Abstract
Understanding of the aetiological basis of thyroid autoimmunity may be gained by studying the early stages of the disease process. We aimed to (1) investigate the relationship between thyroid antibody status and Yersinia enterocolitica (YE) infection in euthyroid subjects and (2) explore the relative importance of genetic and environmental risk factors in the acquisition of YE infection. The association between thyroid antibody status and YE infection was explored using a case–control design. Furthermore, thyroid antibody-positive twins were compared with their thyroid antibody-negative co-twin. In 468 twins, IgA and IgG antibodies to virulence-associated outer-membrane proteins (YOPs) of YE were measured. Of these, 147 were thyroid antibody-positive (cases). A total of 147 age- and gender-matched twins were chosen as controls. The prevalence of YOP antibodies was lower among thyroid antibody-positive individuals than among controls. Yersinia infection was not associated with a positive thyroid antibody status: the odds ratio (with 95% CI) for YOP IgA-ab was 0·66 (0·42–1·05), P = 0·078 and for YOP IgG-ab it was 0·95 (0·60–1·50), P = 0·816. Within discordant twin pairs, the thyroid antibody-positive twin did not have an increased risk of Yersinia infection compared to the thyroid antibody-negative co-twin [odds ratio: YOP IgA-Ab: 0·94 (0·49–1·83), P = 0·866, and YOP IgG-Ab: 1·35 (0·72–2·53), P = 0·345]; 41% (95% CI 10–67% of the liability of being YOP antibody-positive was due to genetic effects. In conclusion, Yersinia infection does not confer an increased risk of thyroid antibodies. The genetic contribution in the acquisition of Yersinia infection is modest.
Keywords: Tgab, thyroid autoimmunity, TPOab, twin studies, Yersinia enterocolitica
Introduction
Yersinia enterocolitica (YE) infection has long been implicated in the pathogenesis of autoimmune thyroid disease (AITD). Humoral as well as cellular immunity against YE in patients with autoimmune thyroid diseases has been demonstrated [1,2], and Wenzel et al. [3,4] have found antibodies against virulent plasmid-encoded proteins (YOP) of YE in a large proportion of patients with AITD. A potential mechanism could be molecular mimicry [5]. According to this hypothesis, the generated Yersinia-specific immune response cross-reacts with thyroid-specific components to cause tissue damage and disease. Various molecular studies have shown epitope homology and cross-reactivity between antigens of YE and the thyroid stimulating hormone-receptor (TSH-R) [6–8]. However, the importance of this mechanism is largely unknown, and clinical studies in support of this hypothesis have been inconclusive [9–13].
Most evidence supporting the role of Yersinia infection in thyroid autoimmunity comes from studying the association of Yersinia infection and overt AITD. It is possible, however, that treatment as well as the disease process itself has an influence on the aetiological components. The presence of autoantibodies to thyroid peroxidase (TPOab) and thyroglobulin (Tgab) in euthyroid subjects is regarded as an early stage in the pathogenesis of overt AITD. These markers of thyroid autoimmunity are under strong genetic influence; however, environmental influences are also of importance [14,15]. Further understanding of the aetiological basis of overt thyroid disease can be gained by focusing on therelationship between thyroid antibody status in healthy individuals and an environmental exposure such as Yersinia infection. Recently, in a cohort of first-degree female relatives of patients with AITD, a significantly higher prevalence of antibodies against YE compared to controls was demonstrated [16]. However, twin pairs discordant for presence of thyroid antibodies offer a unique opportunity to investigate the influence of Yersinia infection by comparing a thyroid antibody-positive twin with a thyroid antibody-negative co-twin. In particular, studies of monozygotic (MZ) twin pairs are valuable because they eliminate and control for many sources of variation (e.g. differences in genetic factors and in intrauterine and early environmental influences) which are inherent in ordinary matched case–control studies.
In the understanding of the aetiology of autoimmune thyroid disease, Yersinia infection is regarded as an environmental exposure. However, many measurable aspects of the environment have a genetic component themselves [17]. The relative impact of genetic and environmental risk factors in the acquisition of YE infection is at present unknown.
Thus, the aim of our twin study was (1) to investigate the relationship between thyroid antibody status and Yersinia infection in euthyroid subjects and (2) to explore the relative importance of genetic and environmental risk factors in the acquisition of YE infection.
Subjects
A representative sample of complete twin pairs was recruited from the population-based Danish Twin Registry [18]. The twins included in the GEMINAKAR study were self-reported as healthy, although individuals with chronic diseases such as low back pain, migraine, etc. were included in the study population. In order to obtain an equal distribution of twin pairs, sampling was stratified according to age, sex and zygosity. The twins in a pair were examined on the same day. Blood samples were drawn between 8 and 9 a.m. after a 12-h fast. A clinical examination was performed and the twins completed health-related questionnaires including questions regarding thyroid disease, smoking habits and medicine intake.
In all, 1512 individuals (756 twin pairs) were examined. Blood samples were available from 736 twin pairs. Serum thyrotrophin (TSH), serum free thyroxine (free T4), serum free triiodothyronine (free T3), serum thyroid peroxidase antibodies (TPOab) and thyroglobulin antibodies (Tgab) were measured. Individuals with self-reported thyroid disease (32 people in 28 twin pairs) and overt biochemical thyroid disease (19 people in 18 twin pairs) were excluded as were their co-twins, leaving 1380 healthy, euthyroid individuals (690 twin pairs). A total of 468 of these individuals, distributed in 68 MZ, 94 dizygotic (DZ) and 72 opposite sex (OS) pairs were informative for Yersinia analysis.
We defined thyroid antibody positivity if the concentration of TPOab and/or if Tgab was above 10 kIU/l and 20 kIU/l, respectively. Of the 468 individuals, 147 subjects fulfilled these criteria and were classified as cases. The 147 cases were distributed in 89 discordant (23 MZ, 40 DZ and 26 OS twin) and 29 concordant (13 MZ, five DZ and 11 OS) twin pairs; 147 twin individuals were selected as age- and gender-matched external controls. Among the controls, 36 complete twin pairs (10 MZ, 18 DZ and eight OS twin pairs) were included due to limited access to serum. To test the consistency of any finding, we also performed all analyses using 60 kIU/l as a cut-off value for both thyroid antibodies.
In the twin case–control design using co-twins as controls, the 29 concordant twin pairs were excluded, leaving 89 twin pairs discordant for presence of thyroid antibodies.
Written informed consent was obtained from all participants and the study was approved by all regional Danish Scientific-Ethical Committees (case file 97/25 PMC).
Assays
Serum TSH was measured using a time-resolved fluoroimmunometric assay (AutoDELFIA hTSH Ultra Kit; Perkin Elmer/Wallac, Turku, Finland), reference range 0·30–4·00 mU/l. Serum free T4 and serum free T3 were determined using the AutoDELFIA FT4 and FT3 (Perkin Elmer/Wallac), respectively. For free T4 the reference range is 9·9–17·7 pmol/l, and for free T3 it is 4·3–7·4 pmol/l. TPOab and Tgab were measured by solid-phase, two-step, time-resolved fluoroimmunoassays (AutoDELFIA TPOab kit and hTgab kit, respectively, Perkin Elmer/Wallac, Turku, Finland). For TPOab the CV analytical was 9·3%, in the range 2·5–258 kIU/l [19]. For Tgab the CV analytical was 9·0%, in the range 5·6–147 kIU/l [19]. Twin pairs were analysed within the same run. All the serum samples were analysed in the same laboratory in Odense.
Zygosity was established by analysis of nine highly polymorphic restriction fragment length polymorphisms and microsatellite markers scattered widely through the genome with an Applied Biosystems AmpFISTR Profiles Plus kit (Perkin-Elmer/Wallac) [20].
Determination of YE status
YOPs of YE serotype 09 were produced by a low calcium (LCR) YE response [21]. YE status was evaluated measuring IgA and IgG. Specific IgG and IgA antibodies against YOPs were demonstrated by immunoblotting with a YOP-ab assay (AID, Strassberg, Germany). Briefly, antigens (25, 34, 36, 37, 39, 40, 46, 48 kDa) are blotted onto nitrocellulose. Next, sera are diluted 1 : 51 in phosphate-buffered saline (PBS)-Tween and incubated with antigen-coated nitrocellulose strips overnight at 22°C. The IgG and IgA antibody–antigen complexes formed are quantified after immunostaining with the AID Scan System. Controls are included in each assay run, using human acute sera (culture-positive YE infection) containing antibodies to the YOPs. Test sera are judged positive if at least three bands (IgG) or one band (IgA) are seen in immunoblotting at a level greater than 15% (IgG) or 10% (IgA) of the reference standards. The interassay variation of the YOP-ab assay is < 3%. The YOP-ab measurement was performed blinded as to thyroid function variables and the presence of TPOab as well as Tgab in the serum samples.
Statistical methods
Case–control study with external controls
In this part of the study, the twins with a positive thyroid antibody status (the cases) were compared with the unrelated control twins. The prevalences of IgA and IgG antibodies against YOPs for each group were estimated. The significance of the differences between groups were analysed using Pearson's χ2 test. Moreover, the relationship between thyroid antibody positivity and Yersinia infection was investigated by a logistic regression analysis. The analyses were performed with and without adjustment for age and were stratified according to zygosity.
Case–control study with co-twin controls
The relationship between thyroid antibody status and the presence of antibodies against YE (IgA as well as IgG, as dichotomous variables) was explored further by analysing the twin pairs discordant for thyroid autoantibodies. In this approach, the cases were twins with a positive thyroid antibody status whereas the thyroid antibody-negative co-twins were controls. These discordant twin pairs were analysed as matched sets with the use of conditional logistic regression analysis. To determine whether risk factors differed according to genetic susceptibility the analyses were stratified according to zygosity.
Twin analyses of Yersinia infection
In these analyses, individuals were considered as Yersinia antibody-positive if at least three bands IgG combined with one band IgA were detected. The classical twin study is based on the assumption that MZ twins are genetically identical, and therefore differences between them are due solely to the environment. DZ twins share on average 50% of their segregating genes, and therefore differences between them are due to a combination of environmental and genetic factors [17,22]. If there is a substantial genetic component in the aetiology of the phenotype, a greater phenotypic similarity in MZ than in DZ twins is to be expected. The similarity in MZ and DZ twins was assessed by probandwise concordance and tetrachoric correlations. Probandwise concordance is defined as the proportion of affected co-twins of probands, expressing the risk that a twin is affected given an affected co-twin, and is directly comparable to disease risk rates in the background population [23]. The calculation of tetrachoric correlations is based on the assumption that there is an underlying normal distribution of liability to acquire a positive Yersinia antibody status [17,24]. The trait becomes manifest when an individual exceeds a given threshold on the liability distribution [17]. The threshold reflects the prevalence of the trait. Using the software program Mx, the effects of age and gender were incorporated into the model as covariates by assuming linear dependence of the thresholds on the covariates. The results were compared to a model without adjustment. The difference in correlations between MZ and DZ twin pairs was assessed with a likelihood ratio test [17].
Model-fitting procedure
Structural equation modelling was used to estimate the magnitude of the genetic and environmental effects on the liability of Yersinia antibody status. The structural equation modelling approach is based on familial relations and was carried out using maximum likelihood methods in Mx. The observed phenotypic variance is decomposed into genetic and environmental contributions [17]. The genetic variance is further subdivided into an additive (A) component and a dominance (D) component. The environmental contribution is divided into a shared/common environmental component (C) and a unique (E) environmental component. Heritability is defined as the proportion of the total variance attributable to total additive genetic variance [17].
C and D are confounded and cannot be estimated simultaneously in a twin study of MZ and DZ twins reared together [17,22,24]. The full model ACE was examined, and the significance of A, C and E were tested by removing them sequentially in specific nested submodels [17]. For comparison among nested models, we used a likelihood ratio test. The difference in log likelihood between a full model and that of a submodel (Δ2LL) is distributed as a χ2 statistic, with the degrees of freedom equal to the difference in the degrees of freedom of the models being compared [17]. A non-significant χ2 statistic (e.g. P > 0·05) means that the model is consistent with the data. Selection of the best-fitting model was carried out using the Akaike information criterion, which is based on a balance between goodness-of-fit and parsimony [17]. AIC corresponds to Δ2LL – 2 × Δd.f. Models with the lowest AIC were preferred.
Models were fitted to the raw data using raw data methods in Mx [25,26]. The effects of specific covariates (as specified earlier) were incorporated into the analyses. According to standard biometric practice [17,24], we assumed equal environment for MZ and DZ twins, no epistasis (gene–gene interaction) and no gene–environment interaction or correlation.
Statistical software
The statistical analyses were carried out using stata 7 [27]. Level of significance was set to 0·05. Univariate quantitative genetic modelling was carried out using Mx [26].
Results
Case–control study with external controls (Table 1)
Table 1.
Basic characteristics of thyroid antibody-positive and -negative individuals.
| Gender | ||||||
|---|---|---|---|---|---|---|
| Case control study, external controls | Females (n) | Males (n) | Age (years)* | TSH (mU/l)* | YOP IgAan (%) | YOP IgGbn (%) |
| Cases (n = 147) | 96 | 51 | 39·7 (11·5) | 2·02 (1·23) | 57 (38·8%) | 72 (49·0%) |
| Controls (n = 147) | 96 | 51 | 41·9 (8·8) | 1·58 (0·81) | 72 (49·0%) | 74 (50·3%) |
Mean values (SD).
Virulence-associated plasmid-encoded proteins of Yersinia Enterocolitica (YOP) IgA-positive: one or more bands.
YOP IgG-positive: three or more bands.
Characteristics are given in Table 1. The prevalence of YOP IgA-ab was 38·8% among thyroid antibody-positive individuals and 49·0% in the control group (P = 0·078), whereas the prevalence of YOP IgG-ab was 49·0% versus 50·3%, respectively (P = 0·816). Using only one twin from each pair as controls, the prevalences were (38·8% versus 53·2%, P = 0·022) for YOP IgA-ab and (49·0% versus 55·0%, P = 0·342) for YOP IgG-ab. The prevalence of YOP IgA-ab and IgG-ab remained lower among thyroid antibody-positive individuals compared with the controls using thyroid antibody levels of 60 kIU/l as a cut-off (31·9% versus 49·0%, P = 0·017) and (38·9% versus 50·3%, P = 0·111), respectively.
Using logistic regression, the odds ratio for the association between thyroid antibody status and YOP IgA-ab was 0·66 (95% CI 0·42–1·05), P = 0·078, and for the association between thyroid antibody status and YOP IgG-Ab it was 0·95 (95% CI 0·60–1·50), P = 0·816. Stratification according to zygosity did not influence the results, and the risk estimates were almost unchanged, adjusting for the effect of age (data not shown). Using a cut-off value of 60 kIU/l for TPOab as well as for Tgab did not significantly change the overall findings (data not shown).
Case–control study with co-twin controls
Among thyroid antibody discordant twin pairs, the odds ratio was 0·94 (95% CI 0·49–1·83, P = 0·866) for the presence of YOP IgA-ab and 1·35 (95% CI 0·72–2·53, P = 0·345) for the presence of YOP IgG-ab. The results were almost unchanged using a thyroid antibody cut-off level of 60 kIU/l. After stratification according to zygosity, neither of the results in these subsamples reached significance and the confidence intervals were wide (data not shown).
Probandwise concordance and tetrachoric correlations (Table 2)
Table 2.
Number of twin pairs, probandwise concordance and tetrachoric correlations for Yersinia outer protein (YOP) positivity according to zygosity and phenotype.
| YOP status | P-value | |
|---|---|---|
| Number of twin pairs | +/+ +/– –/– | |
| MZ | 19 24 25 | |
| DZ | 22 41 31 | |
| Probandwise concordancea | ||
| MZ | 0·61 (0·47–0·75) | 0·172 |
| DZ | 0·52 (0·38–0·64) | |
| Tetrachoric correlation (unadjusted)b | ||
| MZ | 0·44 (0·08–0·71) | 0·280 |
| DZ | 0·19 (− 0·13–0·48) | |
| Tetrachoric correlation (adjustedc)b | ||
| MZ | 0·41 (0·04–0·70) | 0·263 |
| DZ | 0·14 (− 0·19–0·44) | |
MZ: monozygotic; DZ: dizygotic. +/+ and –/–: concordant pairs; ±: discordant pairs; 95% confidence intervals are given in parentheses. Positive Yersinia status was defined as one or more bands of IgA combined with three or more IgG bands.
A bootstrap method was used testing the difference in probandwise concordance (one-sided).
A likelihood ratio test was used testing the difference in tetrachoric correlations between zygosity groups.
Adjustment for age and gender.
The probandwise concordances and the unadjusted as well as the age- and gender-adjusted tetrachoric correlations for Yersinia antibody status are shown in Table 2. No significant differences in either probandwise concordance or tetrachoric correlations were found between MZ and DZ twin pairs.
Biometric modelling (Tables 2 and 3)
Table 3.
Results of model-fitting analyses; the best-fitting model is highlighted in bold; 95% confidence intervals are given in parentheses.
| Genetic component | Environmental components | Goodness-of-fit tests | Comparisons of nested submodels | |||||
|---|---|---|---|---|---|---|---|---|
| Model | A | C | E | − 2LLb | Δ2LLc | Δd.f.d | AIC | Pe |
| ACE | 0·41 (0·00–0·67) | 0·00 (0·00–0·48) | 0·59 (0·33–0·92) | 438·25 | ||||
| AE* | 0·41 (0·10–0·67) | – | 0·59 (0·33–0·90) | 438·25 | 0·000 | 1 | < − 2·00 | 1·000 |
| CE | – | 0·29 (0·05–0·50) | 0·71 (0·50–0·95) | 439·16 | 0·917 | 1 | − 1·083 | 0·338 |
| E | – | – | 1 | 444·84 | 6·591 | 2 | 2·591 | 0·037 |
The best-fitting model. The selection of the best-fitting model was carried out using the Akaike information criterion, which is based on a balance between goodness-of-fit and parsimony. AIC corresponds to Δ2LL –2 × Δd.f. Models with the lowest AIC were preferred.
Adjustment for age and gender;
minus two times log likelihood of data;
the difference in 2LL between the full model and the submodel;
the difference in degrees of freedom between the full model and the submodel;
probability.
No significant differences in either probandwise concordance or tetrachoric correlations were found between MZ and DZ twins. This pattern of correlations suggests the presence of common environmental influences, therefore the ACE model was chosen. Subsequently, reduction to nested submodels was attempted.
The best-fitting model was an AE model, including additive genetic effects and unique environmental effects. Adjusting for age and gender, 41% (95% CI: 10–67%) of the variation in the likelihood of being YOP-positive was due to additive genetic factors (heritability). The remaining 59% (33–90%) could be attributed to unique environmental factors. However, a non-genetic model (the CE model) was also consistent with the data.
Discussion
An association between thyroid autoimmunity and the presence of YE antibodies has long been suggested [4,9,12,13], although results are conflicting. In our study, with focus on the early stages of AITD, antibodies against YE did not confer an increased risk of thyroid autoimmunity. Our findings were strengthened by looking at twin pairs discordant for thyroid autoantibodies. Thus, we suggest that YE has little role in the aetiology of thyroid autoimmunity. Despite the cross-sectional nature of our study, it does provide valuable information. Had we found an association between antibodies against Yersinia and a positive thyroid antibody status among cases and external controls as well as co-twin controls, this would strongly have suggested a causal effect of Yersinia infection and risk of thyroid autoimmunity. In addition, Yersinia infection preceding the appearance of thyroid autoantibodies would suggest a causal relation. The high prevalence of IgA antibodies in our study reflects quite recent Yersinia infection. If Yersinia infection were, indeed, to have a causal effect in the pathogenesis of thyroid autoimmunity, one would expect an association between thyroid antibody status and more chronic stages of Yersinia infection, as reflected by the presence of IgG antibodies, which we did not find either. It should be emphasized that our data do not allow discrimination between subclinical stages of Graves' disease and Hashimoto's thyroiditis. Another limitation is the possibility of selection bias.
Diet (i.e. pork, poultry, cabbage) is believed to be the primary source of Yersinia infection [28], whereas other sources of infection, such as person-to-person transmission, are unlikely [29]. YE is an important bacterial pathogen even in young Danish children [30]; it has been found that the prevalence of IgG seropositivity against YE is around 60% and increases with increasing age [31]. Reflecting the high intake of pork in Denmark, the prevalence of Yersinia infection in the external control group as well as among thyroid antibody-positive individuals is high in our study, but comparable with other Danish studies [28,31].
The reported association between Yersinia infection and the risk of AITD in previous studies could be due at least partly to genetic associations, and it has been suggested that susceptibility genes for AITD may confer a risk for YE infection as well [16]. The use of co-twin controls allowed us to eliminate many sources of variation that are inherent in ordinary case–control studies. In the twin case–control study, the twins share a part of or all their genes. Furthermore, they have generally shared the same childhood environment. If an association were observed using external, but not co-twin, controls the probable explanation for these discrepant results would be confounding by genetic or early environmental influences. We did not find an association between antibodies against Yersinia and a positive thyroid antibody status, either in the ordinary case–control study using external twin controls or in the twin case–control study using co-twin controls. Within discordant twin pairs, the twin with a positive thyroid antibody status did not have a significantly increased risk of Yersinia infection. Complete control for genetic influence is possible in MZ twins. The analyses were therefore repeated separately for MZ and DZ twins, but none of the odds ratios reached significance. We were not able to detect a difference between MZ and DZ twins. However, in these subsamples the number of twin pairs was small. The possibility of a negative finding due to lack of statistical power cannot be excluded, and these sub-analyses are not fully conclusive. However, we find it most likely that factors associated with lifestyle and shared environment or unmeasured environmental exposures contribute to the previously reported associations. Sharing the same home environment would contribute to clustering of Yersinia infection within AITD family relatives, because lifestyle factors such as diet and the consumption of pork would tend to be the same among relatives. On the other hand, assuming that this cultural influence is established mainly in childhood, this confounding effect would be removed at least partly in our study.
In the second part of our study, the probability of acquiring a YE infection was estimated. We did not find any significant differences in probandwise concordances between MZ and DZ twins, and in model-fitting analyses environmental factors accounted for the major part of the variation in the tendency to acquire YE infection, whereas genetic factors accounted only for a restricted proportion of the variation. The mechanism underlying genetic susceptibility involves a complex network of immunological and non-immunological mechanisms. The genetic contribution may include genetic factors involved with the gastrointestinal mucosal barrier to prevent entry of this food-borne pathogen [32]. In addition, the genetic constitution of the host determines not only whether infection takes place, but also the consequences of the infection. Once the pathogen has entered the host it involves the specific response against this pathogen as well as more general mechanisms in the immune response. Finally, genetic factors associated with the Yersinia bacteria itself are in play [33]. Added to this, the bacteria are capable of avoiding the harmful effects of the immune response through interaction effects with the host response [32].
In conclusion, Yersinia infection does not confer an increased risk of thyroid autoimmunity. Morover, the genetic contribution in the acquisition of Yersinia infection is modest.
Acknowledgments
The present work was supported by grants from the Danish Research Agency; the Foundation of 17-12-1981; the Agnes and Knut Mørk Foundation; the Novo Nordisk Foundation; the Foundation of Medical Research in the County of Funen; Else Poulsens Mindelegat; the Foundation of Direktør Jacob Madsen and Hustru Olga Madsen; the Foundation of Johan Boserup and Lise Boserup; the A. P. Møller and Hustru Chastine McKinney Møllers Foundation; the A. P Møller Relief Foundation, the Clinical Research Institute, Odense University and Margareta Markus Charity Foundation London, UK. Perkin Elmer/Wallac, Turku, Finland, kindly provided the kits for determination of TPOab, Tgab, serum TSH, free T4 and free T3. We thank Kirsten Ohm Kyvik and Thorkild I. A. Sørensen who initiated the GEMINAKAR study for their invaluable contribution. Ole Blaabjerg and Esther Jensen are acknowledged for supervising the biochemical thyroid analyses, and technical assistant Detlev Schult for excellent laboratory assistance.
References
- 1.Bech K, Larsen JH, Hansen JM, Nerup J. Yersinia enterocolitica infection and thyroid disorders [Letter] Lancet. 1974;2:951–2. doi: 10.1016/s0140-6736(74)91152-0. [DOI] [PubMed] [Google Scholar]
- 2.Lidman K, Eriksson U, Fagraeus A, Norberg R. Antibodies against thyroid cells in Yersinia enterocolitica infection [Letter] Lancet. 1974;2:1449. doi: 10.1016/s0140-6736(74)90102-0. [DOI] [PubMed] [Google Scholar]
- 3.Wenzel BE, Strieder TM, Gaspar E, Wiersinga WM. Chronic infection with Yersinia enterocolitica in patients with clinical or latent hyperthyroidism. Adv Exp Med Biol. 2003;529:463–6. doi: 10.1007/0-306-48416-1_92. [DOI] [PubMed] [Google Scholar]
- 4.Wenzel BE, Heesemann J, Wenzel KW, Scriba PC. Antibodies to plasmid-encoded proteins of enteropathogenic Yersinia in patients with autoimmune thyroid disease. Lancet. 1988;1:56. doi: 10.1016/s0140-6736(88)91034-3. [DOI] [PubMed] [Google Scholar]
- 5.Albert LJ, Inman RD. Molecular mimicry and autoimmunity. N Engl J Med. 1999;341:2068–74. doi: 10.1056/NEJM199912303412707. [DOI] [PubMed] [Google Scholar]
- 6.Tomer Y, Davies TF. Infections and autoimmune endocrine disease. Baillieres Clin Endocrinol Metab. 1995;9:47–70. doi: 10.1016/s0950-351x(95)80819-1. [DOI] [PubMed] [Google Scholar]
- 7.Prummel MF, Strieder T, Wiersinga WM. The environment and autoimmune thyroid diseases. Eur J Endocrinol. 2004;150:605–18. doi: 10.1530/eje.0.1500605. [DOI] [PubMed] [Google Scholar]
- 8.Heyma P, Harrison LC, Robins-Browne R. Thyrotrophin (TSH) binding sites on Yersinia enterocolitica recognized by immunoglobulins from humans with Graves' disease. Clin Exp Immunol. 1986;64:249–54. [PMC free article] [PubMed] [Google Scholar]
- 9.Corapcioglu D, Tonyukuk V, Kiyan M, et al. Relationship between thyroid autoimmunity and Yersinia enterocolitica antibodies. Thyroid. 2002;12:613–7. doi: 10.1089/105072502320288483. [DOI] [PubMed] [Google Scholar]
- 10.Asari S, Amino N, Horikawa M, Miyai K. Incidences of antibodies to Yersinia enterocolitica: high incidence of serotype O5 in autoimmune thyroid diseases in Japan. Endocrinol Jpn. 1989;36:381–6. doi: 10.1507/endocrj1954.36.381. [DOI] [PubMed] [Google Scholar]
- 11.Arscott P, Rosen ED, Koenig RJ, et al. Immunoreactivity to Yersinia enterocolitica antigens in patients with autoimmune thyroid disease. J Clin Endocrinol Metab. 1992;75:295–300. doi: 10.1210/jcem.75.1.1619022. [DOI] [PubMed] [Google Scholar]
- 12.Resetkova E, Notenboom R, Arreaza G, Mukuta T, Yoshikawa N, Volpe R. Seroreactivity to bacterial antigens is not a unique phenomenon in patients with autoimmune thyroid diseases in Canada. Thyroid. 1994;4:269–74. doi: 10.1089/thy.1994.4.269. [DOI] [PubMed] [Google Scholar]
- 13.Chatzipanagiotou S, Legakis JN, Boufidou F, Petroyianni V, Nicolaou C. Prevalence of Yersinia plasmid-encoded outer protein (Yop) class-specific antibodies in patients with Hashimoto's thyroiditis. Clin Microbiol Infect. 2001;7:138–43. doi: 10.1046/j.1469-0691.2001.00221.x. [DOI] [PubMed] [Google Scholar]
- 14.Brix TH, Hansen PS, Kyvik KO, Hegedüs L. Aggregation of thyroid autoantibodies in first-degree relatives of patients with autoimmune thyroid disease is mainly due to genes: a twin study. Clin Endocrinol (Oxf) 2004;60:329–34. doi: 10.1111/j.1365-2265.2004.01983.x. [DOI] [PubMed] [Google Scholar]
- 15.Hansen PS, Brix TH, Iachine I, Kyvik KO, Hegedüs L. The relative importance of genetic and environmental effects for the early stages of thyroid autoimmunity: a study of healthy Danish twins. Eur J Endocrinol. 2006;154:29–38. doi: 10.1530/eje.1.02060. [DOI] [PubMed] [Google Scholar]
- 16.Strieder TG, Wenzel BE, Prummel MF, Tijssen JG, Wiersinga WM. Increased prevalence of antibodies to enteropathogenic Yersinia enterocolitica virulence proteins in relatives of patients with autoimmune thyroid disease. Clin Exp Immunol. 2003;132:278–82. doi: 10.1046/j.1365-2249.2003.02139.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Neale MC, Maes HH. Methodology for genetic studies of twins and families. Dordrecht, the Netherlands: Kluwer Academic Publishers BV; 2002. [Google Scholar]
- 18.Skytthe A, Kyvik K, Holm NV, Vaupel JW, Christensen K. The Danish Twin Registry: 127 birth cohorts of twins. Twin Res. 2002;5:352–7. doi: 10.1375/136905202320906084. [DOI] [PubMed] [Google Scholar]
- 19.Jensen E, Petersen PH, Blaabjerg O, Hegedüs L. Biological variation of thyroid autoantibodies in fertile women. Thyroid. 2005;15:P13. doi: 10.1515/CCLM.2007.185. [DOI] [PubMed] [Google Scholar]
- 20.Applied Biosystems. Applied Biosystems AmpFISTR Profiles Plus kit, user's manual. Foster City, CA, USA: The Perkin-Elmer Corporation; 1997. [Google Scholar]
- 21.Heesemann J, Eggers C, Schroder J. Serological diagnosis of yersiniosis by immunoblot technique using virulence-associated antigen of enteropathogenic Yersiniae. Contrib Microbiol Immunol. 1987;9:285–9. [PubMed] [Google Scholar]
- 22.Evans DM, Gillespie NA, Martin NG. Biometrical genetics. Biol Psychol. 2002;61:33–51. doi: 10.1016/s0301-0511(02)00051-0. [DOI] [PubMed] [Google Scholar]
- 23.McGue M. When assessing twin concordance, use the probandwise not the pairwise rate. Schizophr Bull. 1992;18:171–6. doi: 10.1093/schbul/18.2.171. [DOI] [PubMed] [Google Scholar]
- 24.Rijsdijk FV, Sham PC. Analytic approaches to twin data using structural equation models. Brief Bioinform. 2002;3:119–33. doi: 10.1093/bib/3.2.119. [DOI] [PubMed] [Google Scholar]
- 25.Hopper JL, Green RM, Nowson CA, et al. Genetic, common environment, and individual specific components of variance for bone mineral density in 10- to 26-year-old females: a twin study. Am J Epidemiol. 1998;147:17–29. doi: 10.1093/oxfordjournals.aje.a009361. [DOI] [PubMed] [Google Scholar]
- 26.Neale MC, Boker SM, Xie G, Maes HH. Mx: statistical modeling. 6. Richmond, VA: Department of Psychiatry; 2002. [Google Scholar]
- 27.Stata Corporation. Stata statistical software, release 7.0. College Station, TX: Stata Corporation; 2001. [Google Scholar]
- 28.Nielsen B, Wegener HC. Public health and pork and pork products. regional perspectives of Denmark. Rev Sci Tech. 1997;16:513–24. [PubMed] [Google Scholar]
- 29.Ethelberg S, Olsen KE, Gerner-Smidt P, Molbak K. Household outbreaks among culture-confirmed cases of bacterial gastrointestinal disease. Am J Epidemiol. 2004;159:406–12. doi: 10.1093/aje/kwh049. [DOI] [PubMed] [Google Scholar]
- 30.Olesen B, Neimann J, Bottiger B, et al. Etiology of diarrhea in young children in Denmark: a case–control study. J Clin Microbiol. 2005;43:3636–41. doi: 10.1128/JCM.43.8.3636-3641.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Linneberg A, Ostergaard C, Tvede M, et al. IgG antibodies against microorganisms and atopic disease in Danish adults: the Copenhagen Allergy Study. J Allergy Clin Immunol. 2003;111:847–53. doi: 10.1067/mai.2003.1335. [DOI] [PubMed] [Google Scholar]
- 32.Acheson DW, Luccioli S. Microbial–gut interactions in health and disease. Mucosal immune responses. Best Pract Res Clin Gastroenterol. 2004;18:387–404. doi: 10.1016/j.bpg.2003.11.002. [DOI] [PubMed] [Google Scholar]
- 33.Darwin AJ. Genome-wide screens to identify genes of human pathogenic Yersinia species that are expressed during host infection. Curr Issues Mol Biol. 2005;7:135–49. [PubMed] [Google Scholar]