Abstract
This study aims to investigate whether the trajectory of dental caries in the life course is associated with rs307355 (TAS1R3) and rs35874116 (TAS1R2) and if there is an epistatic association between rs307355 (TAS1R3) and rs35874116 (TAS1R2). A representative sample of all 5,914 births from the 1982 Pelotas birth cohort was prospectively investigated, and the decayed, missing, and filled teeth (DMF-T) components were assessed at ages 15 (n = 888), 24 (n = 720), and 31 (n = 539) y. Group-based trajectory modeling was used to identify groups with similar trajectories of DMF-T components in the life course. Genetic material was collected, and rs307355 (TAS1R3) and rs35874116 (TAS1R2) were genotyped. Ethnicity was evaluated using ADMIXTURE. Generalized multifactor dimensionality reduction software was used to investigate epistatic interactions. Considering rs307355 (TAS1R3) in the additive effect, the genotype TT was associated with the high decayed trajectory group (odds ratio [OR] = 4.52; 95% confidence interval [CI], 1.15–17.74) and the high missing trajectory group (OR = 3.35; 95% CI, 1.09–10.26). In the dominant effect, the genotype CT/TT was associated with the high decayed trajectory group (OR = 1.64; 95% CI, 1.14–2.35). Allele T was associated with an increased odds of 64% (OR = 1.64; 95% CI, 1.20–2.25) for the decayed component and 41% (OR = 1.41; 95% CI, 1.04–1.92) for the missing component. No associations were observed between rs307355 (TAS1R3) and the filled component. rs35874116 (TAS1R2) was not associated with DMF-T components. Positive epistatic interactions were observed involving rs307355 (TAS1R3) and rs35874116 (TAS1R2) with the decayed component (OR = 1.72; 95% CI, 1.04–2.84). Thus, rs307355 (TAS1R3) genotypes and alleles seem positively associated with the trajectory of decayed and missing components in the life course. Epistatic interaction between rs307355 and rs35874116 may increase the decayed caries trajectory.
Keywords: TAS1R3, TAS1R2, DMF-T, dental caries, longitudinal study, single-nucleotide polymorphism
Introduction
Genome-wide association study has shown that several suggested loci of taste receptor genes are linked to dental caries in the primary dentition (Alotaibi et al. 2021). Sweet tastes are intermediated by the G-protein-coupled receptors superfamily and are mainly expressed in the epithelial cells of the tongue and palate. The G-protein-coupled family includes 3 members: T1R1, T1R2, and T1R3. TAS1R3 and TAS1R2 encode 2 of 3 main proteins of the sweet taste receptor, the T1R3 (taste receptor type 1, member 3) and T1R2 (taste receptor type 1, member 2) (Liao and Schultz 2003). Some studies show that T1R2 and T1R3 act as coreceptors or heterodimers (Montmayeur et al. 2001; Nelson et al. 2001). T1R2 + T1R3 heterodimers function as sweet taste receptors (Zhao et al. 2003; Li 2009; Yoshida and Ninomiya 2016). The ability to perform fine discrimination in sucrose sweetness was found to be reduced in individuals with CT or TT genotypes in rs307355, a single-nucleotide polymorphism (SNP) of TAS1R3 (Fushan et al. 2009). rs307355 is located in a putative regulatory region of TAS1R3 (Fushan et al. 2009). Changes in the promoter of TAS1R3 lead to changes in sweet taste perception (Fushan et al. 2009; Choi et al. 2017). Findings from the luciferase reporter assay show that rs307355 (TAS1R3) affects gene transcription by altering the function, influencing the sweet taste ability of individuals from different ethnicities (Fushan et al. 2009). Koreans with the CT genotype in rs307355 (TAS1R3) presented higher consumption of Soju, a Korean alcoholic beverage that contains a variety of natural or artificial sweeteners (Choi et al. 2017). There is an important association between the genotype CT of rs307355 and dental caries experience (Haznedaroglu et al. 2015). Similarly, TAS1R2 was associated with an increase in taste threshold and preferred intensity of sweet and dental caries (Eriksson et al. 2019).
Although the literature has presented important evidence that rs307355 and rs35874116 are associated with changes in taste perceptions (Fushan et al. 2009; Eriksson et al. 2019), this hypothesis has not been explored in population-based samples with longitudinal monitoring of caries. Moreover, it is unknown whether there are possible interactions between polymorphisms in genes of TAS1R3 and TAS1R2. Caries is a complex, multifactorial disease, and there may be underlying epistatic gene-by-gene interactions. We aimed to investigate the effect of rs307355 (TAS1R3) and rs35874116 (TAS1R2) on dental caries. We have addressed 2 questions: 1) is the trajectory of decayed, missing, and filled teeth (DMF-T) components in the life course positively associated with the rs307355 (TAS1R3) and rs35874116 (TAS1R2) genotypes (homozygotes and heterozygote) and allele? 2) Is there an epistatic interaction between SNPs of rs307355 (TAS1R3) and rs35874116 (TAS1R2) that would increase the caries risk?
Methods
The present study was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) (cohort studies). The ethics approval number is 19551713.9.0000.5317.
Study Design, Setting, and Participants
This birth cohort study was performed in Pelotas, a southern Brazilian city. All live births in the maternity hospitals of Pelotas in 1982 (5,914 children; 99.2%) were identified and included in a perinatal health survey. The Oral Health Study 1997 (OHS-97) was conducted on a sample of this same cohort when subjects were 15 y old (1997). Thus, 900 individuals were randomly selected, and 888 (98.7%) young adults comprised the OHS. The 888 young participants of the OHS-97 were contacted in 2006 (OHS-06) for a new interview and dental examinations. A total of 720 individuals composed the OHS-06 (80% of the initial sample). In 2013 (31 y), 900 individuals from the initial sample were contacted, and all individuals who were located and agreed to continue the study made up the 2013 Oral Health Study (OHS-13). Of the cohort seen at 31 y, 87.9% presented for 3 visits, and 10.9% presented for 2 visits (missed the one at 24 y).
Outcome Variable (Caries Phenotyping)
Dental clinical examinations were carried out by 6 dentists with visual inspection and headlight using a probe and mirror using the decayed, missing, and filled teeth (DMF-T) criteria for identifying caries (World Health Organization [WHO] 2013). No radiographic examination was performed, and we did not consider noncavitated, white spot caries lesions (WHO 2013). Examiners were trained and calibrated according to standardized procedures. Interexaminer reliability was measured by κ statistics, and the lowest κ value observed in this study was 0.65.
The outcome of the present study was the dental caries trajectory of the participants (15, 24, and 31 y) according to DMF-T components. The DMF-T collected at the studied ages was organized as a discrete variable for each DMF-T component. A group-based trajectory modeling was used to identify groups with similar trajectories of each DMF-T component in the life course. The model was estimated with the command “traj” in Stata 16.0 (Silva et al. 2018) to identify the similarity of the trajectory among evaluated individuals. The parameters for the model trajectory were determined based on the maximum likelihood by the quasi-Newton method (Jones and Nagin 2007). The number of trajectories was determined by sequential comparisons of the Bayesian information criterion (BIC) and its fit criteria when substantial differences between the K and K + 1 trajectory model were not produced in the k + 1 model BIC score. Finally, 2 trajectories for each DMF-T component group (low and high) were produced, including the 888 individuals (Fig.).
Figure.
Description of trajectories according to decayed, missing, and filled teeth (DMF-T) and DMF-T components by follow-ups.
Independent Variables
The selection of confounding variables was made through a directed acyclic graph (DAG) (Shrier and Platt 2008), which illustrates the minimum set of variables that must be considered in the adjustment so that all noncausal paths that lead the exposure variables to the outcome of a change in caries risk are adjusted in the analysis. Based on the literature, the following variables had causal relationships in the DAG (ethnicity, SNPs of sweet taste receptor gene, and sugar intake [mediator]; Appendix Fig. 1). The independent variables used in the study were sex (aiming to increase statistical power and residual sample variance) and ethnicity (used to control the possibility of confounding caused by population heterogeneity). Ethnicity was evaluated using ADMIXTURE (Alexander et al. 2009), based on approximately 370,000 SNPs available from the 1982 Pelotas birth cohort compatible with the HapMap and Human Genome Diversity projects for the Brazilian population.
Blood samples were collected by venipuncture at the first examination. DNA and serum were extracted and frozen at −70°C. DNA samples were genotyped using the Illumina HumanOmni2.5-8v1 array, and the SNPs of rs307355 (TAS1R3) and rs35874116 (TAS1R2) were genotyped (Horta et al. 2015).
Power Calculation
Power calculation was performed in OpenEpi (https://www.openepi.com/Menu/OE_Menu.htm). This study has 97% power to detect a difference due to its sample size, an α of 0.05, a 32% prevalence of individuals in the high caries experience category in the nonexposed group (genotype CC), and 45% in the exposed group (genotype CT/TT).
Statistical Methods
Stata statistical package (version 16.0) was used for all statistical analysis (StataCorp). Hardy–Weinberg equilibrium and allele frequency estimation of the population regarding the SNPs were tested using the command “genhw.” The descriptive analysis determined the absolute and relative frequency of independent variables and genotype and allele frequency. The associations between sugar consumption and genotype and DMF-T components were performed with a χ2 test.
Two different analyses were conducted with logistic regression to analyze the association between dental caries by DMF-T component trajectories and SNPs. The primary analysis involved the genotype of individuals. Considering genotype analysis, initially, we assume an additive genetic effect of the T allele in rs307355 (TAS1R3) polymorphism (i.e., homozygous CC individuals = 0, heterozygous CT = 1, and homozygous TT = 2) and C allele in rs35874116 (TAS1R2) (i.e., homozygous TT individuals = 0, heterozygous CT = 1, and homozygous CC = 2). We also performed an analysis considering the dominant effect of allele T (i.e., homozygous CC individuals = 0, heterozygous CT = 1, and homozygous TT = 1) to rs307355 (TAS1R3) and C dominant effect (i.e., homozygous TT individuals = 0, heterozygous CT = 1, and homozygous CC = 1) to rs35874116 (TAS1R2).
Two models were performed for genotype analysis: 1) unadjusted (i.e., no covariates) and 2) adjusted: controlling for ethnicity and sex. A multilevel logistic regression model was used for all allele analysis, considering mixed effects and 2 hierarchical levels: genetic (level 1) and personal (level 2). The same adjustments performed in genotype analysis were performed in the allelic analysis. We also performed genotype and allelic analysis according to the different status of the decayed component in each follow-up.
Linkage disequilibrium analysis was performed, aiming to establish the nonrandom association of alleles. The estimating of D′ was performed using the SHEsis (Shi and He 2005; Wang and Qin 2018). Haplotype analyses were performed using the same software. Generalized multifactor dimensionality reduction (GMDR) software was used to complement the epistatic interactions (i.e., gene–gene interactions). We used the DMF-T component trajectories as outcomes to perform epistatic analysis. Logistic regression models and the genotypes of all SNPs were performed and adjusted by ethnicity and sex.
Results
A total of 539 individuals were assessed in the OHS-13, corresponding to a response rate of 59.9%. The mean (SD) DMF-T was 5.1 (3.7), 5.6 (4.1), and 7.0 (4.5) at 15, 24, and 31 y of age, respectively. The Figure presents the description of trajectories according to DMF-T and DMF-T components by follow-ups. The sample consisted mainly of males (53.8%) and European ethnicity (89.1%) (Appendix Table 1).
Genetic Information
First, we tested whether the 2 SNPs were in Hardy–Weinberg equilibrium (Mario 1999). Indeed, the SNPs rs307355 (TAS1R3) and rs35874116 (TAS1R2) were in Hardy–Weinberg equilibrium (P > 0.05), and the minor allele frequency (MAF) was 0.1475 and 0.3278 in this population, respectively (Appendix Table 2). Therefore, the SNPs were not in linkage disequilibrium (Appendix Fig. 2), and gene–gene interactions could be carried out and the evaluated SNPs in this population will remain constant from one generation to the next in the absence of disturbing factors. The genetic profile (genotypic/allelic) is compared in terms of dental caries trajectory by DMF-T component groups and general characteristics in Table 1. Males and individuals of African ethnicity showed a higher proportion of the T allele of rs307355 (TAS1R3) than female and European individuals. These results suggest a potential residual sample variance and a possible confounding caused by population heterogeneity, which must be controlled in the regressions (by sex and ethnicity).
Table 1.
Description of Genotype and Allele Frequency and Individual Variables.
| Variable | rs307355 (TAS1R3) | ||||
|---|---|---|---|---|---|
| Genotype, n (%) | Allele, n (%) | ||||
| CC | TC | TT | C | T | |
| Sex distribution | |||||
| Male | 270 (77.36) | 72 (20.63) | 7 (2.01) | 612 (87.68) | 86 (12.32) |
| Female | 222 (68.73) | 94 (29.10) | 7 (2.17) | 538 (83.28) | 108 (16.72) |
| Ethnicity | |||||
| European | 463 (75.41) | 141 (22.96) | 10 (1.63) | 1,067 (86.89) | 161 (13.11) |
| African | 29 (50.00) | 25 (43.10) | 4 (6.90) | 83 (71.55) | 33 (28.45) |
| Decayed component trajectory | |||||
| Low | 335 (77.19) | 95 (21.89) | 4 (0.92) | 765 (88.13) | 103 (11.87) |
| High | 158 (66.11) | 71 (29.71) | 10 (4.18) | 387 (80.96) | 91 (19.04) |
| Missing component trajectory | |||||
| Low | 330 (75.86) | 100 (22.99) | 5 (1.15) | 760 (87.36) | 110 (12.64) |
| High | 166 (68.31) | 68 (27.98) | 9 (3.70) | 400 (82.30) | 86 (17.70) |
| Filled component trajectory | |||||
| Low | 224 (67.67) | 99 (29.91) | 8 (2.42) | 547 (82.63) | 115 (17.37) |
| High | 272 (78.39) | 69 (19.88) | 6 (1.73) | 613 (88.33) | 81 (11.67) |
| rs35874116 (TAS1R2) | |||||
| Genotype, n (%) | Allele, n (%) | ||||
| TT | CT | CC | T | C | |
| Sex distribution | |||||
| Male | 167 (48.27) | 146 (42.20) | 33 (9.54) | 480 (69.36) | 212 (30.64) |
| Female | 145 (45.17) | 145 (45.17) | 31 (9.66) | 435 (67.76) | 207 (32.24) |
| Ethnicity | |||||
| European | 288 (47.29) | 263 (43.19) | 58 (9.52) | 839 (68.88) | 379 (31.12) |
| African | 24 (41.38) | 28 (48.28) | 6 (10.34) | 76 (65.52) | 40 (34.48) |
| Decayed component trajectory | |||||
| Low | 208 (47.93) | 187 (43.09) | 39 (8.99) | 603 (69.47) | 265 (30.53) |
| High | 104 (44.64) | 104 (44.64) | 25 (10.73) | 312 (66.95) | 154 (33.05) |
| Missing component trajectory | |||||
| Low | 207 (47.59) | 187 (42.99) | 41 (9.43) | 601 (69.08) | 269 (30.92) |
| High | 109 (44.86) | 110 (45.27) | 24 (9.88) | 328 (67.49) | 158 (32.51) |
| Filled component trajectory | |||||
| Low | 165 (49.85) | 138 (41.69) | 28 (8.46) | 468 (70.69) | 194 (29.31) |
| High | 151 (43.52) | 159 (45.82) | 37 (10.66) | 461 (66.43) | 233 (33.57) |
Gene-by-Gene Interactions and Caries Experience over Time
In order to investigate the genetic influence of the rs307355 (TAS1R3) genotype on caries trajectory by DMF-T components, a logistic regression model reporting odds ratio (OR) was performed (Table 2). We next considered the additive effect of risk alleles (i.e., individuals with 1 or 2 risk alleles are analyzed in separate categories). Only the genotype TT was associated with the high decayed trajectory group (OR = 4.52; 95% confidence interval [CI], 1.15–17.74) and missing tooth trajectory group (OR = 3.35; 95% CI, 1.09–10.26). This result suggests that an increase in dental caries risk is observed only when 2 risk alleles are present in the individuals. In the dominant effect (i.e., individuals with 1 or 2 risk alleles are analyzed in 1 category), the genotype CT/TT was associated with the high decayed trajectory group (OR = 1.64; 95% CI, 1.14–2.35), and no association was observed in the missing trajectory group in adjusted analyses (OR = 1.38; 95% CI, 0.97–1.98). No associations were observed for the filled component. The associations between rs35874116 (TAS1R2) and DMF-T component trajectories were not observed in additive and dominant models.
Table 2.
Logistic Regression Reporting Odds Ratio of the Association between the High Dental Caries Trajectories by Decayed, Missing, and Filled Teeth Components and TAS1R3 rs307355 and TAS1R2 rs35874116 Genotype (in Additive and dominant Effect).
| Additive Effect | OR (95% CI) | P Value | Dominant Effect | OR (95% CI) | P Value |
|---|---|---|---|---|---|
| Decayed component trajectory | |||||
| TAS1R3 rs307355 | |||||
| Unadjusted | Unadjusted | ||||
| CC | 1 | CC | 1 | ||
| CT | 1.59 (1.05–2.40) | 0.025 | CT/TT | 1.74 (1.22–2.46) | 0.002 |
| TT | 5.30 (1.38–20.32) | 0.011 | |||
| Adjusted | Adjusted | ||||
| CC | 1 | CC | 1 | ||
| CT | 1.51 (0.99–2.31) | 0.059 | CT/TT | 1.64 (1.14–2.35) | 0.007 |
| TT | 4.52 (1.15–17.74) | 0.027 | |||
| TAS1R2 rs35874116 | |||||
| Unadjusted | Unadjusted | ||||
| TT | 1 | TT | 1 | ||
| CT | 1.10 (0.75–1.61) | 1.000 | CT/CC | 1.13 (0.83–1.56) | 0.432 |
| CC | 1.29 (0.69–2.43) | 0.700 | |||
| Adjusted | Adjusted | ||||
| TT | 1 | TT | 1 | ||
| CT | 1.09 (0.74–1.69) | 1.000 | CT/CC | 1.13 (0.82–1.55) | 0.473 |
| CC | 1.29 (0.68–2.43) | 0.750 | |||
| Missing component trajectory | |||||
| TAS1R3 rs307355 | |||||
| Unadjusted | Unadjusted | ||||
| CC | 1 | CC | 1 | ||
| CT | 1.35 (0.94–1.94) | 0.101 | CT/TT | 1.46 (1.03–2.06) | 0.034 |
| TT | 3.58 (1.18–10.85) | 0.024 | |||
| Adjusted | Adjusted | ||||
| CC | 1 | CC | 1 | ||
| CT | 1.29 (0.89–1.85) | 0.174 | CT/TT | 1.38 (0.97–1.98) | 0.072 |
| TT | 3.35 (1.09–10.26) | 0.034 | |||
| TAS1R2 rs35874116 | |||||
| Unadjusted | Unadjusted | ||||
| TT | 1 | TT | 1 | ||
| CT | 1.11 (0.76–1.63) | 0.511 | CT/CC | 1.11 (0.81–1.53) | 0.494 |
| CC | 1.11 (0.58–2.09) | 0.708 | |||
| Adjusted | Adjusted | ||||
| TT | 1 | TT | 1 | ||
| CT | 1.10 (0.75–1.61) | 0.568 | CT/CC | 1.10 (0.80–1.51) | 1.000 |
| CC | 1.10 (0.58–2.08) | 0.732 | |||
| Filled component trajectory | |||||
| TAS1R3 rs307355 | |||||
| Unadjusted | Unadjusted | ||||
| CC | 1 | CC | 1 | ||
| CT | 0.57 (0.38–0.86) | 0.004 | CT/TT | 0.58 (0.41–0.81) | 0.002 |
| TT | 0.61 (0.18–2.11) | 0.321 | |||
| Adjusted | Adjusted | ||||
| CC | 1 | CC | 1 | ||
| CT | 0.58 (0.38–1.05) | 0.065 | CT/TT | 0.58 (0.42–1.04) | 0.062 |
| TT | 0.74 (0.20–2.60) | 0.424 | |||
| TAS1R2 rs35874116 | |||||
| Unadjusted | Unadjusted | ||||
| TT | 1 | TT | 1 | ||
| CT | 1.25 (0.87–1.81) | 0.155 | CT/CC | 1.29 (0.95–1.75) | 0.099 |
| CC | 1.44 (0.78–2.67) | 0.226 | |||
| Adjusted | Adjusted | ||||
| TT | 1 | TT | 1 | ||
| CT | 1.27 (0.88–1.84) | 0.593 | CT/CC | 1.30 (0.96–1.78) | 0.086 |
| CC | 1.48 (0.79–2.77) | 0.199 | |||
P value was adjusted by multiple comparisons. Adjusted: ethnicity and sex.
CI, confidence interval; OR, odds ratio.
These results suggest a consistent association between rs307355 (TAS1R3) and decayed and missing components considering different effects (additive or dominant), while no associations are observed for rs35874116 (TAS1R2).
Allelic Findings
We tested the hypothesis that risk alleles might be linked with higher odds of DMF-T components. Thus, we chose an analytical approach that considers 2 hierarchical levels as previously suggested (Yi 2010): genetic (level 1) and personal (level 2). We clustered the alleles in each individual. Allele T of rs307355 (TAS1R3) was associated with an increased odds of 64% (OR = 1.64; 95% CI, 1.20–2.25) for decayed components and 41% (OR = 1.41; 95% CI, 1.04–1.92) for the missing tooth component, corroborating with the genotype results. No associations were observed between the filled component and rs35874116 (TAS1R2) in adjusted analyses (Table 3). Sugar consumption was more strongly associated with the decayed and missing component trajectory than with the rs307355 (TAS1R3) genotype (Appendix Table 3). There were also complementary results in which analyses conducted separately in each follow-up were similar to those obtained with longitudinal data (Appendix Tables 4 and 5). These findings could mean that the observed results have not changed significantly over the years, highlighting the robustness of the present findings. The results from allelic analysis confirm the association between rs307355 (TAS1R3) and decayed and missing components observed in genotype analysis.
Table 3.
Multilevel Logistic Regression Reporting Odds Ratio of the Association between High Dental Caries Trajectories by Decayed, Missing, and Filled Teeth Components and rs307355 (TAS1R3) and rs35874116 (TAS1R2) Alleles.
| Decayed Component Trajectory | |||||
|---|---|---|---|---|---|
| rs307355 (TAS1R3) | OR (95% CI) | P Value | rs35874116 (TAS1R2) | OR (95% CI) | P Value |
| Unadjusted | <0.001 | Unadjusted | |||
| C | 1 | T | 1 | ||
| T | 1.75 (1.28–2.37) | C | 0.11 (0.87–1.42) | 0.397 | |
| Adjusted | 0.002 | Adjusted | |||
| C | 1 | T | 1 | ||
| T | 1.64 (1.20–2.25) | C | 1.12 (0.88–1.42) | 0.371 | |
| Missing Component Trajectory | |||||
| rs307355 (TAS1R3) | OR (95% CI) | P Value | rs35874116 (TAS1R2) | OR (95% CI) | P Value |
| Unadjusted | 0.012 | Unadjusted | 0.545 | ||
| C | 1 | T | 1 | ||
| T | 1.49 (1.09–2.02) | C | 1.07 (0.84–1.36) | ||
| Adjusted | 0.029 | Adjusted | 0.591 | ||
| C | 1 | T | 1 | ||
| T | 1.41 (1.04–1.92) | C | 1.07 (0.80–1.35) | ||
| Filled Component Trajectory | |||||
| rs307355 (TAS1R3) | OR (95% CI) | P Value | rs35874116 (TAS1R2) | OR (95% CI) | P Value |
| Unadjusted | 0.003 | Unadjusted | 0.091 | ||
| C | 1 | T | 1 | ||
| T | 0.63 (0.46–0.85) | C | 0.82 (0.65–1.03) | ||
| Adjusted | 0.060 | Adjusted | 0.078 | ||
| C | 1 | T | 1 | ||
| T | 0.65 (0.47–1.05) | C | 1.23 (0.98–1.56) | ||
P value was adjusted by multiple comparisons. Adjusted (1): ethnicity and sex.
CI, confidence interval; OR, odds ratio.
Epistasis Analysis (Gene–Gene Interaction and Association with DMF-T)
Haplotype analysis (Table 4) was performed, aiming to test the relationship between different alleles and dental caries trajectories. The combinations of allele “C” of rs307355 (TAS1R3) with allele “T” of rs35874116 (TAS1R2) and the allele “T” of rs307355 (TAS1R3) with allele “T” of rs35874116 (TAS1R2) were associated with the decayed trajectory group. The combination of allele “T” of rs307355 (TAS1R3) and “C” of rs35874116 (TAS1R2) was associated with individuals who had more missing teeth in the high caries trajectory group.
Table 4.
Haplotype Analysis of Loci for Hap-Analysis: rs307355 (TAS1R3) and rs35874116 (TAS1R2) and Summary of Generalized Multifactor Dimensionality Reduction Results for Gene–Gene Interactions.
| Haplotype Analysis of Loci for Hap-Analysis | ||||
|---|---|---|---|---|
| Haplotype | High-Trajectory Frequency | Low-Trajectory Frequency | Fisher’s P Value | OR (95% CI) |
| Decayed component trajectory | ||||
| C C | 0.279 | 0.268 | 0.065 | 1.06 (0.82–1.36) |
| C T | 0.531 | 0.614 | 0.003 | 0.71 (0.57–0.89) |
| T C | 0.052 | 0.038 | 0.220 | 1.40 (0.82–2.38) |
| T T | 0.139 | 0.081 | <0.001 | 1.83 (1.28–2.61) |
| Missing component trajectory | ||||
| C C | 0.257 | 0.283 | 0.309 | 0.87 (0.68–1.13) |
| C T | 0.566 | 0.591 | 0.379 | 0.90 (0.72–1.32) |
| T C | 0.068 | 0.027 | <0.001 | 2.61 (1.52–4.47) |
| T T | 0.109 | 0.099 | 0.585 | 1.11 (0.77–1.59) |
| Filled component trajectory | ||||
| C C | 0.297 | 0.244 | 0.061 | 1.30 (0.98–1.66) |
| C T | 0.586 | 0.582 | 0.872 | 1.01 (0.82–1.26) |
| T C | 0.039 | 0.049 | 0.365 | 0.78 (0.46–1.32) |
| T T | 0.078 | 0.125 | 0.062 | 0.59 (0.41–1.02) |
Results were adjusted by ethnicity and sex.
CI, confidence interval; CVC, cross-validation consistency; OR, odds ratio; Te-BA, testing-balanced accuracy; Tr-BA, training-balanced accuracy.
Gene–gene interactions were observed between rs307355 (TAS1R3) and rs35874116 (TAS1R2) (P = 0.034). The combination of these SNPs presented an increased odds ratio of 1.72 (95% CI, 1.04–2.84) of being in the high decayed trajectory group (Appendix Fig. 3). These results suggest a potential epistatic interaction between the tested SNPs. Although the rs35874116 (TAS1R2) has not been directly associated with DMF-T components, it may influence individuals’ risk through interaction with rs307355 (TAS1R3).
Discussion
Our findings corroborate with previous studies (Haznedaroglu et al. 2015) and confirmed the association between rs307355 (TAS1R3) and increased risk of decayed and missing teeth. Our findings showed that genotypes TT and allele T of rs307355 (TAS1R3) were associated with high decayed and missing trajectories in the life course. We have not found a direct association between rs35874116 (TAS1R2) and DMF-T trajectory components and therefore partially disagree with previous studies (Kulkarni et al. 2013; Haznedaroglu et al. 2015; Izakovicova Holla et al. 2015). We did find an important epistatic interaction between rs307355 (TAS1R3) and rs35874116 (TAS1R2) on the decayed trajectory. These findings highlight the complex architecture of the genetic influence of dental caries and show a possible epistatic interaction underlying initial observations of rs35874116 (TAS1R2) (Kulkarni et al. 2013; Haznedaroglu et al. 2015).
Variation in sweet taste perception has been observed in different populations (Ashi et al. 2017), and SNPs on taste genes such as rs3935570 (TAS1R2) and rs17492553 (TAS1R1) were also associated with dental caries experience (Arid et al. 2020; Chisini et al. 2021; Vieira 2021). A systematic review investigating the association between SNPs of taste genes and dental caries experience identified 12 potential SNPs (Chisini et al. 2021). However, the gene TAS1R3 has been presented as mainly responsible for the saccharin-preferring phenotype because it encodes T1R3 (Yoshida and Ninomiya 2016). A study in mice that isolated T1r3 from taster mice showed initial evidence that T1R3 plays an important role only in sweet taste. Nontaster mice that have a mutation in the Sac part of the T1r3 gene became tasters by the transgene insertion and were indistinguishable from control mice (Nelson et al. 2001).
Although we found an association between the TT genotype and T allele of rs307355 (TAS1R3) and the decayed and missing components, we did not observe associations with the filled component. The filled component presents a strong association with high socioeconomic status in our sample. One hypothesis for this result is that restorations, in addition to interim treatment of caries lesions, are performed in cases of dental fractures and esthetic demand. Moreover, Brazilian dentists have been considered highly interventionist, sometimes performing restorations in cases of caries that may be stable or not present (Mialhe et al. 2009; da Silva et al. 2012; Innes and Schwendicke 2017).
Our results are different from a recent SNP analysis in Chinese children (Wu et al. 2022). While no association was observed between rs307355 (TAS1R3) and rs35874116 (TAS1R2) and dental caries in deciduous teeth (Wu et al. 2022), we found an association of rs307355 (TAS1R3) in permanent teeth. Our results corroborate with a cross-sectional study of 184 Turkish children, with ages ranging from 7 to 12 y, in which an association between the heterozygous genotype (TC) of rs307355 (TAS1R3) and dental caries (dft + DMFT) was reported (Haznedaroglu et al. 2015). TC was the most prevalent genotype in the moderate-risk group in that study (i.e., dft + DMFT between 4 and 7). This trend was reproduced in our study, in which 29.7% of individuals from the high-decay trajectory group had the TC allele compared with 21.9% of individuals from the low-decay trajectory group. In adjusted logistic regression considering an additive effect, we did not find an association of genotype TC, although genotype TT was associated with a 4-fold higher chance of being in the high-decay group.
Our findings are influenced by the selection of confounding variables (DAG). As shown in DAG (Appendix Fig. 1), sugar intake is the mediator and not a confounder of the association between SNPs and dental caries: the variable sugar intake is a consequence of SNPs of taste genes (Fushan et al. 2009) and a cause of dental caries (Haznedaroglu et al. 2015). There is no biological plausibility for a direct effect between SNP and caries. Considering causal inference studies, adjusting the association by the mediator is not recommended because the mediator is on the causal path (Mefford and Witte 2012). Additional variables (such as oral hygiene, socioeconomic status, and dental service use) are not caused by SNPs in sweet taste genes. The 3 variables are not considered confounders (Mefford and Witte 2012), and adjusting for such variables can introduce bias (possibility of opening a backdoor pathway) rather than reduce it.
We are aware that our study may have some limitations. Although we found an association between SNP and caries, most studies conclude that genetics is a small component of the susceptibility to dental caries (Chisini, Cademartori, Conde, Costa, et al. 2020; Chisini, Cademartori, Conde, Tovo-Rodrigues et al. 2020). Biological (biofilm, microflora), socioeconomic, and behavioral factors are known to be the main reasons for caries occurrence (Nascimento et al. 2020). Additional limitations are related to the small number of individuals in some categories. Genotype analysis with an additive effect had a low number of individuals and should be evaluated carefully. Interpreting the magnitude of estimates, especially in this group, for the effect of rs307355 (TAS1R3) and DMF-T component trajectory requires caution. Considering this is a longitudinal study, some losses in the follow-ups were observed. The lowest κ value observed in this study was 0.65, which can be considered in the interpretation of the results.
Some strengths of our study also need to be highlighted. We use a well-characterized and large population-based birth cohort. Group-based trajectory modeling was used to identify similar groups of caries experienced in the life course. This study is the first to investigate the influence of genes with a longitudinal caries evaluation. Another strength was the robust analytical approaches used to investigate the gene–gene interaction.
In this study, we observed that the high decayed and missing component trajectory in the life course seems to be positively associated with the rs307355 (TAS1R3) genotypes but not by the rs35874116 (TAS1R2) SNP. In the additive model, the TT genotype was significant, whereas the CT/TT genotype was significant in the dominant mode. In general, the T allele was associated with higher odds of being in the high trajectory of the decayed group. Similarly, the TT genotype in the additive model and T allele were associated with higher odds of being in the high trajectory of the missing group. Epistatic interactions between rs307355 (TAS1R3) and rs35874116 (TAS1R2) may increase the risk of having a high trajectory of decayed teeth, but these need to be replicated in a larger population.
Author Contributions
L.A. Chisini, contributed to conception and design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; F.D.S. Costa, contributed to data acquisition, critically revised the manuscript; B.L. Horta, contributed to data conception, critically revised the manuscript; L. Tovo-Rodrigues, contributed to conception and design, data analysis and interpretation, critically revised the manuscript; F.F. Demarco, contributed to data conception and design, critically revised the manuscript; M.B. Correa, contributed to conception and design, data acquisition, analysis, and interpretation, critically revised the manuscript. All authors gave their final approval and agree to be accountable for all aspects of the work.
Supplemental Material
Supplemental material, sj-docx-1-jdr-10.1177_00220345221138569 for Sweet Taste Receptor Gene and Caries Trajectory in the Life Course by L.A. Chisini, F.D.S. Costa, B.L. Horta, L. Tovo-Rodrigues, F.F. Demarco and M.B. Correa in Journal of Dental Research
Footnotes
A supplemental appendix to this article is available online.
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was conducted in a graduate program supported by CAPES, Brazil. This article is based on data from the study “Pelotas Birth Cohort, 1982” conducted by the Postgraduate Program in Epidemiology at the Universidade Federal de Pelotas with the collaboration of the Brazilian Public Health Association (ABRASCO). From 2004 to 2013, the Wellcome Trust supported the 1982 birth cohort study. The International Development Research Center, World Health Organization, Overseas Development Administration, European Union, National Support Program for Centers of Excellence (PRONEX), the Brazilian National Research Council (CNPq), and the Brazilian Ministry of Health supported previous phases of the study. The Oral Health Study 2013 was supported by the National Council for Scientific and Technological Development (grants #403257/2012-3-FFD and #475979/2013-3-MBC). Some of the authors are researchers supported by CNPq (M.B. Correa, F.F. Demarco, L. Tovo-Rodrigues, B.L. Horta).
ORCID iDs: L.A. Chisini 
https://orcid.org/0000-0002-3695-0361
M.B. Correa
https://orcid.org/0000-0002-1797-3541
References
- Alexander DH, Novembre J, Lange K. 2009. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 19(9):1655–1664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Alotaibi RN, Howe BJ, Chernus JM, Mukhopadhyay N, Sanchez C, Deleyiannis FWB, Neiswanger K, Padilla C, Poletta FA, Orioli IM, et al. 2021. Genome-wide association study (GWAS) of dental caries in diverse populations. BMC Oral Health. 21(1):377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Arid J, Antunes LAA, Koch LFA, Evangelista SS, Vasconcelos KRF, Brancher JA, Gabardo MCL, Milani AJ, Dutra ALT, Antunes LS, et al. 2020. Association of taste receptor gene polymorphisms with dental caries. Braz Oral Res. 34:e055. [DOI] [PubMed] [Google Scholar]
- Ashi H, Lara-Capi C, Campus G, Klingberg G, Lingstrom P. 2017. Sweet taste perception and dental caries in 13- to 15-year-olds: a multicenter cross-sectional study. Caries Res. 51(4):443–450. [DOI] [PubMed] [Google Scholar]
- Chisini LA, Cademartori MG, Conde MCM, Costa FDS, Salvi LC, Tovo-Rodrigues L, Correa MB. 2021. Single nucleotide polymorphisms of taste genes and caries: a systematic review and meta-analysis. Acta Odontol Scand. 79(2):147–155. [DOI] [PubMed] [Google Scholar]
- Chisini LA, Cademartori MG, Conde MCM, Costa FDS, Tovo-Rodrigues L, Carvalho RV, Demarco FF, Correa MB. 2020. Genes and SNPs in the pathway of immune response and caries risk: a systematic review and meta-analysis. Biofouling. 36(9):1100–1116. [DOI] [PubMed] [Google Scholar]
- Chisini LA, Cademartori MG, Conde MCM, Tovo-Rodrigues L, Correa MB. 2020. Genes in the pathway of tooth mineral tissues and dental caries risk: a systematic review and meta-analysis. Clin Oral Investig. 24(11):3723–3738. [DOI] [PubMed] [Google Scholar]
- Choi JH, Lee J, Yang S, Kim J. 2017. Genetic variations in taste perception modify alcohol drinking behavior in Koreans. Appetite. 113:178–186. [DOI] [PubMed] [Google Scholar]
- da Silva RP, Meneghim MC, Correr AB, Pereira AC, Ambrosano GM, Mialhe EL. 2012. Variations in caries diagnoses and treatment recommendations and their impacts on the costs of oral health care. Community Dent Health. 29(1):25–28. [PubMed] [Google Scholar]
- Eriksson L, Esberg A, Haworth S, Holgerson PL, Johansson I. 2019. Allelic variation in taste genes is associated with taste and diet preferences and dental caries. Nutrients. 11(7):1491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fushan AA, Simons CT, Slack JP, Manichaikul A, Drayna D. 2009. Allelic polymorphism within the TAS1R3 promoter is associated with human taste sensitivity to sucrose. Curr Biol. 19(15):1288–1293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Haznedaroglu E, Koldemir-Gunduz M, Bakir-Coskun N, Bozkus HM, Cagatay P, Susleyici-Duman B, Mentes A. 2015. Association of sweet taste receptor gene polymorphisms with dental caries experience in school children. Caries Res. 49(3):275–281. [DOI] [PubMed] [Google Scholar]
- Horta BL, Gigante DP, Goncalves H, dos Santos Motta J, Loret de Mola C, Oliveira IO, Barros FC, Victora CG. 2015. Cohort profile update: the 1982 Pelotas (Brazil) birth cohort study. Int J Epidemiol. 44(2):441, 441a–441e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Innes NPT, Schwendicke F. 2017. Restorative thresholds for carious lesions: systematic review and meta-analysis. J Dent Res. 96(5):501–508. [DOI] [PubMed] [Google Scholar]
- Izakovicova Holla L, Borilova Linhartova P, Lucanova S, Kastovsky J, Musilova K, Bartosova M, Kukletova M, Kukla L, Dusek L. 2015. GLUT2 and TAS1R2 polymorphisms and susceptibility to dental caries. Caries Res. 49(4):417–424. [DOI] [PubMed] [Google Scholar]
- Jones B, Nagin D. 2007. Advances in group-based trajectory modeling and an SAS procedure for estimating them. Sociol Methods Res. 35(4):542–571. [Google Scholar]
- Kulkarni GV, Chng T, Eny KM, Nielsen D, Wessman C, El-Sohemy A. 2013. Association of GLUT2 and TAS1R2 genotypes with risk for dental caries. Caries Res. 47(3):219–225. [DOI] [PubMed] [Google Scholar]
- Li X. 2009. T1R receptors mediate mammalian sweet and umami taste. Am J Clin Nutr. 90(3):733S–737S. [DOI] [PubMed] [Google Scholar]
- Liao J, Schultz PG. 2003. Three sweet receptor genes are clustered in human chromosome 1. Mamm Genome. 14(5):291–301. [DOI] [PubMed] [Google Scholar]
- Mario C. 1999. Hardy-Weinberg equilibrium test and allele frequency estimation. Stata Tech Bull. 8(48):sg110. [Google Scholar]
- Mefford J, Witte JS. 2012. The covariate’s dilemma. PLoS Genet. 8(11):e1003096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mialhe FL, Pereira AC, Meneghim Mde C, Tagliaferro EP, Pardi V. 2009. Occlusal tooth surface treatment plans and their possible effects on oral health care costs. Oral Health Prev Dent. 7(3):211–216. [PubMed] [Google Scholar]
- Montmayeur JP, Liberles SD, Matsunami H, Buck LB. 2001. A candidate taste receptor gene near a sweet taste locus. Nat Neurosci. 4(5):492–498. [DOI] [PubMed] [Google Scholar]
- Nascimento GG, Seerig LM, Schuch HS, Horta BL, Peres KG, Peres MA, Correa MB, Demarco FF. 2020. Income at birth and tooth loss due to dental caries in adulthood: the 1982 Pelotas birth cohort. Oral Dis. 26(7):1494–1501. [DOI] [PubMed] [Google Scholar]
- Nelson G, Hoon MA, Chandrashekar J, Zhang Y, Ryba NJ, Zuker CS. 2001. Mammalian sweet taste receptors. Cell. 106(3):381–390. [DOI] [PubMed] [Google Scholar]
- Shi YY, He L. 2005. SHEsis, a powerful software platform for analyses of linkage disequilibrium, haplotype construction, and genetic association at polymorphism loci. Cell Res. 15(2):97–98. [DOI] [PubMed] [Google Scholar]
- Shrier I, Platt RW. 2008. Reducing bias through directed acyclic graphs. BMC Med Res Methodol. 8:70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Silva FBD, Chisini LA, Demarco FF, Horta BL, Correa MB. 2018. Desire for tooth bleaching and treatment performed in Brazilian adults: findings from a birth cohort. Braz Oral Res. 32:e12. [DOI] [PubMed] [Google Scholar]
- Vieira AR. 2021. Genes and dietary preferences and their impact on caries susceptibility and erosive tooth wear. Monogr Oral Sci. 30:92–96. [DOI] [PubMed] [Google Scholar]
- Wang M, Qin M. 2018. Lack of association between LTF gene polymorphisms and different caries status in primary dentition. Oral Dis. 24(8):1545–1553. [DOI] [PubMed] [Google Scholar]
- World Health Organization (WHO). 2013. Oral health surveys: basic methods—5th edition. Geneva (Switzerland): WHO. [accessed 2022 Oct 27]. https://www.who.int/publications/i/item/9789241548649. [Google Scholar]
- Wu L, Li Z, Zhou J, Ma B, Yu F, Zheng X, Hu X, Ma Z, Su X. 2022. An association analysis for genetic factors for dental caries susceptibility in a cohort of Chinese children. Oral Dis. 28(2):480–494. [DOI] [PubMed] [Google Scholar]
- Yi N. 2010. Statistical analysis of genetic interactions. Genet Res (Camb). 92(5–6):443–459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yoshida R, Ninomiya Y. 2016. Taste information derived from T1R-expressing taste cells in mice. Biochem J. 473(5):525–536. [DOI] [PubMed] [Google Scholar]
- Zhao GQ, Zhang Y, Hoon MA, Chandrashekar J, Erlenbach I, Ryba NJ, Zuker CS. 2003. The receptors for mammalian sweet and umami taste. Cell. 115(3):255–266. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental material, sj-docx-1-jdr-10.1177_00220345221138569 for Sweet Taste Receptor Gene and Caries Trajectory in the Life Course by L.A. Chisini, F.D.S. Costa, B.L. Horta, L. Tovo-Rodrigues, F.F. Demarco and M.B. Correa in Journal of Dental Research