help button home button Endocrine Society JCEM
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Li, H.
Right arrow Articles by Tuomi, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Li, H.
Right arrow Articles by Tuomi, T.
The Journal of Clinical Endocrinology & Metabolism Vol. 88, No. 6 2767-2774
Copyright © 2003 by The Endocrine Society

A Combination of Human Leukocyte Antigen DQB1*02 and the Tumor Necrosis Factor {alpha} Promoter G308A Polymorphism Predisposes to an Insulin-Deficient Phenotype in Patients with Type 2 Diabetes

Haiyan Li, Leif Groop, Anita Nilsson, Jianping Weng and Tiinamaija Tuomi

Diabetes and Endocrine Research Laboratory (H.L., L.G., A.N., T.T.), Department of Endocrinology, Lund University, S-20502 Malmö, Sweden; Department of Endocrinology (J.W.), The First University Hospital, Sun Yat-Sen University, Guangzhou, China 510080; and Department of Internal Medicine (T.T.), Helsinki University Central Hospital, Helsinki FIN-00029, Finland

Address all correspondence and requests for reprints to: Haiyan Li, Ph.D., Wallenberg Laboratory, Department of Endocrinology, Lund University, S-20502 Malmö, Sweden. E-mail: haiyan.li{at}endo.mas.lu.se.


   Abstract
Top
 Abstract
Introduction
 Subjects and Methods
 Results
 Discussion
 References
 
Our previous results have suggested that genes outside the human leukocyte antigen (HLA) class II locus may affect the phenotype of type 2 diabetic patients from families with both type 1 and type 2 diabetes (mixed type 1/2). To study whether the TNF{alpha} gene could be such a modifying gene, we studied TNF{alpha} promoter polymorphisms (G->A substitution at positions -308 and -238) in relation to HLA-DQB1 genotypes in type 2 patients from mixed type 1/2 families or common type 2 diabetes families as well as in patients with adult-onset type 1 diabetes and control subjects. The TNF{alpha}308 AA/AG genotype frequency was increased in adult onset type 1 patients (55%, 69 of 126), but it was similar in type 2 patients from type 1/2 families (35%, 33/93) or common type 2 families (31%, 122 of 395), compared with controls (33%, 95/284; P < 0.0001 vs. type 1). The TNF{alpha}308 A and DQB1*02 alleles were in linkage disequilibrium in type 1 patients (Ds = 0.81; P < 0.001 vs. Ds = 0.25 in controls) and type 2 patients from type 1/2 families (Ds = 0.59, P < 0.05 vs. controls) but not in common type 2 patients (Ds = 0.39). The polymorphism was associated with an insulin-deficient phenotype in the type 2 patients from type 1/2 families only together with DQB*02, whereas the common type 2 patients with AA/AG had lower waist to hip ratio [0.92 (0.12) vs. 0.94 (0.11), P = 0.008] and lower fasting C-peptide concentration [0.48 (0.47) vs. 0.62 (0.46) nmol/liter, P = 0.020] than those with GG, independently of the presence of DQB1*02. In conclusion, TNF{alpha} is unlikely to be the second gene in the HLA area responsible for our previous findings in type 1/2 patients. However, we could show an association between TNF{alpha}308 polymorphism and the phenotype of common type 2 diabetes.


   Introduction
Top
 Abstract
 Introduction
Subjects and Methods
 Results
 Discussion
 References
 
THE PHENOTYPE OF patients with type 2 diabetes is influenced by the family history of diabetes (1, 2). Particularly, a family history of both type 1 and type 2 diabetes was associated with low fasting C-peptide concentration and low frequency of cardiovascular disease (1). On the other hand, family history of type 2 diabetes could influence the features of insulin resistance in patients with type 1 diabetes (3, 4, 5, 6). One natural explanation for these associations could be a shared genetic background between type 1 and type 2 diabetes.

Indeed, sharing type 1 diabetes-associated human leukocyte antigen (HLA) risk haplotypes (either DR4-DQB1*0302 or DR3-DQB1*02) with a type 1 diabetic relative was associated with impaired insulin secretion in response to oral glucose in type 2 diabetic patients (2). Among the second-degree or more distant relatives of type 1 diabetic patients, those sharing a risk haplotype with the type 1 diabetic relative had a lower insulin response, compared with those not sharing such haplotypes. However, no such effect was seen in unrelated patients with similar risk HLA genotypes, suggesting that other genes or combination of alleles on the short arm of chromosome 6 need to be shared.

One such candidate could be the TNF{alpha} gene in the class III region of HLA (7, 8, 9). The gene encodes for a cytokine initially implicated in inflammatory and immunoregulatory actions, which makes it a good candidate gene for type 1 diabetes. A single nucleotide polymorphism (G->A substitution at position -308, G308A) in the promoter region of the TNF{alpha} gene was associated with type 1 diabetes in a manner dependent on HLA-DR3 (10, 11, 12). In Caucasians, the extended DR3-DQB1*02 haplotype containing the A allele correlated with a high TNF{alpha} secretion (11, 13) and conferred a higher relative risk than DR3-DQB1*02 per se did (8, 10, 14, 15, 16). Moreover, TNF{alpha} has also been ascribed a role in the development of obesity, insulin resistance, and possibly type 2 diabetes (17, 18). G308A has been reported to be associated with either increased (19, 20, 21, 22, 23, 24, 25) or decreased (26, 27, 28, 36) insulin resistance in both type 2 diabetic patients and nondiabetic subjects. Moreover, another promoter polymorphism, G238A (G->A substitution at position -238) was associated with a decreased insulin resistance in nondiabetic relatives of type 2 diabetic patients (29). Therefore, we hypothesized that the TNF{alpha} polymorphism could together with the HLA class II locus modify the risk for type 2 diabetes and explain part of the phenotypic heterogeneity in type 2 diabetes, especially in the patients from the mixed type 1/2 diabetes families.


   Subjects and Methods
Top
 Abstract
 Introduction
 Subjects and Methods
Results
 Discussion
 References
 
Subjects

The Botnia study is a population-based study aiming at the identification of genes that increase susceptibility to type 2 diabetes (30). Since 1990, we have recruited 695 families with at least two members with type 2 diabetes from Finland (2). Of them, 100 also had a member with type 1 diabetes (mixed type 1/2 families), and 595 families had only members with type 2 diabetes (common type 2 diabetes families). Families with genetically confirmed maturity-onset diabetes of the young (MODY1, MODY2, MODY3) (31) were excluded. All available family members were invited to participate in the study, and an oral glucose tolerance test (OGTT) was performed to more than 90% of the nondiabetic subjects studied. Diabetes was diagnosed according to the revised World Health Organization criteria (32). All patients with type 1 diabetes had commenced permanent insulin treatment within 6 months after diagnosis and had fasting C-peptide concentrations less than 0.2 nmol/liter. Patients who did not fulfill these criteria were considered to have type 2 diabetes. All type 2 diabetic patients in the present study had been treated with diet and/or oral hypoglycemic agents for at least 1 yr before commencing insulin treatment. They did not have rapid weight loss or ketosis at diagnosis. Informed consent was obtained from all subjects, and the study was approved by the local ethics committee.

The present study included randomly selected unrelated subjects from four groups: 1) type 2 diabetic patients from mixed type 1/2 diabetes families [n = 93, 39 males and 54 females; age (mean ± SD), 64.0 ± 12.8 yr; body mass index (BMI), 27.6 ± 3.8 kg/m2; hemoglobin A1c (HbA1c), 7.6% ± 1.6%, 18% having glutamic acid decarboxylase antibodies (GADab)]; 2) type 2 diabetic patients from common type 2 diabetes families (n = 395, 173 males and 222 females; age, 65.7 ± 9.8 yr; BMI, 28.9 ± 4.7 kg/m2; HbA1c, 7.4% ± 1.5%, 8% having GADab); 3) GADab-negative control subjects without family history of diabetes (n = 284, 129 males and 155 females; age, 55.0 ± 12.9 yr); and 4) patients with adult-onset type 1 diabetes (n = 126, 73 males and 53 females; age at diagnosis, >20 yr, mean, 30.7 ± 8.6 yr; fasting serum C-peptide, 0.0 ± 0.0 nmol/liter, 65% GADab positive) from the Diabetes 2000 Registry in southern Sweden (2).

Genotyping

The TNF{alpha} promoter polymorphisms at positions -308 and -238, both involving a substitution of guanine by adenine, were amplified by PCR according to the method of Day et al. (29). Briefly, mismatch primers creating new restriction sites together with the common alleles were used to amplify genomic DNA, followed by digestion with NcoI (TNF{alpha}308) or MspI (TNF{alpha}238) according to the manufacturer’s instructions (New England Biolabs, Beverly, MA). The PCR products digested with NcoI or MspI were pooled for each individual, separated on a 5% polyacrylamide (49 acrylamide:1 bisacrylamide) gel, and visualized with ethidium bromide, using Gel-Scan 2000 equipment and GelPro software (Corbett Research, Sydney, Australia).

The HLA DQB1 genotypes were typed by using three DQB1 probes to distinguish DQB1 alleles 0201 or 0202(02), 0302, and 0602 or 0603 [0602(3)] alleles (1). The genotypes were presented as 02/0302, 0302/X, 02/X, 0602(3)/X, 0302/0602(3), 02/0602(3), and X/X, where X could mean either a homozygous allele or any allele other than 02, 0302, or 0602(3). To study whether the TNF{alpha}308 polymorphism was associated with DQB1*02 allele, which is in linkage disequilibrium with DR3, the frequency of TNF{alpha}308 AA, AG, and GG genotypes was compared according to whether the subjects had two, one, or no DQB1*02 alleles. The comparison was made in 593 subjects (including 218 control subjects, 67 type 2 patients from the mixed type 1/2 families, 182 type 2 patients from the common type 2 diabetes families, and 126 adult-onset type 1 patients), for whom the full DQB1 genotypes were available (2). Also, subjects positive for DQB1*02 were typed for HLA-DR3 by allele-specific PCR followed by restriction fragment-length polymorphism analysis (2).

Metabolic measurements

An OGTT was performed for all subjects aged more than 15 yr with fasting blood glucose less than 10 mmol/liter and not treated with insulin. After 12 h of overnight fast, the subjects ingested 75 g glucose in a volume of 300 ml. Samples for measurements of blood glucose and serum insulin were drawn at -10, 0, 30, 60, and 120 min, and incremental glucose or insulin areas under the curve were calculated. Blood glucose was measured with a glucose oxidase method using a Glucose Analyzer II (Beckman Instruments, Fullerton, CA). Serum insulin concentrations were measured by RIA (Pharmacia, Uppsala, Sweden) with interassay coefficient of variation of 5%. Fasting serum C-peptide concentrations were measured in duplicate by a RIA with an interassay coefficient of variation of 9% (human C-peptide RIA kit, Linco Research, Inc., St. Charles, MO). The HbA1c concentration was measured by an HPLC with a reference range of 5–7%. Homeostasis model assessment for insulin resistance (HOMA) index was calculated based upon fasting insulin and fasting glucose concentrations (33). GADabs were measured by a radioimmunoprecipitation assay employing 35S-labeled recombinant human GAD65 produced by in vitro transcription/translation (34).

Statistical analysis

Data are presented as median (75%–25% interquartile range) and percentages for categorical variables. The statistical analysis was performed with either Biomedical Data Processing (Los Angeles, CA) or Number Cruncher Statistical Systems (Kaysville, UT) statistical software. The group frequencies were compared using the {chi}2 (with Yates correction when appropriate) or Fisher’s exact tests. The differences in continuous variables between groups were tested by the Mann-Whitney test, the Kruskal-Wallis test for group means, or general linear models ANOVA. Logarithmic transformation of data was used for the covariate analysis. A multiple-regression analysis was carried out using fasting C-peptide concentration or waist to hip (WH) ratio as the dependent variable, and DQB1*02 (presence = 1, absence = 0), TNF{alpha}308 polymorphism (presence = 1, absence = 0), age, BMI, and sex as independent variables. The coefficient of linkage disequilibrium (D) between the TNF{alpha}308 A allele and the DQB1*02 allele was analyzed according to the observed and expected numbers (frequencies) of individuals with and without the A allele, in relation to the presence or absence of the *02 allele. The standardized value of D (Ds) was calculated using the formula: Ds = D/Dmax (Dmax is the theoretical maximum disequilibrium value achievable for the combination of the TNF{alpha}308 A and the DQB1*02 alleles).


   Results
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
Discussion
 References
 
Genotype frequency of TNF{alpha} polymorphisms

G308A polymorphism. The genotype frequencies were similar in control subjects (AA-AG-GG, 4%–29%-67%), type 2 diabetic patients from the mixed type 1/2 families (type 1/2: 3%–32%-65%), and type 2 diabetic patients from the common type 2 diabetes families (common type 2: 3%–28%-69%), but genotypes containing the A allele were more frequent in adult-onset type 1 diabetic patients (AA, 10%, P = 0.035; AG, 45%, P = 0.001 vs. controls; Table 1Go).


View this table:
[in this window]
[in a new window]
  		Table 1. Genotype frequency of TNF{alpha} promoter polymorphisms (G308A and G238A) in patients with adult-onset type 1 diabetes, type 2 diabetic patients from the mixed type 1/2 diabetes families or from the common type 2 diabetes families, and nondiabetic control subjects without family history of diabetes

 
G238A polymorphism. The rare AA genotype was found in only two subjects, in one control subject and one common type 2 patient. The heterozygous AG genotype was present in 6% of the control subjects, 6% of the type 2 patients from type 1/2 families, 5% of the common type 2 patients, and 2% of the adult-onset type 1 patients (Table 1Go).

The presence of TNF{alpha} G308A polymorphism in relation to HLA-DQB1 genotypes

The TNF{alpha}308 polymorphism has been associated with HLA-DR3 (11, 12), which is in linkage disequilibrium with the DQB1*02 allele. Thus, we analyzed the frequency of TNF{alpha}308 AA, AG, and GG genotypes according to whether the subjects had two, one, or no DQB1*02 alleles (see Subjects and Methods), and studied, in subjects with AA/AG or GG genotypes, the frequency of DQB1 genotypes containing the *02 allele (DQB1*02-pos).

In adult-onset type 1 patients, the TNF{alpha}308 AA genotype was present significantly more often in subjects with two than in those with one or no DQB1*02 alleles (67% vs. 4% vs. 0%, P < 0.0001). Conversely, the GG genotype was least common in those with two than in those with one or no DQB1*02 alleles (0% vs. 11% vs. 88%, P < 0.0001) (Fig. 1Go). The TNF{alpha}308 A and DQB1*02 alleles were in strong linkage disequilibrium in type 1 diabetic patients (Ds = 0.81; P < 0.001 vs. Ds = 0.25 in control subjects) but not in common type 2 diabetic patients (Ds = 0.39; P = NS vs. controls) or control subjects, whereas the mixed patients showed an intermediate situation (Ds = 0.59; P < 0.05 vs. controls) (Fig. 1Go). Consequently, in the whole study group, subjects with AA/AG genotypes were more often DQB1*02 positive than subjects with GG genotype (control: 51% vs. 25%; mixed type 1/2: 73% vs. 23%; common type 2: 52% vs. 19%; and adult type 1: 90% vs. 11%; all P < 0.00001, Table 2Go). The proportion of subjects having both AA/AG and DQB1*02 was higher among type 2 patients from type 1/2 families (26%, 24 of 93) than in common type 2 patients (16%, 64 of 395, P = 0.030) and control subjects (17%, 48 of 284, P = 0.057), but it was lower than adult-onset type 1 patients (49%, 62 of 126, P = 0.0002). The increased frequency of DQB1*02 in the AA/AG genotype group reflected the increased frequency of 02/0302 and 02/X genotypes in adult-onset type 1 patients, and it was due to the increased frequency of 02/X and 02/0602(3) genotypes in type 2 diabetic patients and control subjects (Table 2Go).



View larger version (32K):
[in this window]
[in a new window]
  		Figure 1. The frequency of TNF{alpha}308 genotypes according to whether the subjects had two, one, or no HLA DQB1*02 alleles in adult-onset type 1 patients (n = 15, 53 or 58, respectively), type 2 patients from type 1/2 families (n = 2, 27 or 38, respectively), and common type 2 patients (n = 3, 70 or 109, respectively), control subjects (n = 4, 76 or 138, respectively). The coefficient of linkage disequilibrium (Ds) between the TNF{alpha}308 A allele and DQB1*02 allele was 0.81 in adult-onset type 1 patients (P < 0.001 vs. Ds = 0.25 in control subjects), 0.59 in type 2 patients from type 1/2 families (P < 0.05 vs. control subjects), and 0.39 in common type 2 patients (P = NS vs. control subjects).

 

View this table:
[in this window]
[in a new window]
  		Table 2. The frequency of DQB1 genotypes, stratified according to the TNF{alpha}308 polymorphism (AA/AG, GG), in adult-onset type 1 patients, type 2 diabetic patients from the mixed type 1/2 diabetes families or from the common type 2 diabetes families, and nondiabetic control subjects

 
Among the DQB1*02-positive subjects, 74% were DR3 positive (type 1 diabetes 94%, type 2 patients from the mixed 1/2 families 74%, common type 2 patients 65%, control subjects 72%). The linkage disequilibrium between DR3 and TNF{alpha}308A was 0.86 for type 1 diabetes, 0.62 for type 2 patients from the mixed 1/2 families, 0.54 for common type 2 patients, and 0.46 for the control subjects.

Clinical comparison between the TNF{alpha}308 AA/AG and GG genotype groups

To study whether the presence of the G308A polymorphism alone or in association with DQB1*02 influenced the phenotype of patients with type 2 diabetes or nondiabetic subjects, we compared the clinical characteristics in subjects with AA/AG vs. GG, and AA/AG and DQB1*02-pos vs. GG and DQB1*02-neg (DQB1 genotypes without the DQB1*02 allele).

The frequency of GADab did not differ between subjects with TNF{alpha}308 AA/AG and GG (type 1/2, 18% vs. 18%; common type 2, 11% vs. 7%; adult-onset type 1, 68% vs. 61%). It also did not differ between those with or without DQB1*02 (type 1/2, 21% vs. 16%; common type 2, 5% vs. 9%; adult type 1, 66% vs. 64%).

Type 2 patients from the mixed type 1/2 families. The clinical characteristics did not differ between type 2 patients from type 1/2 families with AA/AG and those with GG (Table 3Go). DQB1*02-pos patients had a higher fasting glucose concentration [median (interquartile range), 8.6 (5.1) vs. 7.1 (3.8) mmol/liter, P = 0.021] and a higher glucose area during OGTT [706 (283) vs. 559 (278) mmol/liter, P = 0.060] than the DQB1*02-neg patients did. Fasting insulin, insulin area during OGTT, fasting C-peptide concentration, and the lipid concentrations did not differ among the groups (data not shown). Patients with both AA/AG and DQB1*02 had a higher fasting glucose concentration [8.5 (4.7) vs. 7.1 (2.9) mmol/liter, P = 0.023] and glucose area during OGTT [763 (297) vs. 559 (247) mmol/liter, P = 0.070], compared with the DQB1*02-neg patients with GG, but they did not differ from patients with DQB1*02 alone.


View this table:
[in this window]
[in a new window]
  		Table 3. Clinical characteristics in unrelated type 2 patients from the mixed type 1/2 diabetes families or common type 2 diabetes families with (AA/AG) or without (GG) the TNF{alpha}308 promoter polymorphism

 
Type 2 patients from the common type 2 diabetes families. Common type 2 patients with AA/AG did not differ from those with GG with respect to age, BMI, fasting glucose, fasting insulin, or glucose and insulin responses during OGTT, but they had significantly lower WH ratio (0.92 [0.12] vs. 0.94 [0.11], P = 0.008) and lower fasting C-peptide concentration (0.48 [0.47] vs. 0.62 [0.46] nmol/liter, P = 0.020). Moreover, the women with AA/AG had also lower triglyceride concentration, compared with women with the GG genotype (1.64 [1.08] vs. 1.87 [1.35] mmol/liter, P = 0.023) (Table 3Go). On the contrary, when the patients were stratified according to the DQB1*02 genotype, no differences in fasting C-peptide (0.51 [0.43] vs. 0.62 [0.47] nmol/liter) and triglyceride (1.64 [1.11] vs. 1.78 [1.29] mmol/liter) concentrations or WH ratio 0.93 [0.11] vs. 0.94 [0.11] were seen between those with and without the 02 allele. Common type 2 patients with both AA/AG and DQB1*02 had lower fasting C-peptide (0.46 [0.42] vs. 0.62 [0.46] nmol/liter, P = 0.020) and lower triglyceride (1.54 [1.03] vs. 1.78 [1.29] mmol/liter, P = 0.048) concentrations than those DQB1*02-neg patients with GG (Table 4Go). The results were unchanged when the analysis was restricted to the DR3-positive subset (Table 4Go) or the GADab-positive patients were excluded (fasting C-peptide, P = 0.031; triglyceride P = 0.065). Of note, the fasting C-peptide and triglyceride concentrations in DQB1*02-pos common type 2 patients with AA/AG were as low as those in type 2 patients from type 1/2 families (fasting C-peptide 0.46 [0.41] nmol/liter, triglyceride 1.42 [1.12] mmol/liter).


View this table:
[in this window]
[in a new window]
  		Table 4. Clinical characteristics of the unrelated type 2 patients from the common type 2 diabetes families stratified according to the presence of the A allele of TNF{alpha}308, DQB1*02 and DR3

 
In a multiple regression analysis (see Subjects and Methods), TNF{alpha}308 AA/AG was associated with low fasting C-peptide concentration (regression coefficient = -0.058, P = 0.027) and low WH ratio (regression coefficient = -0.009, P = 0.010; women: regression coefficient = -0.016, P = 0.002), independent of age, BMI, sex, and the presence of DQB1*02 (Table 5Go).


View this table:
[in this window]
[in a new window]
  		Table 5. The association of TNF{alpha}308 AA/AG genotype with low fasting C-peptide concentration or low WH in common type 2 patients

 
Nondiabetic control subjects. The clinical characteristics did not differ between nondiabetic subjects who had TNF{alpha}308 AA/AG or GG genotypes. There also was no difference between those with and without DQB1*02 (data not shown). However, among the DQB1*02-positive group, despite similar age [55.0 (21.4) vs. 58.4 (20.2) yr] and BMI [25.0 (3.5) vs. 25.6 (4.8) kg/m2], subjects with AA/AG (23 males and 25 females) had lower fasting insulin concentration [5.9 (4.4) vs. 7.8 (5.3) mU/liter, P = 0.015] and lower HOMA index for insulin resistance [1.2 (1.1) vs. 1.7 (1.2), P = 0.026] than those with GG (20 males and 28 females). Again, the differences were more marked in women [fasting insulin: 5.7 (4.7) vs. 8.1 (4.8) mU/liter, P = 0.027; HOMA index: 1.2 (1.1) vs. 1.8 (1.2), P = 0.046]. Only three of the DQB1*02 positive subjects without DR3 had AA/AG, which precluded separate analysis of this group. However, when the DR3-negative subjects were excluded from the DQB1*02-positive group, the fasting C-peptide concentration [0.33 (0.27) vs. 0.49 (0.24), P = 0.011] and insulin area under curve during OGTT [3361 (2229) vs. 4772 (5314), P = 0.018] were lower in those with AA/AG, compared with GG. In the DQB1*02-neg group, the clinical characteristics did not differ between subjects with AA/AG and GG (data not shown).


   Discussion
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
References
 
The main aim was to study whether polymorphisms in the TNF{alpha} gene, which have previously been reported to be associated with both type 1 diabetes and subphenotypes of type 2 diabetes, would together with HLA contribute to the insulin-deficient phenotype of type 2 diabetic patients from mixed type 1/2 families. In accordance with a previous study on young-onset type 1 diabetes (11), we found an increased frequency of TNF{alpha}308 genotypes containing the A allele (AA/AG) in patients with adult-onset type 1 diabetes. However, the TNF{alpha}308 genotype frequency was similar in type 2 patients from the mixed type 1/2 families and common type 2 families, compared with the population controls.

The TNF{alpha}308 polymorphism has previously been shown to be in linkage disequilibrium with HLA-DR3 (11, 12) in Europeans. Our data supported this association because the A allele of TNF{alpha}308 was associated with the HLA-DR3 (Ds = 0.61) and somewhat less strongly with the associated DQB1*02 allele (Ds = 0.47; n = 592). The linkage disequilibrium between the TNF{alpha}308 A and DR3 or DQB1*02 alleles was strongest in the adult-onset type 1 patients (Ds 0.86 for DR3 and 0.81 for DQB1*02), followed by the type 2 patients from type 1/2 families (Ds = 0.62 and 0.59), common type 2 patients (Ds = 0.54 and 0.39), and control subjects (Ds = 0.46 and 0.25). These results are in accordance with a previous study showing linkage disequilibrium differences in the chromosome region harboring the HLA class II and TNF{alpha} loci between type 1 diabetic patients and control subjects carrying the risk HLA DRB1-DQB1 genotypes (8). In Scandinavian populations, DQB1*02 occurs mostly in the same haplotype as DR3, although it can also co-occur with DR7, which is considered neutral with respect to the type 1 diabetes risk (35).

The association of TNF{alpha}308 A with type 1 diabetes ascribed to the extended HLA haplotype because of allelic association with DR3-DQB1*02 could result in better defined extended haplotypes and allow the differentiation of more specific association or more stronger disease susceptibility conferred by the extended HLA haplotype than that conferred by TNF{alpha} or HLA class II alleles alone (8, 10, 14, 15, 16). Indeed, the frequency of having both TNF{alpha}308 A and DQB1*02 was significantly higher in adult-onset type 1 diabetic patients than in common type 2 patients or control subjects, and the differences were much more greater than that shown for the A or the *02 allele alone. The same was true for the difference between type 2 patients from type 1/2 families and common type 2 diabetic patients or control subjects carrying both the A and the *02 alleles. However, larger numbers of mixed type 2 patients from type 1/2 families with DR3 are needed to show a statistically significant difference. On the other hand, because both DR3 and DQB1*02 are increased in frequency in type 1 diabetes, it is necessary to also genotype for HLA when attempting to decipher any association between TNF{alpha} and diabetes.

Initially supporting our hypothesis, the type 2 patients from type 1/2 families having both TNF{alpha}308 AA/AG and DQB1*02 (73% of whom had a DR3-DQB1*02-TNF{alpha}308 A) (2) seemed to have a higher glucose area during OGTT than those who had neither the TNF{alpha}308 A allele (GG genotype) or DQB1*02. However, the increased glucose area could not be attributed to the TNF{alpha}308 polymorphism because all DQB1*02-positive type 2 patients from type 1/2 families had a higher glucose area, compared with the DQB1*02-negative patients irrespective of the TNF{alpha}308 genotype. Thus, we could not show an HLA-independent association between TNF{alpha}308 and the phenotype of type 2 patients from type 1/2 families. However, given the high degree of linkage disequilibrium between the TNF{alpha}308 A and the DQB1*02 alleles in type 2 patients from type 1/2 families, a much larger sample size would be needed to exclude a small independent effect of TNF{alpha}.

In common type 2 patients, our data suggested an independent effect of TNF{alpha}308 AA/AG on the insulin-deficient phenotype. In the logistic regression analysis, TNF{alpha}308 AA/AG was associated with low fasting C-peptide concentration and low WH ratio, independent of the presence of DQB1*02. Of note, the DQB1*02-positive and -negative common type 2 patients did not per se differ with respect to these parameters (2). However, there was an additive effect in that patients with both DQB1*02 and AA/AG had lower fasting C-peptide concentration than those with either AA/AG or DQB1*02 alone. In fact, the fasting C-peptide concentration in those with both DQB1*02 and AA/AG was as low as that seen in type 2 patients from type 1/2 families, independent of the presence of GADab. This could reflect a haplotype effect of both DQB1 and TNF{alpha} on the phenotype of common type 2 patients, which could explain that the TNF{alpha}308 polymorphism has previously been associated with low HOMA index in Sardinian (36) but not in Japanese type 2 diabetic patients (37). Indeed, there have been data showing that two (or more) susceptibility loci in the HLA (one in the HLA class II and another in the central MHC region) act epistatically to increase susceptibility of type 1 diabetes. This so-called dosage effect is particularly remarkable for the DR3-DQB1*02 haplotype (7, 38). In Europeans, the A1-B8-DR3-DQB1*02 haplotype, containing the TNF{alpha}308 A allele, is associated with high TNF{alpha} production (11, 13, 39). TNF{alpha} has been proposed to have an active role in pancreatic ß-cell destruction and impaired insulin secretion mediated by overproduction of nitric oxide in the ß-cells (40, 41). Of note, the frequency of GADab did not differ between adult-onset diabetic patients with or without the DQB1*02 allele and those with or without the TNF{alpha}308 A allele; the results were in accordance with the previous studies in Finland and Sweden showing that the association of GADab with DR3 (or DQB1*02) was mainly restricted to the young onset (<20 yr) diabetic patients (42, 43, 44, 45, 46).

After the initial report of Hotamisligil et al. (17, 18, 47) that TNF{alpha} plays a crucial role in the insulin resistance of obesity and type 2 diabetes through its inhibition of insulin signaling, the TNF{alpha} locus has also been linked to obesity (48), and the TNF{alpha}308 polymorphism has been associated with the features of insulin resistance or body fat. The results for nondiabetic subjects are, however, somewhat contradictory. In some studies, the less common allele has been associated with increased fasting insulin concentration, BMI, body fat content, and insulin resistance (20, 21, 22, 23, 24, 25), and in others it has been associated with a decreased BMI (27, 28). The results from our nondiabetic population support the relationship between the DQB1-TNF{alpha} haplotype and a more insulin-deficient phenotype. Among them, subjects with both TNF{alpha}308 AA/AG and DQB1*02 had lower fasting insulin concentration and lower HOMA index for insulin resistance, compared with the others. The differences were more marked in women, which is in agreement with previous studies (22, 23, 27, 28). The contradictory results in the literature regarding the existence of an association between TNF{alpha} and insulin resistance in the nondiabetic subjects could possibly be due to different HLA background or different pattern of linkage disequilibrium between the HLA and the TNF{alpha} alleles of the study populations.

In conclusion, TNF{alpha} is unlikely to be the second gene in the HLA region responsible for our previous findings in type 2 patients from type 1/2 families. However, we could show an association between the TNF{alpha}308 polymorphism and the phenotype of common type 2 diabetes; the A allele was associated with lower WH ratio (mainly in women) and lower fasting C-peptide concentration in common type 2 diabetic patients, independent of the presence of DQB1*02.


   Acknowledgments
 
Peter Almgren and Timo Kanninen are acknowledged for statistical help and the Botnia Research Group for recruiting and clinically studying the subjects.


   Footnotes
 
This work was supported by the Sigrid Juselius Foundation, Påhlsson Foundation, Medical Faculty of the Lund University, Malmö University Hospital, Swedish National Board of Health and Welfare, Swedish Medical Doctors Association, Crafoord Foundation, and Novo Nordisk Foundation. The Botnia study is supported by the Sigrid Juselius Foundation, JDF Wallenberg, EC (BM4-CT95-0662), Swedish Medical Research Foundation, Academy of Finland, Finnish Diabetes Research Society, Swedish Diabetes Association, and Novo Nordisk Foundation.

Abbreviations: BMI, Body mass index; D, coefficient of disequilibrium; Ds, standardized value of D; HbA1C, hemoglobin A1C; HLA, human leukocyte antigen; HOMA, homeostasis model assessment for insulin resistance; OGTT, oral glucose tolerance test; GADab, glutamic acid decarboxylase antibody; WH, waist to hip ratio.

Received March 31, 2002.

Accepted February 27, 2003.


   References
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 References
 

  1. Li H, Isomaa B, Taskinen MR, Groop L, Tuomi T 2000 Consequences of a family history of type 1 and type 2 diabetes on the phenotype of patients with type 2 diabetes. Diabetes Care 23:589–594[Abstract/Free Full Text]
  2. Li H, Lindholm E, Almgren P, Gustafsson A, Forsblom C, Groop L, Tuomi T 2001 Possible human leukocyte antigen-mediated genetic interaction between type 1 and type 2 diabetes. J Clin Endocrinol Metab 86:574–582[Abstract/Free Full Text]
  3. Erbey JR, Kuller LH, Becker DJ, Orchard TJ 1998 The association between a family history of type 2 diabetes and coronary artery disease in a type 1 diabetes population. Diabetes Care 21:610–614[Abstract]
  4. Erbey JR, Robbins D, Forrest KY, Orchard TJ 1999 Low-density lipoprotein particle size and coronary artery disease in a childhood-onset type 1 diabetes population. Metabolism 48:531–534[CrossRef][Medline]
  5. Fagerudd JA, Pettersson-Fernholm KJ, Gronhagen-Riska C, Groop PH 1999 The impact of a family history of type II (non-insulin-dependent) diabetes mellitus on the risk of diabetic nephropathy in patients with type I (insulin-dependent) diabetes mellitus. Diabetologia 42:519–526[CrossRef][Medline]
  6. Makimattila S, Ylitalo K, Schlenzka A, Taskinen MR, Summanen P, Syvanne M, Yki-Jarvinen H 2002 Predictors of intima-media thickness in type I diabetes. Diabetologia 45:711–718[CrossRef][Medline]
  7. Moghaddam PH, Zwinderman AH, de Knijff P, Roep BO, Schipper RF, Van der Auwera B, Naipal A, Gorus F, Schuit F, Giphart MJ 1997 TNFa microsatellite polymorphism modulates the risk of IDDM in Caucasians with the high-risk genotype HLA DQA1*0501-DQB1*0201/DQA1*0301-DQB1*0302. Diabetes 46:1514–1515[Medline]
  8. Hanifi Moghaddam P, de Knijf P, Roep BO, Van der Auwera B, Naipal A, Gorus F, Schuit F, Giphart MJ 1998 Genetic structure of IDDM1: two separate regions in the major histocompatibility complex contribute to susceptibility or protection. Belgian Diabetes Registry. Diabetes 47:263–269[Abstract]
  9. Zavattari P, Lampis R, Motzo C, Loddo M, Mulargia A, Whalen M, Maioli M, Angius E, Todd JA, Cucca F 2001 Conditional linkage disequilibrium analysis of a complex disease superlocus, IDDM1 in the HLA region, reveals the presence of independent modifying gene effects influencing the type 1 diabetes risk encoded by the major HLA-DQB1, -DRB1 disease loci. Hum Mol Genet 10:881–889[Abstract/Free Full Text]
  10. Wilson AG, de Vries N, Pociot F, di Giovine FS, van der Putte LB, Duff GW 1993 An allelic polymorphism within the human tumor necrosis factor {alpha} promoter region is strongly associated with HLA A1, B8, and DR3 alleles. J Exp Med 177:557–560[Abstract/Free Full Text]
  11. Pociot F, Wilson AG, Nerup J, Duff GW 1993 No independent association between a tumor necrosis factor-{alpha} promotor region polymorphism and insulin-dependent diabetes mellitus. Eur J Immunol 23:3050–3053[Medline]
  12. Cox A, Gonzalez AM, Wilson AG, Wilson RM, Ward JD, Artlett CM, Welsh K, Duff GW 1994 Comparative analysis of the genetic associations of HLA-DR3 and tumour necrosis factor {alpha} with human IDDM. Diabetologia 37:500–503[CrossRef][Medline]
  13. Pociot F, Briant L, Jongeneel CV, Molvig J, Worsaae H, Abbal M, Thomsen M, Nerup J, Cambon-Thomsen A 1993 Association of tumor necrosis factor (TNF) and class II major histocompatibility complex alleles with the secretion of TNF-{alpha} and TNF-ß by human mononuclear cells: a possible link to insulin-dependent diabetes mellitus. Eur J Immunol 23:224–231[Medline]
  14. Raum D, Awdeh Z, Yunis EJ, Alper CA, Gabbay KH 1984 Extended major histocompatibility complex haplotypes in type I diabetes mellitus. J Clin Invest 74:449–454
  15. Tienari PJ, Tuomilehto Wolf E, Tuomilehto J, Peltonen L 1992 HLA haplotypes in type 1 (insulin-dependent) diabetes mellitus: molecular analysis of the HLA-DQ locus. The DIME Study Group. Diabetologia 35:254–260[CrossRef][Medline]
  16. Foissac A, Crouau-Roy B, Faure S, Thomsen M, Cambon-Thomsen A 1997 Microsatellites in the HLA region: on overview. Tissue Antigens 49:197–214[Medline]
  17. Hotamisligil GS, Spiegelman BM 1994 Tumor necrosis factor {alpha}: a key component of the obesity-diabetes link. Diabetes 43:1271–1278[Abstract]
  18. Hotamisligil GS, Budavari A, Murray D, Spiegelman BM 1994 Reduced tyrosine kinase activity of the insulin receptor in obesity-diabetes. Central role of tumor necrosis factor-{alpha}. J Clin Invest 94:1543–1549
  19. Heijmans BT, Westendorp RGJ, Droog S, Kluft C, Knook DL, Slagboom PE 2002 Association of the tumour necrosis factor {alpha} -308G/A polymorphism with the risk of diabetes in an elderly population-based cohort. Genes Immunol 3:225–228
  20. Fernandez-Real JM, Gutierrez C, Ricart W, Casamitjana R, Fernandez-Castaner M, Vendrell J, Richart C, Soler J 1997 The TNF-{alpha} gene Nco I polymorphism influences the relationship among insulin resistance, percent body fat, and increased serum leptin levels. Diabetes 46:1468–1472[Abstract]
  21. Herrmann SM, Ricard S, Nicaud V, Mallet C, Arveiler D, Evans A, Ruidavets JB, Luc G, Bara L, Parra HJ, Poirier O, Cambien F 1998 Polymorphisms of the tumour necrosis factor-{alpha} gene, coronary heart disease and obesity. Eur J Clin Invest 28:59–66[CrossRef][Medline]
  22. Hoffstedt J, Eriksson P, Hellstrom L, Rossner S, Ryden M, Arner P 2000 Excessive fat accumulation is associated with the TNF {alpha}-308 G/A promoter polymorphism in women but not in men. Diabetologia 43:117–120[CrossRef][Medline]
  23. Dalziel B, Gosby AK, Richman RM, Bryson JM, Caterson ID 2002 Association of the TNF-{alpha}-308 G/A promoter polymorphism with insulin resistance in obesity. Obes Res 10:401–407[Medline]
  24. Brand E, Schorr U, Kunz I, Kertmen E, Ringel J, Distler A, Sharma AM 2001 Tumor necrosis factor-{alpha}-308 G/A polymorphism in obese Caucasians. Int J Obes Relat Metab Disord 25:581–585[CrossRef][Medline]
  25. Nicaud V, Raoux S, Poirier O, Cambien F, O’Reilly DS, Tiret L 2002 The TNF {alpha}/G-308A polymorphism influences insulin sensitivity in offspring of patients with coronary heart disease: the European Atherosclerosis Research Study II. Atherosclerosis 161:317–325[CrossRef][Medline]
  26. Krikovszky D, Vasarhelyi B, Toth-Heyn P, Korner A, Tulassay T, Madacsy L 2002 Association between G-308A polymorphism of the tumor necrosis factor-{alpha} gene and 24-hour ambulatory blood pressure values in type 1 diabetic adolescents. Clin Genet 62:474–477[CrossRef][Medline]
  27. Walston J, Seibert M, Yen CJ, Cheskin LJ, Andersen RE 1999 Tumor necrosis factor-{alpha}-238 and -308 polymorphisms do not associate with traits related to obesity and insulin resistance. Diabetes 48:2096–2098[Abstract]
  28. Romeo S, Sentinelli F, Capici F, Arca M, Berni A, Vecci E, Mario UD, Baroni MG 2001 The G-308A variant of the tumor necrosis factor-{alpha} (TNF-{alpha}) gene is not associated with obesity, insulin resistance and body fat distribution. BMC Med Genet 2:10[CrossRef][Medline]
  29. Day CP, Grove J, Daly AK, Stewart MW, Avery PJ, Walker M 1998 Tumour necrosis factor-{alpha} gene promoter polymorphism and decreased insulin resistance. Diabetologia 41:430–434[CrossRef][Medline]
  30. Groop L, Forsblom C, Lehtovirta M, Tuomi T, Karanko S, Nissen M, Ehrnstrom BO, Forsen B, Isomaa B, Snickars B, Taskinen MR 1996 Metabolic consequences of a family history of NIDDM (the Botnia study): evidence for sex-specific parental effects. Diabetes 45:1585–1593[Abstract]
  31. Lehto M, Tuomi T, Mahtani MM, Widen E, Forsblom C, Sarelin L, Gullstrom M, Isomaa B, Lehtovirta M, Hyrkko A, Kanninen T, Orho M, Manley S, Turner RC, Brettin T, Kirby A, Thomas J, Duyk G, Lander E, Taskinen MR, Groop L 1997 Characterization of the MODY3 phenotype. Early-onset diabetes caused by an insulin secretion defect. J Clin Invest 99:582–591[Medline]
  32. Alberti KG, Zimmet PZ 1998 Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 15:539–553[CrossRef][Medline]
  33. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC 1985 Homeostasis model assessment: insulin resistance and ß-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28:412–419[CrossRef][Medline]
  34. Tuomi T, Carlsson A, Li H, Isomaa B, Miettinen A, Nilsson A, Nissen M, Ehrnstrom BO, Forsen B, Snickars B, Lahti K, Forsblom C, Saloranta C, Taskinen MR, Groop LC 1999 Clinical and genetic characteristics of type 2 diabetes with and without GAD antibodies. Diabetes 48:150–157[Abstract]
  35. Kockum I, Wassmuth R, Holmberg E, Michelsen B, Lernmark A 1993 HLA-DQ primarily confers protection and HLA-DR susceptibility in type I (insulin-dependent) diabetes studied in population-based affected families and controls. Am J Hum Genet 53:150–167[Medline]
  36. Melis M, Secchi G, Carboni A, Angius MF, Ciccarese M, Maioli M, Tonolo G 2001 TNF-a gene polymorphism is associated with insulin resistance in the Sardinian population. Diabetologia 44(Suppl 1):A92
  37. Ishii T, Hirose H, Saito I, Nishikai K, Maruyama H, Saruta T 2000 Tumor necrosis factor {alpha} gene G-308A polymorphism, insulin resistance, and fasting plasma glucose in young, older, and diabetic Japanese men. Metabolism 49:1616–1618[CrossRef][Medline]
  38. Huang HS, Peng JT, She JY, Zhang LP, Chao CC, Liu KH, She JX 1995 HLA-encoded susceptibility to insulin-dependent diabetes mellitus is determined by DR and DQ genes as well as their linkage disequilibria in a Chinese population. Hum Immunol 44:210–219[CrossRef][Medline]
  39. Wilson AG, Symons JA, McDowell TL, McDevitt HO, Duff GW 1997 Effects of a polymorphism in the human tumor necrosis factor {alpha} promoter on transcriptional activation. Proc Natl Acad Sci USA 94:3195–3199[Abstract/Free Full Text]
  40. Corbett JA, Sweetland MA, Wang JL, Lancaster Jr JR, McDaniel ML 1993 Nitric oxide mediates cytokine-induced inhibition of insulin secretion by human islets of Langerhans. Proc Natl Acad Sci USA 90:1731–1735[Abstract/Free Full Text]
  41. Kwon G, Xu G, Marshall CA, McDaniel ML 1999 Tumor necrosis factor {alpha}-induced pancreatic ß-cell insulin resistance is mediated by nitric oxide and prevented by 15-deoxy-delta12, 14-prostaglandin J2 and aminoguanidine. A role for peroxisome proliferator-activated receptor {gamma} activation and iNOS expression. J Biol Chem 274:18702–18708[Abstract/Free Full Text]
  42. Hagopian WA, Sanjeevi CB, Kockum I, Landin-Olsson M, Karlsen AE, Sundkvist G, Dahlquist G, Palmer J, Lernmark Al 1995 Glutamate decarboxylase-, insulin-, and islet cell-antibodies and HLA typing to detect diabetes in a general population-based study of Swedish children. J Clin Invest 95:1505–1511
  43. Sabbah E, Kulmala P, Veijola R, Vahasalo P, Karjalainen J, Tuomilehto-Wolf E, Akerblom HK, Knip M 1996 Glutamic acid decarboxylase antibodies in relation to other autoantibodies and genetic risk markers in children with newly diagnosed insulin-dependent diabetes. Childhood Diabetes in Finland Study Group. J Clin Endocrinol Metab 81:2455–2459[Abstract]
  44. Sabbah E, Savola K, Kulmala P, Reijonen H, Veijola R, Vahasalo P, Karjalainen J, Ilonen J, Akerblom HK, Knip M 1999 Disease-associated autoantibodies and HLA-DQB1 genotypes in children with newly diagnosed insulin-dependent diabetes mellitus (IDDM). The Childhood Diabetes in Finland Study Group. Clin Exp Immunol 116:78–83[CrossRef][Medline]
  45. Sabbah E, Savola K, Ebeling T, Kulmala P, Vahasalo P, Ilonen J, Salmela PI, Knip M 2000 Genetic, autoimmune, and clinical characteristics of childhood- and adult-onset type 1 diabetes. Diabetes Care 23:1326–1332[Abstract/Free Full Text]
  46. Graham J, Hagopian WA, Kockum I, Li LS, Sanjeevi CB, Lowe RM, Schaefer JB, Zarghami M, Day HL, Landin-Olsson M, Palmer JP, Janer-Villanueva M, Hood L, Sundkvist G, Lernmark A, Breslow N, Dahlquist G, Blohme G; Diabetes Incidence in Sweden Study Group, Swedish Childhood Diabetes Study Group 2002 Genetic effects on age-dependent onset and islet cell autoantibody markers in type 1 diabetes. Diabetes 51:1346–1355[Abstract/Free Full Text]
  47. Hotamisligil GS, Murray DL, Choy LN, Spiegelman BM 1994 Tumor necrosis factor {alpha} inhibits signaling from the insulin receptor. Proc Natl Acad Sci USA 91:4854–4858[Abstract/Free Full Text]
  48. Norman RA, Bogardus C, Ravussin E 1995 Linkage between obesity and a marker near the tumor necrosis factor-{alpha} locus in Pima Indians. J Clin Invest 96:158–162



This article has been cited by other articles:


Home page
 CVIHome page
A. A. Motala, M. Busson, E. M. Al-Harbi, M. A. A. Khuzam, E. M. D. Al-Omari, M. R. Arekat, and W. Y. Almawi
Susceptible and Protective Human Leukocyte Antigen Class II Alleles and Haplotypes in Bahraini Type 2 (Non-Insulin-Dependent) Diabetes Mellitus Patients
Clin. Vaccine Immunol., January 1, 2005; 12(1): 213 - 217.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Li, H.
Right arrow Articles by Tuomi, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Li, H.
Right arrow Articles by Tuomi, T.

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