Abstract
Aim
To evaluate the impact of genetic polymorphisms in uridine 5′-glucuronosylytansferases UGT1A1 and UGT1A3 and iodothyronine-deiodinases types 1 and 2 on levothyroxine (T4; 3,5,3′,5′-triiodo-L-thyronine) dose requirement for suppression of thyrotropin (TSH) secretion in patients with differentiated thyroid cancer (DTC).
Methods
Patients (n = 268) submitted to total thyroidectomy and ablation by 131I, under T4 therapy for at least 6 months were recruited in three public institutions in Brazil. Multivariate regression modelling was applied to assess the association of T4 dosing with polymorphisms in UGT1A1 (rs8175347), UGT1A3 (rs3806596 and rs1983023), DIO1 (rs11206244 and rs2235544) and DIO2 (rs225014 and rs12885300), demographic and clinical variables.
Results
A regression model including UGT1A haplotypes, age, gender, body weight and serum TSH concentration accounted for 39% of the inter-individual variation in the T4 dosage. The association of T4 dose with UGT1A haplotype is attributed to reduced UGT1A1 expression and T4 glucuronidation in liver of carriers of low expression UGT1A1 rs8175347 alleles. The DIO1 and DIO2 genotypes had no influence of T4 dosage.
Conclusion
UGT1A haplotypes associate with T4 dosage in DTC patients, but the effect accounts for only 2% of the total variability and recommendation of pre-emptive UGT1A genotyping is not warranted.
Keywords: Brazilian population, differentiated thyroid cancer, iodothyronine-deiodinases, L-thyroxine, pharmacogenomics, uridine glucuronyltransferases
What is Already Known about this Subject
Inter-individual variation in L-thyroxine (T4) doses required for suppression of thyrotropin (TSH) in patients with differentiated thyroid cancer (DTC) is considerable and multifactorial. Age, gender and body weight explain part of this variability.
To exert its inhibitory control of TSH secretion, T4 must be deiodinated to T3 (3,5,3′-triiodo-L-thyronine) by iodothyronine-deiodinases types 1 (D1) and 2 (D2). Genetic polymorphisms in D1 and D2 have been associated with circulating levels of T3 and T4 and/or TSH, with controversial results.
T4 is metabolized in human liver by UDP-glucuronyltransferases, especially UGT1A1 and UGT1A3 and readily excreted into bile as thyroxine glucuronide. A common polymorphism in UGT1A1 (-53(TA)n, rs8175347) has been associated with the T4 dose required for TSH suppression in DTC patients.
What this Study Adds
Multivariate regression modelling was applied to assess the association of T4 dosing in 268 Brazilian DTC patients with polymorphisms in DIO1 (rs11206244 and rs2235544), DIO2 (rs225014 and rs12885300), UGT1A1 (rs8175347) and UGT1A3 (rs3806596 and rs1983023), in addition to demographic and clinical variables.
A regression model including UGT1A haplotypes, age, gender, body weight and serum TSH concentration accounted for 39% of the inter-individual variation in the T4 dosage. The association of T4 dose with UGT1A haplotype is attributed to reduced UGT1A1 expression and T4 glucuronidation in liver of carriers of low expression UGT1A1 rs8175347 alleles. The DIO1 and DIO2 genotypes had no influence of T4 dosage.
UGT1A haplotypes associate with T4 dosage in DTC patients, but the effect accounts for only 2% of the total variability and recommendation of pre-emptive UGT1A genotyping is not warranted.
Introduction
Thyroid cancer is the most frequent endocrine cancer, occurring in about 5–10% of patients with a thyroid nodule. Thyroid cancers can be classified according to their histopathological characteristics, and the most common variants, namely papillary and follicular tumours, are grouped as differentiated thyroid cancer (DTC). Guidelines for the treatment of intermediate or high risk DTC recommend total thyroidectomy and ablation by 131I followed by long term levothyroxine (T4, 3,5,3′,5′-triiodo-L-thyronine) suppression of thyrotropin (TSH), as the standard of care [1,2]. To exert its inhibitory control of TSH secretion, T4 must be converted into T3 (3,5,3′-triiodo-L-thyronine) via outer (5′)-ring deionidation, catalyzed by iodothyronine-deiodinases types 1 (D1) and 2 (D2). Type 3 iodothyronine deionidase (D3), and to a lesser extent, D1 catalyze the inner (5)-ring deiodination of T3 and T4 into reverse T3 (rT3) and T2, respectively, thus inactivating the thyroid hormone action (Figure 1). D1 is mainly involved in serum T3 production whereas D2 controls the local conversion of T4 into T3 in the hypothalamus and pituitary, and thereby plays a pivotal role in the negative feedback regulation of TSH secretion [3–5]. In humans, D1 and D2 are encoded by the DIO1 and DIO2 genes, which map to chromosomes 1p33-p32 and 14q24.2, respectively.
Figure 1.
Pathways of thyroxine (T4) metabolism. T3, (3,5,3′-triiodo-L-thyronine); rT3, reverse T3; UGTs, UDP-glucuronyltransferases; D1, D2 and D3, iodothyronine deiodinases 1, 2 and 3, respectively
In addition to deiodination, T4 is metabolized in human liver by UDP-glucuronyltransferases (UGTs) and readily excreted into bile as thyroxine glucuronide. Kinetic experiments using recombinant UGT isoforms and liver microsomes revealed that T4 glucuronidation in humans is mediated by UGT1A subfamily enzymes, especially UGT1A1 and UGT1A3 [6–10]. The respective encoding genes, UGT1A1 and UGT1A3, map to chromosome 2q37 and harbour several functional polymorphisms. We reported recently that a common polymorphism in UGT1A1, consisting of a variable dinucleotide repeat within the promoter TATA element (UGT1A1 −53(TA)n, rs8175347) was associated with the T4 dose required for TSH suppression in a cohort of Brazilian DTC patients [11]. A significant trend for decreasing T4 dose with increasing number of copies of the low expression (TA)7 and (TA)8 alleles was observed, and ascribed to reduced T4 glucuronidation in patients harbouring these variant alleles [12]. However, the relatively small difference in T4 doses across the UGT1A1 −53(TA)n groups and considerable overlap of doses among groups, prompted us to suggest that the association between T4 dose and UGT1A1 −53(TA)n genotype should be tested in a larger cohort and extended to polymorphisms in other UGT isoforms involved in T4 glucuronidation in human liver. Accordingly, we recruited 170 additional DTC patients from a different institution and extended our analyses to two common polymorphisms in UGT1A3 (rs3806596, −66T>C and rs1983023, −751T>C), which are major determinants of UGT1A3 haplotypes that impact on the expression of UGT1A3 and the pharmacokinetics and pharmacodynamics of the UGT1A3 preferential substrate, atorvastatin [13,14]. We also examined the influence of polymorphisms in DIO1 (rs11206244, 785C>T, previously D1a-C/T; rs2235544, −34C>A) and DIO2 (rs225014, 274A>G, previously Thr92Ala; rs12885300, −451C>T, previously ORFa-Gly3Asp), which have been associated with circulating concentrations of T3 and T4 and/or TSH, with controversial results [15–20].
Methods
Study cohort and experimental protocol
DTC patients (n = 268) who had been submitted to total thyroidectomy and ablation by 131I and were under T4 therapy for at least 6 months were recruited in three institutions of the Brazilian Public Health System. Two of these are located in Rio de Janeiro (Instituto Nacional de Câncer and the Instituto Estadual de Diabetes e Endocrinologia Luiz Capriglione, n = 98 patients) and the third in São Paulo (A C Camargo Cancer Center, n = 170). The patients from Rio de Janeiro were enrolled in our previous study of the association between the UGT1A1-53(TA)n polymorphism and T4 dose [11]. The experimental protocol of the current study was approved by the Ethics Committees of the three institutions, patients signed an informed consent form to participate and self-identified with one of the official ‘race/colour’ categories adopted by the Brazilian Census, namely branco (meaning White, n = 124), pardo (Brown, n = 123), preto (Black, n = 4), amarelo (Yellow, referring to Asian ancestry, n = 7) and undefined (n = 10). These colour categories are capitalized to highlight their meaning in the context of the Brazil Census. Because of their small number, Black patients were combined with Brown patients for statistical analysis. Exclusion criteria included poor compliance with T4 therapy (assessed by counting the T4 tablets, which were provided free of cost to patients), clinically relevant liver or kidney disease, alcoholism or co-medication with UGT1A1 inducers (anticonvulsants, barbiturates and rifampicin), or any drug that could interfere with TSH secretion (amiodarone, high doses of glucocorticoids, octreotide or bexaterotene) and/or affect T4 gastrointestinal absorption (bile acid binding resins, proton pump inhibitors or calcium carbonate). T4 was provided as Puran T4® (Sanofi-Aventis Ltda, Rio de Janeiro, Brazil) or Synthroid® (Abbott Laboratórios do Brasil Ltda, São Paulo, Brazil), in 25 to 200 μg tablets. Daily doses were adjusted individually to the required level of TSH suppression, determined by risk stratification, according to Tuttle et al. [21]. For high and intermediate risk patients, serum TSH was kept <0.1 mU l−1, whereas the range of 0.1−0.5 mU l−1 was considered appropriate for low risk patients [1,2]. Free T4 and TSH concentrations in serum were measured by chemoluminescence (ADVIA® Centaur® XP Immunoassay System, Siemens Healthcare Diagnostics, Dublin, Ireland).
Genotyping
Standard protocols were used for extraction of genomic DNA from peripheral lymphocytes. The UGT1A1-53(TA)n polymorphism (rs8175347) was assessed by direct sequencing, using the ABI PRISM® 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA), as described previously [11]. Taqman assays were used for allele discrimination of the DIO1 rs2235544 (probe C_15952583_10), DIO1 rs11206244 (C_334342_20), DIO2 rs225014 (C_15819951_10), DIO2 rs12885300 (C_31755153_30), UGT1A3 rs1983023 (C_11473200_10) and UGT1A3 rs3806596 (custom probe AH89MI5) SNPs, using the Fast 7500 Real-Time System (Applied Biosystems, Foster City, CA, USA). Allele and genotype frequency were derived by gene counting. Deviations from Hardy−Weinberg equilibrium were assessed by the goodness-of-fit χ2 test. The χ2 test was used for comparison of allele and genotype frequencies. A P value ≤0.05 was considered significant.
Linkage disequilibrium and haplotype analyses
Pairwise linkage disequilibrium (LD) analyses of polymorphisms in UGT1A, DIO1 and DIO2 genes were conducted online using the CUBEX software [22]. The level of LD between loci was assessed using the coefficients D' and r2. UGT1A haplotypes, comprising UGT1A1 rs8175347, UGT1A3 rs3806596 and UGT1A3 rs1983023, were statistically inferred using the haplo-stats software, version 1.3 [23]. This software attributes a posterior probability value for the diplotype configuration for each individual based on estimated haplotype frequencies. The minimal posterior probability value for inclusion of an individual in the present analyses was set at 0.91, which allowed the inclusion of 253 individuals (94.4% of the overall cohort). Haplotypes are labelled by numbers preceded by ‘Br’ (e.g. HapBr1), a reference to this Brazilian cohort.
Multivariate regression model
To select candidate variables for multiple regression modelling of T4 dosing, we initially examined the independent effects of age, gender, colour, weight, serum T4 and TSH levels, DIO1 and DIO2 genotypes and number of copies (zero, one or two) of UGT1A wild-type haplotype, designated HapBr1. Genetic variables were analyzed according to either codominant or recessive models. In the latter, carriers of one or two copies of the wild-type allele at the DIO1 and DIO2 SNPs and carriers of one or two copies of the UGT1A HapBr1 were contrasted with the respective homozygotes for the variant DIO alleles and non-carriers of UGT1A HapBr1. Discrete covariates entered the analysis transformed as dummy variables to quantify how the individual categories affected T4 daily dose. The effects of discrete variables were described by electing a baseline category and comparing the transformed response predicted at this category with the remaining category or categories of interest. Variables that associated (P < 0.15) with T4 dose in the univariate analyses were considered further in multivariate regression models. Variables that associated (P < 0.05) with T4 dose in successive multivariate analyses were included in the final model. We assessed the goodness of fit of our models based on the statistics r2 and r2 adjusted. The former is a measure of how well the regression line approximates the observed data points. It can also be interpreted as the proportion of the total variance in the observed data that is explained by the model. Since r2 always increases with the number of variables in the model, this measure can be accounted for by adjusting for the number of covariates in the model, yielding the statistics known as the adjusted r2. The relative contribution of each variable in the final regression was estimated by the partial r2 statistics, which measures the degree of association between two random variables, with the effect of a set of controlling random variables removed.
Results
Table 1 summarizes the demographics of the 268 subjects included in the study. Women accounted for 82.5% of the cohort, median age was 46 (interquartile range 37–56) years and most patients (92%) self-identified as White (46.2%) or Brown (45.9%). All self-reported Yellow patients referred to Japanese ancestry. The distribution of the White, Brown plus Black and Yellow patients in the study cohort did not differ significantly (P = 0.12, χ2 test) from the corresponding proportions in the present day population of Southeast Brazil [24].
Table 1.
Demographic and clinical characteristics of Brazilian patients with DTC
| Parameter | Study cohort (n = 268) |
|---|---|
| Gender* | |
| Female | 221 (82.5) |
| Male | 47 (17.5) |
| Self-reported ‘colour/race’† | |
| White | 124 (46.3) |
| Brown | 123 (45.9) |
| Black | 4 (1.5) |
| Yellow | 7 (2.6) |
| Undetermined | 10 (3.7) |
| Age (years)* | 46 (37–56) |
| Weight (kg) | 68 (61–80) |
| Serum TSH (mU l−1) | 0.06 (0.04–0.20) |
| Free T4 (ng ml−1) | 1.50 (1.30–1.69) |
| T4 daily dose (g) | 138 (112–150) |
Data for categorical variables presented as n (%); data for continuous variables presented as median (inter-quartile range).
According to the categories of the Brazilian Census (http://www.ibge.gov.br).
Allele frequency, genotype distribution, linkage disequilibrium and UGT1A haplotypes
Table 2 shows the allele frequency and genotype distribution of the UGT1A1, UGT1A3, DIO1 and DIO2 polymorphisms in the overall cohort. Deviations from Hardy−Weinberg expectations were not observed (P = 0.33−0.79 for the different loci). The three polymorphisms in the UGT1A genes were in strong LD, with pairwise D' values >0.97 and r2s of 0.56 (rs8175347*28 and rs3806596 allele T), 0.72 (rs8175347*28 and rs1983023 allele T) and 0.77 (rs3806596 allele T and rs1983023 allele T). The previously reported [17] extensive LD between rs1120622T and rs2235544A in DIO1 was confirmed in the study cohort (D' = 1.0, r2 = 0.53), whereas no LD (r2 = 0.18) was detected between the rs12885300 and rs225014 SNPs in DIO2. Diplotype analysis of UGT1A loci generated eight haplotypes, of which four had frequencies greater than 1% and together accounted for 98.6% of the overall genetic diversity (Table 3). Haplotypes comprising either the wild-type alleles (HapBr1, frequency 0.516) or the variant alleles at the three loci (HapBr2, 0.350) represented nearly 90% of all haplotypes. The number of patients carrying two, one or no copies of the wildtype HapBr1 haplotype were 64, 134 and 55, respectively.
Table 2.
Frequency distribution of UGT1A1, UGT1A3, DIO1 and DIO2 polymorphisms in Brazilian patients with DTC
| UGT1A1 rs8175347 | UGT1A3 rs3806596 | UGT1A3 rs1983023 | DIO1 rs11206244 C > T | DIO1 rs2235544 C > A | DIO2 rs225014 T > C | DIO2 rs12885300 C > T | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Genotype | Frequency (95% CI) | Genotype | Frequency (95% CI) | Genotype | Frequency (95% CI) | Genotype | Frequency (95% CI) | Genotype | Frequency (95% CI) | Genotype | Frequency (95% CI) | Genotype | Frequency (95% CI) |
| 6/6 | 0.39 (0.33, 0.45) | TT | 0.24 (0.19, 0.30) | TT | 0.32 (0.26, 0.38) | CC | 0.50 (0.44, 0.56) | CC | 0.31 (0.26, 0.37) | AA | 0.35 (0.29, 0.41) | CC | 0.50 (0.44, 0.56) |
| 5/6 | 0.02 (0.01, 0.05) | CT | 0.53 (0.47, 0.59) | CT | 0.51 (0.44, 0.57) | CT | 0.41 (0.35, 0.47) | AC | 0.48 (0.42, 0.54) | AG | 0.53 (0.46, 0.59) | CT | 0.47 (0.41, 0.53) |
| 6/7 | 0.47 (0.41, 0.53) | CC | 0.23 (0.18, 0.28) | CC | 0.17 (0.14, 0.23) | TT | 0.09 (0.06, 0.13) | AA | 0.21 (0.16, 0.26) | GG | 0.12 (0.09, 0.18) | TT | 0.03 (0.02, 0.06) |
| 7/7 | 0.12 (0.08, 0.16) | ||||||||||||
| 7/8 | 0.01 (0.00, 0.03) | ||||||||||||
| Allele | Frequency (95% CI) | Allele | Frequency (95% CI) | Allele | Frequency (95% CI) | Allele | Frequency (95% CI) | Allele | Frequency (95% CI) | Allele | Frequency (95% CI) | Allele | Frequency (95% CI) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 6 (*1) | 0.63 (0.59, 0.67) | T | 0.51 (0.46, 0.55) | T | 0.57 (0.53, 0.61) | C | 0.70 (0.66, 0.74) | C | 0.55 (0.51, 0.60) | A | 0.61 (0.56, 0.65) | C | 0.73 (0.69, 0.77) |
| 5 (*36) | 0.01 (0.00, 0.02) | C | 0.49 (0.45, 0.54) | C | 0.43 (0.39, 0.47) | T | 0.30 (0.26, 0.34) | A | 0.45 (0.40, 0.49) | G | 0.39 (0.35, 0.44) | T | 0.27 (0.23, 0.31) |
| 7 (*28) | 0.35 (0.31, 0.40) | ||||||||||||
| 8 (*37) | 0.004 (0.00–0.01) |
Table 3.
Inferred UGT1A haplotypes
| Haplotype | SNPs | n | Frequency (95% CI) | ||
|---|---|---|---|---|---|
| rs817534 (TA)n | rs3806590 | rs1983023 | |||
| HapBr1 | 6 | T | T | 262 | 0.518 (0.474, 0.561) |
| HapBr2 | 7 | C | C | 177 | 0.350 (0.309, 0.392) |
| HapBr3 | 6 | C | C | 31 | 0.061 (0.044, 0.087) |
| HapBr4 | 6 | C | T | 29 | 0.057 (0.040, 0.081) |
| HapBr5 | 5 | C | C | 4 | 0.008 (0.003, 0.020) |
| HapBr6 | 5 | C | T | 1 | 0.002 (0.0005, 0.01) |
| HapBr7 | 6 | T | C | 1 | 0.002 (0.0005, 0.01) |
| HapBr8 | 7 | T | T | 1 | 0.002 (0.0005, 0.01) |
(TA)n, number of (TATA) repeats. n, number of haplotypes.
Allele frequency data according to self-reported colour are presented in Supplementary Table 1. The only significant deviation from Hardy−Weinberg expectations (P = 0.03) occurred in DIO2 rs12885300 in the Brown and Black group, which showed an excess of heterozygous patients. No significant differences were detected in allele frequency between White vs. Brown plus Black patients. The small number (n = 8) of patients of Japanese descent precluded statistical analyses of their data, but we will mention that the variant UGT1A3 rs1983023T allele was not detected in this group, in contrast with its high frequency in White (0.54) and Brown plus Black patients (0.58).
Regression modelling of T4 dosing
The initial univariate analyses disclosed gender, age, weight, serum T4, serum TSH and the number of copies of the UGTA1 HapBr1 haplotype as being associated (P < 0.15) with T4 dose (Table 4). No association was observed between T4 dose and DIO1 or DIO2 SNPs. In the codominant genetic model, association of T4 dose with UGTA1 was restricted to carriers of zero copies of the HapBr1 haplotype, which is consistent with the results observed under the recessive model (Table 4). Accordingly, the UGT1A haplotype was entered as a recessive trait into multivariate regression modelling, alongside gender, age, weight, serum T4 and serum TSH. The covariates associated (P < 0.05) with T4 daily dose in the final multivariate regression model are listed in Table 4 with their respective regression coefficients and contribution to the model, as measured by the partial r2 statistics. The T4 daily dose requirement increased with increasing weight, decreased with ageing, was inversely associated with the serum TSH concentration, was greater in men than women and was lower in non-carriers of the UGT1A HapBr1 haplotype. The partial r2 statistics showed that weight had the largest influence on T4 dose requirement (r2 = 0.134), followed by age (0.068), gender (0.037), the UGT1A HapBr1 (0.020) and serum TSH (0.013). Based on the adjusted r2 value, the final regression model explained 38.8% of the overall variability in T4 daily dose required for TSH suppression in the study cohort.
Table 4.
Regression modelling of T4 doses in Brazilian patients with DTC
| Co-variates | Univariate regression* | Multivariate regression | |||
|---|---|---|---|---|---|
| Codominant model P value | Recessive model P value | Partial regression coefficient | P value | Partial r2 | |
| Gender† | <0.0001 | 19.78 | <0.0001 | 0.037 | |
| Race/Colour† | |||||
| Brown + Black | 0.39 | ||||
| Yellow | 0.84 | ||||
| Undefined | 0.22 | ||||
| Age | <0.0001 | −0.65 | <0.0001 | 0.067 | |
| Weight | <0.0001 | 0.92 | <0.0001 | 0.134 | |
| Serum T4 | 0.147 | ||||
| Serum TSH | 0.002 | −30.38 | 0.019 | 0.013 | |
| Copies of UGT1A Hap Br1 haplotype | 0.25, 0.01 | 0.01 | −11.83 | 0.003 | 0.020 |
| DIO1 rs2235544 | 0.44, 0.56 | 0.28 | |||
| DIO1 rs11206244 | 0.97, 0.83 | 0.81 | |||
| DIO2 rs225014 | 0.17, 0.92 | 0.49 | |||
| DIO2 rs12885300 | 0.65, 0.49 | 0.53 | |||
| r2 = 0.399 | |||||
| r2 adjusted = 0.388 | |||||
In the codominant model, P values correspond to heterozygous and variant homozygous genotypes for the DIO SNPs and to one or zero copies of UGT1A Hap Br1. In the recessive model, P values correspond to variant homozygous genotypes for the DIO SNPs and to zero copies of UGT1A Hap Br1. Bold P values indicate variables that were entered into multiple regression modelling.
Baseline categories were female gender and White colour.
Prompted by the results reported by Tortolano et al. [19] we examined the association between DIO2 rs225014 and T4 dosage in DTC patients with TSH concentrations in the range of 0.1−0.5 mU l−1, using the recessive genetic model, and obtained a P value 0.63 in the univariate analysis.
Discussion
This is, to our knowledge, the first study of the combined influence of genetic polymorphisms affecting the two principal metabolic pathways of T4, namely glucuronidation and deiodination, on the T4 dose required for TSH suppression in DTC patients. Our study encompassed UGT1A1 and UGT1A3, which play a dominant role in T4 glucuronidation in liver [6–10] and D1 and D2, which play a vital role in the conversion of the prohormone T4, into T3, the active thyroid hormone [3–5]. A two-step (univariate followed by multivariate) regression modelling including polymorphisms in the respective encoding genes, in addition to demographic and clinical covariates was applied to data from a patient cohort recruited in public institutions from the two most populated cities in Brazil. The final regression model, which included weight, age, gender, serum TSH concentration and zero copies of the UGT1A HapBr1 (wild-type) haplotype accounted for 39% of the inter-individual variation in the T4 dose required for TSH suppression. The inclusion of weight, age and gender in the model is consistent with long standing knowledge that these variables are major determinants of T4 dosage in the treatment of thyroid diseases [25–27]. Accordingly, the T4 dose in DTC patients decreased with advancing age, increased with increasing body weight and was greater in men than women. The retention of gender as an independent covariate in the final model suggests that gender-related differences in body weight cannot entirely account for the higher T4 dose requirement in males, but does not exclude the possible influence of differences in body composition, among others, between males and females [28]. The inverse correlation of serum TSH concentration and T4 dosage is consistent with the therapeutic goal of T4-induced TSH suppression in DTC patients.
Among the genetic polymorphisms investigated, only the inferred UGT1A1 haplotypes comprising UGT1A1 rs8175347, UGT1A3 rs3806596 and UGT1A3 rs1983023 were included in the final multivariate model for T4 dose requirement. These SNPs were in tight LD with pairwise D' = 1 and r2 coefficients ranging from 0.56 to 0.77 and two haplotypes, including either the wildtype (HapBr1) or the variant alleles (HapBr2) at the three loci, accounted for 90% of the genetic diversity in the cohort. The multivariate regression modelling showed that non-carriers of UGT1A HapBr1 require significantly lower T4 doses compared with carriers of one or two copies of this haplotype, but no difference between the latter two groups of patients. Non-carriers of the UGT1A HapBr1 harbour the (TA)7 and/or (TA)8 alleles at UGT1A1 rs8175347, and their lower T4 dose requirement is consistent with our previous observation of a trend for decreasing T4 dose with increasing number of copies of these low expression UGT1A1 alleles [11]. Reduced T4 glucuronidation in the liver of patients with low expression UGT1A1 alleles was thought to account for our previous findings and is also likely to account for the present results. However, the most common low expression UGT1A1 −53(TA)7 allele (UGT1A1*28) was tightly linked to the UGT1A3 rs3806596C and rs1983023C variant alleles in non-carriers of the HapBr1 haplotype. These two variant alleles are major markers of the UGT1A3*2 haplotype, which has been linked to higher expression of UGT1A3 and shown to impact on the pharmacokinetics and pharmacodynamics of the UGT1A3 preferential substrate, atorvastatin [13,14]. Although inference of the UGT1A3*2 haplotype requires identification of alleles at other loci, which were not genotyped in our patients, it is reasonable to speculate that carriers of UGT1A3 CC diplotype (HapBr2) would have substantially increased expression of UGT1A3 mRNA and protein, and greater UGT1A3 catalytic activity in liver microsomes, compared with carriers of the TT diplotypes (HapBr1).
Because of the opposite effects of UGT1A1*28 [(TA)7] and admittedly the linked UGT1A3 rs3806596C and rs1983023C alleles on the expression and activity of UGT1A1 and UGT1A3 in human liver, respectively [13,14], we suggest that the extent of UGT1A-mediated glucuronidation will be substrate-dependent, according to the relative contribution of each isoform. The reduced T4 dose requirement for TSH suppression in non-carriers of UGT1A HapBr1 in the present study, suggests a predominant role of UGT1A1 in T4 glucuronidation in human liver, which is consistent with in vitro data [6–10]. UGT1A1 is also the major isoform in the glucuronidation of SN-38, the active metabolite of the chemotherapeutic agent, irinotecan, and UGT1A1 polymorphisms have been shown to associate with SN-38 disposition and toxicity [29]. In contrast to thyroxine and irinotecan, UGT1A3 plays a predominant role in the conversion of atorvastatin into the pharmacologically inactive lactone metabolite, which is enhanced in carriers of the linked UGT1A3*2 and UGT1A1*28 alleles, leading to reduced lipid-lowering effects of atorvastatin [14,15]. UGT-mediated glucuronidation is a major drug metabolizing reaction in humans, and the relative contribution of different enzyme isoforms varies considerably among substrates. Thus, extrapolation of the findings of the present study regarding thyroxine disposition and dosage to other UGT1A substrates, is unwarranted.
The T4 dosage in the Brazilian DCT patients showed no significant association with polymorphisms in genes encoding the D1 and D2 enzymes, namely rs11206244 and rs2235544 in DIO1 and rs225014 and rs12885300 in DIO2. D1 and D2 are predominantly activating enzymes, both converting the pro-hormone T4 into T3 by outer (5′)-ring deiodination. D1, however, has a remarkable preference for reverse T3 (rT3) as substrate, which argues for its role as a scavenger of deiodinate inactive iodothyronines (reviewed in [3]). The tightly linked rs11206244T and rs2235544A alleles have been associated with lower enzymatic activity, relative to their ancestral counterparts, rs11206244C and rs2235544C, respectively [18,30]. Accordingly, carriers of the variant rs11206244T and rs2235544A alleles show reduced serum concentrations of free T3 and higher concentrations of free T4 and free rT3, resulting in lower T3 : T4 and T3 : rT3 ratios and a higher rT3 : T4 ratio, but no effect no serum TSH concentration [18,31]. Furthermore, a study in DTC patients under chronic T4 treatment detected no influence of DIO1 rs11206244 on the set point of the hypothalamus-pituitary-thyroid axis [19]. The latter observation is consistent with the lack of association between this polymorphism (and the linked rs2235544 SNP) and T4 dosage in our study cohort. Of note, this result was verified using either the codominant or the recessive genetic model for the regression analyses.
We observed no association between DIO2 rs225014 or rs12885300 polymorphisms on T4 dose requirement for TSH suppression in DTC patients, whether adopting codominant or recessive genetic models. In the case of rs225014, this is consistent with previous results in patients with DTC or Hashimoto thyroiditis [18–20]. By contrast, Torlontano et al. [19]. reported that DTC patients, homozygous for rs225014C, needed a higher T4 dose compared with carriers of the wild-type rs225014T allele. The latter observation applied only to patients with serum TSH concentrations in the range of 0.1−0.5 mU l−1 (‘near-suppressed’ group) but not in the ‘suppressed group’ (TSH <0.1 mU l−1). However, regression analyses of the near-suppressed group in our cohort, using a recessive genetic model, did not verify these findings. Possible reasons for the discordant results of Torlonano et al. [19] were discussed by Heemstra et al. [20], and the balance of evidence is consistent with the lack of significant association between DIO2 rs225014 and T4 dosage in DTC patients.
The other DIO2 SNP examined in our study, rs12885300C>T, has shown inconsistent effects in previous studies in vitro [32,33] and in vivo [1,17,18]. In patients on T4 treatment, Panicker et al. [17]. observed no correlation between this SNP and circulating concentrations of thyroid hormones, whereas a weaker negative feedback of free T4 on TSH in DTC patients homozygous for the variant rs12885300T, compared with carriers of the wild-type allele, was reported by Hoftijzer et al. [18]. These authors, however, were uncertain as to the clinical significance of this finding. Accordingly, in our DTC patients, no association between the rs12885300C>T genotype and T4 dosage was observed.
In conclusion, we investigated the association of clinical, demographical and genetic variables with T4 dose required for TSH suppression in Brazilian patients with DTC. Multivariate regression analyses confirmed significant associations of T4 dosage with weight, age, gender, serum TSH concentration and linked polymorphisms in UGT1A3 and UGT1A1 genes. A regression model comprising these covariates explained 39% on the inter-individual dose variability in the study cohort. Polymorphisms in DIO1 and DIO2 had no effect on T4 dose requirement for TSH suppression.
Competing Interests
There are no competing interests to declare.
This study was supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Financiadora de Estudos e Projetos (FINEP), Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ), and Instituto Nacional de Ciência e Tecnologia em Oncogenomica (INCITO).
Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web-site:
Distribution of UGT1A1, UGT1A3, DIO1 and DIO2 polymorphisms in Brazilian DTC patients, according to self-reported ‘colour/race’
References
- 1.American Thyroid Association (ATA) Guidelines Taskforce on Thyroid Nodules and Differentiated Thyroid Cancer. Cooper DS, Doherty GM, Haugen BR, Kloos RT, Lee SL, Mandel SJ, Mazzaferri EL, McIver B, Pacini F, Schlumberger M, Sherman SI, Steward DL, Tuttle RM. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid. 2009;19:1167–1214. doi: 10.1089/thy.2009.0110. [DOI] [PubMed] [Google Scholar]
- 2.Pacini F, Castagna MG, Brilli L, Pentheroudakis G ESMO Guidelines Working Group. Thyroid cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2012;23(Suppl. 7):vii110–119. doi: 10.1093/annonc/mds230. [DOI] [PubMed] [Google Scholar]
- 3.Panicker V. Genetics of thyroid function and disease. Clin Biochem Rev. 2011;32:165–175. [PMC free article] [PubMed] [Google Scholar]
- 4.Maia AL, Goemann IM, Meyer EL, Wajner SM. Deiodinases: the balance of thyroid hormone: type 1 iodothyronine deiodinase in human physiology and disease. J Endocrinol. 2011;209:283–297. doi: 10.1530/JOE-10-0481. [DOI] [PubMed] [Google Scholar]
- 5.Arrojo E, Drigo R, Fonseca TL, Werneck-de-Castro JP, Bianco AC. Role of the type 2 iodothyronine deiodinase (D2) in the control of thyroid hormone signaling. Biochim Biophys Acta. 2013;1830:3956–3964. doi: 10.1016/j.bbagen.2012.08.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Findlay KA, Kaptein E, Visser TJ, Burchell B. Characterization of the uridine diphosphate-glucuronosyltransferase-catalyzing thyroid hormone glucuronidation in man. J Clin Endocrinol Metab. 2000;85:2879–2883. doi: 10.1210/jcem.85.8.6715. [DOI] [PubMed] [Google Scholar]
- 7.Yoder Graber AL, Ramírez J, Innocenti F, Ratain MJ. UGT1A1*28 genotype affects the in-vitro glucuronidation of thyroxine in human livers. Pharmacogenet Genomics. 2007;17:619–627. doi: 10.1097/FPC.0b013e3280121fe9. [DOI] [PubMed] [Google Scholar]
- 8.Tong Z, Li H, Goljer I, McConnell O, Chandrasekaran A. In vitro glucuronidation of thyroxine and triiodothyronine by liver microsomes and recombinant human UDP-glucuronosyltransferases. Drug Metab Dispos. 2007;35:2203–2210. doi: 10.1124/dmd.107.016972. [DOI] [PubMed] [Google Scholar]
- 9.Yamanaka H, Nakajima M, Katoh M, Yokoi T. Glucuronidation of thyroxine in human liver, jejunum, and kidney microsomes. Drug Metab Dispos. 2007;35:1642–1648. doi: 10.1124/dmd.107.016097. [DOI] [PubMed] [Google Scholar]
- 10.Kato Y, Ikushiro S, Emi Y, Tamaki S, Suzuki H, Sakaki T, Yamada S, Degawa M. Hepatic UDP-glucuronosyltransferases responsible for glucuronidation of thyroxine in humans. Drug Metab Dispos. 2008;36:51–55. doi: 10.1124/dmd.107.018184. [DOI] [PubMed] [Google Scholar]
- 11.Vargens DD, Neves RR, Bulzico DA, Ojopi EB, Meirelles RM, Pessoa CN, Prado CM, Gonçalves PA, Leal VL, Suarez-Kurtz G. Association of the UGT1A1-53(TA)n polymorphism with L-thyroxine doses required for thyrotropin suppression in patients with differentiated thyroid cancer. Pharmacogenet Genomics. 2011;21:341–343. doi: 10.1097/FPC.0b013e3283448d19. [DOI] [PubMed] [Google Scholar]
- 12.Beutler E, Gelbart T, Demina A. Racial variability in the UDP-glucuronosyltransferase 1 (UGT1A1) promoter: a balanced polymorphism for regulation of bilirubin metabolism? Proc Natl Acad Sci U S A. 1998;95:8170–8174. doi: 10.1073/pnas.95.14.8170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Riedmaier S, Klein K, Hofmann U, Keskitalo JE, Neuvonen PJ, Schwab M, Niemi M, Zanger UM. UDP-glucuronosyltransferase (UGT) polymorphisms affect atorvastatin lactonization in vitro and in vivo. Clin Pharmacol Ther. 2010;87:65–73. doi: 10.1038/clpt.2009.181. [DOI] [PubMed] [Google Scholar]
- 14.Cho SK, Oh ES, Park K, Park MS, Chung JY. The UGT1A3*2 polymorphism affects atorvastatin lactonization and lipid-lowering effect in healthy volunteers. Pharmacogenet Genomics. 2012;22:598–605. doi: 10.1097/FPC.0b013e3283544085. [DOI] [PubMed] [Google Scholar]
- 15.Peeters RP, van den Beld AW, Attalki H, Toor H, de Rijke YB, Kuiper GG, Lamberts SW, Janssen JA, Uitterlinden AG, Visser TJ. A new polymorphism in the type II deiodinase gene is associated with circulating thyroid hormone parameters. Am J Physiol Endocrinol Metab. 2005;289:E75–81. doi: 10.1152/ajpendo.00571.2004. [DOI] [PubMed] [Google Scholar]
- 16.Mentuccia D, Proietti-Pannunzi L, Tanner K, Bacci V, Pollin TI, Poehlman ET, Shuldiner AR, Celi FS. Association between a novel variant of the human type 2 deiodinase gene Thr92Ala and insulin resistance: evidence of interaction with the Trp64Arg variant of the beta-3-adrenergic receptor. Diabetes. 2002;51:880–883. doi: 10.2337/diabetes.51.3.880. [DOI] [PubMed] [Google Scholar]
- 17.Panicker V, Cluett C, Shields B, Murray A, Parnell KS, Perry JR, Weedon MN, Singleton A, Hernandez D, Evans J, Durant C, Ferrucci L, Melzer D, Saravanan P, Visser TJ, Ceresini G, Hattersley AT, Vaidya B, Dayan CM, Frayling TM. A common variation in deiodinase 1 gene DIO1 is associated with the relative levels of free thyroxine and triiodothyronine. J Clin Endocrinol Metab. 2008;93:3075–3081. doi: 10.1210/jc.2008-0397. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hoftijzer HC, Heemstra KA, Visser TJ, le Cessie S, Peeters RP, Corssmit EP, Smit JW. The type 2 deiodinase ORFa-Gly3Asp polymorphism (rs12885300) influences the set point of the hypothalamus-pituitary-thyroid axis in patients treated for differentiated thyroid carcinoma. J Clin Endocrinol Metab. 2011;96:E1527–1533. doi: 10.1210/jc.2011-0235. [DOI] [PubMed] [Google Scholar]
- 19.Torlontano M, Durante C, Torrente I, Crocetti U, Augello G, Ronga G, Montesano T, Travascio L, Verrienti A, Bruno R, Santini S, D'Arcangelo P, Dallapiccola B, Filetti S, Trischitta V. Type 2 deiodinase polymorphism (threonine 92 alanine) predicts L-thyroxine dose to achieve target thyrotropin levels in thyroidectomized patients. J Clin Endocrinol Metab. 2008;93:910–913. doi: 10.1210/jc.2007-1067. [DOI] [PubMed] [Google Scholar]
- 20.Heemstra KA, Hoftijzer HC, van der Deure WM, Peeters RP, Fliers E, Appelhof BC, Wiersinga WM, Corssmit EP, Visser TJ, Smit JW. Thr92Ala polymorphism in the type 2 deiodinase is not associated with T4 dose in athyroid patients or patients with Hashimoto thyroiditis. Clin Endocrinol (Oxf) 2009;71:279–283. doi: 10.1111/j.1365-2265.2008.03474.x. [DOI] [PubMed] [Google Scholar]
- 21.Tuttle RM, Leboeuf R, Shaha AR. Medical management of thyroid cancer: a risk adapted approach. J Surg Oncol. 2008;97:712–716. doi: 10.1002/jso.21010. [DOI] [PubMed] [Google Scholar]
- 22. CubeX: Cubic Exact Solution. Available at http://www.oege.org/software/cubex/ (last accessed 10 February 2014)
- 23. Haplo-stats softaware. Available at http://www.mayo.edu/research/labs/statistical-genetics-genetic-epidemiology/software (last accessed 10 February 2014)
- 24. Instituto Brasileiro de Geografia e Estatística. Available at http://www.sidra.ibge.gov.br/bda/tabela/listabl.asp?z=t&c=262 (last accessed 10 February 2014)
- 25.Rosenbaum RL, Barzel US. Levothyroxine replacement dose for primary hypothyroidism decreases with age. Ann Intern Med. 1982;96:53–55. doi: 10.7326/0003-4819-96-1-53. [DOI] [PubMed] [Google Scholar]
- 26.Cunningham JJ, Barzel US. Lean body mass is a predictor of the daily requirement for thyroid hormone in older men and women. J Am Geriatr Soc. 1984;32:204–207. doi: 10.1111/j.1532-5415.1984.tb02003.x. [DOI] [PubMed] [Google Scholar]
- 27.Sawin CT, Herman T, Molitch ME, London MH, Kramer SM. Aging and the thyroid. Decreased requirement for thyroid hormone in older hypothyroid patients. Am J Med. 1983;75:206–209. doi: 10.1016/0002-9343(83)91192-0. [DOI] [PubMed] [Google Scholar]
- 28.Santini F, Pinchera A, Marsili A, Ceccarini G, Castagna MG, Valeriano R, Giannetti M, Taddei D, Centoni R, Scartabelli G, Rago T, Mammoli C, Elisei R, Vitti P. Lean body mass is a major determinant of levothyroxine dosage in the treatment of thyroid diseases. J Clin Endocrinol Metab. 2005;90:124–127. doi: 10.1210/jc.2004-1306. [DOI] [PubMed] [Google Scholar]
- 29.Innocenti F, Ratain MJ. Pharmacogenetics of irinotecan: clinical perspectives on the utility of genotyping. Pharmacogenomics. 2006;7:1211–1221. doi: 10.2217/14622416.7.8.1211. [DOI] [PubMed] [Google Scholar]
- 30.Peeters RP, van Toor H, Klootwijk W, de Rijke YB, Kuiper GG, Uitterlinden AG, Visser TJ. Polymorphisms in thyroid hormone pathway genes are associated with plasma TSH and iodothyronine levels in healthy subjects. J Clin Endocrinol Metab. 2003;88:2880–2888. doi: 10.1210/jc.2002-021592. [DOI] [PubMed] [Google Scholar]
- 31.van der Deure WM, Hansen PS, Peeters RP, Uitterlinden AG, Fenger M, Kyvik KO, Hegedüs L, Visser TJ. The effect of genetic variation in the type 1 deiodinase gene on the interindividual variation in serum thyroid hormone levels: an investigation in healthy Danish twins. Clin Endocrinol (Oxf) 2009;70:954–960. doi: 10.1111/j.1365-2265.2008.03420.x. [DOI] [PubMed] [Google Scholar]
- 32.Gereben B, Kollár A, Harney JW, Larsen PR. The mRNA structure has potent regulatory effects on type 2 iodothyronine deiodinase expression. Mol Endocrinol. 2002;16:1667–1669. doi: 10.1210/mend.16.7.0879. [DOI] [PubMed] [Google Scholar]
- 33.Coppotelli G, Summers A, Chidakel A, Ross JM, Celi FS. Functional characterization of the 258 A/G (D2-ORFa-Gly3Asp) human type-2 deiodinase polymorphism: a naturally occurring variant increases the enzymatic activity by removing a putative repressor site in the 5′ UTR of the gene. Thyroid. 2006;16:625–626. doi: 10.1089/thy.2006.16.625. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Distribution of UGT1A1, UGT1A3, DIO1 and DIO2 polymorphisms in Brazilian DTC patients, according to self-reported ‘colour/race’