Department of Medicine Mount Sinai School of Medicine New York, New York 10029
Address all correspondence and requests for reprints to: Dr. T. F. Davies, Box 1055, Mount Sinai Medical Center, 1 Gustave L. Levy Place, New York, New York 10029.
![]() |
Introduction |
---|
![]() ![]() ![]() |
---|
No one has seriously doubted the inherited nature of Graves disease and Hashimotos disease since their familial occurrence was first observed many years ago (1, 2, 3). The autoimmune thyroid diseases (AITD), however, do not show Mendelian inheritance. While the degree of penetrance of genes may be variable, a Mendelian pattern can be discerned using classic genetic methods requiring careful observation of large family trees or large numbers of families. Because the AITDs do not show this pattern of inheritance they are referred to as "complex diseases," similar to Type 1 and Type 2 insulin dependent diabetes, hypertension, and osteoporosis, to name but a few. In these complex conditions it has been difficult to separate environmental influences from genetic susceptibility, and a large body of scholarship has been devoted to devising methods for their discernment (4).
In simple theory, monozygotic twins should tell us a great deal about the genetics of a disease. Environmental influences should affect both monozygotic and dizygotic pairs to an equal extent, while monozygotic twins should demonstrate the influence of genetics. Unfortunately, life is not so simple, and recent authors have been jumping to all sorts of inappropriate conclusions from the study of twins with Graves disease, even questioning the importance of genetic susceptibility in the AITDs (5). Although twin studies are of great value, it is very unlikely that the immune repertoires of identical twins are the same. Immunoglobulin genes and T-cell receptor genes, which feed the immune repertoire, are unique in that they undergo the phenomenon of somatic recombination and somatic mutation, in which nucleotide additions and deletions occur at random within clones of B cells and T cells (6). This is true for twins, within inbred strains of mice, and presumably would be true for cloned animals as well. Hence, twins may have fundamentally distinct immune repertoires, even without differing environmental exposures. There is now overwhelming evidence for a bias in the use of T-cell receptor V genes by T cells involved in AITD (7, 8, 9, 10, 11) in agreement with most other autoimmune diseases (12). Therefore, the degree of concordance for AITD in twin pairs can only be a rough guide to the genetic contribution to AITD. Reports clarifying the concordance rate for Graves disease in twins at 22% for monozygotic and 0% for dizygotic (13), although less than previously reported, are exciting confirmatory data for a major genetic component, which was measurable despite the background "noise" contributed by the somatic recombination events referred to earlier. Indeed, the fact that such twin data continue to show a strong genetic component is direct evidence that the unknown susceptibility genes involved are unlikely to be strongly associated with the T-cell and B-cell somatic recombination process. This reduces the likelihood that the immunoglobulin and T-cell receptor genes themselves have major roles to play in genetic susceptibility to AITD; rather they are likely to be only passively involved, in keeping with recent observations (14).
One well-researched gene region that may influence the ultimate outcome of somatic recombination by affecting the selection of B cells and T cells for survival, is the human leukocyte antigen (HLA) gene region. While this multigene region on chromosome 6 is identical in monozygotic twins, it has also been thought to be informative in "HLA-identical" siblings. The concordance rate for AITD in such HLA class II identical siblings was been found to be only 7% (15) suggesting that the HLA region was not intimately involved in susceptibility to AITD. However, the designation of "HLA-identical" is only correct according to the quality of the HLA typing performed at the time the work was performed, and this is an area where there have been major technical advances. The molecular characterization of the HLA gene region has allowed detailed PCR amplification techniques to provide more precise typing of the many HLA alleles. This approach has confirmed, in many studies, that Graves disease is associated with characteristic HLA allotypes (15, 16, 17, 18). The data for Hashimotos disease have been less convincing, probably because of the great problems with defining a single phenotype for autoimmune thyroiditis (for examples and discussion, see refs. 19, 20, 21, 22). Indeed, heterogeneity in a disease can play havoc with all genetic studies and can often be found as the likely cause of disparity between different reports.
In this months JCEM, Heward et al. (23) provide a well-designed contribution to this literature by defining an HLA haplotype (DRB110304-DQB1102-DQA110501) that showed preferential transmission with Graves disease between parents and affected children, but not with unaffected children. This information leaves no doubt that the HLA region is involved in susceptibility to AITD and again highlights the generic role of DQA110501 in autoimmune disease (17). The argument about HLA involvement in AITD, however, has hinged more essentially upon the degree of the contribution to overall susceptibility for which the HLA genes are responsible. For example, 24% of the controls in the study by Heward et al. showed the same susceptibility haplotype as the patients with Graves disease. Moreover, 53% of the Graves disease patients did not express the haplotype. What do we learn from this? In my opinion there is only one conclusion. Such an HLA haplotype contributes to AITD susceptibility but is not necessary for disease development (24). This conclusion explains why classical linkage analysis of families with AITD has failed to demonstate linkage with the HLA region (25, 26, 27). In contrast, such an approach has been highly successful for Type 1 diabetes where the HLA contribution to susceptibility has been shown to be much greater than in AITD (28, 29). Linkage analysis, rather than restricting itself to a particular haplotype (such as the Graves disease haplotype of Heward et al.) looks at the cosegregation of any possible allele at the locus being studied (such as HLA). It detects linkage when any of these alleles consistently cosegregates with the disease. Of course the cosegregating allele has to be the same within each family but can be different between families. It is not possible, therefore, to dismiss the lack of demonstrated linkage between HLA and AITD as unimportant. In fact it is highly informative and indicates that we should be looking hard for non-HLA genes contributing to overall genetic susceptibility to AITD(30).
Interesting progress has also been made in regard to non-HLA genes in AITD. A haplotype of the CTLA-4 gene, coding for a molecule that controls the degree of the immune response, has now been consistently associated with Graves disease (31, 32) and Hashimotos disease (33), but again has not been found to be linked to AITD in our own studies (14). The CTLA-4 gene has also been associated with other autoimmune diseases, such as Type 1 diabetes and Addisons disease (32, 33), and is thought to be, therefore, another small contributor to the propensity for developing autoimmune disease. This suggests that AITD susceptibility could be made up of multiple small contributing genetic influences.
Genes detected in association studies as giving a low relative risk (risk ratios < 35), such as HLA and CTLA-4 in AITD, may contribute no more than 5% each to overall genetic susceptibility (34). Hence, 1020 genes may be influencing the expression of AITD. If so, none of these genes would be detectable by classical linkage analysis, which may only detect important genes giving risk ratios of more than 5 and contributing more than 10% to genetic susceptibility (24). To some degree this explains the popularity of simple association studies using classical methods as well as transmission disequilibrium testing between parents and children as used by Heward et al. (35). However, with the multiple genetic loci thought to be associated with Type 1 diabetes (36), at least two major genes were found, one in the HLA region and another shown to be the insulin gene VNTR (37). Hence, it is likely that a mixture of a few highly significant genes may also contribute the major proportion of AITD susceptibility. If so, what are the likely major genes involved in AITD ? Initially, many of us working in the field felt that one of the well-known thyroid antigens must be involved. Now it appears that the thyroglobulin, thyroid peroxidase, and TSH receptor genes have all failed to show linkage with AITD (14, 38, 39, 40), although there are no data for the iodine symporter. Additional candidate genes, particularly immune related, continue to be explored, but such an approach may take many years to detect important genes, as there are so many potential candidates.
Hence, we and others have been pursuing a whole genome screen using microsatellite markers. To date, in a series of 56 AITD families, we have found evidence of linkage in patients with Graves disease for loci on chromosomes 14 and 20 (38, 39, 42) and also on the X chromosome (41). The latter raises the intriguing question of the role of X-inactivation mosaics as a fundamental cause of autoimmune disease (43). The search to identify what genes, their shape and their color, inhabit these loci and contribute to the cause of thyroid autoimmunity is one of the most exciting areas of thyroid research today. Those naysayers who claim that all these "studies are disappointing and generate frustration for those within the field" (5) should keep watching...
![]() |
Acknowledgments |
---|
![]() |
Footnotes |
---|
Received August 11, 1998.
Accepted August 11, 1998.
![]() |
References |
---|
![]() ![]() ![]() |
---|