Antibodies to Glutamic Acid Decarboxylase in Syrian and Jordanian Type 1 Diabetes Patients and their Siblings

Introduction

There is general agreement that type 1 diabetes mellitus (insulin-dependent diabetes mellitus, IDDM) is an autoimmune disease.^1-2 The factors that induce the selective destruction of the insulin-producing islet cells of the pancreas and sustain this autoimmune process are unknown.³ There is increasing evidence that a close interaction exists between environmental agents and genetic susceptibility.⁴

Although autoreactive T-cells rather than autoantibodies are responsible for the destruction of ß-cells,⁵ the detection and characterization of type 1 diabetes-associated autoantibodies is of great help in prediction of the disease.⁶ Identification and characterization of the autoantibody-autoantigen system in IDDM is of crucial and fundamental importance in prediction and potential immunotherapy.⁷

Currently, cytoplasmic islet cell antibodies (ICAs),⁸ insulin autoantibodies,⁹ glutamic acid decarboxylase antibodies (anti-GAD)^10-11 and 37-k antigen antibodies are potentially useful humoral markers for such prediction.¹² However, which autoantigen is the primary target and which arises as a secondary phenomenon is still unclear.¹³ The attempt to identify the earliest events in the autoimmune process suggests that GAD is one of the first and most important autoantigens.^14-15

PATIENTS AND METHODS

Sera were obtained from 180 patients with type 1 diabetes, 123 of their siblings, and from 178 from subjects not suffering from diabetes or any autoimmune disease (control group). The type 1 diabetes group consisted of 85 Syrian patients (46 males, 39 females) with a mean age of 13.6±5.9 years (mean±SD) and duration of diabetes of 3.9±2.3 years, and 95 Jordanian patients (47 males, 48 females) with a mean age of 13.9±6.5 and duration of diabetes of 4.3±2 years. The patients were recruited from the Syrian and Jordanian National Centers for Diabetes, and from other diabetes clinics in both countries. The sibling group consisted of 45 Syrian type 1 diabetes siblings (20 males, 25 females) with mean age of 11.3±6.1, and 78 Jordanian type 1 diabetes siblings (41 males, 37 females) with a mean age 12.3±7.1. The control group consisted of 78 healthy Syrian (mean age 9.9±4.3 years) and 100 healthy Jordanian (mean age of 7.8±4.5 years) subjects with no history of diabetes and/or autoimmune diseases.

Sera were analyzed for anti-GAD antibody using the enzyme-linked immunosorbent assay (ELISA) kit. With the kit, we measured the concentration of GAD-specific antibodies, and not arbitrary units. The results are given with reference to the GAD specific human monoclonal islet cell antibody clone 3 (MICA 3) and were calibrated to human IgG concentration. Samples with anti-GAD concentrations <32 ng/ml were considered antibody negative, those with values between 32-50 ng/ml were considered in the “grey area” of the test system and samples with values >50 ng/ml were considered anti-GAD positive.

Data are presented as mean±SD. Statistical significance was determined by the correlation coefficient and chi-square test (χ2). P values less than 0.05 were considered statistically significant.

RESULTS

The prevalence of anti-GAD was 34.1% in Syrian and 49.5% in Jordanian type 1 diabetes patients (P=0.03) (Table 1). In the siblings, the prevalence was 20.0% in Syrians and 23.0% in Jordanians (P=0.6). In the control groups the rate was 1.3% in Syrians and 2.0% in Jordanians (P=0.7). Anti-GAD prevalence rates were significantly higher in Syrian and Jordanian type 1 diabetes patients and their siblings than in their corresponding control groups (Table 1). Prevalence rates were also different between Syrian and Jordanian type 1 diabetes patients (P=0.03). There were no significant differences between males and females within nationalities for either the type 1 diabetes patients or their siblings (Table 1).

Table 1. Prevalence of anti-GAD in Syrian and Jordanian populations.

Although the prevalence of anti-GAD for type 1 diabetes patients with clinical diabetes of less than two years duration was 88.8% in both national groups, no significant correlation was found between the mean titer of anti-GAD and the clinical duration of diabetes. We also found no correlation between sex and the presence of anti-GAD in either Syrian or Jordanian type I diabetes patients or their siblings.

The mean age of both Syrian and Jordanian type 1 diabetes patients who had anti-GAD was 12.9±5.3 years, and the mean age of their siblings who had anti-GAD was 10.33±4.79. In the type 1 diabetes patients the highest prevalence of anti-GAD was among the age group 13 to 16 years, while in the siblings it peaked in two age groups, 5 to 8 and 13 to 16 years (Figure 1). When subjects positive for anti-GAD were classified according to anti-GAD titer, 58.8% of the type 1 diabetes patients and 38% of the siblings had a titer over 90 ng / ml (Figure 2).

DISCUSSION

Our data showed that anti-GAD was positive in a considerable number of type 1 diabetes patients, irrespective of the time since clinical onset. This finding is concordant with most other studies.^15,16 In addition, we found no correlation between sex and the positivity of anti-GAD, a finding also reported in several other studies.¹⁷ Two reports found a female preponderance of anti-GAD, the first a report from Australia (75% females vs 63% males, P=0.03)¹⁸ and the second from Sweden (79% females vs 63% males, P<0.001.)¹⁹

The highest frequency of anti-GAD in type 1 diabetes patents was in the age group 13 to 16 years, while in the IDDM siblings it was highest in the two age groups 5 to 8 and 13 to 16 years (Figure 1). No correlation, however, was found between age and anti-GAD titers. All these findings suggest that anti-GAD can be a potentially useful predictive marker for type 1 diabetes.^16-19

To the best of our knowledge, this is the first study to determine the prevalence of anti-GAD not only in Syrian and Jordanian populations but in the rest of the Arab World and the Near East as well. Anti-GAD prevalence rates were significantly higher in Syrian and Jordanian type 1 diabetes patients and siblings than in their corresponding control groups (P<0.001, P <0.0001, P=0.002, respectively). Since the control groups were well-matched by age, sex and geographic area, our data imply that the anti-GAD would achieve a sufficiently predictive value to identify individuals at high risk of developing type 1 diabetes and to subsequently institute preventive therapy whenever such therapy becomes available. Another advantage of anti-GAD is that it is simple to perform and to standardize. By contrast, the islet cell antibody (ICA) test is difficult to standardize and is cumbersome and time-consuming. Nevertheless, several studies suggest that while the anti-GAD may have a higher specificity for type 1 diabetes prediction than the ICA test, the combination of both tests increases the sensitivity for type 1 diabetes prediction.^20-22

While we found a similar prevalence of anti-GAD in the Syrian and Jordanian type 1 diabetes siblings, it was surprising to find a significant difference in the prevalence of anti-GAD between the Syrian and Jordanian Type 1 diabetes patient groups (P=0.03). This observation may be explained by the difference in the socioeconomic levels of the type 1 diabetes groups we studied. However, the similar results in the siblings and their related control groups in both populations do not support this explanation.

The wide geographic differences in the incidence and prevalence of Type 1 diabetes prompted us to compare our results with those of others (Table 2). The highest prevalence of anti-GAD had been among the Finnish (82%), Swedish (70%),²³ and European, and Australian (74%)(15) populations, while the lowest prevalence was among Koreans.²⁴ Our data showed a comparable prevalence of anti-GAD between Syrian and Japanese populations,²⁴ and between Jordanians and Thais.²⁵ Ethnic and / or environmental factors may be the cause of differences in anti-GAD prevalence observed in different populations.²³

Table 2. Prevalence of anti-GAD in different populations.

Additionally, insulin-dependent diabetes mellitus is associated with certain HLA Class II alleles. Most population studies have shown that HLA-DR3 and -DR4 alleles are associated with an increased risk of developing type 1 diabetes. In the last decade much evidence has accumulated suggesting that the HLA-DQ alleles, which are in linkage disequilibrium with DR, rather than the DR locus, are more closely associated with increased susceptibility for type 1 diabetes. More recent studies have shown that there is a positive association between anti-GAD and DQ alleles.^14-26,27 Thus it would be important to determine the susceptible alleles that correlate highly with the positivity of anti-GAD among Syrian and Jordanian populations. This issue is currently being investigated.

The most important finding in our study was that Syrian and Jordanian IDDM siblings had one of the highest reported prevalence rates of anti-GAD (20% and 23% respectively). Furthermore, we found a high frequency of type 1 diabetes siblings with an anti-GAD titer of 32 to 50 ng/ml, which has been considered a grey area (Figure 2). However, while autoimmune reaction against pancreatic ß-cells may occur in many children without developing overt diabetes, in some cases type 1 diabetes may develop many years later. How many of these anti-GAD positive siblings will go on to develop diabetes is a question that can be answered only after constant and long-term follow up of these subjects. Until that issue is clarified, it seems appropriate to screen for the presence of anti-GAD in siblings of Type 1 diabetes patients.

One year after the initial screening, we redetermined anti-GAD in non-diabetic siblings who originally tested positive for anti-GAD (Table 3). We found that the levels of anti-GAD titer had increased in three subjects. The first one had normal fasting blood sugar (FBS) and plasma insulin levels. The second subject had an insulin level at the lower limit (4.2 μ/mL), and the third one had an elevated FBS (245.5 mg/dL) and a decreased plasma insulin level (2.8 μ/mL). This last sibling was diagnosed with type 1 diabetes and started on insulin therapy.

Table 3. Levels of anti-GAD at beginning of study and after one year in 11 type 1 diabetes siblings with positive anti-GAD.

Levels of anti-GAD titer decreased in three type 1 diabetes siblings. This finding is not in agreement with those of Tuomilehto²⁸ and Whittingham.²⁰ In the Tuomilehto study the levels of anti-GAD were determined in multiple samples collected over a period of 10 years before the clinical onset of type 1 diabetes. Tuomilehto found that none of the anti-GAD positive subjects became negative during the subsequent follow up in the pediatric period.²⁸ In the Whittingham study (a follow-up to the Tuomilehto study), the prevalence of anti-GAD was 83% at the beginning of the study. During the 10 year pre-diabetic period the prevalence of anti-GAD was 25% in subjects who developed type 1 diabetes.²⁹ Moreover, the levels of anti-GAD in those same subjects did not change during the follow-up period. The discordant results between our study and that of Tuomilehto may be due to the anti-GAD assay methodology. We used ELISA technique while Tuomilehto used radioimmunoprecipitation assay technique. Differences in the groups may be another reason for this discrepancy. It is worth mentioning that there is a general agreement that anti-GAD provides the best way to predict which noninsulin dependent or type 2 patients would later need regular insulin treatment.³⁰ Most studies suggest that patients diagnosed with type 2 diabetes with a positive GAD index would be later reclassified as having type 1 diabetes.^31-33 The determination of anti-GAD prevalence in such Jordanian and Syrian patients will be one of our targets in the coming years.

Measurement of anti-GAD is a valuable screening tool to detect individuals at risk for developing type 1 diabetes, especially type 1 diabetes siblings. Anti-GAD testing would thus be useful for the implementation of preventive therapy for type 1 diabetes, whenever an effective and acceptable treatment becomes available.