Estrogen receptor-alpha polymorphisms and predisposition to TMJ disorder.

The haplotype with the highest log-likelihood was the [GC]. The overall model showed that the presence of [GC] was significantly higher in the TMJD with pain compared with control group ( P =.0028) and when the TMJD groups were put together (TMJD total = TMJD no pain + TMJD with pain) the P value was.0012 ( Table 3 ). The odds ratio analysis showed that patients carrying the haplotype [GC] increased significantly the risk to develop signs of TMJD by 2.51 times when the TMJD no pain group was compared with control group (OR = 2.51, 95% CI = 1.267 to 4.97). The same haplotype [GC] increased significantly in 3.203 times the risk to develop TMJD with pain when TMJD with pain group was compared with control group (OR = 3.203, 95% CI = 1.633 to 6.284). However, the risk type [CG] was not significantly associated with higher risk to develop symptoms of TMJD when the TMJD no pain group was compared with TMJD pain group (OR = 1.276, 95% CI = 0.712 to 2.288).

The PvuII and XbaI polymorphic sites are located on intron 1, and whether these sites have functional consequences is unknown. However, polymorphisms on introns could affect mRNA production, as these sites may contain transcriptional regulatory sequences. 52 For example, the Sp1 polymorphism, located in the first intron of the collagen-type I α -1 gene, is known to change the mRNA transcription, leading to decreased bone mineral density and increased fracturerisk. 50 Similarly, the PvuII–XbaI polymorphic sites on the first intron of the ER α gene could influence gene expression. Additionally, a recent report showed that the PvuII polymorphism is located within a potential bMyb binding site, which regulates transcription efficacy of a reporter gene. 25 Alternatively, the location of a variable length of the (TA)n VNTR in the promoter of the ER α gene could also affect gene transcription. Furthermore, previous studies have shown that a VNTR in proximity to a promoter can have a significant influence on transcriptional regulation. 18 However, other polymorphic sites in the estrogen receptor gene might similarly influence TMJD predisposition. Our results show that the presence of genotype [GC] haplotype in the ER α gene may indicate that TMJD is related to receptor dysfunction. Although the exact mechanism is unclear, 3 hypotheses can be proposed about the association between the ER α gene and TMJD predisposition. First, the haplotype [GC] may directly affect the level of ER α through transcriptional regulation. Alternatively, the haplotype [GC] may be in linkage disequilibrium with an exonic polymorphism that has a direct influence on ER α protein function. Finally, the polymorphism may be linked with another unidentified mutation adjacent to the ER α gene affecting the estrogen action on chondrocytes 44 and/or the action of estrogen on monocytes and macrophages altering the release of inflammatory mediators such as cytokines, IL1, 15 IL6, 34 tumor necrosis factor- α (TNF- α ), 36 , 42 and nitric oxide. 30 , 46 Also, our findings showed no differences in haplotype frequencies between the TMJD no pain group when compared with the TMJD pain group, but differences emerged when comparing both TMJD groups to the control group. One possible interpretation of these results is that this rare haplotype is associated with the mechanism of TMJ pathology but is not associated with pain related to TMJD. The presence of 1 risk haplotype [GC] was twofold higher in the combined compared with control group, and the presence of 2 risk haplotypes was only observed in the TMJD group and not in the control group ( Table 3 ). Obviously, additional studies are warranted in order to replicate these findings. The clinical implications of these findings, while speculative, suggest that estrogen action may be of importance after a trauma in the temporomandibular joint. For example, this polymorphism could influence inflammatory responses by altering estrogen’s effects on chondrocytes, which could lead to an increased risk of developing TMJD. Also, we cannot rule out that this polymorphism may also influence the level of estrogen in brain areas responsible for pain processing and also could alter functioning of the opioid system, consequently increasing the pain perception. 49 , 55 As reported for Zubieta et al, 55 when gonadal steroid levels are low, women are less capable of suppressing deep pain by the activation of the μ -opioid system. Moreover, administration of exogenous estrogen increased pain-related μ -opioid receptor binding and reduced pain sensitivity. 44 Recently, Kang et al (2007), 19 studying the association of Pvu II and Xba I polymorphisms and pain susceptibility in female temporomandibular joint osteoarthritis (OA), showed that TMJ OA patients carrying the [GC] haplotype were found to have a significantly higher risk of moderate or severe pain compared to those without the GC haplotype, suggesting that ER- α genotype may be associated with pain susceptibility in female TMJ OA patients. The present study is limited by a small sample size and by the admixed Brazilian population, which limits the generalizability of the data. Additionally, the present genetic study includes only data from articular pain but not muscle pain, and whether similar findings would emerge among patients with myalgia is not known. Considering that estrogen receptors are also present in the muscles, 54 additional research including individuals with muscle pain is warranted. Similar to Kang et al (2007), 19 within the limitations of this pilot study, our results suggest that women carrying the [GC] haplotype at the locus A–351G in their ER α gene have 3.2 times greater susceptibility to TMJD than those who do not carry this genotype. Therefore, this paper supports previous studies regarding the influence of estrogen in TMJD. Further studies are needed to replicate these results and to determine whether ER α is associated with individual differences in pain perception among women with TMJD.

A total of 300 women, from 18 to 44 years old, were enrolled from the patient pool from the Integrated Dental Clinic at the Dental College of Piracicaba, Brazil. The participants were interviewed and asked to complete a detailed questionnaire on personal and family medical history, TMD symptoms, and age of onset, frequency, severity, and treatment and its response. Next, the Research Diagnostic Criteria for Temporomandibular Disorders (RDC/TMD) questionnaire 10 was completed for each participant. This research was approved by the Brazilian Institutional Review Board, and informed consent was obtained prior to commencement. Participants were divided into 3 groups with 100 patients in each group. Age and self-reported racial group are described in Table 1 . The inclusion criteria for group 1 (signs and symptoms of TMJD) and group 2 (only signs of TMJD without pain) were presence of disc displacement (DD) with reduction (R); without reduction with limited opening (WRLO); without reduction and without limited opening (WRWLO). The exclusion criteria were presence of pain/inflammation elsewhere in the body, or taking contraceptives or any other medication. The difference between groups 1 and 2 was that group 1 had the added inclusion criterion of arthralgia for more than 6 consecutive months, whereas for group 2 the presence of arthralgia was a criterion for exclusion. For the control group all participants were healthy, without any signs or symptoms of TMJD. The control group was age matched to the TMJD groups. The sampling of buccal epithelial cells was performed as previously described by Trevilatto and Line. 48 Briefly, individuals rinsed their mouths with mouthwash, containing 5 mL 3% glucose solution for 1 minute. After mouthwash, a sterile wood spatula was used to scrape oral mucosa. The tip of the spatula was then inserted into the retained mouthwash solution. Buccal epithelial cells were pelleted by centrifugation at 5000 g for 10 minutes. The supernatant was discarded and the cell pellet resuspended in 500 μ L of extraction buffer solution [10 mM Tris-HCl (pH 7.8), 5 mM ethylenediamine tetra-acetic acid (EDTA), and 0.5% sodium dodecyl sulphate (SDS)]. Finally, the samples were frozen at −20°C until used for DNA extraction. After thawing, the samples were incubated overnight (ON) with 100 ng/mL proteinase K (Sigma Chemical Co, St Louis, MO) at 37°C with agitation. DNA was then purified by sequential phenol/chloroform extraction and salt/ethanol precipitation. DNA was dissolved in 200 μ L TE buffer [10 mM Tris (pH 7.8), 1 mM EDTA]. The concentration was estimated by measurements of OD 260 nm. The sequence from T-397G locus in the ER α gene promoter was PCR amplified with primers 5′–GATATCC AGGGTTATGTGGCA–3′ (forward) and 5′– AGGTGTTGCC TATTATATTAACCTTGA – 3′ (reverse). PCR was carried out in a total volume of 50 μ L, containing 500 ng genomic DNA, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl 2 , 1 μ M of each primer, 200 μ M of each dATP, dCTP, dGTP, and dTTP, and 2.5 U Taq DNA polymerase (Amersham Pharmacia Biotech AB, Uppsala, Sweden). The solution was incubated for 5 minutes, 40 seconds at 95°C, followed by 35 cycles of 1 minutes, 40 seconds at 95°C; 1 minute, 40 seconds at 61°C; and 1 minute, 40 seconds at 72°C; with a final extension of 72°C for 7 minutes, 40 seconds. A fragment of A-351G of the ER α gene promoter was PCR amplified with primers 5′– GATATCCAGGGTTA TGTGGCA – 3′ (forward); 5′– AGGTGTTGCCTATTATATT AACCTTGA – 3′ (reverse). PCR was carried out in a total volume of 50 μ L, containing 500 ng of genomic DNA, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl 2 , 1 μ M of each primer, 200 μ M each dATP, dCTP, dGTP, and dTTP, and 4 U Taq DNA polymerase (Amersham Pharmacia Biotech AB). The solution was incubated for 5 minutes, 40 seconds at 95°C, followed by 35 cycles of 1 minute, 40 seconds at 95°C; 1 minute, 40 seconds at 61°C; and 1 minute, 40 seconds at 72°C; with a final extension of 72°C for 7 minutes, 40 seconds. A 3.5- μ L aliquot of ER α (T-397C) PCR products was digested with 4 units of XbaI (New England Biolabs, Inc, Beverly, MA) in a final volume of 20 μ L. For analysis of ER α (A-351G) polymorphism a 3.5- μ L aliquot of ER α PCR products was digested with 0.6 units of Hga I (New England Biolabs, Inc) in a final volume of 20 μ L. The digestions were carried out at 37°C for 16 hours. The sizes of the fragments generated by the restriction enzymes were previously described. 3 , 24 , 51 The total amount aliquot of the digest was mixed with 3 μ L of loading buffer and electrophoresed on a 10% vertical nondenaturing polyacrylamide gel at 20 mA. The gel was silver stained by DNA Silver Staining Kit (Amersham Pharmacia Biotech AB). All statistical analyses in this article were performed by using SAS v9.1. A P value of less than.05 was considered statistical significant. The genotype and allele frequencies of single polymorphisms were calculated by direct counting and then dividing by the number of subjects to produce genotype frequency. The significance of the differences in observed frequencies of each polymorphism in test groups was assessed by χ 2 test. Our 2 SNPS are genotyped and each SNP has 2 alleles ( A, G, and T, C respectively). Therefore there were 4 possible haplotypes: [AT], [AC], [GT], and [GC]. The frequencies of the 4 haplotypes can be denoted by P [AT] , P [AC] , P [GT] and P [GC] . Assuming Hardy-Weinberg equilibrium, 9 genotypes and their frequencies can are constructed in Table 2 based on the frequencies of haplotypes. Table 2 lists all 9 genotypes at 2 SNPs, possible diplotypes within each genotype, and their frequencies. Note that all genotypes except the double heterozygote have only 1 known diplotype. For the double heterozygote AG/TC , there are 2 possible unknown diplotypes [GC][AT] and [AC][GT]. Let p denote the relative frequency of diplotype [GC][AT] within the genotype GA/CT, therefore p can be expressed as in Liu et al. 26 The frequencies of subjects in each genotype {10,29,16,34,173,18,10,8,2} are also listed as the observations. The log-likelihood function of the 4 unknown haplotype frequencies ( P [GC] , P [GT] , P [AC] , P [AT] ) given 9 observed genotypes (G) and observation frequencies (O) is then written as previously reported 26 : Log L ( P [ GC ] , P [ GT ] , P [ AC ] , P [ AT ] ∣ G , O ) ∝ 2 ∗ 10 ∗ log ( P [ GC ] ) + 29 ∗ log ( P [ GC ] P [ GT ] ) + 2 ∗ 16 ∗ log ( P [ GT ] ) + 34 ∗ log ( P [ GC ] P [ AC ] ) + 1 ∗ 173 log ( P [ GC ] P [ AT ] + P [ AC ] , P [ AT ] ) + 18 ∗ log ( P [ GT ] P [ AT ] ) + 2 ∗ 10 ∗ log ( P [ AC ] ) + 8 ∗ log ( P [ AC ] P [ AT ] ) + 2 ∗ 2 ∗ log ( P [ AT ] ) An iterative Expectation-Maximization algorithm developed by Liu et al 26 can be implemented to estimate the unknown haplotype frequencies. In the E step, the relative frequency of diplotype [GC][AT] within the genotype GA/ CT is calculated by function (1) denoted by π̂ In the M step, the haplotype frequencies can be calculated using: P ^ [ GC ] = ( 2 ∗ 10 + 29 + 34 + 173 ∗ π ^ ) / 2 ∗ 300 , P ^ [ GT ] = [ 2 ∗ 16 + 29 + 18 + 173 ∗ ( 1 − π ^ ) ] / 2 ∗ 300 , P ^ [ AC ] = [ 2 ∗ 10 + 34 + 8 + 173 ∗ ( 1 − π ^ ) ] / 2 ∗ 300 , P ^ [ AT ] = [ 2 ∗ 18 + 8 + 173 ∗ ( 1 − π ^ ) ] / 2 ∗ 300. The E step and M step are taken repeatedly until the estimates converge. After haplotype frequencies were obtained from the EM algorithm, haplotypic effects can be estimated by recategorizing the 10 diplotypes into 3 diplotype groups ([GC] is the risk reference haplotype): [GC][GC] - no risk presence reference haplotype; [ GC ] [ GC ¯ ] − 1 risk presence reference haplotype; and [ GC ¯ ] [ GC ¯ ] double presence risk reference haplotype. The frequencies of haplotypes formed by T-397C and A-351G polymorphisms [AC], [AT], [GC], and [GT] were calculated by using the EM algorithm. The χ 2 was used to test the association between the haplotypes and test groups. Four χ 2 tests were calculated for each haplotype [AC], [AT], [GC], and [GT] by using the 3 recategorized groups: double presence of risk reference haplotypes, 1 risk presence reference haplotype and no risk presence reference haplotype ( Table 3 ). The risk reference haplotype can then be determined from the 4 haplotypes [AC], [AT], [GC], and [GT] by selecting the highest log-likelihood value of the χ 2 tests. Furthermore, the risk associated with haplotypes was calculated as the odds ratio with 95% confidence intervals.