| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Laboratory for Immunogenetics, Ludwig Maximilians University (K.M.F., E.K., E.D.A.), and the Third Medical Department, Krankenhaus München-Schwabing, and Diabetes Research Institute (K.M.F., A.-G.Z.), 80804 Munich, Germany
Address all correspondence and requests for reprints to: Priv.-Doz. Dr. Anette-G. Ziegler, Krankenhaus München-Schwabing, Diabetes Research Institute, Kölner Platz 1, 80804 Munich, Germany. E-mail: anziegler{at}lrz.uni-muenchen.de
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Autoimmune type 1 diabetes is also associated with specific human leukocyte antigen (HLA) DR3- and DR4-containing haplotypes, and the risk is particularly high for DR3/4 heterozygous or DR4/4-DQB1*0302/*0302 homozygous individuals (reviewed in Refs. 2, 3). Although almost all pediatric IDDM patients have DR3 and/or DR4, these haplotypes are less prevalent in patients with diabetes onset after 15 yr of age (4). Furthermore, it has been suggested that DR3 is associated with a more slowly progressive form of IDDM (5). Few data on genetic markers in gestational diabetes exist (6, 7, 8, 9, 10). The aim of the present study was to analyze the frequency of HLA class II alleles and genotypes in patients with GDM and to determine whether HLA markers can predict the development of islet autoantibodies and IDDM postpartum. We found that GDM per se is not associated with specific HLA class II alleles, that the proportion of GDM women with islet autoantibodies is highest in those who carry the DR3 or DR4 allele, and that women with a DR3 or DR4 genotype have a higher risk to develop IDDM after pregnancy compared to non-DR3/non-DR4 patients.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
From 1989–1996, 481 Caucasian women with GDM were recruited in
the context of the multicenter prospective study BABYDIAB in different
hospitals of Germany. They were tested for islet autoantibodies at
delivery and were followed for the development of IDDM or NIDDM
postpartum. All subjects were recruited consecutively on admission to
delivery in a nonselective fashion, as described in detail previously
(1). Diagnosis of GDM was made according to the German Diabetes
Association, using an oral glucose tolerance test with a 75-g glucose
load. GDM was diagnosed if 2 of 3 capillary blood glucose values
exceeded the following limits: more than 5.0 mmol/L (fasting) before
oral glucose tolerance test, more than 10.6 mmol/L after 60 min, and
more than 8.9 mmol/L after 120 min. All participating centers (1) were
asked to follow the therapeutic guidelines for the treatment of GDM
recommended by the Diabetes and Pregnancy Study Group of the German
Diabetes Association to ensure a uniform therapeutic strategy among the
participating hospitals. In accordance with these guidelines, insulin
therapy was required if capillary blood glucose levels, during a diet
regimen for at least 1 week, exceeded the following limits: more than
5.0 mmol/L before and more than 7.8 mmol/L at 60 min and more than 6.7
mmol/L at 120 min after meals. After delivery, patients were invited to
participate in a follow-up examination at 9 months, 2 yr, and 5 yr
postpartum, when oral glucose tolerance tests were performed and venous
blood samples for HLA typing were requested. Of all 481 women enrolled,
332 participated in at least 1 follow-up visit with an oral glucose
tolerance test postpartum (Fig. 1
). Of
those, 184 women also accepted blood drawing for HLA typing. They were
included into the analysis of this manuscript (median age, 31.4 yr). Of
all of the women enrolled, 149 were lost to follow-up because they had
moved to an unknown address or refused participation. As shown in Fig. 1, there was no major difference in islet autoantibody frequency or
frequency of IDDM, NIDDM, or impaired glucose tolerance (IGT) between
GDM women with or without HLA typing. Furthermore, no differences in
baseline characteristics (age at delivery, body mass index before
pregnancy, or gestation duration) between women who were HLA typed
compared to those for whom HLA typing was not available were found (z
test). Of the 184 women included in the study, 120 were treated with
diet alone [White classification (11) GDM-A], and 64 were treated
with insulin (White GDM-B) during pregnancy. Postpartum diabetes was
diagnosed according to WHO criteria (2 h blood glucose level, >200
mg/dL = >11.1 mmol/L). Diabetes was classified as insulin
dependent (IDDM) in the presence of ketoacidosis or as requiring
insulin treatment not only to control hyperglycemia and symptoms but to
prevent the spontaneous occurrence of ketoacidosis. This classification
was based upon clinical evaluation and not influenced through antibody
measurement or HLA typing. Due to instructions given to us by the
ethical approval board of the Bayerische Landesärztekammer,
clinicians and patients were not informed about these results. Women
with IGT postpartum (n = 12) were not considered diabetic in this
analysis. The median follow-ups from the day of delivery to postpartum
diabetes onset and to the last contact with the patient were 0.8 yr
(0–3.5 yr) and 3.0 yr (1.1–7.7 yr), respectively. Informed consent
was obtained from all patients who participated in the study.
|
HLA class II allele frequencies were also determined in 254 nondiabetic unrelated control subjects from the same ethical background and region of Germany (median age, 39.5 yr; 50% male).
HLA typing
HLA class II alleles, HLA-DRB1, HLA-DQA1, and HLA-DQB1, were determined using PCR-amplified DNA and nonradioactive oligonucleotide probes (sequence-specific oligonucleotide typing). The isolation of the genomic DNA was performed from fresh peripheral blood cells (ethylenediamine tetraacetate-blood) by the salting out procedure described in the 11th Workshop report (12). Genomic DNA (1 µg) was amplified by PCR in a 100-µL reaction volume using 25 pmol primers (12), 20 nmol deoxy-NTPs, 2.5 U Taq polymerase, and buffer [10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, 1.5 mmol/L MgCl2, and 0.001% gelatin]. For the dot blot hybridization the amplified DNA (2 µL) was spotted onto nylon membranes (Boehringer Mannheim, Mannheim, Germany) and hybridized with 15 DRB generic and 17 DQA- and 20 DQB-specific oligonucleotides, respectively (12). The hybridized oligonucleotides, 3'-end labeled with digoxygenin-11–2',3'-dideoxyuridine-5'-triphosphate (Boehringer Mannheim) were detected using anti-digoxigenin-AP Fab (Boehringer Mannheim), and visualization was achieved with the chemiluminescent substrate CSPD (Boehringer Mannheim) (13). The HLA-DRB1, HLA-DQA1, and HLA-DQB1 alleles were named according to the nomenclature of 1996 (14). The alleles DQA1*03011 and *0302 were indistinguishable by sequence-specific oligonucleotide typing and were therefore referred to as DQA1*03. Also, the alleles DQB1*0201 and *0202 were indistinguishable and were referred to as DQB1*02. In 179 women, complete HLA genotyping (DR, DQA, and DQB) was available. In five women (one with elevated islet autoantibodies), complete typing could not be obtained (one with missing DQA, one with missing DR and DQA, two with missing DQB, and one with missing DR alleles).
Islet autoantibody determination
Autoantibodies to glutamic acid decarboxylase (GADA), the protein tyrosine phosphatase IA-2 (IA2A), and islet cells (ICA) were analyzed as previously described (1, 15). The upper limit of normal was defined by the 99th percentile of antibody levels determined in nondiabetic women at delivery (GADA, 13 U; IA2A, 5 U; ICA, 5 Juvenile Diabetes Foundation units) (1). Antibody assays were entered into the international Antibody Proficiency Program (organized by N. MacLaren, Gainesville, FL). The GADA assay achieved a specificity of 100% and a sensitivity of 94% (Third Workshop), the IA2A assay achieved specificities of 100% and 100% (First Workshop), and the ICA assay achieved specificities of 100% and 67% (NinthWworkshop), respectively.
Statistical analysis
Allele and genotype frequencies were determined for patients and the control group. If only one allele could be detected, we assumed that the subject was homozygous for this allele. P values were determined by Fisher’s exact test and were corrected for the number of comparison made (Pc; DRB1, Pc = puncorr x 11; DQA1, Pc = puncorr x 9; DQB1, Pc = puncorr x 14). Kaplan-Meier life table analysis was used to determine the cumulative risk of diabetes postpartum. Follow-up started with the day of delivery and ended with diabetes or the last contact with the patient. Five women with GDM remained insulin dependent from the day of delivery, with a calculated follow-up time of zero (life table curves for these cases start below 100%). Differences in survival experiences were determined by log-rank test in the Statistical Package for Social Science (SPSS, Inc., Chicago, IL). Ninety-five percent confidence intervals (CI) for cumulative risk were calculated from the SE thereof.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Table 1 summarizes allele
frequencies of HLA-DRB1, -DQA1, and -DQB1 in women with GDM compared to
control subjects. No significant differences in the distribution of
allele frequencies were observed between cohorts.
|
Of women with GDM, 43 had at least 1 elevated islet autoantibody
(GADA, IA2A, or ICA) at delivery. In this group a significant increase
in allele frequency of DRB1*03 was observed compared to that in the
control population (23.8% vs. 10.6%;
Puncorr = 0.001;
Pc = 0.02; Table 1
). After stratification
for DR3, the frequency of DR4 was also significantly increased
(Puncorr = 0.007). In autoantibody-negative
women, the frequency distribution of HLA class II alleles was similar
to that in control subjects (Table 1). Of all islet
autoantibody-positive women with GDM, 24 (55.8%) had elevated GADA at
delivery. In this population IDDM-associated alleles were even more
prevalent (DR3, 32.6%; DR4, 32.4%; DQA1*03, 34.8%; DQB1*0302,
31.2%; Pc = 0.002, 0.009, 0.004, and
0.0003, respectively; Table 1). Twenty-five of the women with islet
autoantibodies (59.5%) and 19 (82.6%) of those with GADA had a
genotype with at least 1 DR3 or DR4 allele.
HLA class II allele and genotype frequencies in women with GDM who developed diabetes postpartum
Twenty-four women with GDM developed IDDM during follow-up
postpartum (Table 2; median age 31
years). In this subgroup of women with postpartum IDDM, a higher
frequency of DRB1*03 was detected compared to that in control subjects
(30.4% vs. 10.6%; Pc = 0.005;
Table 1). After stratification for DR3, a significant increase in DR4
in GDM women who developed IDDM postpartum was also observed
[uncorrected P (Puncorr) =
0.02]. Seventeen (73.9%) of the 24 women who developed IDDM
postpartum had a genotype with at least 1 DR3 or DR4 allele (Table 2);
none, however, was DR3/4 heterozygous. All but 1 women with DR4 had the
DR4-DQB1*0302 haplotype. Six women with GDM who developed IDDM
postpartum had a DR2-containing genotype and four of those included the
"protective" DQB1*0602 allele rarely found in patients with IDDM
(16). In the group of patients with GDM who developed NIDDM postpartum
(median age, 32 yr), no specific allele was elevated (Table 1).
|
Of all women with GDM, 80 had a DR3- or DR4-containing genotype,
whereas 102 were typed as non-DR3/non-DR4. The cumulative risk to
develop IDDM within 2 yr postpartum was 20.9% (95% CI, 12–30) in
women with DR3 or DR4 compared to 6.5% (95% CI, 1–12) in women
without a DR3- or DR4-containing genotype (P = 0.002;
Fig. 2). In DR3- or DR4-positive women
with GDM who were treated with insulin during pregnancy (GDM-B), the
2-yr risk of IDDM was 47.7% (95% CI, 29–67) compared to 13.9% (95%
CI, 1–27) in DR3- and DR4-negative GDM-B women (P =
0.007; Fig. 3A). The risk of IDDM in
women with GDM who were treated with diet during pregnancy in contrast
was relatively low, and no difference between genotypes was observed
(Fig. 3B).
|
|
compares the cumulative life
table risk of IDDM postpartum relative to the presence or absence of a
DR3 or DR4 genotype in GDM women with or without islet autoantibodies
at delivery. The cumulative life table risk was increased in women with
islet autoantibodies (Fig. 4, A vs. B) and particularly in
those women who had islet autoantibodies and, in addition, a DR3- or
DR4-containing genotype (Fig. 4A; risk at 2 yr postpartum, 44.3%; 95%
CI, 25–64), and it was virtually absent in women who were negative for
both antibodies and a DR3-, DR4-containing genotype (Fig. 4B; risk at 2
yr postpartum, 2.8%; 95% CI, 0–7). However, when comparing the
cumulative risk of antibody-positive patients relative to the presence
or absence of a DR3 or DR4 genotype and the presence or absence of
GADA, no increase in risk was observed in GADA-positive patients with
DR3- or DR4-containing genotypes (Fig. 5), indicating that GADA by themselves
were the strongest predictor of IDDM risk. Altogether 92% of women
with GDM and progression to IDDM postpartum had islet autoantibodies
and/or a DR3- or DR4-containing genotype; the 2-yr life table risk for
this group was 22.6% (95% CI, 15–30) compared to 2.8% (95% CI,
0.1–7) in women without islet autoantibodies and without a DR3-,
DR4-containing genotype (P = 0.0001).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Other studies have also reported that the distribution of HLA class II alleles in women with GDM or IGT during pregnancy does not differ from that in pregnant women with normal glucose tolerance (6, 10, 17). As GDM is a heterogeneous disorder predisposing to the development of IDDM and NIDDM postpartum, it is not surprising that there is no increase in specific HLA alleles, in particular not of those associated with susceptibility to IDDM. Furthermore, women with GDM who develop IDDM later in life represent a subgroup of IDDM patients with a relatively slowly evolving form of diabetes and late onset of the disease. In these patients, in whom clinical symptoms are often not so pronounced, it has been shown that the prevalence of DR3- or DR4-positive genotypes is lower than that in patients with type 1 diabetes with onset of diabetes before puberty (8, 18). Caillat-Zucman and co-workers reported that 97% of children with onset of IDDM before 15 yr of age had a DR3- or DR4-containing genotype compared to 74% of patients with onset after 35 yr of age (4). In the present study, we found that the allele frequency of DR3 was significantly increased in women who developed IDDM postpartum compared to the background prevalence in the general population. We also observed that 74% of women who developed postpartum IDDM had a DR3- or DR4-positive genotype. These results are in accordance with previous studies (4, 6, 7, 8, 9, 10) and are remarkably similar to those obtained in the study of Caillat-Zucman, who investigated patients without preceding GDM. They indicate that the subgroup of patients with GDM who develop IDDM postpartum have typical HLA genotypes of late-onset IDDM.
Not only is the distribution of HLA alleles different between patients with childhood or adult-onset IDDM but so is the antibody profile. In adult patients with newly diagnosed IDDM or pretype 1 diabetes, GADA are the most prevalent antibodies, whereas in young subjects, IA-2A and insulin autoantibodies are more frequent (19). In women with GDM, we have previously shown that GADA are the most sensitive single antibody markers to predict the development of IDDM postpartum (1). We here demonstrate that in GDM, as found in IDDM (19, 20), they are strongly associated with the presence of DR3 in particular, but also with DR4 and DQB1*0302.
Islet autoantibodies measured at delivery in women with GDM can predict the onset of IDDM postpartum with a risk correlating to the number of autoantibodies present in the patient’s serum (1). In this study, HLA markers of IDDM were also associated with an increased risk of progression to IDDM postpartum. This was particularly so for women who had required insulin treatment during pregnancy (GDM-B), in whom the risk of disease within 2 yr after pregnancy was 50% when they carried a DR3- or DR4-containing genotype. This suggests that HLA typing for the presence of DR3 or DR4 in women with GDM of White B could be used as an alternative to islet autoantibody testing to assess IDDM risk postpartum. Of women with postpartum IDDM, 74% had a DR3- or DR4-positive genotype, and a similar proportion (71%) had elevated islet autoantibodies. By combining antibody measurement and determination of susceptible HLA alleles (DR3, DR4), 92% of patients with progression to IDDM were identified, suggesting that the combination of both markers may increase sensitivity. Moreover, women without antibodies and without a DR3 or DR4 genotype had a low risk of developing IDDM (2.8%), indicating that HLA typing in antibody-negative women may be useful for excluding IDDM risk. In women with positive islet autoantibodies, in contrast, additional HLA typing was less informative. Here, the presence or absence of GADA seems to be the most important factor that determines diabetes risk independent of HLA genes. This implies that HLA provides susceptibility for antibodies, but has no major effect on the rate of progression to IDDM once antibodies are detectable.
We conclude that specific HLA alleles predispose to the development of diabetes-associated autoantibodies and type 1 diabetes postpartum. The measurement of antibodies, in particular GADA, appears to be the single most useful screen in women with GDM, but when islet autoantibody testing is not available, HLA typing for the presence or absence of DR3 or DR4 may be a useful strategy in women who required insulin during pregnancy. In islet autoantibody-negative women, HLA typing may also be useful to increase either sensitivity (presence of HLA-DR3 or -DR4) or specificity (absence of HLA-DR3 or -DR4).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received December 28, 1998.
Revised March 12, 1999.
Accepted March 23, 1999.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. Lobner, A. Knopff, A. Baumgarten, U. Mollenhauer, S. Marienfeld, M. Garrido-Franco, E. Bonifacio, and A.-G. Ziegler Predictors of Postpartum Diabetes in Women With Gestational Diabetes Mellitus Diabetes, March 1, 2006; 55(3): 792 - 797. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Narendran, E. Estella, and S. Fourlanos Immunology of type 1 diabetes QJM, August 1, 2005; 98(8): 547 - 556. [Full Text] [PDF]
|
|
|
|
 |
|
 L. L. Moore, M. L. Bradlee, M. R. Singer, K. J. Rothman, and A. Milunsky Chromosomal Anomalies among the Offspring of Women with Gestational Diabetes Am. J. Epidemiol., April 15, 2002; 155(8): 719 - 724. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
