl2), deoxyribonucleotide triphosphates (2 mM each), primers (0.4 μM each per reaction)m and 1.25 U of AmpliTaq Gold DNA polymerase (Applera, Foster City, Calif). All PCR amplifications were performed in a Gene Amp PCR System 9700 Thermal Cycler (Applera). Bisulfite-modified CpGenome universal methylated DNA (Intergen Company) was used as a positive control for methylated alleles, and bisulfite-modified DNA from normal human lymphocytes was used as a positive control for unmethylated alleles. Negative PCR controls without DNA were included in every PCR run. Ten microliters of each PCR amplification were loaded directly onto 2% agarose gels, stained with ethidium bromide, and visualized by ultraviolet transillumination. Previous studies have demonstrated that this method has high sensitivity: It detects 1 methylated nucleotide in 1000 unmethylated nucleotides.20
Table 1. Primer Sequences for Methylation-specific Polymerase Chain Reaction Analysis
|MGMT M||TTCGACGTTCGTAGGTTTTCGC||GCACTCTTCCGAAAACGAAACG||80||59||Rosas et al., 200118|
|MGMT U||TTTGTGTTTTGATGTTTGTAGGTTTTTGT||AACTCCACACTCTTCCAAAAACAAAACA||94||59||Rosas et al., 200118|
|p16 M||TTATTAGAGGGTGGGGCGGATCGC||ACCCCGAACCGCGACCGTAA||150||60||Rosas et al., 200118|
|p16 U||TTATTAGAGGGTGGGGTGGATTGT||CAACCCCAAACCACAACCATAA||151||60||Rosas et al., 200118|
|DAP-K M||GGATAGTCGGATCGAGTTAACGTC||CCCTCCCAAACGCCGA||114||59||Rosas et al., 200118|
|DAP-K U||GGAGGATAGTTGGATTGAGTTAATGTT||CAAATCCCTCCCAAACACCAA||116||59||Rosas et al., 200118|
|E-cadherin M||TTAGGTTAGAGGGTTATCGCGT||TAACTAAAAATTCACCTACCGAC||116||57||Herman et al., 199619|
|E-cadherin U||TAATTTTAGGTTAGAGGGTTATTGT||CACAACCAATCAACAACACA||97||53||Herman et al., 199619|
For each patient, the starting date of follow-up was defined as the date either laryngeal or hypopharyngeal cancer was diagnosed. The end of follow-up was on December 31, 2000; on the date a patient departed from the study area; or on the date of death, whichever occurred first. Analyses were done separately for all causes of death (overall mortality), and only deaths from upper aerodigestive cancers (cause-specific mortality; International Classification of Diseases, 9th Revision [ICD9] codes 140–150 and 161) were considered as the event of interest. In the latter analysis, we treated deaths from other causes as censored events.
SPTs were identified according to previously described criteria.13 In brief, these criteria emphasized that an SPT must be clearly distinct from the first tumor in terms of location but not necessarily in terms of the time of appearance. Nonmelanoma skin carcinoma (ICD9 code 173) was not considered an SPT. The follow-up for occurrence of SPT ended on December 31, 2000; on the date a patient departed from the study area; on the date of death; or on the date an SPT was diagnosed, whichever occurred first.
We fitted an unconditional logistic regression model that treated the methylation status of gene as a dependent variable to study the association between hypermethylation in the 4 genes of interest and the clinical and demographical characteristics in our group of patients with laryngeal or hypopharyngeal cancer. The results from that unconditional logistic regression model are presented as odds ratios (ORs) and 95% confidence intervals (95% CIs). A Cox proportional-hazards model, which was adjusted for clinical and demographic covariates, was used to estimate the hazard ratios (HRs) and 95% CIs for overall and cause-specific mortality and for the occurrence of SPTs according to the hypermethylation status of the p16, MGMT, DAP-K, and E-cadherin genes. We studied the joint effect of hypermethylation of 2 genes on cause-specific and overall survival for all possible combinations of genes (total, 6 combinations) by introducing interaction terms into the Cox proportional-hazards model.
The mean patient age at enrollment was 60.4 years (±9.89 years). The mean duration of follow-up was 8.1 years, and the maximum follow-up duration was 21 years. Among the 235 patients who were studied, 200 patients (85%) died during follow-up, and the remaining 35 patients (15%) remained alive at the end of follow-up. Among the patients who died, 134 patients had cancer of the upper aerodigestive tract specified as the underlying cause of death on the death certificate, and 66 patients had another cause of death.
The analysis for promoter methylation was successful in 126 patients (54%) for p16, in 212 patients (90%) for MGMT, in 213 patients (91%) for DAP-K, and in 190 patients (81%) for E-cadherin. The proportion of patients with hypermethylation was 44% for p16, 27% for MGMT, 42% for DAP-K, and 43% for E-cadherin. There was no association between hypermethylation status of any of the 4 genes and demographic or lifestyle characteristics of the patients, although there was a suggestion that hypermethylation in p16 and MGMT was more frequent in light smokers than in heavy smokers (Table 2). There also was an inverse correlation between hypermethylation in MGMT and lymph node involvement (OR, 0.18; 95% CI, 0.04–0.83). No correlation was observed in hypermethylation between any combinations of the 4 genes.
Table 2. Odds Ratios for Promoter Hypermethylation in the p16, MGMT, DAP-K, and E-cadherin Genes According to Clinical and Demographic Characteristics of Patients With Laryngeal/Hypopharyngeal Cancer
| Unskilled workers||44||31.8||1||—||76||26.3||1||—||76||43.4||1||—||67||43.3||1||—|
| Skilled workers||79||48.1||1.93||0.86–4.35||130||27.7||1.08||0.55–2.13||131||42||0.96||0.54–1.73||119||42||0.93||0.50–1.73|
|Site of tumor|
| Larynx||91|| ||1||—||153|| ||1||—||154|| ||1||—||134|| ||1||—|
| Epilarynx||13|| ||0.42||0.11–1.60||19|| ||0.48||0.14–1.71||19|| ||0.89||0.33–2.40||18|| ||1.00||0.35–2.80|
| Hypopharynx||21|| ||0.58||0.19–1.76||39|| ||0.55||0.20–1.49||39|| ||0.73||0.34–1.57||37|| ||0.79||0.36–1.76|
|Average cigarette smoking, cigarettes/d|
| ≤15||17|| ||1||—||35|| ||1||—||34|| ||1||—||30|| ||1||—|
| >15||100|| ||0.32||0.09–1.07||167|| ||0.44||0.19–1.05||169|| ||0.86||0.40–1.88||151|| ||1.78||0.73–4.36|
|Average alcohol drinking, g/d|
| ≤40||30|| ||1||—||48|| ||1||—||48|| ||1||—||41|| ||1|| |
| >40||96|| ||1.17||0.46–2.98||164|| ||0.54||0.25–1.17||165|| ||0.82||0.40–1.68||149|| ||0.87||0.40–1.90|
|Lymph node stage|
No increase in the risk of cause-specific or overall mortality was observed according to the hypermethylation status of any of the genes studied (Table 3) or for any other combination of hypermethylated genes. Table 4 shows the HRs of developing an SPT according to hypermethylation of the 4 genes under study. In total, 29 SPTs were observed in the current study, including 9 SPTs of the lung and 7 SPTs of other sites of the upper aerodigestive tract. There was no evidence that hypermethylation of any gene increased the risk of developing an SPT at any site.
Table 3. Hazard Ratios of Cause-specific Mortality and All-cause Mortality According to Hypermethylation Status of the p16, MGMT, DAP-K, and E-cadherin Genes
Table 4. Hazard Ratios of Developing a Second Primary Tumor Among Patients With Laryngeal/Hypopharyngeal Cancer According to Hypermethylation Status of the p16, MGMT, DAP-K, and E-cadherin Genes
Promoter hypermethylation of cancer-related genes, as alternative pathways to mutation for gene silencing in neoplastic cells, has been described as a crucial event in the development of head and neck cancer, and reports suggest that it affects the prognosis for these patients.6 However, most reports have been limited by small sample sizes, short follow-up, and investigation of a single gene.5, 8
Strengths of the current study include the relatively large sample size, the long follow-up, and the availability of information on clinical and lifestyle factors collected during the enrollment of these patients in the frame of a multicentric, case-control study. Limitations of our study include that we could not visualize the gene methylation profile in 9% to 46% of samples that were tested for the different genes. Despite the efficient amplification of β-globin, the integrity of DNA obtained from archival PET samples also may have been affected by manipulations (ie, bisulfite modification), giving rise to lower methylation-specific PCR amplification efficiency, especially for larger amplicons. Moreover, we do not have information regarding treatment to study whether the role of hypermethylation on survival differs according to the type of treatment received. Thus, because we observed an inverse relation between lymph node involvement and hypermethylation in MGMT, it is plausible that the inactivation of MGMT through hypermethylation reduces the ability of laryngeal and hypopharyngeal tumors to invade the lymph nodes after alkylating chemotherapy. Indeed, it has been reported that, although MGMT hypermethylation alone, without treatment with an alkylating agent, is not a good prognostic factor per se, it improved survival in patients with malignant astrocytoma, glioma, and diffuse large B-cell lymphoma when they were treated with alkylating chemotherapeutic agents.8, 22
To study the prognostic relevance of hypermethylation on laryngeal and hypopharyngeal cancer, we selected 4 genes (p16, MGMT, DAP-K, and E-cadherin) that reportedly are hypermethylated frequently in head and neck tumors5, 7–9, 23, 24 and are involved in cell proliferation control, DNA repair, and invasiveness. For p16 (27–36%),5, 24–26MGMT (29–32%),5, 8, 21E-cadherin (35–42%),5, 23, 24 and DAP-K (18–75%),5, 18, 21, 26, 27 we observed a frequency of hypermethylation similar to that reported by other authors.
No direct association between promoter hypermethylation in any of the 4 genes under study and age, sex, occupation, alcohol drinking, or tumor stage was observed, in agreement with recently reported data5, 8; whereas an inverse association between the number of cigarettes smoked per day and hypermethylation in p16 and MGMT was suggested. Similarly, other studies have observed higher hypermethylation frequencies in MGMT among never smokers28 and in p16 among former smokers,25 possibly suggesting that smoking causes loss of functioning in Tp53 and p16 by mechanisms other than promoter hypermethylation.
In our cohort of patients with laryngeal and hypopharyngeal cancer, we did not observe any correlation between mortality and hypermethylation of any of the 4 investigated genes, either alone or in combination. Similarly, some previous investigators did not observe any effect of either p16 hypermethylation or MGMT hypermethylation on the prognosis for patients with head and neck cancer29, 30 and esophageal cancer11; whereas the results from other studies5, 8 suggested a role of MGMT hypermethylation in the survival of patients with head and neck cancer. An association between aberrant gene methylation and the development of SPTs has been suggested.31, 32 We did not observe any association between the development of SPTs after laryngeal/hypopharyngeal cancer and the tested genes. This is in line with recent reports29, 33 in which an independent origin of SPTs from primary tumors or further acquisition of mutations leading to the development of SPTs were assumed.
In conclusion, the results of the current study confirmed that hypermethylation of cancer-related genes is a frequent event in laryngeal and hypopharyngeal cancer and that it can be measured in archival PET samples. However, our data from a large series of patients with long follow-up did not suggest any association between hypermethylation of the 4 genes that were investigated and survival in patients with cancer of the larynx and hypopharynx.
For their very kind collaboration, we thank Drs. A. Andrion, G. Bussolati, L. Gubetta, G. Mazzucco, R. Navone, G. Palestro, M. Papotti, S. Taraglio, and B. Torchio from the Pathology Departments in Turin, Italy; Dr. M. E. Ardanaz Aicua from Institute of Public Health of Navarra in Pamplona, Spain; and Drs. F. Garcia-Bragado, F. J. Pardo, and A. Puras from the Pathology Departments in Pamplona, Spain. We also thank Dr. Lorenzo Richiardi and Gilles Ferro for their useful suggestions.