|Home | About | Journals | Submit | Contact Us | Français|
The overall 5-year survival rate of approximately 60% for head and neck cancer patients has remained essentially unchanged over the past 30 years. MicroRNA-137 (miR-137) plays an essential role in cell cycle control at the G1/S phase checkpoint. However, aberrant miR-137 promoter methylation observed in squamous cell carcinoma of the head and neck (SCCHN) suggests a tumor-specific molecular defect that may contribute to disease progression.
The goal of this study is to assess, in formalin-fixed paraffin-embedded tumor tissue, the association between miR-137 promoter methylation and survival (both overall and disease-free) and with prognostic factors including stage, tumor size, nodal positivity, tumor grade and surgical tumor margin positivity.
Promoter methylation status of miR-137 was ascertained by methylation-specific PCR and detected in 11/67 SCCHN patients (16.4%), with no significant differences according to site (oral cavity, pharynx, larynx). Methylation of the miR-137 promoter was significantly associated with overall survival (Hazard Ratio = 3.68, 95% Confidence Interval: 1.01–13.38) but not with disease-free survival or any of the prognostic factors evaluated.
This study indicates that miR-137 is methylated in tumor tissue from pharyngeal and laryngeal squamous cancers, in addition to oral squamous cell carcinoma; and that miR-137 promoter methylation has potential utility as a prognostic marker for SCCHN.
Head and neck cancer describes a heterogeneous group of malignancies occurring in the upper aerodigestive tract, the majority (93%) of which are squamous cell carcinomas (SCCHN).1 The overall 5-year survival of approximately 60% among head and neck cancer patients has been virtually unchanged over the past 3 decades.2 Despite therapeutic advances in cancer treatment, there has been no significant improvement in laryngeal cancer survival and only a slight gain in survival has been observed for cancers of the oral cavity and pharynx.1 As treatments improve and become more targeted, there is a need to identify biomarkers that correlate with prognosis and/or prognostic factors in an attempt to understand mechanisms driving aggressive tumor behavior and to identify patients who are at greatest risk for mortality.
MicroRNAs are small, non-coding RNA molecules that negatively regulate target gene expression through degradation and translational inhibition, with potentially hundreds of target mRNAs.3 Alterations in the expression levels of microRNAs occur in most, if not all, types of cancer4–19, including SCCHN.4–6, 18 The potential prognostic value of microRNAs is suggested by their frequent involvement in proliferation, differentiation and apoptotic pathways. Altered expression of individual microRNAs has been associated with survival4, 6, 20–31, tumor stage27, 32, grade30, lymph node metastasis4, 24, 27 and vascular invasion4, 32 in multiple tumor types. An estimated 10% of microRNAs are controlled by DNA methylation33, with aberrant methylation being one source of microRNA dysregulation. Associations between microRNA promoter methylation and overall25, 34, 35 and disease-free34 cancer survival have been reported.
MicroRNA-137 (miR-137) is located on chromosome 1p21.3 and lies across a large CpG island.16 Promoter methylation of miR-137 has been described in several solid tumors,36–38 including a report36 that it is frequently methylated in tumor tissue from oral squamous cell carcinoma (OSCC). A significantly higher frequency of miR-137 promoter methylation in oral rinse samples from SCCHN patients than from controls was recently reported by our group39. Cyclin-dependent kinase 6 (Cdk6) has been identified as a target of miR-137, which interacts with Cyclin D1 to phosphorylate Rb allowing the cell to progress through the G1/S phase checkpoint.16, 36 Experimental studies have shown that transfection of miR-137 to deficient OSCC cell lines results in a reduction of Cdk6 levels and increased cell cycle arrest at the G1 phase.36 Loss of control of the G1/S-phase checkpoint could reduce the capacity of the cell to repair DNA damage prior to synthesis, potentially resulting in genomic instability. More recently, additional targets of miR-137 have been identified, including Mib1, Jard1b, VKORC1 and Ezh2.40–43 These targets also suggest that changes in miR-137 expression may impact a switch from proliferation to differentiation.41
Known prognostic factors for SCCHN include stage1, tumor size and lymph node metastasis,44, 45 tumor grade46 and surgical tumor margin positivity.47–49 The primary objective of this study is to evaluate the prognostic value of miR-137 promoter methylation in SCCHN by assessing its association with overall and disease-free survival, as well as known prognostic factors for the disease.
This study was conducted as a part of the University of Pittsburgh Head and Neck Cancer Specialized Program of Research Excellence (SPORE). The study population consisted of consecutively diagnosed adult patients (≥ 18 years) at the University of Pittsburgh Medical Center (Pittsburgh, PA) between September 2007 and April 2009 with primary squamous cell carcinoma of the oral cavity, pharynx or larynx and no prior history of cancer. Archival tumor tissue was available for 67/102 (65.7%) patients meeting these criteria. Patients for whom cancer tissue was available did not significantly differ in terms of age, gender, cancer site or stage in comparison to patients without available tissue. IRB approval was obtained under the University of Pittsburgh Head and Neck Cancer SPORE for sample collection and use of patient data. All subjects provided written informed consent for participation in this study.
Upon enrollment, patients completed an epidemiologic questionnaire providing detailed demographic and behavioral information. Clinical data were extracted at the time of diagnosis and recorded in the University of Pittsburgh Head and Neck Oncology Registry. Patients entered into the registry were prospectively followed for the ascertainment of vital status through February 8, 2010. Formalin-fixed paraffin-embedded (FFPE) tumor tissue was obtained and evaluated by a board-certified pathologist to verify that representative sections were used. Tissue samples used in this study were collected prior to administration of radiation or chemotherapy.
DNA was extracted from tumor tissue with the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA). Sodium bisulfite modification of the resultant DNA was performed using the EZ DNA Methylation Kit (Zymo Research, Orange, CA) per the manufacturer’s protocol. Methylation status of miR-137 was determined by methylation-specific PCR (MSP), as previously described.39 PCR products were analyzed by separation on high-resolution 4% agarose E-gels (Invitrogen, Carlsbad, CA) and visualized and digitally captured with an EDAS 290 high-performance ultraviolet transilluminator using 1D 3.6 software (Kodak, Rochester, NY). Each MSP reaction contained a fully methylated and unmethylated bisulfite converted human DNA (Qiagen, Valencia, CA) as a positive and negative control.
Total tissue microRNA was extracted from FFPE tissue scrolls using the RecoverAll Total Nucleic Acid Isolation kit (Ambion, Austin, TX) according to the manufacturer’s instructions for RNA isolation. Briefly, FFPE tissue scrolls were deparaffinized using xylene and ethanol followed by protease digestion, and finally nucleic acid isolation. DNA was removed from the RNA sample by a final DNase digestion step. The final volume of the RNA (Total RNA containing microRNA) was 60μl. RNA concentration was measured using the NanoDrop-1000 (Thermo Fisher Scientific, Waltham, MA) and the RNA integrity was determined using the Bioanalyzer 2100 (Agilent, Santa Clara, CA).
cDNA synthesis for individual microRNA validation was performed according to manufacturer's recommendation. Briefly, RNA was reverse transcribed using the MicroRNA reverse transcription kit (Applied Biosystems) in combination with the stem-loop primer for miR-137 and the endogenous control U6snRNA (Applied Biosystems). 5 μl of total RNA (2.0 ng/μl) was combined with 3 μl RT primer (5×), 1.5 μl RT buffer containing MgCl2 (10×), 1.0 μl MultiScribe Reverse Transcriptase (50 U/μl), 0.15 μl dNTPs (25 mM each), and 0.19 μl RNase inhibitor (0.25 U/μl) in a total reaction volume of 10 μl. Reactions were run on the Applied Biosystems Gene Amp PCR System 9700 in a 96-well plate using the following parameters: 16°C for 30 min, 42°C for 30 min, 85°C for 5 min, 4°C hold. The single tube microRNA expressions were detected using TaqMan chemistry according to the manufacturer's protocol (Applied Biosystems). 1.0 μl of cDNA was combined with 1.0 μl 20X TaqMan MicroRNA assay and 10.0 μl 2X TaqMan Universal PCR MasterMix II, No AmpErase UNG in a final volume of 20 μl. The reactions were run in duplicate in a 96-well plate on the 7900HT Real-Time PCR System according to manufacturer's recommendations (Applied Biosystems). Under our experimental conditions, we analyzed 4 endogenous control miRNAs (U6snRNA, RNU 43, RNU 48 and miR-26b) to determine the most appropriate microRNA for normalization. We were looking for the most consistent and highly abundant microRNA among these controls (not shown). Therefore, we normalized our data using U6 to generate expression values. Expression values are reported relative to a calibrator (Ambion’s First Choice Epidermal Carcinoma Human RNA).
Descriptive statistics for the study population were generated by miR-137 promoter methylation status for demographic, behavioral and clinical characteristics. Categorical variables were compared using Fisher’s exact test. The Skewness-Kurtosis test was used to evaluate normality of continuous variables.50 In the case of non-normality, the Mann-Whitney U test was used for comparison; otherwise a two-sample t-test was used.
Exact logistic regression51 was used to model each prognostic factor of interest, conditioned on select covariates to adjust for potential confounding. Prognostic factors for SCCHN included tumor stage at diagnosis, comparing advanced (stage III [T3 N0 M0 or T1-3 N1 M0] or IV [T4 or > N2 or M1]) to local (AJCC stage group I [T1 N0 M0] or II [T2 N0 M0]); tumor size (T classification), comparing T3/T4 to T1/T2; lymph node positivity (N classification), comparing positive (N1 or greater) to negative (N0); tumor grade, comparing poorly differentiated to well/moderately differentiated; and surgical tumor margins, restricted to patients receiving surgical intervention, comparing patients with positive to patients with negative margins. The primary exposure of interest was miR-137 promoter methylation (methylated/unmethylated). Gender, smoking status (ever/never), daily fruit and vegetable consumption (above/below median) and tumor site were considered as covariates for each model based on the univariate results and predetermined biological importance. Exact logistic regression results were compared to the results from asymptotic models using bootstrap variance estimators to assess comparability.
Univariate Kaplan-Meier estimates were generated for 2-year overall and disease-free survival by miR-137 promoter methylation status. Differences in survival by methylation status were assessed using the log-rank test. In the analysis of disease-free survival, patients who died prior to recurrence were considered to be censored at death.
Multivariable Cox proportional hazards models52 provided estimates of overall and disease-free survival, adjusted for age, stage at diagnosis and tumor site. The proportional hazards assumption was tested for each model using an approach based on the slope of scaled Schoenfeld’s residuals as a function of time.53 P-values ≤0.05 were considered to be statistically significant throughout.
Promoter methylation of miR-137 was detected in tumor tissue from 16.4% (11/67) of SCCHN patients. Thirty-two (32) of the 67 tumor tissue samples were available for further analysis of miR-137 expression, including eight (8) that presented with miR-137 promoter methylation (Figure 1). The eight samples with miR-137 promoter methylation (marked by the star in Figure 1) had little or no expression of miR-137. Conversely, four of the 32 tumor tissue samples evaluated had extremely elevated expression of miR-137, none of which were positive for miR-137 promoter methylation.
SCCHN patients with miR-137 promoter methylation-positive tumors ate fewer daily servings of fruits and vegetables (P = 0.005) and were more likely to be women (P = 0.13) relative to those with unmethylated tumors (Table 1). Although not statistically significant, 10 of the 11 patients with miR-137 promoter methylation, 90.9% used both alcohol and cigarettes compared to 58.9% of unmethylated patients.
When considered by tumor site, miR-137 promoter methylation was detected in 14.8% (4/27) of oral cancers, 22.7% (5/22) of pharyngeal cancers and 11.1% (2/18) of laryngeal cancers (Table 2). The proportions of patients with miR-137 promoter methylation did not differ significantly by tumor site (P = 0.71). Univariate comparisons of SCCHN prognostic factors by miR-137 methylation status showed no other significant differences between patients with miR-137 methylation-positive and methylation-negative tumors.
Neither the crude nor adjusted exact logistic regression models show any significant associations between miR-137 promoter methylation and stage at diagnosis, tumor size, nodal positivity, tumor grade or surgical tumor margin positivity for SCCHN, although confidence intervals are wide (Table 3). Similar results were obtained from the asymptotic logistic regression models using a bootstrap variance estimator (not shown).
The average overall follow-up time was 15.2 months (median = 14.8; 10th percentile = 7.8 months; 90th percentile = 22.9 months), with 14 deaths. The average follow-up for disease-free survival was 13.8 months (median = 13.1; 10th percentile = 4.3 months; 90th percentile = 22.6 months), with 9 recurrences.
In univariate analyses, SCCHN patients with miR-137 promoter methylation had a significantly lower overall survival rate (P = 0.046) compared to those with unmethylated tumors (Figure 2a). No difference was observed for disease-free survival between patients with and without miR-137 promoter methylation (P = 0.63; Figure 2b).
In the multivariable Cox proportional hazards analysis adjusting for age, stage at diagnosis and tumor site (Table 4), SCCHN patients with tumors positive for miR-137 promoter methylation had a significantly poorer overall survival (HR = 3.68, 95% CI: 1.01–13.38) compared to those with unmethylated tumors. Promoter methylation of miR-137 was not significantly associated with disease-free survival.
Dysregulation of microRNA expression has been correlated with outcome or prognostic factors for many different cancer types.4, 6, 20–30, 32 Aberrant promoter methylation is one mechanism through which microRNA expression can be suppressed. There have been recent reports in the literature of associations of promoter methylation of individual microRNAs and cancer prognosis,25, 34, 35 including correlation of promoter methylation of miR-124a, another microRNA that targets Cdk6, with poorer overall and disease-free survival in acute lymphocytic leukemia patients.34 Here, we report the detection of miR-137 promoter methylation in tumor tissue of 16.4% of SCCHN patients, occurring in oral, pharyngeal and laryngeal cancers; and describe an association of miR-137 promoter methylation with poorer overall survival among SCCHN patients. To our knowledge, this is the first study to evaluate the potential prognostic value of promoter methylation of a microRNA in SCCHN.
The observed association between miR-137 promoter methylation and overall survival among SCCHN patients may relate, in part, to the involvement of miR-137 in cell-cycle control through regulation of Cdk6. Loss of miR-137 expression results in reduced ability of the cell to arrest at the G1 phase, increasing proliferation,36,16 which may lead to accumulation of DNA damage thus enhancing genomic instability. Another potential contributor is the possible role of miR-137 in cellular differentiation or a switch from proliferation to differentiation.41 Expression levels of miR-137 are reported to be elevated in neuronal differentiation and decreased in poorly differentiated gliomas,16 although it is presently unknown if this generalizes to other histologies. Contrary to this point, we found no association of miR-137 promoter methylation with tumor grade, although low statistical power limits our ability to detect such an association. In spite of its purported involvement in cell cycle control, we also found no significant associations of miR-137 promoter methylation with other SCCHN prognostic factors, including stage, tumor size, nodal involvement or surgical tumor margin positivity.
Strengths of this study include high-quality prospective data collection; employment of methylation-specific PCR, a relatively inexpensive and sensitive method to detect DNA methylation for ascertainment of miR-137 methylation status; and the use of surgical tumor tissue taken prior to initiation of radiation and/or chemotherapy, which precludes potential bias relating to treatment effects. Another strength is our use of exact logistic regression modeling to obtain more accurate inferences for small sample sizes.51
This study also has several limitations. The relatively small sample size and proportion of miR-137 promoter methylation-positive samples limits our power to detect associations with prognostic factors. The present study has power ranging from 10%–18% to detect an association with an OR of 2.0. Another limitation is the relatively short follow-up time; future studies with longer follow-up time are needed to confirm our observed association between miR-137 promoter methylation and overall survival and further assess the association with disease-free survival. Another aspect that needs to be addressed in future studies is the correlation of promoter methylation with miR-137 expression with larger sample sizes. However, we did find that all the available samples that presented with miR-137 promoter methylation (8) had little or no expression of miR-137 whereas four of the 32 tumor tissue samples available for analysis had elevated expression of miR-137, yet none of these samples had miR-137 promoter methylation. Complicated modes of gene regulation such as this have been observed previously. For example, MAD2 is down-regulated in the absence of promoter methylation54 and RIZ1 regulation requires both promoter methylation and histone acetylation.55 Despite methylation-specific PCR being a tried-and-true sensitive method for detection of promoter methylation,56 it is not quantitative. MicroRNA-137 methylation-positive tumors, as detected by methylation-specific PCR, may contain a low percentage of methylated alleles, which may not be sufficient to produce a substantial reduction of miR-137 expression, and this could dilute the magnitude of an association between methylation and prognosis/outcome.
The results of this study suggest that miR-137 promoter methylation is a relatively common occurrence in SCCHN, occurs across all sites, and may have value as a prognostic biomarker for the disease. However, future research efforts should focus on quantitative methylation analysis; further evaluation of the prognostic value of miR-137 promoter methylation in expanded cohorts with longer follow-up; and assessment of the combined effect of loss of multiple tumor suppressors involved in regulation of the G1/S phase checkpoint. Continued efforts to identify such novel prognostic biomarkers or biomarker panels are crucial in reaching the greater goal of understanding the biology behind aggressive tumor behavior, and ultimately improving survival and reducing mortality from head and neck cancer.
We would like to thank Jennifer Hetrick for promptly providing us with patient follow-up data, facilitating the survival analysis in this manuscript. This work was supported by grants from the American Cancer Society [RSG-05-246-01-GMC] and the National Institutes of Health [CA148629] to RWS; the National Institutes of Health [CA132385] to ET and RWS; the National Institutes of Health [CA097190] to JRG; and the Clinical and Translational Science Institute Multidisciplinary Pre-doctoral Fellowship Program, awarded through the Clinical and Translational Science Institute and the Institute for Clinical Research Education at the University of Pittsburgh [5TL1RR024155-03] to SML.
Support: This work was supported by grants from the American Cancer Society [RSG-05-246-01-GMC] and the National Institutes of Health [CA148629] to RWS; the National Institutes of Health [CA132385] to ET and RWS; the National Institutes of Health [CA097190] to JRG; and the Clinical and Translational Science Institute Multidisciplinary Pre-doctoral Fellowship Program, awarded through the Clinical and Translational Science Institute and the Institute for Clinical Research Education at the University of Pittsburgh [5TL1RR024155-03] to SML.
Disclosures: The authors have no conflicts of interest or financial disclosures to declare.