Genetic variants in selected pre-microRNA genes and the risk of squamous cell carcinoma of the head and neck

Authors

  • Zhensheng Liu MD, PhD,

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Guojun Li MD, PhD,

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    2. Department of Head and Neck Surgery, The University ofTexas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Sheng Wei MD, PhD,

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Jiangong Niu PhD,

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Adel. K. El-Naggar MD,

    1. Department of Pathology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Erich M. Sturgis MD, PhD,

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    2. Department of Head and Neck Surgery, The University ofTexas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Qingyi Wei MD, PhD

    Corresponding author
    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    2. Program in Human and Molecular Genetics, The University of Texas Graduate School of Biomedical Sciences, Houston, Texas
    • Department of Epidemiology, Unit 1365, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030
    Search for more papers by this author
    • Fax: (713) 563-0999


Abstract

BACKGROUND:

Single nucleotide polymorphisms (SNPs) in microRNAs (miRNAs) may alter the processing, transcription, and expression of miRNAs and, thus, may contribute to cancer development. The authors hypothesized that common polymorphisms in pre-miRNAs are associated individually and (more likely) collectively with the risk of squamous cell carcinoma of the head and neck (SCCHN).

METHODS:

The authors genotyped 4 common polymorphisms in pre-miRNAs: Homo sapiens miRNA 146a (hsa-mir-146a) (reference SNP 2910164 [rs2910164]; guanine to cytosine [G→C]), hsa-mir-149 (rs2292832; guanine to thymine [G→T]), hsa-mir-196a2 (rs11614913; C→T), and hsa-mir-499 (rs3746444; adenine to guanine [A→G]) in 1109 patients with SCCHN (cases) and in 1130 cancer-free patients (controls) in a non-Hispanic white population that was frequency-matched by age and sex. Univariate and multivariate logistic regression models were used to calculate crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs).

RESULTS:

Of the 4 SNPs that were studied, the hsa-mir-499 AG and GG genotypes were associated with a reduced risk of SCCHN (OR, 0.83; 95% CI, 0.69-0.99). When the 4 SNPs were combined according to putative risk genotype, the number of observed risk genotypes was associated with an increased risk of SCCHN in a dose-response manner with ORs of 1.0, 1.20, and 1.40 for individuals who had 0 or 1 risk genotypes, 2 or 3 risk genotypes, and 4 risk genotypes, respectively (Ptrend = .037). Specifically, the risk was 1.23-fold (95% CI, 0.98-fold to 1.56-fold) for individuals with 2 to 4 risk genotypes and 1.40-fold (95% CI, 1.02-fold to 1.92-fold) for individuals who had 4 risk genotypes compared with individuals who had 0 or 1 risk genotypes. This risk was more pronounced in men and in patients with oropharyngeal cancer.

CONCLUSIONS:

The combined risk genotypes of 4 common SNPs in pre-miRNAs were associated significantly with a moderately increased risk of SCCHN. Larger studies are needed to validate the current findings. Cancer 2010. © 2010 American Cancer Society.

Squamous cell carcinoma of head and neck (SCCHN), including cancers of the oral cavity, pharynx, and larynx, are the sixth most common cancers worldwide.1 In the United States, approximately 48,010 new cases were diagnosed and resulted in 11,260 deaths in 2009.2 Although tobacco and alcohol are the known risk factors for SCCHN, only a fraction of smokers and drinkers develop SCCHN, suggesting the existence of genetic susceptibility to this disease. Indeed, numerous studies have demonstrated that genetic variations or single nucleotide polymorphisms (SNPs) have been associated with the risk of developing SCCHN.3-5

MicroRNAs (miRNAs) are small, single-stranded, nonprotein-coding RNAs of about 22 nucleotides. To date, hundreds of miRNA molecules have been identified in the human genome that play key roles in a broad range of physiologic and pathologic processes.6, 7 Although their biologic functions largely remain unclear, recent studies have demonstrated that miRNAs may function as tumor suppressors and/or oncogenes.8-12 It was demonstrated that aberrant expression of miRNAs was related to the etiology, diagnosis, and prognosis of many cancers, including SCCHN.13-21 For example, the let-7 miRNA (mir) family member (mir-let-7), like mir-16, mir-18, mir-21, mir-146, and miRNA-211, was highly expressed in SCCHN cell lines and tumor tissues,17, 22 and high miR-211 expression may be associated with disease progression and poor outcomes in patients with oral cancer,15 whereas mir-342, mir-346, and mir-373 had low expression in SCCHN cell lines.16, 17

Although the role of miRNA genetic variants in cancer susceptibility largely remains unknown, the importance of miRNA SNPs has been implicated in many cancers. It is well known that common SNPs in miRNAs and SNPs within their targets (miRNA-binding SNPs) may affect miRNA target expression and functions and, thus, may contribute to cancer risk.23, 24 Studies have demonstrated that polymorphisms of miRNAs are associated with the risk of lung cancer,25, 26 bladder cancer,27 thyroid cancer,28 renal cell carcinoma,29 colon cancer,30 and breast cancer.18 For example, a Homo sapiens miRNA 196a2 polymorphism (hsa-mir-196a2) was associated with survival in patients with nonsmall cell lung cancer,25 whereas hsa-mir-196a2 (cytosine to thymine [C→T]) and hsa-mir-499 (adenine to guanine [A→G]) variant genotypes were associated with an increased risk of breast cancer.31 These associations appear to be biologically plausible, because it has been demonstrated that a pre-mir-146a polymorphism may affect the miRNA expression.28

To our knowledge, few studies have investigated miRNA expression in head and neck cancer tissues. Previous studies demonstrated that hsa-146a was overexpressed in SCCHN cell lines and tumor tissues and that hsa-mir-149 was down-regulated in squamous cell carcinoma of the tongue.15, 16, 32 In addition, a recent study indicated that the hsa-mir-146a G→C, hsa-mir-149 G→T, hsa-mir-196a2 C→T and hsa-mir-499 A→G polymorphisms were associated with a risk of lung cancer. Because tobacco smoke is a major risk factor for both lung cancer and SCCHN, and given the important role of miRNAs in the development of cancers, including SCCHN, we hypothesized that SNPs in these miRNAs also may contribute to a susceptibility to SCCHN. To test this hypothesis, we genotyped and analyzed associations of the 4 previously described common polymorphisms in pre-miRNAs25 with the risk of SCCHN in a hospital-based, case-control study of 1109 patients with SCCHN (cases) and a control group of 1130 cancer-free individuals in a non-Hispanic white population.

MATERIALS AND METHODS

Study Participants

The participants were recruited from an ongoing SCCHN study as described previously.33 Briefly, the group of cases was comprised of patients who had newly diagnosed, untreated SCCHN histologically confirmed at The University of Texas M. D. Anderson Cancer Center between October 1999 and October 2007. Patients with second SCCHN primary tumors, primary tumors of the nasopharynx or sinonasal tract, primary tumors outside the upper aerodigestive tract, cervical metastases of unknown origin, or any histopathologic diagnosis other than SCCHN were excluded. Of the patients who were eligible for the case group, the response rate was approximately 93%, and only a few minority patients were recruited. Consequently, the current analysis included 1109 non-Hispanic white patients with primary tumors of the oral cavity (n = 326; 29.4%), oropharynx (n = 566; 51%), or larynx/hypopharynx (n = 217; 19.6%).

The individuals who were eligible for the control group had a response rate of 85%, and 1130 cancer-free controls were recruited from hospital visitors who accompanied patients to the clinics but who were not seeking medical care and were genetically unrelated to patients in the enrolled case group or to each other. First, we surveyed potential control participants at the clinics by using a short questionnaire to determine their willingness to participate in research studies and to obtain demographic information for frequency matching to the cases by age (±5 years) and sex. After obtaining written, informed consent, we interviewed each eligible individual to collect additional information about risk factors, such as tobacco smoking and alcohol use, and to obtain a 1-time, 30 mL blood sample for biomarker tests. The research protocol was approved by The University of Texas M. D. Anderson Cancer Center Institutional Review Board.

Genotyping

From each blood sample, a leukocyte cell pellet that was obtained from the buffy coat by centrifugation of 1 mL of the whole blood was used for DNA extraction with the Qiagen DNA Blood Mini kit (Qiagen, Valencia, Calif) according to the manufacturer's instructions. DNA purity and concentration were determined by spectrophotometer measurement of absorbance at 260 nm and 280 nm, respectively. For genotyping, we chose the 4 miRNA SNPs that have been reported as important,25 because each of these miRNAs targeted several important genes (eg, hsa-mir-146-3p targeting macrophage-stimulating 1 receptor [MST1R], menage a trois homolog 1 [MNAT1], mucin 1 [MUC1], interleukin 6 [IL6], and autophagy-related 7 homolog [ATG7]; hsa-mir-146-5p, which targets IL1 receptor-associated kinase 1 [IRAK1], general transcription factor IIH polypeptide 4 52 kDa [GTF2H4], centrin EF-hand protein 2 [CETN2], cytochrome P450 family 11 subfamily A polypeptide 1 [CYP11A1], and tumor necrosis factor receptor-associated factor 6 [TRAF6]; hsa-mir-196a-3p, which targets lymphocyte-specific protein 1 [LSP1], kinesin family member 20A [KIF20A], proteasome 26S subunit/non-ATPase 10 [PSMD10], BCL2-associated athanogene [BAG], and arachidonate 15-lipoxygenase [ALOX15]; hsa-mir-196a-5p, which targets homeobox B8 [HOXB8], v-myc myelocytomatosis viral oncogene avian homolog [MYC], 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 1 [PFKFB1], TOX high-mobility group box family member 3 [TOX3], and protein expressed in nonmetastatic cells 4 [NME4]; hsa-mir-499-3p, which targets G protein-coupled receptor 1 [GPR1], fat tumor suppressor homolog 1 [FAT], nibrin [NBN], IL1 receptor-like 1 [IL1RL1], and B-cell 2-like 14 apoptosis facilitator [BCL2L14]; and hsa-mir-499-5p, which targets GPR1, leukocyte cell-derived chemotaxin 1 [LECT1], potassium intermediate/small conductance calcium-activated channel subfamily N member 3 [KCNN3], protein PC4 and splicing factor-arginine/serine-rich 1 [SFRS1], interacting protein 1 [PSIP1], and retinol binding protein 2 [RBP2]). The polymerase chain reaction (PCR)-restriction fragment length polymorphism method was used to amplify the fragments that contained polymorphisms of hsa-mir-146a (G→C; reference SNP 2910164 [rs2910164]), hsa-mir-196a2 (C→T; rs11614913), or hsa-mir-499 (A→G; rs3746444), as described previously. For the hsa-mir-149 polymorphism (G→T; rs2292832), the primers were newly designed with a forward primer (5′-GTTTCTGGGAGAATTGAGG-3′) and a reverse primer (5′-GGAATCGTTTGAATCTGGAG-3′). The PCR profile included an initial melting step ay 94°C for 5 minutes, then 35 cycles at 94°C for 60 seconds, at 53°C for 45 seconds, and at 72°C for 1 minute, and a final extension step at 72°C for 10 minutes. These primers generated a PCR product of 250 base pairs (bp) (the TT genotype) that was digested by PvuII (New England Biolabs, Ipswitch, Mass) into fragments of 122 bp and 40 bp for CC and into fragments of 250 bp, 122 bp, and 40 bp for CT. The genotyping assays for 10% of the samples were repeated, and the results were 100% concordant.

Statistical Analysis

Differences in select demographic variables, smoking, and alcohol consumption between SCCHN cancer cases and controls were evaluated by using the chi-square test. Associations between the risk of SCCHN and miRNA SNP genotypes were estimated by computing odds ratios (ORs) and their 95% confidence intervals (CIs) from both univariate and multivariate logistic regression analyses. Those individuals who had smoked >100 cigarettes in their lifetime were defined as ever smokers, as defined traditionally in epidemiological studies; those who had quit smoking for >1 year previously were considered former smokers; and the remaining individuals were considered current smokers. Individuals who drank alcoholic beverages at least once weekly for more than 1 year in previous years were defined as ever drinkers; of these, those who had quit drinking for more than 1 year previously were defined as former drinkers, and the others were defined as current drinkers. The ORs and their 95% CIs for the miRNA genotypes were calculated by logistic regression analysis with adjustment for age (in years), sex, smoking status, and alcohol use. In logistic regression analysis, the miRNA genotypes were recorded as a dummy variable. Considering potential joint effects of the 4 miRNA polymorphisms on SCCHN risk, we evaluated associations between the risk of SCCHN and the combined genotypes of these polymorphisms. Stratified analysis was used to estimate the risk for subgroups by age, sex, smoking status, drinking status, and tumor site. All statistical analyses were performed with the SAS software (version 9.1.3; SAS Institute, Inc., Cary, NC).

RESULTS

The distributions of selected characteristics of the cases and controls are presented in Table 1. There were no significant differences in the distributions of age and sex between cases and controls (P = .466 and P = .727, respectively), and the mean age (±standard deviation) was similar for cases (57.2 ± 11.1 years) and controls (56.8 ± 11.0 years; P = .898). There were more ever-smokers and ever-drinkers among cases than among controls (72.3% vs 51.1%, respectively, and 72.7% vs 56.3%, respectively; P < .001 for both). However, all of these variables were adjusted further for any residual confounding effect in later multivariate logistic regression analyses.

Table 1. Frequency Distribution of Selected Variables in Patients With Squamous Cell Carcinoma of the Head and Neck (Cases) and Cancer-Free Controls
CharacteristicCases, n=1109Controls, n=1130Pa
No.%No.%
  • a

    Determined using a 2-sided chi-square test.

  • b

    Included both patients with laryngeal cancer (n=174) and patients with hypopharyngeal cancer (n=43).

Age, y     
 ≤5030127.131828.1.466
 51-5729026.227023.9 
 >5751846.754248 
Sex     
 Women27224.527023.9.727
 Men83775.586076.1 
Smoking status     
 Never30727.755348.9<.001
 Former38034.341236.5 
 Current4223816514.6 
Alcohol use     
 Never30327.349443.7<.001
 Former24221.818216.1 
 Current56450.945440.2 
Tumor site     
 Oral cavity32629.4   
 Oropharynx56651   
 Larynx/hypopharynxb21719.6   

The genotype and allele frequencies of the hsa-mir-146a (G→C, rs2910164), hsa-mir-149 (G→T, rs2292832), hsa-mir-196a2 (C→T, rs11614913), and hsa-mir-499 (A→G, rs3746444) SNPs and their associations with the risk of SCCHN are summarized in Table 2. All genotype distributions of the 4 SNPs in the control group were in agreement with Hardy-Weinberg equilibrium (P = .449 for hsa-mir-146a, P = .271 for hsa-mir-149, P = .737 for hsa-mir-196a2, and P = .441 for hsa-mir-499), and there was no overall difference in genotype distributions between cases and controls. However, compared with the hsa-mir-499 AA genotype, the variant AG and GG + AG genotypes were associated with a statistically significant decreased risk of SCCHN (AA genotype: adjusted OR, 0.80; 95% CI, 0.66-0.97; P = .023; AG and GG + AG variant: adjusted OR, 0.83; 95% CI, 0.69-0.99; P = .040), and this risk was not observed for other 3 SNPs (Table 2).

Table 2. Frequency Distribution of Pre-MicroRNA Genotypes and Their Associations With the Risk of Squamous Cell Carcinoma of the Head and Neck
GenotypeNo. (%)PbAdjusted OR (95% CI)
CasesControlsa
  • OR indicates odds ratio; CI, confidence interval; hsa-mir-146a, Homo sapiens microRNA 146a; G, guanine; C, cytosine; hsa-mir-149, Homo sapiens microRNA 149; T, thymine; hsa-mir-196a2, Homo sapiens microRNA 196a2; hsa-mir-499, Homo sapiens microRNA 499; A, adenine.

  • a

    The observed genotype frequency among individuals in the control group was in agreement with Hardy-Weinberg equilibrium (p2+2pq+q2=1: chi-square=0.486 [P = .4486] for hsa-mir-146a G→C; chi-square=1.211 [P = .271] for hsa-mir-149 G→T; chi-square=0.113 [P = .737] for hsa-mir-196a2 C→T; and chi-square=0.594 [P = .441] for hsa-mir-499 A→G).

  • b

    P values were calculated from 2-sided chi-square tests for either genotype distribution or allele frequency.

  • cAdjusted for age, sex, smoking status, and alcohol use in a logistic regression model.

All patients1109 (100)1130 (100)  
 hsa-mir-146a: G→C    
 GG630 (56.8)655 (58).8341.00
 CG411 (37.1)405 (35.8) 1.01 (0.84-1.22)
 CC68 (6.1)70 (6.2) 1.00 (0.69-1.45)
 CC/CG479 (43.2)475 (42) 1.01 (0.85-1.21)
 C allele frequency0.2470.241.670 
hsa-mir-149: G→T    
 GG580 (52.3)586 (51.8).7791.00
 GT441 (39.8)445 (39.4) 0.97 (0.81-1.16)
 TT88 (7.9)99 (8.8) 0.87 (0.63-1.21)
 TT/GT529 (47.7)544 (48) 0.95 (0.80-1.13)
 T allele frequency0.2780.285.637 
hsa-mir-196a2: C→T    
 CC350 (31.6)383 (33.9).4041.00
 CT565 (50.9)545 (48.2) 1.17 (0.96-1.42)
 TT194 (17.5)202 (17.9) 1.03 (0.79-1.33)
 TT/CT759 (68.4)747 (66.1) 1.13 (0.94-1.36)
 T allele frequency0.4300.420.509 
hsa-mir-499: A→G    
 AA745 (67.2)710 (62.8).0651.00
 AG309 (27.9)366 (32.4) 0.80 (0.66-0.97)
 GG55 (4.9)54 (4.8) 1.01 (0.67-1.51)
 GG/AG364 (32.8)420 (37.2) 0.83 (0.69-0.99)
 G allele frequency0.1890.210.081 

Because each of the other 3 nonsignificant SNPs appeared to have some minor effect on the risk of SCCHN, next, we performed a combined analysis of all 4 SNPs to evaluate potentially modifying effects of the combined genotypes on the risk of SCCHN (Table 3). We categorized the combined putative risk (OR, >1.0) of each genotype from all SNPs into a new variable according to the number of risk genotypes (for the protective genotype [OR, <1.0], we reversed the reference group). We observed that individuals who had 4 risk genotypes had the highest risk (adjusted OR, 1.45; 95% CI, 0.82-2.55), although the trend test was not statistically significant (Ptrend = .070). In trichotomized groups of 0 or 1 risk genotypes, 2 or 3 risk genotypes, and 4 risk genotypes, only individuals who had 4 risk genotypes had a significantly increased risk of SCCHN (adjusted OR, 1.40; 95% CI, 1.02-1.92) compared with individuals who had 0 or 1 risk genotypes, but the trend in risk for 3 groups (0 or 1 risk genotypes, 2 or 3 risk genotypes, and 4 risk genotypes) was statistically significant (Ptrend = .037). To facilitate further stratified analysis, we dichotomized all participants into groups that had 0 or 1 risk genotypes, and 2 to 4 risk genotypes; and individuals who had 2 to 4 risk genotypes also had a significantly increased risk of SCCHN (adjusted OR, 1.23; 95% CI, 0.98-1.56) compared with individuals who had 0 or 1 risk genotypes.

Table 3. Association of Combined Homo sapiens Pre-MicroRNA 146a (hsa-146a), hsa-196a2, hsa-149, and hsa-499 Genotypes With the Risk of Squamous Cell Carcinoma of the Head and Neck
Risk GroupCases, n=1109Controls, n=1130OR (95% CI)a
No.%No.%CrudeAdjusted
  • OR indicates odds ratio; CI, confidence interval.

  • a

    Adjusted for age, sex, smoking status, and alcohol use in a logistic regression model.

Combined genotype      
 0302.7343ReferentReferent
 114112.717115.10.94 (0.55-1.60)1.04 (0.59-1.84)
 240236.339434.91.16 (0.69-1.93)1.26 (0.74-2.14)
 337533.838534.11.10 (0.66-1.84)1.23 (0.72-2.11)
 416114.514612.91.25 (0.73-2.14)1.45 (0.82-2.55)
 P for trend    .152.070
Trichotomized      
 0-117115.420518.1ReferentReferent
 2-377770.1779691.20 (0.95-1.50)1.20 (0.95-1.52)
 416114.514612.91.32 (0.98-1.79)1.40 (1.02-1.92)
 P for trend    .065.037
Dichotomized      
 0-117115.420518.1ReferentReferent
 2-493884.692581.91.22 (0.97-1.52)1.23 (0.98-1.56)

Next, we stratified data on dichotomized risk genotypes by age group, sex, smoking status, alcohol status, and cancer histology. We estimated the ORs associated with 2 to 4 risk genotypes (as the risk group) compared with 0 or 1 risk genotypes (as the reference group) with adjustment for the aforementioned variables (Table 4). When age was categorized into 2 groups based on the median age of the controls, younger individuals (aged ≤57 years) with 2 to 4 risk genotypes exhibited a borderline significantly higher risk (adjusted OR, 1.34; 95% CI, 1.00-1.87) compared with the reference group. The ORs associated with having 2 to 4 risk genotypes were more evident for men (adjusted OR, 1.29; 95% CI, 1.00-1.69) than for women and were more evident for never-smokers (adjusted OR, 1.41; 95% CI, 1.00-2.08) than for ever-smokers. Finally, the histology-specific risk was more pronounced for patients with oropharyngeal cancer (OR, 1.32; 95% CI, 1.00-1.76) than for patients with cancers of the oral cavity and larynx/hypopharynx.

Table 4. Stratification Analysis of the Risk of Squamous Cell Carcinoma of the Head and Neck Associated With Homo sapiens Pre-MircoRNA 146a (hsa-146a), hsa-196a, hsa-149, and hsa499 Combined Genotypes
CharacteristicNo. of Cases (%)No. of Controls (%)Adjusted OR (95% CI)a
Referent Group: 0-1 GenotypeRisk Group: 2-4 GenotypesReferent Group: 0-1 GenotypeRisk Group: 2-4 GenotypesReferent Group: 0-1 GenotypeRisk Group: 2-4 Genotypes
  • OR indicates odds ratio; CI, confidence interval.

  • a

    Adjusted for age, sex, smoking status, and alcohol use in a logistic regression model.

Age, y      
 ≤778 (13.2)513 (86.8)100 (17)488 (83)Referent1.34 (1.00-1.87)
 >5793 (18)425 (82)105 (19.4)437 (80.6)Referent1.17 (0.84-1.62)
Sex      
 Men128 (15.3)709 (84.7)160 (18.6)700 (81.4)Referent1.29 (1.00-1.69)
 Women43 (15.8)229 (84.2)45 (16.7)225 (83.3)Referent1.06 (0.65-1.73)
Smoking status      
 Never43 (14)264 (86)103 (18.6)450 (81.4)Referent1.41 (1.00-2.08)
 Ever128 (16)674 (84)102 (17.7)475 (82.3)Referent1.16 (0.86-1.57)
Alcohol use      
 Never42 (13.9)261 (86.1)89 (18)405 (82)Referent1.36 (0.91-2.04)
 Ever129 (16)677 (84)116 (18.2)520 (81.8)Referent1.18 (0.88-1.57)
Tumor site      
 Oral cavity53 (16.3)273 (83.7)205 (18.1)925 (81.9)Referent1.06 (0.75-1.51)
 Oropharynx83 (14.7)483 (85.3)205 (18.1)925 (81.9)Referent1.32 (1.00-1.76)
 Larynx/hypopharynx35 (16.1)182 (83.9)205 (18.1)925 (81.9)Referent1.12 (0.72-1.75)

DISCUSSION

In the current study, we examined associations between 4 previously reported, important polymorphisms (pre-miRNA hsa-mir-146a, hsa-mir-149, hsa-mir-196a2, and hsa-mir-499)25 and the risk of SCCHN in a non-Hispanic white population. To the best of our knowledge, this is the first and largest study of the role of these pre-miRNA polymorphisms in the etiology of SCCHN. Of the 4 polymorphisms, we observed that only the hsa-mir-499 was associated significantly with risk of SCCHN; however, we did observe an effect of the combined risk genotypes of all the 4 polymorphisms on the risk of SCCHN in a dose-response manner. For example, individuals who had 4 risk genotypes had a significant 40% increased risk of SCCHN compared with individuals who had 0 or 1 genotypes, and the risk was more pronounced in younger subgroups, in men, and in never-smokers. These findings suggest that these 4 pre-miRNA polymorphisms may have a joint effect on the risk of SCCHN, although the exact mechanism by which these variants may influence the risk of SCCHN is not clear. It is possible that this variant either may be functional itself, or it may be in linkage disequilibrium with other functional variants that are involved in the etiology of SCCHN. Such possibilities need to be explored in future studies.

Recent studies have demonstrated that miRNAs may play an important role in human carcinogenesis, and SNPs located either in the pre-miRNAs or within miRNA-binding sites are likely to affect the expression of the miRNA targets and, thus, may contribute to the susceptibility to cancer.24, 34 For example, 1 study indicated that expression of the pre-miRNA-146a variant C allele was 1.9-fold lower than expression of the G allele and that the amount of mature miRNA-146a variant C allele was 1.8-fold lower than the amount of the G allele, suggesting that the pre-miRNA-146a G→C substitution may lead to reduced amounts of mature miRNA-146a; however, the hsa-mir-146a GC heterozygous genotype was associated with an increased risk of papillary thyroid carcinoma, whereas the variant homozygous CC genotype was protective.28 These data also suggest that the C allele may be tumor-specific in cancers with different etiologies.

In the current study, we observed that frequencies of the hsa-mir-146 GG genotype (58%), CG genotype (35.8%), and CC genotype (6.2%) in 1130 non-Hispanic controls were consistent with the frequencies of 58.4%, 35.5%, and 6.1%, respectively, observed by in Jazdzewski et al in 901 Caucasian controls.28 Although, to our knowledge, there have been no previous reports of an association between the hsa-mir-146a polymorphism and the risk of SCCHN, a recent study reported that pre-miRNA-146 was overexpressed in head and neck cancer cell lines and tumor tissues.16, 22 Our results suggest that hsa-mir-146a rs2910164 variant C genotypes may not play a major role in the development of SCCHN. Similarly, other studies of associations between the hsa-mir-146a rs2910164 variant C genotypes and the risk of bladder cancer, renal cancer, and breast cancer failed to identify any overall association in Caucasian27, 29 or Chinese populations.31

To date, few studies have examined the effect of the hsa-mir-196a2 rs11614913 C→T polymorphism on human cancers. One study reported that carriers of the hsa-mir-196a2 CC genotype had reduced survival among patients with nonsmall cell lung cancer in a Chinese population.25 In another study of 1009 patients with breast cancer and 1093 healthy Chinese population controls, the same authors also reported that the hsa-mir-196a2 CC and CC/CT genotypes were associated with a significantly increased risk of breast cancer.31 More recently, Hoffman et al reported that has-mir-196a2 TC and TT were associated significantly with a decreased risk of breast cancer.35 Our hsa-mir-196a2 C→T genotype data from 1109 patients with SCCHN and 1130 controls in a non-Hispanic white population had a distribution of hsa-mir-196a2 CC, CT, and TT genotypes (33.9%, 48.2%, and 17.9%, respectively) in controls similar to that reported in the HapMap database (31%, 50%, and 19%, respectively); however, we did not observe an association between the variant genotypes and the risk of SCCHN. Consistent with our current study, other studies also failed to observe an association between the hsa-mir-196a2 polymorphism and the risk of bladder cancer and renal cancer in Caucasian populations.27, 29

To date, few epidemiologic studies have investigated the association between the hsa-mir-499 rs3746444 A→G SNP and the risk of cancer and survival. In 1 Chinese study, the hsa-mir-499 rs3746444 A→G genotypes were not associated with the risk of nonsmall cell lung cancer or survival.25 In another case-control study of 1009 breast cancer cases and 1093 controls in a Chinese population, the same research team reported that individuals who had the hsa-mir-499 variant GG and GG/AG genotypes had a significantly increased risk of breast cancer compared with individuals who had the AA genotype.31 However, to our knowledge, no reported studies have investigated the association between hsa-mir-499 A→G genotypes and cancer risk in Caucasian populations. In the current study, however, we observed that the hsa-mir-499 variant AG and GG/AG genotypes were associated with a significant, moderately reduced risk of SCCHN, suggesting that the hsa-mir-499 variant G genotypes may play a role in the etiology of SCCHN.

Finally, we observed that the combined effects of these 4 polymorphisms (hsa-146a, hsa-196a, hsa-149, and has-499) were associated with the risk of SCCHN in a risk-genotype, dose-response manner, particularly for individuals who had 4 risk genotypes compared with individuals who had 0 or 1 risk genotypes. This finding implies that the combined genotypes of these pre-miRNAs jointly may have a significant effect on the risk of SCCHN. Furthermore, when comparing the group with 2 to 4 risk genotypes and the group with 0 or 1 risk genotypes, the risk of SCCHN was higher in never-smokers than in ever-smokers, indicating that the risk in nonsmokers may be determined genetically in the absence of exposure to smoking. This risk appeared to be more evident for oropharyngeal cancers than for oral cavity cancers and laryngeal/hypopharyngeal cancers. It has been demonstrated that human papillomavirus infection plays a major role in the etiology of oropharyngeal cancer,36, 37 suggesting that these pre-miRNAs may have an interaction with oncogenic proteins. However, this hypothesis needs to be tested in future studies.

The limitations of this study also exist. Possible selection bias cannot be ruled out, because this was a hospital-based, case-control study, and the cases and controls were not selected from the same population. Also, our analysis was limited to non-Hispanic white individuals, so it is uncertain whether the results are generalizable to other populations. However, by matching on age, sex, and ethnicity, potential confounding factors may be have been minimized. To confirm the role of these polymorphisms in cancer risk will require additional larger studies in different populations and other types of cancer.

In summary, in this case-control study, we observed some evidence of an association between the hsa-mir-499 polymorphism and the risk of SCCHN in a non-Hispanic white population. Our results also suggest that the 4 SNPs may have a joint effect on the risk of SCCHN, especially among men, never-smokers, and patients with oropharyngeal cancer. However, because this is the first study to our knowledge concerning the combined effects of pre-miRNA polymorphisms on the risk of SCCHN, additional, larger replication studies will be needed to confirm the current results.

Acknowledgements

We thank Margaret Lung and Kathryn L. Tipton for their assistance in recruiting the patients; Min Zhao, Jianzhong He, and Kejin Xu for their laboratory assistance; and Li-E Wang for data management.

CONFLICT OF INTEREST DISCLOSURES

Funded in part by National Institutes of Health (NIH) grants R01 ES 11740-07 and R01 CA 131274-01 to Q.W., P50 CA097007 to S.L., and NIH grant P30 CA 016672 to The University of Texas M. D. Anderson Cancer Center.

Ancillary