| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Autoimmune Disease Unit, Cedars-Sinai Research Institute and University of California School of Medicine (J.C.J., J.G., B.R., S.M.M.), Los Angeles, California 90048; the Child Study Center and Departments of Genetics and Psychology, Yale University School of Medicine (D.L.P.), New Haven, Connecticut 06520; the Departments of Medicine (M.Z., J.M.M.) and Psychiatry (J.A.E.), University of Miami School of Medicine, Miami, Florida 33101; the Departments of Pathology and Molecular Microbiology and Immunology, The Johns Hopkins University (C.L.B., N.R.R.), Baltimore, Maryland 21205; and the Department of Pediatrics, Medical College of Georgia (W.H.H.), Augusta Georgia 30912
Address all correspondence and requests for reprints to: Dr. Sandra M. McLachlan, Cedars-Sinai Medical Center, 8700 Beverly Boulevard, Suite B-131, Los Angeles, California 90048.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
2(4)
= 20.67; P < 0.0032), and the most parsimonious
model includes a major gene effect. More importantly, evidence for
genetic transmission was obtained for the phenotype defined by the
ratio of inhibition by subdomain Fab B1:B2. Thus, for this ratio
(reflecting recognition of the B domain), the no transmission model was
rejected 2(4) = 63.59; P
< 0.000008). Moreover, the polygenic hypothesis could be rejected, but
not the major locus hypothesis, suggesting that major genes might be
involved in familial transmission of this trait. In conclusion, our findings suggest that autoantibody recognition of the TPO immunodominant region and the TPO B domain is genetically transmitted. These data may open the way to the identification by candidate analysis or positional cloning of at least one gene responsible for the development of Hashimoto’s thyroiditis.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The hallmark of the humoral response in Hashimoto’s thyroiditis is autoantibodies to thyroid peroxidase (TPO), previously known as the thyroid microsomal antigen (reviewed in Ref. 1). Such autoantibodies are also present in the majority of Graves’ patients. The ability of an individual to develop TPO autoantibodies appears to be vertically transmitted (2) and is consistent with autosomal dominant transmission of a major gene on a polygenic background (3).
TPO, the major enzyme involved in thyroid hormone synthesis, is a large
(
210-kDa), homodimeric, membrane-associated glycoprotein expressed
on the surface of thyroid follicular cells (reviewed in Ref. 1).
Polyclonal autoantibodies to TPO in patients’ sera can be
quantitatively fingerprinted in terms of the spectrum of epitopes that
they recognize in an immunodominant region on the molecule (4–8;
reviewed in Ref. 9). Thus, as illustrated for selected sera (Fig. 1
), TPO autoantibody epitopic
fingerprints can focus on the A or B domain within this immunodominant
region. Other sera interact equally with both domains or with
individual subcomponents of these domains. Remarkably, the TPO
autoantibody epitopic fingerprint of an individual is stable, without
epitope spreading, at least over 15 yr (7). Fingerprints are conserved
even during the boost in autoantibody levels during the postpartum
period (6). This lack of autoantibody epitopic spreading contrasts with
the spreading described for T cell (10) and some B cell (11, 12)
epitopes, but is consistent with observations on acetylcholine receptor
autoantibodies in myasthenia gravis (13). The basis for TPO
autoantibody epitopic fingerprint conservation is not known. However,
fingerprint analyses of families with juvenile Hashimoto’s
thyroiditis probands (7) raised the possibility of genetic control of
this phenotype.
|
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The Old Order Amish families included in this investigation are part of a larger sample being followed in a genetic linkage study of bipolar affective disorder (14). The decision to study autoimmune thyroid disease in these families was not based on any hypothesis regarding the association of thyroid disease and bipolar affective disorder (and none was found) but, rather, on the availability of these kindreds. When initial testing of a randomly selected family showed a number of individuals positive for TPO autoantibodies, family members subsequently recruited into the bipolar study were tested for the presence of TPO autoantibodies.
Sera from 124 of 368 (33.7%) Old Order Amish (205 women and 163 men) were previously found to be positive for TPO autoantibodies, as detected by RIA (TMAb RIA Kit, Kronus, San Clemente, CA) (15). Seven families (I-VII) comprising at least 2 generations of TPO autoantibody-positive members were identified. These included 43 women and 20 men with TPO autoantibodies sufficiently high (>10 U/mL) for epitopic fingerprinting (see below). Within these families, 28 subjects (20 women and 6 men) could not be fingerprinted because of TPO autoantibody levels below 10 U/mL. An additional 88 members (36 women and 52 men; 48%) were negative for TPO autoantibodies. In Amish families III and IV, additional serum samples were analyzed from 6 individuals once and from 8 individuals twice 2–5 yr after obtaining the initial sample. Replacement therapy with L-T4 (Synthroid, Knoll Pharmaceutical Co., Mt. Olive, NJ) was initiated in 6 of these 14 individuals during the course of the study.
TPO autoantibody epitopic fingerprints previously described for 17 members of 4 juvenile Hashimoto’s thyroiditis families (12 women and 5 men) (7) were included in the segregation analysis. Unlike the Amish families, selection of these juvenile Hashimoto’s families (total of 17 women and 8 men) was based on identification of a proband. This particular set of families was drawn from a larger group included in an earlier study of thyroid autoantibodies (16, 17) because sera from multiple individuals were available over a 13-yr period.
Preparation of TPO autoantibody Fab for fingerprinting studies
Four TPO autoantibodies that map the four components in the immunodominant region were used in these studies, SP1.4, WR1.7, TR1.8, and TR1.9 (18). For simplification, the subdomains recognized by these autoantibodies were previously renamed A1, A2, B1, and B2, respectively (7). The antibodies were isolated, cloned, and expressed as Fab using combinatorial Ig gene libraries (18, 19, 20). As described previously (4), Fab synthesis by XL1-blue cells was induced with 1 mmol/L isopropyl-thio-galacto-pyranoside (Sigma Chemical Co., St. Louis, MO). Fab were affinity purified on protein G-Sepharose (Pharmacia Biotech, Piscataway, NJ), and their concentrations were determined by SDS-PAGE (21).
Epitopic fingerprinting of serum TPO autoantibodies
Epitopic fingerprinting of individual sera was performed by allowing recombinant Fab to compete for serum autoantibody binding to [125I]TPO (4). As described previously (7), quantitation of competition by the Fab was most accurate when sera bound 15–20% of the radiolabeled TPO. Therefore, for each serum sample we first determined the dilution required to give about 15% radiolabeled TPO binding. Duplicate serum aliquots were incubated (1 h, room temperature) with [125I]TPO (20,000 cpm) in a total volume of 200 µL. Subsequently, protein A (Pansorbin, Calbiochem, La Jolla, CA; 50 µL) was added to precipitate the antigen-antibody complex, and incubation was continued (30 min). After washing and centrifugation (25 min, 1400 x g, 4 C), supernatants were removed by aspiration, and radioactivity in the pellets was counted to determine the percentage of [125I]TPO bound.
Competition studies were performed by incubating appropriately diluted patient’s serum with [125I]TPO in the absence or presence of TPO-specific Fab (4 x 10-8 mol/L), either separately or as a pool. Antigen-IgG complexes were precipitated with protein A. Fab lack the CH2 domain of the Fc region and are not precipitated by protein A. Sera from different individuals within a family or from the same individual on different occasions were analyzed in duplicate in the same assay. Specific binding was calculated by subtraction of [125I]TPO binding by control serum without TPO autoantibodies (2–5% of the total counts per min).
Expression of data
The data are expressed in three different ways: 1) percent inhibition by the four Fab combined; 2) percent inhibition by each Fab individually, that is percent inhibition by A1, A2, B1, or B2 Fab; and 3) ratio of percent inhibition by Fab to different subdomains, A1:A2 ratio (calculated from percent inhibition by Fab A1/percent inhibition by Fab Fab A2), B1:B2 ratio (percent inhibition by Fab B1/percent inhibition by Fab B2), and A2:B1 ratio (percent inhibition by Fab A2/percent inhibition by Fab B1). These parameters were considered to correspond to eight phenotypes. Investigation of these phenotypes was based on our previous observations (4, 5, 6, 7). To avoid repetition, the background for this focus is provided with data for the segregation analyses.
Segregation analyses
To test the hypothesis for transmission of the above phenotypes
within families and whether transmission is consistent with genetic
modes of inheritance, complex segregation analyses were performed using
the unified model as implemented in the computer program POINTER (22).
These analyses are designed to test a series of nested hypotheses
regarding the possible genetic mechanisms that could be important for
the expression of a particular trait. The maximum likelihood of the
data under each hypothesis is compared to the maximum likelihood of a
competing hypothesis using a 2 test.
The unified model (as incorporated in POINTER) has four major parameters: 1) q, the frequency of the putative major susceptibility allele; 2) D, the degree of dominance of that allele; 3) H, the heritability of the polygenic component contributing to the expression of the trait under examination; and 4) T, the effect of the major susceptibility allele in the population. Evidence for transmission is assessed by comparing the likelihood of the hypothesis of no transmission with the likelihood of the hypothesis that transmission is due to a single gene with polygenic background (the so-called mixed model hypothesis). If twice the difference in log likelihood is not significantly different, the hypothesis of no transmission cannot be rejected. That is, there is no evidence for vertical transmission, and no further comparisons are made. On the other hand, if the difference in log likelihoods is significant, the hypothesis of no transmission is rejected (that is, there is evidence for vertical transmission), and further comparisons are warranted. The next two comparisons examine hypotheses that 1) a major gene alone (no polygenic background) is important for the expression of the phenotype; and 2) polygenic inheritance alone (no major locus) is consistent with the mode of transmission.
Following the same logic, evidence for the hypothesis of major locus transmission is assessed by comparing the likelihood of this model to that of the mixed model. If the difference in twice the log likelihood is not significant, the major locus hypothesis cannot be rejected. If at the same time, a similar comparison between the polygenic hypothesis and the mixed model hypothesis can be rejected, the data suggest that the most parsimonious mode of inheritance includes a major locus with no polygenic background. If both major locus and polygenic hypotheses can be rejected, then the most parsimonious mode of transmission includes both a major locus and a polygenic background.
In all cases, competing genetic hypotheses were evaluated by taking
twice the difference between the maximum log likelihood estimates for
each model. This difference in log likelihoods is distributed as a
2 statistic with degrees of freedom equal to the
difference in the number of estimated parameters in the two models.
Thus, four parameters are estimated when fitting the Mendelian mixed
model, and only one is estimated when fitting the polygenic model.
Twice the difference between the log likelihoods for the mixed and
polygenic model is distributed as a 2 with 3 degrees of
freedom.
Family ascertainment
The ascertainment of families can introduce a bias into the estimation of the likelihoods of genetic hypotheses. As described above, the seven Amish families were part of a larger group studied for bipolar disorder who were found to be positive for TPO autoantibodies, and their selection did not involve a proband. However, the four juvenile Hashimoto families had been selected on the basis of a proband (16, 17). Consequently, the segregation analyses performed included a correction for ascertainment of these four probands.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Pedigrees for Amish family I, subfamily IVb, and family III (Fig. 2) provide examples of the fingerprint
data obtained. In family I, TPO antibodies in the mother are
predominantly inhibited by Fab B1 and B2 (indicating preferential
recognition of the B domain), whereas TPO antibodies in the father are
predominantly inhibited by A1 and A2 Fab (indicating preferential
recognition of the A domain). Similarly, TPO antibodies in their
daughter preferentially recognize the B domain (B domain bias), whereas
those in their son are biased toward the A domain. The siblings of the
father have varied fingerprints, one with comparable inhibition by A1,
A2, B1, and B2 Fab (recognition of both A and B domains) and the other
with inhibition mainly by B1 Fab (predominant recognition of the B
domain). In contrast, a nuclear family from pedigree IVb illustrates
similar TPO epitopic fingerprints in all members studied. Data on a
much larger pedigree is illustrated for family III (Fig. 2). An
interesting aspect of this pedigree includes the predominant inhibition
by B1 Fab in all 4 TPO autoantibody-positive individuals in nuclear
family IIIb. In 14 individuals among families III and IV, serum was
available on more than 1 occasion over a 2- to 5-yr interval. In all
cases, TPO autoantibody fingerprints were conserved, even in
individuals receiving replacement therapy (see Materials and
Methods), as illustrated for TPO autoantibody-positive members of
nuclear families IIIa and IIIb (Fig. 3).
|
|
Analyses were performed using eight phenotypes as follows.
1) Percent inhibition by the combination of Fab A1, A2, B1, and B2. This is the inhibition produced by the four-Fab combination on the ability of an individual to recognize the TPO immunodominant region. Because the immunodominant region is recognized by all individuals and by more than 80% of autoantibodies of sera in each individual (4, 5, 6, 7), this phenotype is similar (but not identical) to TPO antibody positivity.
2) Percent inhibition by Fab A1, A2, B1, or B2. These values
indicate the ability of TPO autoantibodies in a particular individual
to recognize the A1, A2, B1, or B2 subdomain, as described above and
illustrated in Fig. 2
for some Amish families and in Ref. 7 for the
four juvenile Hashimoto families.
3) Ratio of inhibition by Fab to different subdomains, namely the A1:A2 ratio, the B1:B2 ratio, and the A2:B1 ratio (for calculations, see Materials and Methods)
Previously, we observed a correlation between the percent inhibition for A1 and A2 Fab as well as for B1 and B2 Fab, but not for A2 and B1 Fab (4). Thus, the B1:B2 ratio (or the A1:A2 ratio) is a measure of the ability of an individual to recognize the B domain (or the A domain), whereas the A2:B1 ratio is essentially a control ratio.
The data were analyzed in terms of competing genetic models in a hierarchical fashion (see Materials and Methods), and unless otherwise specified, the significance levels are corrected for multiple testing of eight phenotypes.
Inhibition by the four-Fab phenotype
For inhibition by the four-Fab phenotype, the model of no
transmission can be rejected (2(4) = 20.67;
P < 0.0032). Furthermore, the no major locus
hypothesis can be rejected (2(3) =
15.14; P < 0.016), but not the no polygenic background
model (Table 1). The most parsimonious
model includes a major gene effect, but it is not possible to
distinguish between specific single gene models (data not shown).
|
Inhibition by Fab A1, A2, B1, or B2 alone phenotype
No evidence was obtained for familial/genetic transmission of
inhibition by the individual Fab A1, A2, or B2 separately. Moreover,
although there was evidence suggesting familial/genetic transmission of
the inhibition by B1 Fab alone phenotype
(2(4) = 14.17; P < 0.007),
this value is not statistically significant after correction for
multiple testing (P < 0.056).
Ratio of inhibition by Fab to different subdomains (A1:A2, B1:B2, and A2:B1)
Consistent with the lack of a relationship between recognition of
the A2 and B1 subdomains (4), no evidence was obtained for transmission
of the phenotype A2:B1 inhibition ratio. Evidence suggestive of
transmission was obtained for the phenotype A1:A2 inhibition ratio
before (2(4) = 13.01; P <
0.011), but not after, correction for multiple testing
(P < 0.088). Consequently, the conservative
interpretation is that there is no conclusive evidence for transmission
of recognition of the TPO A domain.
With respect to the phenotype B1:B2 inhibition ratio, the no
transmission model was rejected (2(4) =
63.59; P < 0.000008; after correction for multiple
testing). Furthermore, the no major locus hypothesis could be rejected
(2(3) = 61.40; P <
0.000008; corrected for multiple tests), but the no polygenic
background hypothesis could not be rejected (Table 2). The dominant, additive, and recessive
models could all be rejected when compared to the general Mendelian
major locus hypothesis (data not shown). The data are consistent with
the hypothesis that a single major gene with intermediate dominance is
involved in the expression of the phenotype defined by the ratio of
inhibition by domain B1:B2 Fab.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Autoantibody responses to a large protein autoantigen such as TPO are polyclonal (reviewed in Ref. 1) and are likely to be controlled by several groups of genes. Consequently, it may be necessary to simplify further the phenotype defined by the ability of an individual to develop autoantibodies. For TPO autoantibodies, phenotypic simplification is now feasible based on our ability to determine the epitopic fingerprints of an individual (4, 5, 6, 7, 8). In the present study, we investigated TPO autoantibody epitopic fingerprints in 80 members of 11 families. The majority of fingerprints (63) are from 7 multiplex Amish kindreds. The Amish families are unusual in that they include large numbers of related TPO autoantibody-positive individuals whose serum levels are sufficiently high (15) to permit epitopic fingerprinting. However, the fingerprints of the Amish resemble those of other patients with autoimmune thyroid disease (4, 5, 6, 7), and their conservation, even in individuals receiving T4 replacement therapy, is also in accordance with our previous findings (6, 7). The other 17 fingerprints used in the analysis were drawn from 4 families with a juvenile Hashimoto thyroiditis proband (34). This disease of early onset occurs in a population in which, like the Amish families reported herein, many parents and siblings have thyroid autoantibodies (16, 35).
The results from complex segregation analysis provide support for the concept that TPO epitopic fingerprints are inherited. Thus, our data suggest that the ability of an individual to recognize the TPO immunodominant region as a whole and the B domain of this region in particular is under genetic control. Clearly, these findings should be extended and confirmed for other families as well as with respect to the specific modes of transmission. However, an important question raised by these data is whether any of the candidate genes involved in immune responses plays a role in TPO autoantibody epitopic inheritance.
One candidate locus, shown to control some antibody responses in humans, is unlikely to be involved. Major histocompatibility complex (MHC) human leukocyte antigen genes control autoantibody responses to epidermal cadherin in pemphigus vulgaris (36) and to DNA topoisomerase I in systemic sclerosis (37). However, we found no association between TPO autoantibody fingerprints and MHC class I or II polymorphisms in juvenile Hashimoto thyroiditis families (7). Furthermore, not all antibody responses in humans are controlled by the MHC. For example, antibody responses to a major malarial antigen are not associated with human leukocyte antigen DRB, DQA, or DQB alleles (38).
Some autoimmune diseases have been associated with polymorphisms of the Ig gene constant regions, Gm (heavy chain) and Km (light chain) allotypes (reviewed in Ref. 39). In particular, associations have been observed between Gm allotypes and susceptibility to autoimmune thyroid disease (40, 41, 42, 43). On the other hand, segregation and/or linkage analysis appear to exclude a role for Ig heavy chain genes, as measured by Gm allotypes (44) or Fc region microsatellite markers (45), in the genetic basis for autoimmune thyroid disease. However, some autoimmune responses are associated with polymorphisms of the variable (rather than constant) regions of Ig genes (46, 47). Furthermore, the human Ig heavy chain variable (VH) gene repertoire appears to be genetically controlled and unaltered by chronic stimulation (48). Consequently, as shown in other studies (49), polymorphisms of the variable regions may need to be investigated to determine whether the Ig heavy chain locus plays a role in autoimmune thyroid disease or TPO autoantibody production.
The intriguing possibility that Ig gene polymorphism may be involved in
the transmission of TPO autoantibody epitopic recognition is supported
by observations of an association between genes encoding TPO-specific
recombinant Fab and recognition of the TPO A and B domains. Thus,
TPO-Fab with 
light chains derived from the O12 germline gene
interact with the A domain of the TPO immunodominant region (18, 50).
Conversely, recognition of the B domain is associated with non-O12
light chain genes (either or 
) (8, 18, 50, 51, 52). light chains
derived from O12 genes are common in the expressed repertoire (53, 54),
and gene polymorphism is limited (53). On the other hand, heavy
chains derived from polymorphic germline genes, such as hv1263 (55) and
hv3005 (46), appear to be used by autoantibodies that recognize the TPO
B domain (8, 18). The most compelling evidence for genetic transmission
of TPO epitopic fingerprints was for phenotypes involving the B domain
of the TPO immunodominant region. Taken together, these observations
point the way for future studies.
In conclusion, complex segregation analysis provides evidence that autoantibody recognition of the TPO immunodominant region and the B domain of this region is genetically transmitted. Recently, considerably progress has been made in defining genetic loci linked to the hyperthyroidism of Graves’ (but not Hashimoto’s) disease (45, 56, 57). Consequently, as we suggested previously (7), focusing on TPO autoantibody epitopes may open the way to the identification by candidate analysis or positional cloning of at least one gene responsible for the development of Hashimoto’s thyroiditis.
|
Footnotes |
|---|
2 Equal contribution was made by both authors.
3 Present address: Endocrinology, Veterans Administration Medical
Center and the University of California, San Francisco, California
94121. Recipient of a University of California-San Francisco Molecular
Medicine Training Program Fellowship Award.
Received November 20, 1997.
Revised March 23, 1997.
Revised October 26, 1998.
Accepted January 19, 1999.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
gene repertoire of IgM+ B cells. J Clin
Invest. 99:1614–1627.[Medline]
This article has been cited by other articles:
 |
|
 |
|
  N. Cakan and D. Kamat Commentary Clinical Pediatrics, March 1, 2006; 45(2): 192 - 195. [PDF]
|
|
|
|
 |
|
 J. Guo, P. N. Pichurin, J. C. Morris, B. Rapoport, and S. M. McLachlan "Naked" Deoxyribonucleic Acid Vaccination Induces Recognition of Diverse Thyroid Peroxidase T Cell Epitopes Endocrinology, August 1, 2004; 145(8): 3671 - 3678. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Tomer and T. F. Davies Searching for the Autoimmune Thyroid Disease Susceptibility Genes: From Gene Mapping to Gene Function Endocr. Rev., October 1, 2003; 24(5): 694 - 717. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. M. Allen, W.-C. Hsueh, M. M. Sabra, T. I. Pollin, P. W. Ladenson, K. D. Silver, B. D. Mitchell, and A. R. Shuldiner A Genome-Wide Scan for Autoimmune Thyroiditis in the Old Order Amish: Replication of Genetic Linkage on Chromosome 5q11.2-q14.3 J. Clin. Endocrinol. Metab., March 1, 2003; 88(3): 1292 - 1296. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Vaidya, P. Kendall-Taylor, and S. H. S. Pearce The Genetics of Autoimmune Thyroid Disease J. Clin. Endocrinol. Metab., December 1, 2002; 87(12): 5385 - 5397. [Full Text] [PDF]
|
|
|
|
 |
|
 J. Guo, S. M. McLachlan, and B. Rapoport Localization of the Thyroid Peroxidase Autoantibody Immunodominant Region to a Junctional Region Containing Portions of the Domains Homologous to Complement Control Protein and Myeloperoxidase J. Biol. Chem., October 18, 2002; 277(43): 40189 - 40195. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Tomer, D. A. Greenberg, G. Barbesino, E. Concepcion, and T. F. Davies CTLA-4 and Not CD28 Is a Susceptibility Gene for Thyroid Autoantibody Production J. Clin. Endocrinol. Metab., April 1, 2001; 86(4): 1687 - 1693. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
