Autoimmunity, February 2008; 41(1): 11–18

Autoantibodies in type 1 diabetes

CRAIG E. TAPLIN & JENNIFER M. BARKER

Barbara Davis Center for Childhood Diabetes, University of Colorado Health Sciences Center, Aurora, CO, USA

Abstract
Autoimmunity Downloaded from informahealthcare.com by University of Guelph on 05/06/12

Diabetes mellitus is a disorder characterized by hyperglycemia in both the fasting and post-prandial states. The two most
common forms of diabetes mellitus, type 1 and type 2 (previously called juvenile-onset and adult-onset, respectively),
comprise the vast majority of cases. Type 1 diabetes (T1DM) has been shown to be a disease characterized by immune-
mediated destruction of the insulin-secreting cells of the pancreas; it comprises the majority of cases of diabetes seen in
childhood and approximately, 5– 10% of all cases of diabetes mellitus in the USA and perhaps accounts for an even higher
percentage in those nations with lower rates of obesity. The process of beta-cell destruction, marked by the production of
autoantibodies to the beta-cell, occurs over many years and ultimately results in metabolic abnormalities first manifested as
impaired glucose tolerance and then progressing to symptomatic hyperglycemia. It has been reported that approximately 50%
of the genetic risk for T1DM can be attributed to the HLA region. The highest risk HLA-DR3/4 DQ8 genotype has been
shown to be highly associated with beta-cell autoimmunity. The first antibodies described in association with the development
of T1DM were islet cell autoantibodies (ICA). Subsequently, antibodies to insulin (IAA), glutamic acid decarboxylase (GAA
For personal use only.

or GAD) and protein tyrosine phosphatase (IA2 or ICA512) have all been defined. The number of antibodies, rather than the
individual antibody, is thought to be most predictive of progression to overt diabetes.

Keywords: Diabetes mellitus, autoantibodies, hyperglycemia, B cells

Diabetes mellitus is a disorder characterized by an even higher percentage in those nations with lower
hyperglycemia in both the fasting and post-prandial rates of obesity [6]. It is one of the most prevalent
states. Although it is a more heterogeneous and chronic diseases of childhood.
nuanced disorder than suggested by the traditional A model for the development of T1DM was
dichotomous classification, the two most common proposed by Eisenbarth [7] over 20 years ago, which
forms of diabetes mellitus, type 1 and type 2 describes T1DM as a chronic, progressive auto-
(previously called juvenile-onset and adult-onset immune disorder. In this model, a subset of subjects at
respectively), comprise the vast majority of cases. In genetic risk for diabetes experience an as yet undefined
adults, type 2 diabetes (T2DM) is the most common insult which initiates the process of beta-cell destruc-
form in the western world [1]; its pathophysiological tion, marked by the production of autoantibodies to
hallmark is insulin resistance, which may or may not the beta-cell. This destruction occurs over many years
be associated with absolute insulin deficiency. As such and ultimately results in metabolic abnormalities first
it is not the subject of this review. Type 1 diabetes manifested as impaired glucose tolerance and then
(T1DM) has been shown to be a disease characterized progressing to symptomatic hyperglycemia.
by immune-mediated destruction of the insulin-
secreting cells of the pancreas; it comprises the
Genetics of autoimmune b-cell destruction
majority of cases of diabetes seen in childhood [2 – 4]
and approximately, 5 –10% of all cases of diabetes Most people with T1DM do not have a family history
mellitus in the USA [5] and perhaps accounts for of the disease; nonetheless there is clearly a genetic

Correspondence: J. M. Barker, Barbara Davis Center for Childhood Diabetes, Box A140, Building M20, 1775 N. Ursula Street, P.O. Box
6511, Aurora, CO 80045-6511. Tel: 1 303 724 6710. Fax: 1 303 724 6779. E-mail: [email protected]

ISSN 0891-6934 print/ISSN 1607-842X online q 2008 Informa UK Ltd.

DOI: 10.1080/08916930701619169
 12 C. E. Taplin & J. M. Barker

predisposition for developing beta-cell autoimmunity More recently, this field has been advanced with the
[8 –10] with first degree relatives having a lifetime risk identification of the beta-cell antigens glutamic acid
of developing T1DM of 6 vs. 0.4% in the general decarboxylase (GAD/GAA) and protein tyrosine
population. In addition, there is a strong, approxi- phosphatase (IA2/ICA512).
mately 50%, concordance in identical twins for the For most autoimmune disorders and for basic
development of T1DM [11,12]. immunologic research ELISA assays employing plate
Large population studies have established certain bound antigen are the standard. Multiple workshops
human leukocyte antigen (HLA) haplotypes that are utilizing sera from patients with T1DM and mouse
associated with a high risk for developing T1DM, in strains have demonstrated that standard ELISA formats
particular certain HLA-DR and DQ genes. Indeed, it lack sensitivity and specificity compared to radioassays
has been reported that approximately 50% of the [19]. Highly sensitive radioimmunoassays have been
genetic risk for T1DM can be attributed to the HLA developed for antibodies directed against GAA and IA2
region [13]. The highest risk HLA-DR3/4 DQ8 as well as antibodies to insulin (IAA; Figure 1).
genotype has been shown to be highly associated with
beta-cell autoimmunity [13,14]. Where those at high
IAA, GAA and IA2 antibodies
risk for diabetes have been identified at birth and
Autoimmunity Downloaded from informahealthcare.com by University of Guelph on 05/06/12

followed prospectively for the development of diabetes Insulin and its precursor, proinsulin, are the only
autoantibodies, the highest risk for persistently positive known beta-cell specific autoantigens [17]. It had long
autoantibodies and progression to diabetes is found in been recognized that treatment with exogenous forms
those with the highest risk HLA haplotype [15]. of insulin could induce antibodies directed against
these “foreign” peptides (e.g. bovine insulin) [20], but
in 1983 the discovery of anti-insulin antibodies (in
Autoantibodies and T1DM
patients with T1DM prior to the administration of
The first antibodies described in association with the exogenous insulin) was made [21]. Subsequently, a
development of T1DM were islet cell autoantibodies large number of studies have demonstrated that anti-
(ICA) [16]. Subsequently, antibodies to insulin (IAA), insulin antibodies are present for years before the
For personal use only.

glutamic acid decarboxylase (GAA or GAD) and development of T1DM [22] and correlate inversely
protein tyrosine phosphatase (IA2 or ICA512) have all with the age at which T1DM develops [17]. Genetics
been defined. These antibodies can be present years clearly play a role; with higher IAA levels associated
before the development of clinically significant hyper- with DR4 [23] and DQ8 genotypes [24].
glycemia, and the presence of persistently positive and Glutamic acid decarboxylase (GAD) is the enzyme
multiple antibodies is highly predictive for the develop- responsible for GABA synthesis within the nervous
ment of T1DM. This will be further explored below. system and islet cells and is expressed not only within
human beta-cells, but also alpha and delta cells.
Antibodies directed against GAD were first documented
Islet cell autoantibodies (ICA)
in patients with the neurological disorder stiff-man
The association of T1DM and Addison’s disease, syndrome [17] and were subsequently found in patients
another autoimmune endocrine disorder, provided the with new-onset T1DM as well as in patients with
first clues to its etiology as an autoimmune disease. autoimmune polyendocrinopathy syndrome type 2
Lymphocytic infiltration of the adrenal cortex had been (APS2).
documented in patients with Addison’s disease and was Further development in the field led to the
thought to be evidence for a specific autoimmune insult discovery of an additional islet cell antigen related to
directed at the adrenal cortex [17]. Also, new-onset the protein tyrosine phosphatase IA2, also termed
diabetic patients whose islets were available for ICA512 [25]. This islet cell antigen was found to bind
histological examination showed evidence of a lym- antibodies with a high positive predictive value for the
phocytic process within the islet; so-called “insulitis”. development of T1DM (Table I).
As a result of the recognition of this association, the
discovery of cytoplasmic islet cell autoantibodies (ICA)
Diabetes autoantibody standardization
was made [16]. ICA’s are detected by reaction of sera
with sections of human pancreas and then staining for With the increasing utility of measuring diabetes
these autoantibodies. Some patients with Addison’s autoantibodies worldwide in both clinical and research
disease had high titers in their serum of ICA. These settings, and with the advent of large multicenter
ICA’s are directed against certain islet cell surface and studies utilizing diabetes antibodies to predict risk, it
cytoplasmic antigens. Quantification of ICA has been became important to standardize assay methodology.
standardized to Juvenile Diabetes Foundation (JDF) To this effect, workshops were convened first for ICA
units [17]. Higher levels of these units have been [26,27] and followed by multi-center studies upon the
associated with an increased risk for T1DM in relatives description of GAD, IA2 and IAA autoantibodies.
of subjects with T1DM [18]. The Diabetes Antibody Standardization Program
 Autoantibodies in diabetes 13
Autoimmunity Downloaded from informahealthcare.com by University of Guelph on 05/06/12

Figure 1. General outline of fluid phase 96-well plate assays for autoantibodies. In above example I125 -insulin utilized, but assay format
identical for GAD65, ICA512 [17].

(DASP) [28], established in collaboration with the measurements into the classification of sera as positive
Centers for Disease Control (CDC), reported in 2003 or negative, inter-laboratory concordance, sensitivity
For personal use only.

results from 46 laboratories in 13 countries using and specificity were improved. Workshops such as
coded sera from 50 type 1 diabetics and 50 controls these have also defined and standardized units of
(blood donors). Inter-laboratory concordance was measurement using WHO international reference
found to be high for GAD and IA2 when radiobinding reagents with titers expressed in WHO units/ml.
assays were used; however concordance between
laboratories measuring IAA was inferior with low
sensitivity to detect IAA in patient sera. Recently, Other islet autoantibodies
DASP refined the potential utility of IAA assays [29].
Other islet autoantigens have been, and continue to
Concordance was high for antibody affinity, but low
be, discovered that are associated with T1DM.
for antibody titer in the classification of patients vs.
Utilizing sera from pre-diabetic individuals and in
control sera. However, by incorporating affinity
animal models, antibodies have been detected against
Table I. Subset of biochemically characterized autoantigens [17]. glycolipid antigens [17]. Also, antibodies against the
molecule ICA69 which is beta-cell specific in the rat
Sensitivity (%) Comment but not in humans and has an identical sequence to
Insulin 49–92 Higher levels cow’s milk related peptide p69, have been discovered
young children [30]. Anti-ICA69 antibodies are present in diabetic
Glutamic acid 84 Higher sensitivity sera, but assays have not been developed that reliably
decarboxylase adult onset distinguish diabetic from control sera in a way that is
(GAD) type 1A
more useful than the currently used IAA, GAA and
ICA512/IA-2 74 Tyrosine
phosphatase IA2 assays. ICA12 (the transcription factor SOX13)
like molecule binds to molecules in the sera of diabetic patients, but
IA-2b/phogrin 61 Tyrosine antibodies developed against this antigen are present
phosphatase in only a small proportion of new onset patients [17].
like molecule
Carboxypeptidase H was discovered in screening
Carboxypeptidase H 10 Infrequent
GLIMA38 14–38 Not yet sequenced studies of pre-diabetic sera, and while not islet specific,
GM2-1 ? Ganglioside: it is an abundant islet protein [17]. Radioassays for
chromatography antibodies directed against this antigen have not yet
assay been developed. Several other autoantigens have been
ICA69 ? Western blot assay
partially characterized such as osteopontin, molecules
with poor specificity
ICA12 ,20 Diabetes relatedness expressed in the kidney and islet such as nephrin and
requires further study filtrin, CD38 and heat shock protein 65 [17]. While
research continues into the utility of detectable
 14 C. E. Taplin & J. M. Barker

antibodies against the above antigens, they are not in of transient antibody positivity had been positive for
use clinically or in long-term prospective trials more than one antibody. It is possible that these patients
following patients over time for diabetes risk and will go on to reacquire islet autoantibodies years or
prevention. Thus the most widely used autoantibodies decades later, and further data with longer follow-
in T1DM remain ICA, IAA, GAA and IA2. up periods will be required to define this risk.
Nonetheless, these patients represent a potentially
illuminating avenue for the investigation of possible
Antibodies and progression to T1DM
factors involved in “switching off” islet autoimmunity.
The development and refinement of assays for islet
specific autoantibodies has led to intense investigation
Pattern of development of diabetes-related autoantibodies
into their temporal relationship to the development of
diabetes and thus the predictive value of measuring Antibodies to islet cell antigens tend to appear
these antibodies in patients prior to the onset of sequentially rather than simultaneously. In an early
hyperglycemia. With regard to the development of analysis from DAISY, Yu and colleagues [33] analyzed
T1DM, researchers have investigated whether they samples from 155 young first degree relatives of subjects
develop simultaneously or sequentially, the number of with T1DM along with a separate group of older first
Autoimmunity Downloaded from informahealthcare.com by University of Guelph on 05/06/12

positive antibodies and diabetes risk, whether these degree relatives for IAA, IA2 and GAA. Samples were
antibodies are transient or persistent and their analyzed yearly until one antibody became positive, and
predictive value in higher risk groups such as first then at 3–6 monthly intervals. In the younger cohort, of
degree relatives of people with T1DM. those who seroconverted for at least one antibody, none
Prospective population based studies of the natural showed evidence of simultaneous positivity for multiple
history of T1DM such as the Diabetes Autoimmunity antibodies. Of the total of 26 seroconversion events in
Study in the Young (DAISY) [15], the Finnish DIPP both cohorts, in only three cases did more than one
study [31], and the German BabyDIAB study [32] antibody appear at the same time. In these three
have emphasized the importance of these diabetes patients, the interval between most recent serum
related autoantibodies to the development of diabetes collections was 1.8 and 6 years in two, raising doubt as
For personal use only.

and have clarified characteristics of their appearance. to whether these antibodies truly appeared “simul-
The appearance of T1DM autoantibodies usually taneously”. The timing of the acquisition of a second or
precedes the onset of diabetes by months or years, and third autoantibody varied greatly from months to years.
may occur as early as the first year of life [15]. The authors examined genotype as a factor in the tempo
of sequential antibody acquisition and found no
difference between those with a high risk HLA DR3/4
False and transiently positive diabetes related
haplotype and those with the lower risk HLA status.
autoantibodies
In this study, all those who went on to develop diabetes
Analysis of data from the DAISY study has shown that developed more than one antibody prior to diagnosis,
a significant proportion of antibodies positive on the with most progressing to diabetes within 2 years of
initial assay do not confirm on repeat blinded aliquots acquiring the final antibody, though one child was
of the same serum sample and are therefore falsely persistently positive for three antibodies for 8 years
positive [15]. The rates of false positive antibodies are before presenting with symptoms of hyperglycemia.
0.1% for ICA512, 0.5% for GAA/GAD65 and 1.1% Results such as this suggest that humoral autoimmunity
for IAA. For IAA, false positive results were of lower to islet antigens is acquired in a sequential fashion over
titer than those who were confirmed positive. This months or years, rather than as one “catastrophic” event
observation demonstrates the need for confirmatory in time. Of course this does not exclude the possibility
samples when an individual is found to be positive for that the precipitating “insult” occurred at one point in
a diabetes related antibody. time, nonetheless, evidence suggests that beta-cell
In addition to false positive autoantibodies, subjects destruction is chronic and progressive.
may also have transient autoantibodies, that is, Evidence also shows that of the three commonly
confirmed positive on at least one sample but negative assayed autoantibodies, either GAA or IAA tend to
on subsequent sampling. It has been shown in several appear first, followed by IA2. In the abovementioned
cohorts (e.g. DIPP, DAISY) that antibodies confirmed study, of those patients who eventually expressed more
to be positive may disappear months or years later. than one autoantibody, 2/3 acquired GAA first and
Indeed, in the DAISY study where rigorous study design 1/3 acquired IAA first.
and blinded assay repetition accounted for and excluded
false positives, 50 of 112 confirmed positives were found
Number of diabetes related autoantibodies is associated
to be transient. Almost half of those were positive at
with risk for T1DM
more than one visit before antibodies waned. None went
on to develop diabetes during the study period (a mean As inferred from the above research, the number
follow-up period of 4 years) and only two of the 50 cases of antibodies, rather than the individual antibody,
 Autoantibodies in diabetes 15

is thought to be most predictive of progression to overt The level of IAA has been associated with persistence
diabetes. Verge et al. analyzed sera from 45 new onset of autoimmunity and diabetes. Higher levels of IAA are
patients with T1DM and a large number of first degree more likely to be confirmed positive, remain positive on
relatives and controls [34]. All but one of those with subsequent samples, and differentiate pre-diabetic
new onset T1DM had at least one antibody (98%) and patients from those persistently non-diabetic [15].
more than 3/4 had at least two antibodies at the time of Higher GAA levels have also been shown to differen-
diagnosis. None of the control subjects were positive for tiate those with persistent islet autoantibodies from
more than one antibody. In this high-risk group, disease those who are transiently positive [15], and in a
free survival was no different between those who combined cohort of high risk individuals from
acquired IAA first and those who acquired GAA first. Germany and the Bart’s Oxford (BOX) cohort, IA-2
Taking this concept further, the same titers in the upper three quartiles had significantly
group quantified the 5-year risk for developing higher risk (79% over 10 years) than the lowest quartile
T1DM in those people with a first degree relative [36]. They also found the highest IAA quartile had a
with T1DM [35]. They analyzed 882 first degree similar 10 year risk for diabetes that was significantly
relatives, measuring GAA, IAA, and IA2 and followed different than the lowest three quartiles (77 vs. 37% at
these subjects for up to 11 years. The risk for developing 10 years). A correlation between GAA titer and
Autoimmunity Downloaded from informahealthcare.com by University of Guelph on 05/06/12

T1DM was significantly increased for each extra diabetes risk was not present in this study.
autoantibody acquired. For those with two or more Similarly, the affinity of insulin autoantibodies for
antibodies, the 3-year risk was 39% and 5-year risk insulin has been associated with diabetes development.
68%. Those with all three antibodies had a 100% 5-year In subjects positive for insulin autoantibodies, those
risk (Figure 2). Later data from DAISY has shown that with autoantibodies with higher affinity for insulin have
during a mean 4 year follow-up period (range 0.17 – 9 a higher risk for diabetes. Achenbach and coworkers
years), 41% (24 of 58) patients persistently positive for analyzed the affinity of anti-insulin autoantibodies for
at least one antibody developed overt T1DM [15]. children followed prospectively [24]. A high percentage
An interesting further question relates to how the of the children who went on to develop multiple anti-
acquisition of islet autoimmunity relates to family islet autoantibodies or to progress to diabetes expressed
For personal use only.

history. Offspring of diabetic fathers carry a higher risk high affinity IAA autoantibodies.
for acquiring IAA, GAA and IA2 when compared to Characteristics of IAA have been well studied and
being born to a mother with T1DM [8]. To date, this IgG subtypes of this islet autoantibody have been
observation has not been completely explained. related to the development of diabetes. Achenbach and
colleagues in Europe followed 180 first-degree relatives
of diabetic probands, and found that diabetes risk was
Characteristics of diabetes-related autoantibodies
higher in those who had IAA subtypes IgG2, IgG3 and
are associated with an increased risk for T1DM
IgG4 than those without these IgG subclasses(36).
Other autoantibody related factors have been associ- They found the same IgG subclasses of IA-2 also
ated with development of diabetes including level, predicted a higher risk, but no relationship was found
affinity for antigen and antigen subtype. between GAD IgG subclass and diabetes risk.

Figure 2. The diabetes-free survival of first-degree relatives, according to the number of autoantibodies (ab) present at baseline, considering
IAAs, GAAs, and ICA512bdcAAs [35]. Copyright 1996 American Diabetes Association. Diabetes. Vol 45, 1996;926–933. Reprinted with
permission from the American Diabetes Association.
 16 C. E. Taplin & J. M. Barker

The association of diabetes-related autoantibodies and Other organ specific autoantibodies and T1DM
abnormalities of glucose tolerance prior to the development
The association between T1DM and autoimmune
of T1DM
adrenal failure led to the discovery of immune-
Although researchers have been able to develop reliable mediated beta-cell destruction and the development
estimates of medium to long term diabetes risk based of assays to detect islet autoantibodies. Indeed,
on the appearance of autoantibodies, little has been T1DM often occurs together with a spectrum of
known until recently about the relationship of changes autoimmune diseases, including autoimmune thyroid
in glucose tolerance to the appearance of antibodies. disease, celiac disease and Addison’s disease. Multi-
The Type 1 Diabetes Prediction and Prevention Study organ autoimmune disease may cluster together in the
(DIPP) performed in Finland, which has the highest same patient, for example in the well-described
incidence of T1DM in the world, found that first phase autoimmune polyendocrine syndrome type 2
insulin responses (FPIR) on intravenous glucose (APS2). Recent studies have further defined the
tolerance testing (IVGTT) were below the 5th prevalence and associations of individual organ-
percentile of normal in over 40% of children at the specific autoantibodies in individuals with T1DM.
time their seroconversion to ICA positivity was T1DM may also be a component of the rare
detected [36,37]. This study measured antibodies at Autoimmune polyendocrine candidiasis ectodermal
Autoimmunity Downloaded from informahealthcare.com by University of Guelph on 05/06/12

maximum intervals of 12 months, and IVGTT’s were dystrophy (APECED) syndrome, although it is not
performed at a mean of 0.6 years after the positive result one of the major components.
in almost all children. Correlating with previously
described data on titer and diabetes risk, those with Thyroid autoantibodies
higher ICA titers (greater than 20 JDFU) had
significantly lower FPIR than those with ICA titers Autoimmune thyroid disease is common in the general
less than 20 JDFU. Multiregression analysis showed population and its association with T1DM is well
that low FPIR on IVGTT was associated with multiple described [40]. The hallmark is the expression of
autoantibody positivity and high titers of IAA. antibodies to thyroid peroxidase (TPOAb) and
thyroglobulin (TGAb). The prevalence of either
For personal use only.

Approximately, half of those with a low FPIR went on

to develop diabetes at a mean interval of 2.2 years after TPOAb or TGAb in a cohort of 814 individuals
they converted to positive ICA. with T1DM was 29% (median duration of diabetes
From these studies it can be concluded that 3.4 years, median age 14.8 years) and, as in the general
autoantibodies tend to develop sequentially, with population, a female preponderance of thyroid
GAA or IAA appearing first with IA2 later on. The autoantibodies has been observed [41]. Duration of
progression to multiple antibody positivity may be diabetes is also significantly associated with TPO or
rapid, particularly in the young, but in general develops TG antibodies. TPOAb appears to confer a greater
over months and years. Most intuitively, the presence risk for the development of clinical thyroid disease,
of antibodies that bind islet antigens with high affinity, with or without TGAb, than does TGAb alone.
and multiple antibody positivity, are highly predictive
for the eventual onset of overt T1DM, and in those Tissue transglutaminase antibodies
with all three antibodies, the 5 year risk approaches
100%, at least in those with a high risk genotype. In the same study, 10% were positive for tissue
tranglutaminase antibody (TTGAb), a marker of
celiac disease or gluten sensitive enteropathy. Inter-
Diabetes autoantibodies-bystanders or active players? estingly in contrast to other published data [42], there
were no statistically significant associations between
It is widely accepted that autoimmune islet cell TTGAb and sex, age or duration of T1DM. In those
destruction is a T cell dependent process. Despite the biopsied (19 of 82), 84% had positive histological
importance and utility of autoantibodies for disease findings. About 27 of the 82 positive for TTGAb had
prediction, it is likely that anti-islet autoantibodies do at least one other of either TPOAb, TGAb or 21-
not by themselves cause beta-cell destruction. Indeed, hydroxylase antibody. Subjects with higher levels of
this may be inferred by the lack of diabetes in infants TTGAb by radioimmunoassay are more likely to have
born to antibody positive mothers, where antibodies a small intestinal biopsy consistent with Celiac disease.
are transplacentally acquired [38]. ICA, anti-insulin,
and anti-GAD autoantibodies can be detected in the
21-Hydroxylase antibodies
serum of infants of antibody-positive mothers in the
absence of diabetes in the child, including those Addison’s disease is less common in patients with
infants born to mothers who develop overt T1DM T1DM than autoimmune thyroid disease or celiac
during pregnancy. Also, type 1 diabetes has been disease, with 21-hydroxylase antibodies (21-OHAb)
reported in a patient with agammaglobulinemia where present in 1 – 2% of diabetics in cross-sectional studies
B cells are severely depleted [39]. [43]. 21-OHAb’s are significantly associated with
 Autoantibodies in diabetes 17

thyroid antibodies in patients with T1DM and tend to [5] Engelgau MM, Geiss LS, Saaddine JB, Boyle JP, Benjamin
higher prevalence in patients with a longer duration of SM, Gregg EW, et al. The evolving diabetes burden in the
United States. Ann Intern Med 2004;140:945–950.
T1DM. The risk for 21-OHAb appears to be
[6] Bruno G, Runzo C, Cavallo-Perin P, Merletti F, Rivetti M,
selectively increased for those who are thyroid Pinach S, et al. Incidence of type 1 and type 2 diabetes in adults
antibody positive, and of those patients positive for aged 30–49 years: The population-based registry in the
adrenal antibodies in the DAISY study, 38% had province of Turin, Italy. Diabetes Care 2005;28:2613–2619.
clinical disease [41]. The development of adrenal [7] Eisenbarth GS. Type I diabetes mellitus. A chronic auto-
disease in the cohort positive for 21-OH autoanti- immune disease. N Engl J Med 1986;314:1360–1368.
[8] Harjutsalo V, Reunanen A, Tuomilehto J. Differential
bodies has been extensively reviewed [44,45]. Adrenal transmission of type 1 diabetes from diabetic fathers and
dysfunction appears to develop over time and can be mothers to their offspring. Diabetes 2006;55:1517– 1524.
first identified by an increase in plasma renin activity [9] Rich SS. Mapping genes in diabetes. Genetic epidemiological
and ACTH, prior to the development of overt cortisol perspective. Diabetes 1990;39:1315–1319.
and aldosterone deficiency. [10] Tuomilehto J, Podar T, Tuomilehto-Wolf E, Virtala E.
Therefore, there is a high prevalence (approximately, Evidence for importance of gender and birth cohort for risk
of IDDM in offspring of IDDM parents. Diabetologia 1995;
1/3 of patients) of a second organ-specific autoantibody 38:975–982.
in patients with T1DM, with TPOAb being the most [11] Kyvik KO, Green A, Beck-Nielsen H. Concordance rates of
Autoimmunity Downloaded from informahealthcare.com by University of Guelph on 05/06/12

common. Researchers have sought to define the HLA insulin dependent diabetes mellitus: A population based study
DR/DQ genotypes associated with antibody “subspe- of young Danish twins. BMJ 1995;311:913 –917.
cificity” in T1DM. For example, 21-OHAb positivity [12] Redondo MJ, Rewers M, Yu L, Garg S, Pilcher CC, Elliott RB,
et al. Genetic determination of islet cell autoimmunity in
has been found in 16% of those with T1DM and the
monozygotic twin, dizygotic twin, an non-twin siblings of
DRB1*0404-DQ8, DR3-DQ2 genotype [41]. This patients with type 1 diabetes: Prospective twin study. BMJ
compares with the prevalence of Addison’s disease of 1 1999;318:698 –702.
in 10,000 in the general community, providing strong [13] Nerup J, Platz P, Andersen OO, Christy M, Lyngsoe J, Poulsen
evidence that HLA genes play a role in the expression of JE, et al. HL-A antigens and diabetes mellitus. Lancet 1974;2:
non-islet cell organ-specific autoantibodies in individ- 864– 866.
[14] Lambert AP, Gillespie KM, Thomson G, Cordell HJ, Todd
uals with T1DM.
For personal use only.

JA, Gale EA, et al. Absolute risk of childhood-onset type 1

diabetes defined by human leukocyte antigen class II genotype:
A population-based study in the United Kingdom. J Clin
Conclusion
Endocrinol Metab 2004;89:4037 –4043.
In recent years, researchers have significantly advanced [15] Barker JM, Barriga KJ, Yu L, Miao D, Erlich HA, Norris JM,
et al. Prediction of autoantibody positivity and progression to
our knowledge of the underlying autoimmune pro-
type 1 diabetes: Diabetes autoimmunity study in the young
cesses leading to pancreatic beta-cell destruction, the (DAISY). J Clin Endocrinol Metab 2004;89:3896– 3902.
antibodies that are produced as this process develops [16] Bottazzo GF, Florin-Christensen A, Doniach D. Islet-cell
and the temporal relationship between the two. As the antibodies in diabetes mellitus with autoimmune polyendo-
knowledge in this field has evolved, so have become crine deficiencies. Lancet 1974;2:1279–1283.
apparent potential “windows of opportunity” to [17] Yu L, Eisenbarth GS. Type 1 diabetes: molecular, cellular and
clinical immunology. Humoral Autoimmunity. Online edition,
intervene in the pathway of beta-cell destruction that
2 February 2007.
are being aggressively explored in many centers [18] Bonifacio E, Bingley PJ, Shattock M, Dean BM, Dunger D,
worldwide. In addition, we have come to understand Gale EA, et al. Quantification of islet-cell antibodies and
a great deal about associated autoimmune diseases that prediction of insulin-dependent diabetes. Lancet 1990;335:
can occur outside the pancreas in patients with T1DM, 147– 149.
diseases also marked by organ-specific antibodies in [19] Greenbaum CJ, Palmer JP, Kuglin B, Kolb H. Insulin
autoantibodies measured by radioimmunoassay methodology
those same patients at risk for, or with, T1DM.
are more related to insulin-dependent diabetes mellitus than
those measured by enzyme-linked immunosorbent assay:
Results of the Fourth International Workshop on the
References Standardization of Insulin Autoantibody Measurement. J Clin
[1] Geiss LS, Pan L, Cadwell B, Gregg EW, Benjamin SM, Endocrinol Metab 1992;74:1040–1044.
Engelgau MM. Changes in incidence of diabetes in US adults, [20] Berson SA, Yalow RS. Quantitative aspects of the reaction
1997–2003. Am J Prev Med 2006;30:371–377. between insulin and insulin-binding antibody. J Clin Invest
[2] Taplin CE, Craig ME, Lloyd M, Taylor C, Crock P, Silink M, 1959;38:1996 –2016.
et al. The rising incidence of childhood type 1 diabetes in New [21] Palmer JP, Asplin CM, Clemons P, Lyen K, Tatpati O, Raghu
South Wales, 1990–2002. Med J Aust 2005;183:243 –246. PK, et al. Insulin antibodies in insulin-dependent diabetics
[3] Karvonen M, Viik-Kajander M, Moltchanova E, Libman I, before insulin treatment. Science 1983;222:1337–1339.
LaPorte R, Tuomilehto J. Incidence of childhood type 1 [22] Ziegler AG, Vardi P, Gross DJ, Bonner-Weir S, Villa-Komaroff
diabetes worldwide. Diabetes Mondiale (DiaMond) Project L, Halban PA, et al. Production of insulin antibodies by mice
Group Diabetes Care 2000;23:1516–1526. rejecting insulin transfected cells. J Autoimmun 1989;2:
[4] Haynes A, Bower C, Bulsara MK, Jones TW, Davis EA. 219– 227.
Continued increase in the incidence of childhood type 1 [23] Pugliese A, Bugawan T, Moromisato R, Awdeh ZL, Alper CA,
diabetes in a population-based Australian sample (1985– Jackson RA, et al. Two subsets of HLA-DQA1 alleles mark
2002). Diabetologia 2004;47:866–870. phenotypic variation in levels of insulin autoantibodies in first
 18 C. E. Taplin & J. M. Barker

degree relatives at risk for insulin-dependent diabetes. J Clin specificities determines risk of type I diabetes. J Autoimmun
Invest 1994;93:2447–2452. 1996;9:379–383.
[24] Achenbach P, Koczwara K, Knopff A, Naserke H, Ziegler AG, [35] Verge CF, Gianani R, Kawasaki E, Yu L, Pietropaolo M,
Bonifacio E. Mature high-affinity immune responses to Jackson RA, et al. Prediction of type I diabetes in first-degree
(pro)insulin anticipate the autoimmune cascade that leads to relatives using a combination of insulin, GAD, and
type 1 diabetes. J Clin Invest 2004;114:589– 597. ICA512bdc/IA-2 autoantibodies. Diabetes 1996;45:926–933.
[25] Payton MA, Hawkes CJ, Christie MR. Relationship of the [36] Achenbach P, Warncke K, Reiter J, Naserke HE, Williams AJ,
37,000- and 40,000-M(r) tryptic fragments of islet antigens in Bingley PJ, et al. Stratification of type 1 diabetes risk on the
insulin-dependent diabetes to the protein tyrosine phospha- basis of islet autoantibody characteristics. Diabetes 2004;53:
tase-like molecule IA-2 (ICA512). J Clin Invest 1995;96: 384– 392.
1506–1511. [37] Keskinen P, Korhonen S, Kupila A, Veijola R, Erkkila S,
[26] Boitard C, Bonifacio E, Bottazzo GF, Gleichmann H, Savolainen H, et al. First-phase insulin response in young
Molenaar J. Immunology and Diabetes Workshop: Report on healthy children at genetic and immunological risk for type I
the Third International (Stage 3) Workshop on the standard- diabetes. Diabetologia 2002;45:1639 –1648.
isation of cytoplasmic islet cell antibodies. Held in New York, [38] Koczwara K, Bonifacio E, Ziegler AG. Transmission of
October 1987 Diabetologia 1988;31:451–452. maternal islet antibodies and risk of autoimmune diabetes in
[27] Bonifacio E, Boitard C, Gleichmann H, Shattock MA, offspring of mothers with type 1 diabetes. Diabetes 2004;53:
Molenaar JL, Bottazzo GF. Assessment of precision, con- 1– 4.
cordance, specificity, and sensitivity of islet cell antibody [39] Martin S, Wolf-Eichbaum D, Duinkerken G, Scherbaum WA,
Autoimmunity Downloaded from informahealthcare.com by University of Guelph on 05/06/12

measurement in 41 assays. Diabetologia 1990;33:731–736. Kolb H, Noordzij JG, et al. Development of type 1 diabetes
[28] Bingley PJ, Bonifacio E, Mueller PW. Diabetes Antibody
despite severe hereditary B lymphocyte deficiency. N Engl J
Standardization Program: First assay proficiency evaluation.
Med 2001;345:1036–1040.
Diabetes 2003;52:1128 –1136.
[40] Umpierrez GE, Latif KA, Murphy MB, Lambeth HC, Stentz
[29] Achenbach P, Schlosser M, Williams AJ, Yu L, Mueller PW,
F, Bush A, et al. Thyroid dysfunction in patients with type 1
Bingley PJ, et al. Combined testing of antibody titer and
diabetes: A longitudinal study. Diabetes Care 2003;26:
affinity improves insulin autoantibody measurement: Diabetes
1181–1185.
Antibody Standardization Program. 2007. Clinical Immu-
[41] Barker JM, Yu J, Yu L, Wang J, Miao D, Bao F, et al.
nology 2007;122:85–90.
Autoantibody “subspecificity” in type 1 diabetes: Risk for
[30] Li L, Hagopian WA, Brashear HR, Daniels T, Lernmark A.
organ-specific autoimmunity clusters in distinct groups.
Identification of autoantibody epitopes of glutamic acid
decarboxylase in stiff-man syndrome patients. J Immunol Diabetes Care 2005;28:850–855.
For personal use only.

1994;152:930 –934. [42] Cerutti F, Bruno G, Chiarelli F, Lorini R, Meschi F, Sacchetti

[31] Nejentsev S, Sjoroos M, Soukka T, Knip M, Simell O, C, et al. Younger age at onset and sex predict celiac disease in
Lovgren T, et al. Population-based genetic screening for the children and adolescents with type 1 diabetes: An Italian
estimation of type 1 diabetes mellitus risk in Finland: Selective multicenter study. Diabetes Care 2004;27:1294 –1298.
genotyping of markers in the HLA-DQB1, HLA-DQA1 and [43] Yu L, Brewer KW, Gates S, Wu A, Wang T, Babu SR, et al.
HLA-DRB1 loci. Diabet Med 1999;16:985–992. DRB1*04 and DQ alleles: Expression of 21-hydroxylase
[32] Ziegler AG, Hummel M, Schenker M, Bonifacio E. autoantibodies and risk of progression to Addison’s disease.
Autoantibody appearance and risk for development of J Clin Endocrinol Metab 1999;84:328–335.
childhood diabetes in offspring of parents with type 1 diabetes: [44] Betterle C, Scalici C, Presotto F, Pedini B, Moro L, Rigon F,
The 2-year analysis of the German BABYDIAB Study. et al. The natural history of adrenal function in autoimmune
Diabetes 1999;48:460–468. patients with adrenal autoantibodies. J Endocrinol 1988;117:
[33] Yu L, Rewers M, Gianani R, Kawasaki E, Zhang Y, Verge C, 467– 475.
et al. Antiislet autoantibodies usually develop sequentially [45] Betterle C, Volpato M, Rees SB, Furmaniak J, Chen S,
rather than simultaneously. J Clin Endocrinol Metab 1996;81: Zanchetta R, et al. II. Adrenal cortex and steroid 21-
4264–4267. hydroxylase autoantibodies in children with organ specific
[34] Verge CF, Gianani R, Kawasaki E, Yu L, Pietropaolo M, autoimmune diseases: Markers of high progression to clinical
Chase HP, et al. Number of autoantibodies (against insulin, Addison’s disease. J Clin Endocrinol Metab 1997;82:
GAD or ICA512/IA2) rather than particular autoantibody 939– 942.