Download Genetic Susceptibility to the Development of Autoimmune Disease

ClinicalScience (1997) 93,479-491 (Printed in Great Britain)
479
Editorial Review
Genetic susceptibility to the development of autoimmune disease
Joanne HEWARD and Stephen C. L. GOUGH*
University of Birmingham, Edgbaston, Birmingham B /52 T H , U.K., and *Birmingham Heartlands Hospital,
Bordesley Green East, Birmingham B9 555, U.K.
1. Autoimmune diseases are common conditions
which appear to develop in genetically susceptible
individuals, with expression of disease being modified by permissive and protective environments.
Familial clustering and data from twin studies provided the impetus for the search for putative loci.
Both the candidate gene approach in populationbased case-control studies and entire genome
screening in families have helped identify susceptibility genes in a number of autoimmune diseases.
2. After the first genome screen in type 1 (insulindependent) diabetes mellitus it seems likely that
most autoimmune diseases are polygenic with no
single gene being either necessary o r sufficient for
disease development. Of the organ-specific autoimmune diseases, genome screens have now been
completed in insulin-dependent diabetes mellitus
and multiple sclerosis. Furthermore, the clustering
of autoimmune diseases within the same individuals
suggests that the same genes may be involved in the
different diseases. This is supported by data showing that both HLA (human leucocyte antigen) and
CTLA-4 (cytotoxic T-lymphocyte-associated-4)appear
to be involved in the development of insulindependent diabetes mellitus and Graves’ disease.
3. Genome screens have also been completed in
some of the non-organ-specific autoimmune diseases including rheumatoid arthritis, inflammatory
bowel disease and psoriasis. Many candidate genes
have also been investigated although these are
predominantly in population-based case-control
studies.
4. Substantial progress has been made in recent
years towards the identification of susceptibility loci
in autoimmune diseases. The inconsistencies seen
between case-control studies may largely be due to
genetic mismatching between cases and controls in
small datasets. Family-based association studies are
being increasingly used to confirm genetic linkages
and help with fine mapping strategies. It will, however, require a combination of biology and genetics,
as has been necessary with the major histocompatibility complex in insulin-dependent diabetes mellitus, to identify primary aetiological mutations.
INTRODUCTION
Autoimmune diseases occur when an immune
response is directed against a specific organ or a
number of organs and systems within an individual.
Failure in immune tolerance, where tissue is recognized as ‘foreign’ instead of ‘self leads to disturbed
function and often failure of a specific organ or
organs. Autoimmune diseases can be separated into
two groups - those that are ‘organ-specific’ and
those that are ‘non-organ-specific’. Organ-specific
autoimmune diseases include, for example, type 1
(insulin-dependent) diabetes mellitus, Graves’
disease, Hashimoto’s thyroiditis and multiple sclerosis. Non-organ-specific diseases include systemic
lupus erythematosus, rheumatoid arthritis, juvenile
chronic arthritis, psoriasis and inflammatory bowel
disease.
Autoimmune diseases are common within the
general population and although the causes are
largely unknown, it appears that development of
disease in genetically susceptible individuals can be
modified in permissive and protective environments.
Evidence for the role of genetic factors in these
diseases has been provided by looking at disease
clustering within families and concordance rates in
both monozygotic and dizygotic twins. As monozygotic twins share an identical genetic makeup, the
higher concordance rates in these twins compared
with dizygotic twins (see individual diseases for
details) are to be expected if genetic components
are involved in the disease process. However, concordance rates for the various autoimmune diseases
Key words: autoimmunity, cytotoxic T-lymphocyte-associated4 gene, human leucocyte antigen. inflammatory bowel disease, insulin-dependent diabetes mellitus, juvenile chronic
arthritis, multiple sclerosis, psoriasis, systemic lupus erythematosus, thyroid disease.
Abbreviations: CTM-4 gene, cytotoxic T-lymphocyte-associated4 gene; HlA, human l e u c q t e antigen; hsp, heat-shock protein; IDDM, insulin-dependent diabetes mellitus;
IL-Ira, interleukin- I receptor antagonist: JCA,juvenile chronic arthritis; LMP, large multifunctional protease; MBP, myelin basic protein; MHC, major histocompatibility complex;
MS, multiple sclerosis; SLE, systemic lupus erythematosus; TNF, tumour necrosis factor; TSHR. thyroid-stimulating hormone receptor; VNTR, Mn’able number of tandem repeats.
Correspondence:D r S. C. L. Cough.
480
J. Heward and S. C.L. Gough
are well below 100% suggesting that susceptibility to
disease is not only due to genetic components but
that other factors, such as the environment, play a
role.
The multifactorial nature of autoimmune diseases
makes the study of causative factors problematic.
After pioneering work in insulin-dependent (type 1)
diabetes mellitus it seems likely that most autoimmune diseases are ‘polygenic’ implying the
involvement of many genes [l].As no single gene
has been shown to be either necessary for the
development of the disease or sufficient to cause it,
common disease genes are, therefore, referred to as
encoding susceptibility. The involvement of more
than one susceptibility gene in the disease process
means that the effect of a single gene is likely to be
small. This concept highlights the greatest difficulty
in attempting to ‘hunt-out’ susceptibility disease
genes, namely establishing sufficiently sized datasets.
There are two approaches in humans which have
been used to try and identify susceptibility genes of
common diseases: first the candidate gene approach
and second entire genome screening. A number of
candidate genes have been hypothesized and tested,
predominantly in population-based case-control
studies. Case-control studies attempt to show the
association of a gene to disease in a ‘disease-free’
population compared with a ‘diseased’ population.
As we shall see later these have yielded contrasting
results when the same candidate has been tested in
different datasets. Such studies requiring hundreds
of patients and controls have, in practice, often been
performed on too few subjects. It is highly likely,
therefore, that non-genetic false positives or association artefacts have arisen because of random variation as the result of a chance event. It is also
impossible to eliminate population stratification [2]
from such studies and it remains unclear as to which
individuals are most suitable to act as a control
group. An alternative to the population-based casecontrol study is the family-based study which eliminates many of the problems mentioned above. The
testing of transmission of alleles of candidate genes
from parents to affected and unaffected siblings
using intrafamilial association tests such as the
transmission disequilibrium test are tests not only of
allelic association but also genetic linkage [3]. Based
on the laws of simple Mendelian inheritance, the
transmission disequilibrium test assumes that a
heterozygous parent for a genetic variant will transmit each allele with equal frequency to any offspring. A statistically significant excess of
transmission of the candidate allele to a group of
offspring with disease provides evidence of allelic
association. In order to detect single-gene effects,
large numbers of families with parents are needed.
While the family approach is more robust, collection
of familial DNA samples is much more difficult than
simply collecting patients and controls as many of
the autoimmune diseases do not manifest themselves until the patient is middle aged, by which time
one or both parents may be deceased. The collection process is also more time-consuming involving
home visits to parents and relatives often at weekends in out of work hours. However, this type of
study is particularly important in revealing transmission of disease genes through families and is a sensitive method for detecting susceptibility genes in
polygenic diseases [4-61.
The second approach of entire genome screening
and linkage analysis makes no assumption of the
gene or genes involved in the development of the
disease, disease gene frequencies or mode of
inheritance of genes. Studies already performed in
type 1 diabetes [7, 81, multiple sclerosis [9-111,
psoriasis [12, 131 and inflammatory bowel disease
[14] have made use of genetic markers (microsatellites) from existing genetic maps. The markers have
been analysed in datasets of large numbers of
families in which DNA is available from two
affected siblings with or without parental DNA.
Genetic linkage of a marker to disease is said to be
present when there is a significant excess of alleles
shared identical by descent, when parental DNA is
available for analysis, and identical by state when
parental DNA is not available. Results from the
completed genome searches have found regions of
linkage, some of which reassuringly coincide with
previous candidate loci including human leucocyte
antigen (HLA). They have also confirmed that the
common multifactorial diseases are polygenic in
nature.
Clinicians have long since noted that it is not
uncommon to diagnose more than one autoimmune
disease in a single patient; for example, autoimmune
thyroid disease is frequently seen in patients with
type 1 diabetes. Autoimmune diseases also seem to
cluster within families. This has been confirmed by
the 5th Genetic Analysis Workshop [15] with the
finding of at least one case of autoimmune thyroid
disease in 40% of type 1 diabetic families, a result
supported by data from the British Diabetic Association Warren Repository where 22% of type 1 diabetic families had at least one case of autoimmune
thyroid disease (S. C. Bain, personal communication). These results imply that the same genes may
be involved in different autoimmune diseases. The
remainder of this review presents an overview of
results obtained to date.
ORGAN-SPECIFIC AUTOIMMUNE DISEASES
Insulin-dependent diabetes rnellitus
Insulin-dependent diabetes mellitus (IDDM) is
characterized by T-cell-mediated autoimmune
destruction of the pancreatic /?-cells leading to a
deficiency in insulin production resulting in a lifelong requirement of insulin replacement by injection. The disease may also lead to the development
of complications including eye disease (diabetic
retinopathy), renal disease (diabetic nephropathy),
48 I
Genetic susceptibility to autoimmune disease
nerve damage (neuropathy) and vascular disease
(including peripheral vascular disease, coronary
artery disease and cerebrovascular disease). It is a
common childhood condition affecting around 1 in
300-400 children with a concordance rate in monozygotic twins of 36% [16, 171. Genetic susceptibility
to type 1 diabetes has been established from candidate gene and whole genome screen studies and has
recently been reviewed [l, 181. The major findings
only, therefore, will be summarized here.
makes it extremely difficult to distinguish the ‘true’
susceptibility allele from alleles in linkage disequilibrium. For example, in Caucasians DR3 and DR4
along with associated DQ alleles are increased in
patients with diabetes but it has been almost impossible to define which of these alleles is responsible
for association of the HLA region with disease. In
order to overcome difficulties due to linkage disequilibrium, studies of many different ethnic groups
have been performed using the hypothesis that the
‘true’ susceptibility allele of the MHC class I1 region
will be present in all races. Results from studies
combining both genetic and biological approaches
have shown conclusively that the HLA DQA1,
DQBl and DRBl encode primary aetiological
determinants in the MHC diabetic locus. However,
in some ethnic groups, additional susceptibility loci
must lie outside the class I1 region because DR3, for
example, can be stratified by class I polymorphisms
[191. Genuine differences in associations between
different genes in the MHC and type 1 diabetes may
be seen in various ethnic groups because of differences in allele frequencies at the susceptibility locus
and at other interacting loci that may be unlinked.
Cross-ethnic group comparisons are therefore only
informative when results are positive. A negative
result in one ethnic group does not mean that positive results in other ethnic groups are false.
Other genes within the MHC that have been suggested as candidate genes for IDDM determinants
include the transporter associated with antigen processing (TAP) genes [20, 211 and the large multifunctional protease (LMP)genes [22]. TAP is a
Major histocompatibility complex (MHC) class II region
Major histocompatibility complex (MHC) class I1
molecules are assembled in the endoplasmic reticulum and are transported through the Golgi and
trans-Golgi reticulum to the cytosol where they take
up degraded antigen, and are transported to the cell
surface to present the antigen to CD4+ T-cells.
These T-cells attack foreign antigen presented by
the MHC class I1 molecules either directly or by
activating cells such as macrophages. This interaction between MHC class I1 molecules and T-cells
makes the MHC class I1 region a strong candidate
region for diseases which develop as a result of
T-cell-mediated autoimmunity.
The HLA class I1 region of the MHC on chromosome 6p21 (IDDMI) (Fig. 1) accounts for 35% of
familial clustering in IDDM and is by far the major
susceptibility locus [l]. Many studies have been performed on the DRB1, DQAl and DQBl regions
which are in strong linkage disequilibrium [19]. This
Class Il region
TAP
DP
DN R
B2A2BlAl
A 3
DM LMP
LMP
9 2 A2 9 3 91 Al
21921 B
AB
DR
DQ
RTAP DO
91 92 9 3
99 A
I000
Class I11 region
c4
Bf
TNF
HSP7O
C2
2
1
2000
A B
Class I region
B
C
X
E
J
A
H
G
F
Fig. 1. MHC class I, II and 111 genes. TAP, transporter associated with antigen processing; LMP, large multifunctional protease; HSP70, heat-shock protein 70; TNF,
tumour necrosis factor.
402
J. Heward and S. C.L. Gough
heterodimer composed of TAP1 and TAP2 which
translocates degraded antigens into the endoplasmic
reticulum for antigen processing by MHC class I
molecules. TAP is polymorphic and situated adjacent to the HLA-DQ locus. Results from many different races including Sardinians, Finnish,
Norwegians and Japanese have shown no primary
association between TAP and IDDM and the differences in frequency of the TAP alleles appear to be
the result of linkage disequilibrium with DQ [23,
241. LMP2 and LMP7 are subunits of a subset of
proteasomes which are large molecular assemblies
with multiproteolytic activities believed to degrade
damaged and unwanted cellular proteins. Recent
work by Deng et al. [22] indicates a role for LMP7
in IDDM with an increase in allele A being
observed in both case-control and family studies.
The authors claim an association independent of
linkage disequilibrium with the D R D Q region. Subsequent work by Undlien et al. [25] has, however,
shown no independent association of either LMP2
or LMP7 polymorphisms with susceptibility to
IDDM. The discrepancy in results between these
two studies appear to be due to the failure of Deng
et al. [22] to include the information from DR4 subtyping.
A combination of biology and genetics has been
necessary to determine primary aetiological determinants in the MHC. Such an approach, which has
also been adopted for the insulin gene locus, will
undoubtedly be necessary at other loci to pinpoint
aetiological variants.
unknown, insulin gene expression appears to be
influenced by the VNTR genotype [26].
insulin gene region
Other genes
The insulin gene (INS) region (IDDMZ) on
chromosome llp15.5 contributes around 10% of the
familial clustering seen in type 1 diabetes [l].
Having originally been postulated as a candidate
gene for type 2 (non-insulin-dependent) diabetes,
allelic variation at the INS VNTR (variable number
of tandem repeats) was first shown to be associated
with type 1 diabetes in case-control studies and later
linked in family-based studies (for review see [26]).
IDDM2 has been mapped to a 4.1 kb region surrounding INS [27] and after the identification of 13
common polymorphic sites, the susceptibility locus
has been shown to lie within the VNTR [5]. There
are three classes of VNTR, classes I, I1 and I11 averaging 570, 1200 and 2200 bp respectively. Although
susceptibility to type 1 diabetes may be conferred by
the shorter class I VNTR alleles, population studies
show that class I11 VNTR alleles are dominantly
protective and are associated with a 60-70% reduction in the risk of developing the disease [26]. Further interest in this region relates to an observation
that transmission of predisposing INS VNTR alleles
to type 1 diabetic offspring is related to parent of
origin [26]. Although the mechanism linking variation at the INS VNTR and type 1 diabetes remains
Recent genome-wide searches have identified 14
further putative loci as contributing to type 1 diabetes [l, 181. Additional family studies are under
way to confirm these loci, and fine mapping strategies are ongoing to ultimately identify specific
mutations causing disease. Table 1 lists the current
replicated IDDM loci.
Cytotoxic T-lymphocyte-associated-4 gene
As previously mentioned, certain susceptibility
genes may be common to more than one autoimmune disease. This could explain the clustering of
different autoimmune diseases within families and
the same individuals. The cytotoxic T-lymphocyteassociated4 (CTLA-4) gene on chromosome 2q33
was first identified as a candidate gene in Graves’
disease [28] but is an equally strong candidate for
other T-cell-mediated autoimmune diseases. Recent
studies in vifro and in v i m have shown that CTLA-4
may downregulate T-cell function (for review see
[29]). Further work has shown that CTLA-4 encodes
a T-cell receptor that may inhibit CD28-dependent
interleukin-2 production [30]. A recent study by Nistic0 et al. [31], examining an A to G polymorphism
in the leader peptide of exon 1 of the CTLA-4 gene,
showed an increased frequency of the G allele in a
Belgian case-control study and increased transmission of the G allele to diabetic offspring in Italian
and Spanish family studies. Failure to replicate in
additional family sets from the U.K., U.S.A. and
Sardinia is most likely to be the result of heterogeneity between the different ethnic groups studied.
Increasing evidence in Graves’ disease, however,
suggests that CTLA-4 is a susceptibility locus for
autoimmune disease and therefore may well play a
role in type 1 diabetes.
Graves’ disease
Graves’ disease is an autoimmune disease of the
thyroid gland which is characterized by an overactive
thyroid gland (hyperthyroidism), a diffuse goitre and
in some cases ophthalmopathy and pretibial myxoedema. The frequency and severity of the symptoms,
however, varies between individuals. Overactivity of
the gland is the direct result of the stimulating effect
of an autoantibody directed against the thyroidstimulating hormone receptor (TSHR) [32]. The
disease usually presents in the fourth decade of life
and is reported to be 7-10 times more common in
women than men. The reason for the female preponderance is unknown but as with other organspecific autoimmune diseases a number of hypotheses
Genetic susceptibility to autoimmune disease
Table 1. IDDM susceptibility loci. INS, insulin gene; GCK, glucokinase
gene. See text for references.
Locus name
Chromosome location
IDDMI [HM)
IDDM2 [INS)
IDDM3
IDDM4
6p2 I
I lp21
lDDM5
IDDM6
IDDM7
IDDMB
IDDM9
IDDMI0
IDDMI I
CTIA-4 [IDDMIZ)
IDDMI3
GCK
1%
I lq13
$25
18q
2q3 I
6q27
3q21-q25
lop1 1.2-ql1.2
14q24.3-q3 I
2q33
2q34
7P
have been put forward. First, susceptibility loci may
be located on the sex chromosomes. Currently there
is no published evidence to substantiate this.
Second, females may have an increased immune
responsiveness. Third, observed differences may be
related to differences in sex hormones. Testosterone
suppresses and oestrogens exacerbate experimental
autoimmune thyroiditis. Therefore, a stronger
genetic influence may be required to overcome the
inhibitory effects of testosterone for Graves’ disease
to develop in males.
As with type 1 diabetes, evidence for the existence
of genetic susceptibility associated with permissive
and protective environments is supported by epidemiological data [33]. Although most cases of Graves’
disease occur in unrelated individuals (sporadic),
10-20% appear to cluster within families. The concordance rate in identical (monozygotic) twins
appears to be in excess of 30%, but falls well short
of 100%. The concordance rate in non-identical
twins is around 3-9%. The dissection of susceptibility genes in Graves’ disease is not as advanced as
that seen in type 1 diabetes and most reports have
concentrated on candidate genes in populationbased case-control studies.
MHC class II region
The target of the autoimmune process in Graves’
disease is the follicular cell. This, and activated
lymphocytes in patients with Graves’ disease exhibit
aberrant expression of MHC class I1 antigens
including the DR antigen. The MHC HLA region
on chromosome 6 is therefore an obvious candidate
for Graves’ disease.
Almost all studies examining the relationship
between the HLA region on chromosome 6 and
autoimmune thyroid disease are population-based
case-control studies. Most have found associations
with specific alleles of the class I1 region. It is likely
that inconsistent results from early studies and those
483
from more recent reports describing independent
associations within the class I1 region, differences
between male and female populations and age at
onset of disease have arisen because of datasets with
too few subjects.
While some of the reported differences between
populations from different races and geographical
locations may also be the result of a chance finding
in a small dataset, it is highly likely that differences
do exist [34, 351, as has been seen in the larger
family studies in type 1 diabetes [l].
There is increasing evidence supporting an association between Graves’ disease and HLA-DR3 at
least in Caucasian populations. However, DR3 is in
strong linkage disequilibrium with DQB1*0201 and
DQA1*0501, both of which are strongly associated
with Graves’ disease [36, 371. As with type 1 diabetes it will be extremely difficult to determine the
primary susceptibility locus. Studies implicating
independent HLA associations and differences
between males and females need to be repeated
with larger numbers of subjects before meaningful
conclusions can be drawn.
There are very few family-based studies available
for review [38-401. Those that have been reported
have failed to replicate, with linkage analysis, results
obtained in population-based case-control studies
[38]. The reason for this is almost certainly the
result of too few families within the datasets with
studies lacking the power to detect linkage. Moreover, the largest reported study to date also included
families of different ethnic backgrounds [38]. However, a combined segregation and linkage analysis of
patients with Graves’ disease and associated autoantibody status showed linkage to HLA-DR [39].
A more recent family-based study implicates a
role for DPBl in distinguishing autoantibody-positive family members who develop Graves’ disease
(DR17-DQ2-DPB1*0101) from those who would
remain euthyroid (DR17-DQ2-DPB1*0401) [40]. As
approximately 10% of the general population have
thyroid autoantibodies, but not clinical thyroid dysfunction, replication of the findings of Ratanachaiyavong et al. [40] in further datasets is needed to
determine whether HLA typing will help identify
autoantibody-positive individuals likely to develop
thyroid disease.
CTLA-4 gene
Yanagawa et al, [28] looked at exon 3 of the
C T U - 4 gene region which contains an (AT)n
repeat in the 3’ untranslated region and observed 21
alleles ranging in size from 88 to 134 bp. A significant excess of allele 106 was observed in patients
with Graves’ disease in a case-control study. The
result was most evident in females with protective
HLA haplotypes (DQA1*0201 positive/DQA1*0501
negative). Although the number of AT repeats may
be important for mRNA stability, the disease sus-
484
J. Heward and S. C.L. Gough
ceptibility mutation within CTLA-4 gene remains
unknown.
As previously mentioned, Nistico et al. [31]
examined an A to G polymorphism in exon 1 (which
is in linkage disequilibrium with allele 106 in exon 3)
of CTLA-4 and found association of the G allele
with IDDM. This result was replicated in a casecontrol study of Hong Kong Chinese patients with
Graves’ disease [31]. Similar associations have since
been reported in Caucasian Graves’ subjects from
Germany, Canada [41] and the U.K. [42], indicating
that this may be a common gene to both type 1 diabetes and Graves’ disease.
Thyrotropin (thyroid-stimulating hormone) receptor
gene
The fact that Graves’ disease develops as a result
of antibodies stimulating the TSHR [32], makes the
TSHR gene an obvious candidate for the disease.
Polymorphism at the first position of codon 52
(C52-A52) changing a proline into a threonine has
been described [43]. Although early case-control
studies report association of this polymorphism with
disease [44], this result has not been replicated in
other datasets [45]. Combined segregation and linkage analysis using three microsatellite markers
within the TSHR gene introns in a family-based
study failed to provide evidence for genetic linkage
to Graves’ disease [46]. Although this represents one
of the largest family studies in Graves’ disease to
date, too few affected sib-pairs were available to
completely exclude an effect on disease susceptibility. From our knowledge of type 1 diabetes it is
likely that Graves’ disease is a polygenic disorder in
which each gene contributes between 5 and 10% to
genetic susceptibility. It is unlikely, therefore, that
datasets the size of that used by De Roux et al. [46]
will be of sufficient size to exclude most genetic
determinants to Graves’ disease.
Tomer et al. [47] recently reported linkage analysis of eight candidate gene regions including the
TSHR gene on chromosome 14q31. This study
looked at 109 family members from 19 Caucasian
North American and Italian families, which included
only 14 patients with Graves’ disease and 32 with
Hashimoto’s thyroiditis. The microsatellite Dl4S8l
gave the highest positive LOD score, although this
marker is a considerable distance from the TSHR
gene (approximately 25 centiMorgans). This result,
which needs replicating, places a Graves’ susceptibility gene (GD-1) in a similar region to ZDDMll, but
it is unlikely that either of these loci correspond to
the TSHR, and therefore it is unlikely that the
TSHR is the gene in this region responsible for conferring primary susceptibility to disease.
Interleukin- I receptor antagonist gene
Interleukin-1 (IL-1) can modify the functions of
thyroid cells in v i m . It is produced in the mono-
nuclear infiltrate by thyroid cells and there is evidence that IL-1 produces fibroblast activation in
thyroid-associated ophthalmopathy. The interleukin1 receptor antagonist (IL-lra) is a 22-25kDa protein that is related to IL-lcr and IL-1p. Although
IL-lra competes with IL-lcr and p to occupy cell-surface receptors, it does not stimulate signal transduction thus inhibiting IL-1 action.
A VNTR in intron 2 of the ZL-lru gene (ILlRN)
gives rise to five alleles [48]. The two common
alleles, ILlRN*l (which has four copies of an 86 bp
repeat) and ILlRN*2 (which has two copies of the
same repeat) account for 95% variability at this
locus. ILlRN*2 is a marker of chronic inflammatory
disease and is increased in patients with Graves’
disease [48]. This could be due to linkage disequilibrium with other genes on chromosome 2 or it
could be functionally significant as each 86 bp repeat
has possible transcription factor binding sites. This
association has only been detected in a case-control
study. Until the association of ILlRN*2 is shown in
family studies, population stratification and genetic
mismatching between cases and controls cannot be
excluded as a likely explanation. No association has
been found between ILlRN*2 and thyroid antibody
levels, or other clinical features of Graves’ disease.
Hashimoto’s thyroiditis
Hashimoto’s thyroiditis is characterized by hypothyroidism, a diffuse goitre and the presence of
autoantibodies to thyroglobulin and thyroid peroxidase. Few genetic studies have been performed and
those that have generally look at the HLA region.
Again, conflicting results have been obtained [38,
49-51]. The first studies performed in Caucasians
showed non-significant increases in DR3 and DQw2
and no association with DPB in a case-control study
[49]. These results were not replicated in a subsequent case-control study that found no association
with DRBl or DQBl but an association with
DQA1*0402 [50]. The same authors also performed
a meta-analysis of several previous studies and
demonstrated weak, positive associations between
disease and DR3 and DR4, suggesting that their
lack of association with DRBl and DQBl could be
due to sample size. A family-based study found DR5
(DRll+DR12) to be increased in family members
but all LOD scores were negative leading the
authors to suggest that it is unlikely that Hashimoto’s thyroiditis is linked to the HLA region in Caucasians [38]. However, as previously mentioned, the
numbers of subjects in this study were small. A
Japanese case-control study showed an increase in
DRB4*0101 and HLA-A2 with the presence of both
alleles conferring increased risk of disease [51].
DQA1*0102 was decreased suggesting a putative
protective role. Further studies need to be performed to elucidate the true role of HLA in this
disease along with further genetic studies looking at
the roles of other candidate genes.
Genetic susceptibility to autoimmune disease
In addition to showing allelic association of a
microsatellite of C T U - 4 (allele 106) with Graves’
disease, association has also been reported with
Hashimoto’s thyroiditis [42]. This association was
reported, however, in a small case-control study and
needs replicating in family-based studies.
Multiple sclerosis
Multiple sclerosis (MS ) is a chronic inflammatory
demyelinating disease of the central nervous system
that results in a number of sensory and motor
neurological manifestations. The disease probably
results from T-cell-mediated destruction of the myelin sheath in genetically susceptible individuals.
Epidemiological evidence provides good support
for a genetic basis to the development of MS.
Increased familial risks range from 30% for monozygotic twins [52] to 3-4% for first-degree relatives
[53]. By screening 15 000 individuals with MS, Ebers
et al. [54] reported an increased frequency of
disease among first-degree biological relatives. The
frequency of MS among first-degree non-biological
relatives (adopted relatives), however, was no
greater than that of the expected population prevalence. These data imply no detectable shared
environmental effect on disease aetiology. In the
only study of conjugal pairs of MS the crude recurrence in children rose to 1 in 17 [%], which is significantly higher than reported population-based risks
for offspring of single affected parents (1 in 200),
demonstrating the importance of the inheritance of
genetic susceptibility loci from both parents.
MHC class II region
Much attention has been paid to the roles of the
genes within this region in susceptibility to MS in
many different races and, as in the other diseases,
reviewed results are conflicting [56, 571. Associations
with DR2 have been reported in numerous Caucasian populations. Increases in DQAl*O102 and
DQB1*0602have also been observed with 96% of MS
patients in one study having a DQA allele with glutamine at residue 34 and a DQB allele encoding DQB
chains sharing long polymorphic stretches [58]. However, no excess of DQB1*0302 and DQB1*03032
was observed in patients with MS; these alleles
share hypervariable regions with DQB1*06, implying
that if MS is related to particular DQB alleles it
does not appear to be due to the hypervariable
regions of these alleles. Further studies have placed
associations to DQBl sequences and position 34 of
DQA as secondary to linkage disequilibrium with
the haplotype DRB1*1501-DQA1*0102-DQB1*0602
due to the absence of over-representation of the
DQA/DQB heterodimers in DR15-negative patients
[591*
405
Myelin basic protein
The myelin basic protein (MBP) gene is on
chromosome 18 and consists of seven exons. A study
looking at two tetranucleotide repeats 5‘ to exon 1
of the MBP gene in Shanghai Chinese revealed that
no allele was associated with MS [60]. However
when the (TGGA), polymorphism 5’ to the MBP
was studied in Danish patients with MS, three different band patterns were observed with the 450 bp
band being significantly increased in patients with
MS [61]. Further studies in Danish patients with MS
have revealed other polymorphisms in this region
which were also associated with MS [62]. These data
from case-control studies suggest that polymorphisms of the MBP gene may play a role in MS but
further studies in families are required for confirmation.
Other susceptibility loci
Several genome screens have recently been completed in families with MS. Sawcer et al. [9] found
linkage at chromosome 17q22 and 6p21 (MHC)
along with several other regions using affected sibpairs, but could not replicate susceptibility loci in
extra datasets. The Multiple Sclerosis Genetics
Group using affected sib-pairs and affected relative
pairs identified 19 regions that may be important in
MS including the MHC but no locus generated
overwhelming evidence of linkage [lo]. Finally,
Ebers et al. [ l l ] identified five loci on chromosomes
2,3,5,11 and X but found no evidence of linkage in
the HLA region. However, a marker just outside the
HLA region showed significant evidence for linkage
disequilibrium in all datasets studied. These studies,
although providing candidate regions which can be
followed up in future genetic studies, highlight the
difficulties faced by geneticists when attempting to
replicate the findings of a polygenic trait in different
populations.
Other diseases
Other organ-specific autoimmune diseases include
myasthenia gravis and Addison’s disease. The autoimmune process in myasthenia gravis results in postsynaptic blockade of neuromuscular conduction by
autoantibodies directed against the acetylcholine
receptor. There is a 40% concordance rate between
monozygotic twins and an increased incidence of
disease in relatives of patients with myasthenia
gravis [63]. An increased frequency of autoimmune
thyroid disease has been reported in families of
patients with myasthenia gravis providing further
evidence for clustering of autoimmune diseases [64].
Candidate genes for myasthenia include the MHC
[65], immunoglobulin genes [66], T-cell antigen
406
J.Heward and S. C.L. Gough
receptor genes [67] and the acetylcholine receptor
gene. The role of the gene encoding the CI subunit of
the acetylcholine receptor has been investigated
[68]. Two polymorphic sites, H B and BB, within the
first intron have been identified. In a family-based
study, the HB*14 allele was consistently transmitted
to the affected offspring from the parent implicating
a dominant role for the allele in disease susceptibility [68]. Importantly, this result has been replicated
by Heckmann et al. [69].
Addison’s disease results from the autoimmune
destruction of the adrenal glands leading to dysfunction of steroidogenesis. The disease can develop
independently or as part of autoimmune polyglandular syndrome type I1 in which it occurs with other
autoimmune diseases including IDDM and autoimmune thyroid disease. Associations have been
found with HLA-B8, DRB1*0301, DQA1*0501 and
DQB1*0201 [70-721. The major autoantigen in
Addison’s disease, steroid 21-hydroxylase, has a
functional gene (CYP21B) and a pseudogene
( C W 2 I A ) located in the HLA class I11 region on
chromosome 6 [73]. A primary role for polymorphic
sites within these genes is difficult to ascertain as a
result of strong linkage disequilibrium with other
genes within the HLA region including class I1 loci.
NON-ORGAN-SPECIFICAUTOIMMUNE DISEASES
Systemic lupus erythematosus
Systemic lupus erythematosus (SLE) is characterized by many abnormal immune reactions and clinical symptoms. Autoantibodies are produced against
double-stranded DNA, intracellular ribonucleoproteins, haematological cells and phospholipids. Autoantibody profiles differ among patients but the same
pattern is consistent in individuals. Some of the
autoantibodies produced are directly pathogenic, for
example, anti-erythrocyte, antiphospholipid and
antiplatelet antibodies, while the antinuclear antibodies can cause disease due to the tissue damage
that results from the immune system trying to
neutralize their actions. SLE can occur at any age
but the second to the fifth decade of life is the most
common time for the disease to manifest itself.
Although 10-12% of patients have first-degree relatives with SLE [74] affected family members usually
present at a similar time point rather than age suggesting the involvement of external environmental
triggers [75]. Although a number of environmental
stimuli have been postulated including sex hormones, UV light, viral infections and diet, genetic
susceptibility is highly likely with 1.7-3% of firstdegree relatives of patients developing SLE compared with 0.2-0.3% in the general population [74,
751. Concordance rates also support the involvement
of genetic factors in SLE with a rate of between 24
and 69% in monozygotic twins compared with 2-9%
in dizygotic twins [76].
MHC class II region
Early reports looking at the involvement of HLA
in SLE linked DR2 and DR3 to disease but later
reports indicated that these alleles have an increased
association with the production of autoantibodies
rather than with the disease itself [75]. The production of antiRo (Ro is a small nuclear and cytoplasmic ribonucleoprotein of unknown function) and
antiLa (La is a transcriptional termination factor for
RNA polymerase) correlate with the presence of
DR2, DR3, DQ1 (DQ5 and DQ6) and DQ2 [77,
781. Autoantibodies to Ro are found in 25-50% of
patients with SLE and autoantibodies to La usually
accompany them. The highest levels of antiRo and
antiLa antibodies are found in patients who are
heterozygous for D Q 5 D Q 6 and DQ2 suggesting
that the production of these autoantibodies is more
dependent on D Q than D R [77, 791. Other autoantibodies are present in patients with SLE but the
HLA associations with these are less clear. Antiphospholipid antibodies are found in 34-44% of
SLE patients and an increased frequency of
DR7DR4-positive patients carry anticardiolipin
antibodies [80]. These are in linkage disequilibrium
with DRB4 suggesting that this may be the primary
association. Studies of different races have identified
a-myriad of HLA associations with SLE: DQAl*
0501 (in linkage disequilibrium with DR3) being
strongly associated in Scandinavians [81], DR15
(subtype of DR2) in Southern Chinese [82] and
DR3 and DR2 in Germans [83]. These results indicate that both DR3 and DR2 (primarily 1501 and
1601) and their associated D Q alleles play a role in
SLE. Increases have also been observed in
DPB1*0101 but again this is in linkage disequilibrium with DR3. Unfortunately all results to date
have been obtained from population-based casecontrol studies and need to be confirmed in familybased studies.
Heat-shock protein genes
Heat-shock proteins (hsp) are highly conserved
proteins synthesized after stressful stimuli. The
expression of hsp90 is found to be increased in the
mononuclear cells of about 25% of patients with
SLE and antibodies to this protein are detected in
patients with SLE. Those patients with increased
antibody production are more likely to have renal
disease and low C3. Another heat-shock protein
thought to play a role in SLE is hsp70. The hsp70-1,
hsp70-2 and hsp-horn genes produce products with
highly similar sequences but they differ in their
regulation. Hsp70-2 encodes a protein functionally
relevant to antigen processing and has been implicated in autoimmune disease in Caucasians. Associ-
Genetic susceptibility to autoimmune disease
ation of a polymorphism (A to G transition) in the
coding region of the hsp70-2 gene with SLE in African Americans independent of DR3 or the C4A
deletion has been reported in a case-control study
[84]. This awaits confirmation in families.
Tumour necrosis factor gene
Tumour necrosis factor (TNF) is an inducible
cytokine with a broa; range of actions including
increased HLA class I and I1 expression and
increased B- and T-cell proliferation. Macrophages
are the major source of TNF but it is also present in
other cells such as skin cells and B- and T-cells. A
polymorphism (guanosine to adenosine substitution)
in the promoter region, giving rise to a rare TNF
gene allele, has been found to be increased in
patients with SLE in a case-control study, but is
probably due to linkage disequilibrium with DR3
[85]. This is supported by the observation that TNF
is more strongly associated with autoantibody production than with the disease itself but only when in
association with DR3. This suggests that the TNFa
polymorphism, TNF2, plays a role in susceptibility
to disease on a DR3 haplotype but not independently.
Rheumatoid arthritis
Rheumatoid arthritis was originally described as a
chronic inflammatory disease of peripheral joints
but is now recognized as a chronic or subacute systemic inflammatory disorder. Concordance rates in
monozygotic twins vary between 12 and 30% and a
2-3-fold excess of disease in females has been
reported. The HLA region has been linked to
disease and is thought to account for half of familial
rheumatoid arthritis and a fifth of rheumatoid arthritis in the population [86]. HLA-DRB1*04 is associated with disease in Caucasians with the
DRBl*O401/0404 genotype carrying a higher risk of
development of more severe forms of the disease
especially in young men [87]. A more recent study
[88] found that DRB1*04 or DRB1*01/04 was only
related to disease in seropositive patients and that
rheumatoid factor was a better predictor of disease
severity than HLA subtype, suggesting that the
effect of these alleles on severity of disease may be
linked to seropositivity. The positive association with
disease of DRB1*04 and DRBl*Ol has been replicated but once more no relation between these
alleles and severity of disease was noted [89]. Again
small numbers in some of these studies may explain
differences in results. Reported associations
between a polymorphism of TNF, T-cell receptor a
and loci are probably all secondary to the role
played by HLA-DRB1*04 [go]. Recently rheumatoid
arthritis has been linked to the N M P l gene (a
macrophage resistance gene) on chromosome 2q35
487
[91] and a genome screen has confirmed linkage of
rheumatoid arthritis to HLA [92].
Inflammatory bowel disease
The chronic inflammatory bowel diseases including Crohn’s disease and ulcerative colitis have
unknown aetiology but evidence suggests that
genetic factors play a role in predisposition to these
diseases. Although these diseases predominantly
involve the bowel, other systems are also affected,
including liver, joints and skin. The HLA region has
been implicated in ulcerative colitis with an affected
sib-pair study providing evidence for linkage with
the HLA-DRB1 region [93]. The same study showed
association of HLA-DRB1*0103 and DRB1*12 with
ulcerative colitis but no association of Crohn’s
disease with any HLA genes. These findings also
indicated that the HLA-DR3-DQ2 haplotype predicts extensive ulcerative colitis, which is at odds
with earlier work indicating that this haplotype has a
protective role in the disease [94]. A recent genome
screen [14], however, has identified a susceptibility
locus for Crohn’s disease on chromosome 16. This
result has since been replicated in a further dataset
[951.
Psoriasis
Psoriasis is an inflammatory skin disease characterized by red, scaly skin patches usually on the
scalp, elbow and knees. Different types of arthropathy are also seen in patients with psoriasis which at
times can be virtually indistinguishable from
rheumatoid disease. The usual age of onset is
15-30 years of age and it affects 2% of the population. The disease has a large genetic component
with concordance rates in monozygotic twins of
65-70% compared with 15-20% in dizygotic twins.
The risk to first-degree relatives is between 18 and
23% and inheritance of the disease fits an autosomal
recessive model in these relatives [96]. Environmental factors are also thought to play a role and
these include streptococcal infection and stress.
MHC class I1 region
The disease has been subdivided into two types;
type I has an early onset and a positive family
history and type I1 has a late onset and is sporadic
with no family history. Type I psoriasis is associated
with the Caucasian extended HLA haplotype
Cw6-B57-DRB1*0701-DQA1*0201-DQB1*0303
with
the class I antigens (Cw6-B57) being associated to a
much higher extent than the class I1 alleles
(DRB1*0701-DQA1*0201-DQB1*0303) [971, Suggesting that a gene for familial psoriasis is associated
with the class I side of the extended haplotype. It is
still unknown as to whether the susceptibility gene
488
J. Heward and S. C.L. Gough
lies within the HLA class I genes or another gene in
close linkage disequilibrium.
DQA1*0501-DQB1*0301 and DRBl*08-DQAl*
0401-DQB1*0402 as susceptibility alleles and
DRB1*07, DQA1*0201 and DQB1*0201 as protective alleles.
Other genes
A microsatellite marker on chromosome 17q was
shown to be in linkage with psoriasis in a large
extended white American family [12], although
studies performed in extended kindreds from
Northern Europe have failed to replicate this result
[98, 991. A further locus on chromosome 4q has
been reported in extended families from Ireland and
the UK [13]. This finding awaits replication. In a
recent genome-wide search using largely the 260
microsatellite markers employed in the genome
screen of type 1 diabetes, four regions of preliminary linkage were identified on chromosomes 2, 8
and 20 and significant linkage was demonstrated
with markers from the MHC region at 6p21 [loo].
As in type 1 diabetes this comprehensive screen
indicates that a gene or genes located within the
MHC region are conferring the greatest single effect
on the susceptibility to disease.
Juvenilechronic arthritis
Juvenile chronic arthritis (JCA) is an inflammatory disease starting in children under the age of
16 years and lasting at least 3 months. There are two
major types of JCA, pauciarticular and polyarticular,
depending on the clinical presentation and number
of joints affected. A further subgroup has a systemic
onset with fever and a rash. Thirty-five percent of
patients who start with pauciarticular JCA go on to
develop polyarticular JCA, which leads to severe
joint destruction and disability.
MHC class II region
The associations of JCA with the HLA region
tend to vary according to the type of JCA being
studied. Those with persistent pauciarticular JCA have
been reported as having increased DRBl*1301,
DRB1*0801, DQB1*0603,DQBl*04 and DPB1*0201
and decreased DRB1*0701 [101, 1021. Those with
polyarticular JCA are reported to show an association with DRB1*0801, DQB1*04 and DPB1*0301
with a decrease in DRB1*04 [loll. A further study
into seropositive and seronegative polyarticular JCA
confirmed the roles of DRB1*0801, DQB1*04 and
DPB1*0301 in this subgroup of the disease [103].
Studies performed on the subgroup with systemic
onset JCA have implicated DRB1*04 as playing a
major role in this disease. Those patients with polyarticular JCA that had a pauciarticular onset are
thought to have an independent risk from
DQA1*0101 [loll. Studies on the early onset pauciarticular subgroup have identified DRB1*11-
Other genes
Other genes implicated in JCA include the T-cellreceptor variable gene and the IL-1 gene. The
T-cell-receptor gene has an M chain (TCRa) and a j?
chain (TCRb). The Tcrb gene complex is on
chromosome 7 and is composed of variable, diversity, joining and constant regions. The variable
region (Tcrb-V) has 20 subfamilies each with
between 1 and 10 members. Studies have identified
a polymorphism which correlates to the Tcrb-V6.1
gene which is associated with JCA in patients with
the DQA1*0101 allele [104]. Polymorphism (C to T
change at position 889 ) in the promoter of IL-lci
has been shown to be increased in JCA [105]. This
polymorphism was increased in DPZpositive
patients and decreased in DRS-positive patients.
In common with many of the diseases reviewed,
most of the studies reported to date in JCA are
population-based case-control studies on small
numbers of subjects. Although the age of presentation makes the collection of families possible, there
is a considerable degree of phenotypic heterogeneity. Any genetic analysis will therefore have to
take this into account.
CONCLUSIONS
Genetic susceptibility to the development of autoimmune disease is a complex subject with many different genes and their products interacting with each
other and external stimuli. A summary of results
found to date can be seen in Tables 2 and 3. Certain
gene regions including HLA and probably CTLA-4
are likely to cause susceptibility to more than one
autoimmune disease thereby helping to explain the
clustering of diseases within the same families and
individuals.
Conflicting results have been obtained in many
different datasets. This is partially due to the different racial and ethnic origins of the subjects but
also the inadequate numbers studied and poor
phenotypic characterization in heterogeneous diseases. Inconsistencies between population-based
case-control studies are largely the result of genetic
mismatching between cases and controls, highlighting the need for genetic studies to be carried
out primarily in families. Many studies have used
numbers that are too small to detect single-gene
effects in polygenic diseases. This applies to both
case-control and family-based studies.
The combined approach of genome-wide searches
in families, and targeting candidate genes in both
family studies and population-based case-control
Genetic susceptibility to autoimmune disease
489
Table 2. Summary of organ-specific autoimmyne disease susceptibility loci. TAP, transporter assou’ated with antigen processing;
IMP, large multifunctional protease; INS, insulin gene. See text for references.
~
HlA
TAP
LMP
INS
crlA-4
IDDM1-I 0
It-Ira
~~
IDDM
Graves’ disease
Hashimoto’s thFoiditis
Myastheniagravis
Addison’s Disease
MS
t
t
t
t
t
t
t
t
t
t
t
t
t
t
t
Thptropin receptor
Iggenes
T-cell receptor
Acet$choline receptor
Steroid 21-hydroxylase
MBP
Proteolipidprotein
t
t
t
t
t
t
t
t
Table 3. Summary of non-organ-specific autoimmune disease susceptibilii lod. TAP, transporter associated with antigen
pmcesing; IMP, large multifunctional protease. See text for references.
SLE
HLA
t
TAP
IMP
T-cell receptor
Complement
Heat-shock proteins
TNF
It-I
t
t
t
t
Psoriasis
Inflammatory bowel disease
t
t
t
Rheumatoid arthritis
t
t
t
t
studies, has a h w e d geneticists to make substantial
progress towards the identification of susceptibility
loci in autoimmune diseases. The greatest advances
have been made in IDDM although progress is
being made in MS, inflammatory bowel disease,
psoriasis and rheumatoid arthritis. Family-based
association studies such as the transmission disequilibrium test are being increasingly used to confirm
linkages and fine map genes in positional cloning
strategies in IDDM, and may be the way forward for
other autoimmune diseases. Intrafamilial association
studies have the advantage over population-based
case-control studies of avoiding artefactual associations due to population stratification. Furthermore,
the addition of diallelic markers to existing human
genome maps will facilitate the use of tests such as
the transmission disequilibrium test at an earlier
stage, and remove the need for the difficult acquisition of large numbers of families with multiple
affected siblings.
ACKNOWLEDGMENTS
We wish to acknowledge our colleagues including
Tony Barnett, Jayne Franklyn and John Todd, and
the Wellcome Trust and Lilly Industries for support.
REFERENCES
I. Todd JA. Genetic anaiysis of type I diabetes using whole genome approaches
(Review). Proc Nad Acad Sci USA 1995; 92: 8560-5.
2. Gough SCL, Saker PJ, Pritchard LE, et al. Mutation of the glucagon receptor
gene and diabetes melliius in the UK: association or founder effect. Hum Mol
Genet 1995; 4: 1609- 12.
3. Spielman RS, McGinnis RE, Ewens WJ. Transmission test for linkage
disequilibrium: the insulin gene region and insulindependent diabetes mellitus
(IDDM). hj Hum Genet 1993; 52: 506-16.
4. Copeman J, Cucca F, H e m e C, et al. Linkage disequilibrium mapping of a type
I diabetes susceptibilitygene (IDDM7) to chromsome 2q3 I-q33. Nature Genet
1995; 9 80-5.
5. Bennet ST,Lucassen AM, Gough SCL, et al. Susceptibility to human type I
diabetes at lDDM2 is determined by tandem repeat variation at the insulingene
minisatellite locus. Nature Genet 1995; 9 284-92.
6. Memman T, Twells R Meniman M. et al. Evidence by allelic associationdependent methods for a type I diabetes plygene (IDDM6) on chromosome
18q2I. Hum Mol Genet 1996; 6 1003-10.
7. Davis 1, Kawaguchi Y, Bennet S,et al. A genome-wide search for human type I
diabetes susceptibilitygenes. Nature (London) 1994; 371: 130-6.
8. Hashimoto L, Habita C, Beressi J, et al. Genetic mapping of a susceptibilitylocus
for insulin dependent diabetes mellitus on chromosome I Iq. Nature (London)
1994; 371: 161-4.
9. Sawcer S, Jones HB, Feakes R, et al. A genome screen in multiple sclerosis
reveals susceptibility loci on chromosome 6p2l and 17q22. Nature Genet 1996;
13: 464-8.
10. The Multiple Sclerosis Genetics group. A complete genome screen for multiple
sclerosis underscores a role for the major histocompatibility complex. Nature
Genet 1996; 13: 469-71.
I I. Eben GC, Kuhy K, Bulman DE, et al. A full genome search in multiple
sclerosis. Nature Genet 1996; 13: 472-6.
12. Tomfohrde j, Silverman A, Barnes R et al. Gene for familial psoriasis
490
J. Heward and S. C. L. Gough
susceptibility mapped to the distal end of human chromosome 17q. Science
(Washington DC) 1994; 264: 1141-5.
13. Matthews D, Fry L, Powles A, et al. Evidence that a locus for familial psoriasis
maps to chromosome 4q. Nature Genet 1996; 1 4 23 1-3.
14. Hugot JP. Laurent-Puig P, Gower-Rousseau C, et al. Mapping of a susceptibility
locus for Crohn’s disease on chromosome 16. Nature (London) 1996; 3 7 9
821-3.
IS. Payami H, Joe S, Thomson G. Autoimmune thyroid disease in type I diabetic
families. Genet Epidemiol 1989; 7: 83-5.
16. Olmos P, A’Hern R Heaton 4 et al. The significance of the concordance rate
for Type I (insulindependent) diabetes in identical twins. Diabetologia 1988
3 I:747-50.
17. Kyvik KO, Green A, Beck-Nielsen H. Concordance rates of insulin dependent
diabetes rnellitus: a population based study of young Danish twins. Br Med
J 1995;311:913-17.
18. Gough SCL. Genetics of insulin dependent diabetes mellitus. In: Shield JPH,
Baum ID, eds. Clinical paediatrics - childhood diabetes. London: Bailliere’s,
1996 593-605.
19. Cucca F, Todd JkHLA susceptibility to type I diabetes: methods and
mechanisms. In: Browning MJ, McMichael AJ, eds. H W M H C genes, molecules
and functions. Oxford, U.K: BIOS Scientific Publishers, 1996: 383-406.
20. Colonna M. Allelic variants of the human putative peptide transporter involved
in antigen processing. Proc Natl Acad Sci USA 1992; 8 9 3932-6.
21. Caillat-Zucman S, Bertin E, Timsit J, Boitard C, h a n R Bach IF. Protection
from insulin-dependent diabetes mellitus is linked to a peptide transporter gene.
Eur J lmmunol 1993; 23: 1784-8.
22. Deng GY, Muir A, Maclaren NK, She JX. Association of lMP2 and lMP7 genes
within the major histocompatibility complex with insulin dependent diabetes
mellitus: population and family studies. Am J Hum Genet 1995; 5 6 528-34.
23. Caillat-Zucman S, Daniel S, Timsit J, Garchon H, Boitard C. Bach J. Family study
of linkage disequilibrium between TAP2 transporter and HIA closs II genes:
absence of TAP2 contribution to associationwith insulin dependent diabetes
mellitus. Hum lmmunol 1995; 44:80-7.
24. Ronningen KS. Linkage disequilibrium between TAP2 variants and HLA class II
alleles; no primary association between TAP2 vatiants and insulin dependent
diabetes mellitus. Eur J lmmunol 1993; 23: 1050-6.
25. Undlien DE, Akselsen HE, loner G, et al. No independent associations of LMP2
and LMP7 polymorphisms with susceptibility to develop IDDM. Diabetes 1997;
4 6 307- 12.
26. Bennet ST, Todd JkHuman type I diabetes and the insulin gene: principles of
mapping polygenes. Annu Rev Genet 1996 3 0 343-70.
27. Lucassen A, Julier C, Beressi J,-P., et al. Susceptibility to insulin dependent
diabetes mellitus maps to a 4.1 kb segment of DNA spanning the insulin gene
and associated VNTR Nature Genet 1993; 4 305- 10.
28. Yanagawa T, Hidaka Y. Guimmes V, Soliman M, DeGroot L J.CTLA4 gene
polymorphism associated with Graves’ disease in a Caucasian population. J Clin
Endocrinol Metab 1995; 8 0 41-5.
29. BluestoneJkIs CTLA4 a master switch for peripheral T cell tolerance?
Jlmmunol 1997; 128 1989-93.
30. Walanus TL, Bakker CY, BluestoneJkCTLA-4 ligation blocks
CD-28-dependent T cell activation. J Exp Med 1996; 183 2541-50.
31. Ninico L, Bunetti, Pritchard LE, et al. The CTLA4 region of chromosome 2q33
is linked to, and associated with, type I diabetes. Hum Mol Genet 1996; 5:
1075-89.
32. Meek JC, JonesAE,Lewis UJ.Vanderlaan WP. Characterisationof the long
acting thyroid stimulator of Graves’ disease. Proc Natl Acad Sci USA 1961; 5 2
342-9.
33. Stennky V, Kozma L. Balm C, et al. The genetics of Graves’ disease: HLA and
disease susceptibility. J Clin Endocrinol Metab 1985; 61: 735-40.
34. Katsuren E, Awata T. Matsumoto C, Yamamoto K. HLA class II alleles in
Japanese patients with Graves’ disease: weak associations of HLA-DR and -DQ.
EndocrJ 1994; 41: 599-603.
35. Cavan DA, Penny MA, Jacobs KH, et al. The HLA associationwith Graves’
disease is sex-specific in Hong Kong Chinese subjects. Clin Endocrinol 1994; 4 0
63-6.
36. Yanagawa Y, Mangkabruks 4 Chang YE, et al. Human histocompatibility
leukocyte antigen-DQAl*OSOI allele associated with genetic susceptibilityto
Graves’ disease in a Caucasian population. J Clin Endocrinol Metab 1993; 7 6
1569-74.
37. Earlow ABT, Wheatcroft N, Watson P, Weetman AP. Association of HLADQAI*0501 with Graves’ disease in English Caucasian men and women. Clin
Endocrinol 1996; 44: 73-7.
38. Roman SH, Greenberg D, Rubinstein P, Wallenstein S. Davies TF. Genetics of
autoimmune thyroid disease: lack of evidence for linkage to HLA within families.
J Clin Endocrinol Metab 1992; 7 4 496-503.
39. Shields DC, RatanachaiyavongS, McGregor AM, Collins A Morton NE.
Combined segregation and linkage analysis of Graves’ disease with a thyroid
autoantibody diathesis. Am J Hum Genet 1994; 5 5 540-54.
40. RatanachaiyavongS, McGregor AM. HLA-DP8I polymorphisms on the MHCextended haplotypes of families of patients with Graves’ disease: two distinct
HLA-DR17 haplotypes. EurJ Clin Invest 1994; 2 4 309-15.
41. Donner H, Rau H, Walfish PG, et al. CTLA4 alanine-17 confers genetic
susceptibilityto Graves’ disease and to type I diabetes mellitus. J Clin
Endocrinol Metab 1997; 8 2 143-6.
42. Kotsa K, Watson PF, Weetman AP. A CTLA4 gene polymorphism is associated
with both Graves’ disease and autoimmune hypothyroidism. Clin Endocrinol
1997; 4 6 551-4.
43. Bahn RS. Dutton CM. Heufelder AE. Sarkar G. Agenomic point mutation in the
extracellular domain of the thyrotropin receptor in patientswith Graves’
ophthalmopathy. J Clin Endocrinol Metab 1994; 7 8 256-60.
44. Cuddihy RM. Schaid DS, Bahn RS. Multivariate analysis of HLA loci in
conjunction with a thyrotropin receptor codon 52 polymorhpism in conferring
risk of Graves’ disease. Thyroid 1996; 6 26 1-5.
45. Watson PF, French A, Pickerill AP, Mcintosh RS, Weetman AP. Lack of
association between a polymorphism in the coding region of the thyrotropin
receptor gene and Graves’ disease. J Clin Endocrinol Metab 1995; 8 0 1032-5.
46. de Roux N, Shields DC, Misrahi M, RatanachaiyavongS, McGregor AM,
Milgrom E. Analysis of the thyrotropin receptor as a candidate gene in familial
Graves’ disease. J Clin Endocrinol Metab 1996; 81: 3483-6.
47. Tomer Y. Earbesino G, Keddache M, Greenberg DA, Davies TF. Mapping of a
major susceptibilitylocus for Graves’ disease (GD-I) to chromosome 14q31.
J Clin Endocrinol Metab 1997; 8 2 1645-8.
48. Blakemore Al, Watson PF, Weetman AP, Duff GW. Association of Graves’
disease with an allele of the interleukin-I receptor antagonist gene. J Clin
Endocrinol Metab 1995; 8 0 I I 1-15.
49. Tandon N. Zhang L, Weetman AP. HLA associations with Hashimoto’s
thyroiditis. Clin Endocrinol 1991; 3 4 383-6.
50. Jenkins D, Penny M A Fletcher ]A, et al. HLA class II gene polymorphism
contributes little to Hashimotof thyroiditis. Clin Endocrinol 1992; 37: 141-5.
5 I. Wan XL, Kimura A, Dong RP, Honda K, Tamai H, Sasazuki T. HLA-A and
DRB4 genes in controlling the susceptibilityto Hashimoto’s thyro
lmmunol 1995; 4 2 I3 1-6.
52. Ebers GC, 8ulman DE, Sadovnick AD,et al. A population-based study of
multiple sclerosis in twins. N Engl J Med 1986; 315: 1638-42.
53. Sadovnick AD,Baird PA, Ward RH. Multiple sclerosis: updated risks for
relatives. Am J Med Genet 1988; 29: 533-41.
54. Ebers GC, Sadovnick AD,Risch NJ,Group CCS. A genetic basis for familial
aggregation in multiple sclerosis. Nature (London) 1995; 377: 150- I.
55. Robertson NP, ORiordan JI,ChatawayJ. et al. Offspring recurrence rates and
clinical chatacteristics of conjugal multiple sclerosis. Lancet 1997; 349: 1587-90.
56. Cullen CG, Middleton D, Savage DA, Hawkins S. HLA-DR and -DQ DNA
genotyping in multiple sclerosis patients in Northern Ireland. Hum lmmunol
1991; 30: 1-6.
57. Marrosu MG, Munoni F. M u m MR et al. Sardinian multiple sclerosis is
associated with HLA-DR4 a serologic and molecular analysis. Neurol I988 3 8
1749-53.
58. Spurkland A, Ronnongen KS, Vandvik B, Thorsby E, Vartdal F. HLA-DQAI and
DQB I genes may jointly determine susceptibilityto develop multiple sclerosis.
Hum lmmunol 1991; 3 0 69-75.
59. Haegert DG, Francis GS. HLA-DQ polymorphisms do not explain HLA class II
associations with multiple sclerosis in two Canadian patient groups. Neurol
1993; 4 3 1207-1 0.
60. Kelly M A Zhang Y, Mijovic CH, Chou K-Y. Barnett AH, Francis D k Genetic
susceptibility to multiple sclerosis in the Shanghai Chinese is not linked to the
myelin basic protein microsatellite.J Clin Pathol Mol Pathol 1995; 4 8 M I I 1-2.
61. lbsen SN, Clausen ]A. A repetitive DNA sequence 5’ to the human myelin basic
protein gene may be linked to MS in Danes. Acta Neurol Scand 1996; 93:
236-40.
62. lbsen SN, Clausen ]A. Genetic susceptibility to multiple sclerosis may be linked
to ploymorphism of the myelin basic protein gene. J Neurol Sci 1995; I 3 I:
96-8.
63. Murphy J, Murphy SF. Myasthenia gravis in identical twins. Neurology 1986; 3 6
78-80.
64. Kenin-Stonar L, Metcalf R 4 Dyer PA, Kowalska G, Ferguson I, Hams R
Genetic factors in myasthenia gravis: a family study. Neurology 1988; 38:
38-42.
Genetic susceptibility to autoimmune disease
65. Fria D, Hermann C Jr,Naeim F, Smith GS, Walford RL. HL-A antigens in
myasthenia gravis. Lancet 1974; I: 240-2.
66. Nako Y, Myazaki T, Ota K et al. Gm allotypes in myastheniagravis. Lancet
1980; I:677-80.
67. Oksenberg JR, Sherritt M, Begovitch AB, et al. T-cell receptor V alpha and C
alpha alleles associated with muliple sclerosis and myasthenia gravis. Proc Natl
Acad Sci USA 1989; 86: 988-92.
68. Garchin HJ, Djabiri F, Viard JP, Gajdos P. Bach IF. Involvement of human muscle
acetylcholine receptor a subunit gene (CHRNA) in susceptibilityto myasthenia
gravis. Proc Natl Acad Sci USA 1994; 9 I:4668-72.
69. Heckmann JM, Morrison KE, Emeryk-Szajewska B, et al. Human muscle
acetylcholine receptor alpha-subunit gene (CHRNAI) associationwith
autoimmune myasthenia gravis in black, mixed-ancestryand Caucasian subjects.
J Autoimm 1996; 9 175-80.
70. Maclaren NK, Riley W. Inherited susceptibilityto autoimmune Addison's disease
linked to human leukocyte antigens DR3 andlor DR4, except when associated
with type I autoimmune polyglandular syndrome. J Clin Endocrinol Metab 1986;
6 2 455-9.
71. Weetman AP, Zhang L, Tandon N, Edwards OM. HIA associations with
autoimmune Addison's disease. Tissue Antigens I991; 38: 3 1-3.
72. PartanenJ. Peterson P, Westman P, Aranko S. Krohn K. MHC class II and 111 in
Addison's disease. MHC alleles do not predict autoantibody specificity and no
independent role for 21-hydroxylase gene polymorphism in disease
susceptibility. Hum lmmunol 1994; 41: 135-40.
73. Winqvist 0, Karlsson FA, Kampe 0.21 -hydroxjase, a major autoantigen in
idiopathic Addison's disease. Lancet 1992; 3 3 9 1559-62.
74. Hochberg MC. The application ofgenetic epidemiology to systemic lupus
erythematosus. J Rheumatol 1987; 14: 867-9.
75. Kaplan D. The onset of disease in twins and siblings with systemic lupus
erythematosus. J Rheumatol 1984; II:648-52.
76. Deapen DM, Escalante A, Weinrib L, et al. A revised estimate of twin
concordance in systemic lupus erythematosus. Arth Rheum 1992; 3 5 3 11-18,
77. Hamilton RG, Harley JB, Bias WB, et al. Two Ro (SS-A) autoantibody responses
in systemic lupus erythematosus: correlation of HLA-DR/DQ specificities with
quantitative expression of Ro (SS-A) autoantibody. Arth Rheum 1988; 31:
496-505.
78. Harley JB, Alexander EL, Arnett FC, et al. Anti-RolSSA and anti-La/SSB in
patients with Sjogrens syndrome. Arth Rheum 1986; 2 9 196-206.
79. Harley JB, Reichlin M, Arnett FC. et al. Gene interactions at HLA-DQ enhances
autoantibody production in primary Sjogrens syndrome. Science (Washington
DC) 1986; 232 1145-7.
80. Savi M. Fenaccioli GF, Neri TM, et al. HLA-DR antigens and anticardiolipin
antibodies in Northern Italian systemic lupus etythematosus patients. Arth
Rheum 1988; 31: 1568-70.
81. Skarsvag S,Hansen KE. Holst A, Moen T. Distribution of HLA class II alleles
among Scandanavian patients with systemic lupus erythematosus (SLE): an
increased risk of SLE among non [DRBI*03, DQA180501, DQB180201I class II
homozygotes?Tissue Antigens 1992; 4 0 128-33.
82. Doherty DG, Ireland R, Demaine AG, et al. Major histocompatibility complex
genes and susceptibilityto systemic lupus erythematosus in Southern Chinese.
Arth Rheum I992; 3 5 641-5.
83. Yao Z, Kimura A, Hartung K, et al. Polymorphism of the DQAI promotor
region (QAP) and DRBI, QAP. DQAI, DQBl haplotypes in systemic lupus
erythematosus. Eur J lmmunogenet 1993; 2 0 259-66.
84. Jarjour W, Reed AM, Gauthier J, Hunt S, Winfield JB. The 8.5-kb Pstl allele of
the stress protein gene, Hsp 70-2. An independent risk factor for systemic lupus
erythematosus in African Americans. Hum lmmunol 1996; 4 5 59-63.
49 I
85. Wilson AG, Gordon C, di Giovine FS, et al. A genetic association between
systemic lupus erythematosus and tumour necrosis factor alpha. Eur J lmmunol
1994;24: 191-5.
86. Hasstedt SJ, Clegg DO, lngles L, Ward RH. HLA-linked rheumatoid arthritis.
Am J Hum Genet 1994; 5 5 738-46.
87. MacGregor A, Ollier W, Thomson W, Jawaheer D, Silman A. HLA-DRBI*0401/
0404 genotype and rheumatoid arthritis: increased association in men, young
age at onset, and disease severity. J Rheumatol 1995; 2 2 1032-6.
88. Suarez- Almazoe ME, Tao S. Mousta~ahF, Russell As, Maksymowych W. HIADRI, DR4, and DRBI disease related subtypes in rheumatoid athritis.
Association with susceptibilitybut not severiy in a city wide community based
study. J Rheumatol 1995; 2 2 2027-33.
89. Hall FC, Weeks DE, Camilleri JP, et al. Influence of the HLA-DRBI locus on
susceptibilityand severity in rheumatoid arthritis. Q J Med 1996; 89: 821-9.
90. Mu H. Charmley P, King MC, Criswell IA. Synergy between T cell receptor
beta gene polymorphism and HIA-DR4 in susceptibilityto rheumatoid arthritis.
Arth Rheum 1996; 3 9 93 1-7.
91. Shaw MA, Clayton D, Atkinson SE, et al. Linkage of rheumatoid arthritis to the
candidate gene NRAMPI on 2q35. J Med Genet 1996; 33: 672-77.
92. Hardwick LJ, Walsh S, Butcher S,et al. Genetic mapping of susceptibilityloci in
the genes involved in rheumatoid arthritis. J Rheumatol 1997; 24: 197-8.
93. Satsangi J, Welsh KI, Bunce M, et al. Contribution of genes of the major
histocompatibility complex to susceptibility and disease phenotype in
inflammatory bowel disease. Lancet 1996; 347: I212- 17.
94. Heresbach D, Colombel F, Danze PM, Semana G. The HLA DRB1'0301DQB I*020 I haplotype confers protection against inflammatory bowel disease.
Am J Genet 1996; 9 I: 1060.
95. Ohmen ID, Yang HY,Yamamoto KK, et al. Susceptibility locus for inflammatory
bowel disease on chromosome I 6 has a role in Crohn's disease but not in
ulcerative colitis. Hum Mol Genet 1996; 5 1679-83.
96. Swanbeck G, lnerot A, Martinsson T, Wahlstrom J. A population genetic study
of psoriasis. Br J Dermatol 1994; 131: 32-9.
97. Schmitt-Egenolf M, Eiermann TH, BoehnckeWH, Stander M, Steny W. Familial
juvenile onset psoriasis is associated with the human leukocyte antigen (HLA)
class I side of the extended haplotype CW~-B~~-DRBI*O~O~-DQAI*O~OIDQBI*0303: a population and family based study. J Invest Dermatol 1996; 106
711-14.
98. Matthews D, Fry L, Powles A, Weissenbach J, Williamson R Confirmation of
genetic heterogeneity in familial psoriasis. J Med Genet 1995; 3 2 546-8.
99. Nair Rp, Guo SW,JenischS. et al. Scanning chromosome I 7 for psoriasis
susceptibility: lack of evidence for a distal 17q locus. Hum Hered 1995: 4 5
2 19-30.
100. Trembath RC, Clough RL, RosbothamJL, et al. Identification of a major locus
on chromosome 6p and evidence for further disease loci revealed by a two
stage genome-wide search in psoriasis. Hum Mol Genet 1997; 6 813-20.
101. Ploski R, Vinje 0, Ronningen KS, et al. H I A class II alleles and heterogeneityof
juvenile rheumatoid arthritis. Arth Rheum 1993: 3 6 465-71.
102. Paul C, Schoenwald U, Truckenbrodt H, et al. HLA-DP/DR interaction in early
onset pauciarticularjuvenile chronic arthritis. lmmunogenetics 1993; 37: 442-8.
103. Barron KS, Silverman ED, Gonzales JC, Gwerbach D, ReveilleID. DNA analysis
of HLA-DR, DQ and DP alleles in children with polyarticular juvenile
rheumatoid arthritis. J Rheumatol 1992; 1 9 1611-15.
104. Luyrink L, Gabriel CA, Thompson SD, et al. Reduced expression of a human Vl/
6.1 T cell receptor allele. Proc Natl Acad Sci USA 1993; 9 0 4369-73.
105. McDowell TL, Symonds]A, Ploski R Forre 0, Duff G. A genetic association
between juvenile rheumatoid arthritis (JRA) and interleukin I alpha (ILla)
polymorphism. Arth Rheum 1995; 3 8 221-8.