|
|
| |||||||||||||||||||||||||||||||||||||||
Background Epidemiologic studies have implicated the ingestion of cow's milk in the pathogenesis of insulin-dependent diabetes mellitus (IDDM). Moreover, in a recent study, 100 percent of patients with new-onset IDDM had antibodies against bovine serum albumin (BSA), with a majority directed against a 17-amino-acid BSA peptide (ABBOS). Cellular immune mechanisms are thought to be the principal mediators of pancreatic beta-cell destruction in IDDM.
Methods We measured the responses of peripheral-blood mononuclear cells to BSA and ABBOS or serum IgG anti-BSA antibodies (by particle-concentration fluorescence immunoassay) in 71 patients with IDDM, 55 subjects at various degrees of risk for IDDM, 36 patients with other autoimmune disorders (chronic autoimmune thyroiditis, rheumatoid arthritis, and systemic lupus erythematosus), and 48 normal subjects.
Results The responses of peripheral-blood mononuclear cells to BSA or ABBOS were positive in 2 of 24 patients with new-onset IDDM, 1 of 25 first-degree relatives of patients with IDDM who were negative for islet-cell antibodies, 2 of 30 first-degree relatives of patients with IDDM who were positive for islet-cell antibodies, 1 of 28 patients with established IDDM, and 1 of 29 normal subjects. Similarly, anti-BSA antibodies were not detected significantly more often in patients with new-onset IDDM (3 of 31, 10 percent) than in normal subjects (1 of 37, 3 percent; P = 0.32). However, many patients with autoimmune disease and subjects at increased risk for IDDM had anti-BSA antibodies (frequency, 10 to 31 percent).
Conclusions Anti-BSA antibodies may reflect a general defect in the process of immunologic tolerance associated with a predisposition to autoimmunity rather than immunity specific to beta cells. The absence of cellular immunity to BSA and ABBOS in IDDM does not support a role for this antigen in the pathogenesis of the disorder.
Case-control and retrospective questionnaire studies designed to uncover potential environmental agents that could trigger IDDM have implicated the early nutritional intake of infants. The risk of IDDM is reportedly lower in breast-fed than in formula-fed infants, especially infants who are breast-fed for long periods4,5,6,7. In one case-control study, early exposure (before three months of age) to cow's milk or solid foods increased the subsequent risk of IDDM8. Furthermore, studies of the frequency of diabetes in rodents fed defined diets,9,10 comparisons of geographic data on the consumption of cow's milk with the frequency of IDDM,11 and sporadic reports of antibodies against bovine serum albumin (BSA) in patients with IDDM12,13 have implicated the ingestion of cow's milk as a possible environmental triggering agent. Interest in this association was heightened by a report that nearly all patients with new-onset IDDM had serum IgG anti-BSA antibodies and that a majority of these antibodies were directed against a 17-amino-acid BSA peptide (ABBOS)14.
Multiple studies in humans and nonobese diabetic mice suggest that cellular immune mechanisms are the principal mediators of islet-cell destruction1,2. Therefore, we studied the proliferative responses of peripheral-blood mononuclear cells to BSA or ABBOS in order to explore further the possible pathogenic role of anti-BSA immunity in this disease. In addition, since the aforementioned study14 did not include subjects at various levels of risk for IDDM (i.e., first-degree relatives of patients with IDDM who are positive or negative for islet-cell antibodies) or patients with other organ-specific or generalized autoimmune disease, we measured anti-BSA antibodies in such subjects to determine the specificity of anti-BSA immunity for IDDM.
Methods
Patients
Blood samples were obtained from 210 subjects involved in ongoing studies of the natural history of IDDM15,16 and were analyzed for antibodies, cellular immune activity, or both. The subjects included 48 normal subjects (mean [±SD] age, 32 ±8 years; 24 men and 24 women); 55 first-degree relatives of patients with IDDM, 25 who were negative for islet-cell antibodies (age, 24 ±15 years; 14 men and 11 women) and 30 who were positive for islet-cell antibodies (age, 32 ±16 years; 14 men and 16 women); 43 patients with newly diagnosed IDDM (<3 months' duration) (age, 17 ±11 years; 23 male and 20 female patients); 28 patients with established IDDM (duration of disease, 12 ±7 months; age, 16 ±7 years; 12 male and 16 female patients); 13 patients with chronic autoimmune thyroiditis (duration of disease, 3 ±5 months; age, 19 ±15 years; 3 male and 10 female patients); 16 patients with systemic lupus erythematosus (duration of disease, 9 ±6 years; age, 39 ±11 years; 2 men and 14 women); and 7 patients with juvenile rheumatoid arthritis (age, 22 ±18 years; 1 male and 6 female patients). IDDM was defined according to the World Health Organization criteria17. The patients with chronic autoimmune thyroiditis, systemic lupus erythematosus, or rheumatoid arthritis were being treated for their respective disease, and none had IDDM. The study was approved by the University of Florida Institutional Review Board, and informed consent was obtained from the subjects or their parents.
Islet-Cell Antibodies
Islet-cell antibodies were assessed by indirect immunofluorescence with frozen human pancreas16. A test was considered to be positive for islet-cell antibody when the intensity of the fluorescence and the pattern of staining of the undiluted serum equaled or exceeded those of a standard serum sample that had been calibrated to contain 10 Juvenile Diabetes Foundation (JDF) units18. All positive serum samples were expressed in JDF units determined by comparing the end-point dilution of each positive sample with a standard dilution curve for the reference serum sample.
Assay for IgG Anti-BSA Antibodies
IgG anti-BSA antibodies were measured by particle-concentration fluorescence immunoassay19. BSA was covalently conjugated to carboxylated polystyrene beads (IDEXX, Portland, Me.) with 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (Bio-Rad, Hercules, Calif.). The plates were standardized with the use of a mouse IgG1kappa anti-BSA specific monoclonal antibody (University of Florida Hybridoma Facility, Gainesville). For the serum samples from patients, 20 microl of serum diluted 1:100 was incubated with 20 microl of BSA-conjugated polystyrene beads (0.125 percent wt/vol) in a well of a Fluoricon assay plate (Baxter, Mundelein, Ill.). All reagents were prepared in 0.02 M TRIS buffer, pH 8.0, containing 0.15 M sodium chloride, 1 percent Tween-20, and 0.1 percent chicken ovalbumin. After 10 minutes, the unbound proteins were removed by vacuum filtration and the BSA-conjugated polystyrene beads were washed three times with the same buffer. The beads were then incubated for 10 minutes with 20 microl of fluorescein-conjugated F(ab')2 fragments of affinity-purified goat antihuman IgG Fc antibodies (6.5 µg per milliliter) (Jackson Immuno Research Laboratories, West Grove, Pa.), and washed three more times before binding activity was assessed with a Pandex Screen machine (Baxter, Deerfield, Ill.). All values were expressed as the mean of duplicate incubations, and the interassay and intraassay coefficients of variability were 10 percent or less. A response was considered to be positive if the level of fluorescence activity was 2 SD or more above the mean value for serum from 37 normal subjects. The specificity for anti-BSA antibodies was determined by the parallel preincubation of serum for 30 minutes with buffer solutions as described above or with buffer containing 1 percent BSA or 1 percent ovalbumin. Specificity was considered to be present if 75 percent or more of the fluorescence activity of serum preincubated with BSA was lost, and the values did not differ by more than 10 percent from those for serum preincubated with ovalbumin.
Analysis of the Proliferative Response of Peripheral-Blood Mononuclear Cells
Two methods were used to assess cellular proliferation. In the first method, 1 x 105 peripheral-blood mononuclear cells isolated from heparin-treated whole blood by Ficoll-Hypaque density centrifugation were cultured in round-bottomed 96-well tissue-culture trays in RPMI 1640 medium containing 10 percent human AB+ serum in an atmosphere of 95 percent air and 5 percent carbon dioxide. In the second method, heparin-treated whole blood was centrifuged at 500 x g for 10 minutes, and phosphate-buffered saline was added to the cell pellet before peripheral-blood mononuclear cells were separated by Ficoll-Hypaque density centrifugation. The isolated cells were washed twice with phosphate-buffered saline and then cultured (2 x 105 per well) in round-bottomed 96-well tissue-culture trays in serum-free HL-1 medium (Ventrex Labs, Portland, Me.). The two protocols differed in that the first method is a standard assay for the determination of cellular immunity20 that we have previously found to identify the responses of peripheral-blood mononuclear cells to glutamate decarboxylase in patients with IDDM,21 whereas the second method involves more stringent removal of potential sources of exogenous BSA or ABBOS peptide. For both methods, the cells were continuously incubated for seven days in triplicate or quadruplicate cultures with the following antigens: 1, 5, 10, 25, or 50 µg of BSA per milliliter (radioimmunoassay grade fraction V, Sigma Chemical, St. Louis); 1, 5, 10, 25, or 50 µg of ABBOS per milliliter; 10 µg of human insulin per milliliter; or 1 or 5 µg of tetanus toxoid per milliliter. As a control measure, the cells were also incubated for four days with 1 µg of phytohemagglutinin per milliliter. The ABBOS peptide was synthesized by the peptide-synthesis facility at the University of Florida or kindly donated by Dr. Hans-Michael Dosch (method 2). In both assay systems, 18 hours before the cells were harvested, 1 micro Ci of [3H]thymidine was added to each well. The incorporation of thymidine into proliferating cells was assessed by Matrix 96 beta-particle counting (Packard Instruments, Meriden, Conn.) or liquid-scintillation counting, and the mean value of each triplicate or quadruplicate culture was determined. Antigen-specific cellular proliferation was expressed as the stimulation index: the mean number of counts per minute incorporated in the presence of antigen divided by the mean number of counts per minute incorporated in its absence (medium alone). A response was considered to be positive if the stimulation-index value was 2 SD or more above the mean value for cultures of cells from normal subjects. Only subjects whose peripheral-blood mononuclear cells responded to mitogenic stimulation by phytohemagglutinin were included in these analyses, since this activity indicated a viable cellular preparation (i.e., suitable for the detection of antigen-specific responses). Analyses of differences between study groups were performed by both analysis of variance and Fisher's exact tests (two-tailed).
Results
Cellular Immunity to BSA or ABBOS
The number or proportion of subjects whose cells proliferated in response to BSA or ABBOS in method 1 did not differ significantly between the study groups (Figure 1). One normal subject had a positive response to BSA, and positive responses to ABBOS occurred in only 1 of 14 relatives of patients with IDDM who were negative for islet-cell antibodies (7 percent) and 1 of 13 patients with new-onset IDDM (8 percent). The mean (±SD) stimulation indexes in response to BSA and ABBOS were, respectively, 1.2 ±0.3 and 1.5 ±0.8 in the normal subjects, 1.3 ±1.5 and 1.1 ±0.5 in the relatives of patients with IDDM who were negative for islet-cell antibodies, 1.2 ±0.4 and 1.0 ±0.2 in the relatives of patients with IDDM who were positive for islet-cell antibodies, 1.0 ±0.3 and 0.9 ±0.2 in the patients with new-onset IDDM, and 0.8 ±0.3 and 0.8 ±0.2 in the patients with established IDDM.
|
Overall, 2 of 24 patients with new-onset IDDM, 1 of 25 first-degree relatives of patients with IDDM who were negative for islet-cell antibodies, 2 of 30 first-degree relatives who were positive for islet-cell antibodies, 1 of 28 patients with established IDDM, and 1 of 29 normal subjects had positive responses to BSA or ABBOS. The results of tests using cells collected at a different time were similar in both assays in the subjects tested. No concentration of BSA or ABBOS increased the responses in any diabetes-associated groups. Although synthetic peptides might not have the correct tertiary conformation for antigenic presentation and stimulation, the failure of BSA, which contains ABBOS and should undergo internal antigen processing in antigen-presenting cells, to induce proliferation further indicates a lack of responsiveness to ABBOS.
With the use of either assay method, most subjects responded to tetanus toxoid. In addition, the mean response to tetanus toxoid was not significantly different between the study groups (P = 0.23). However, the mean stimulation index of both the patients with new-onset IDDM (20.6 ±15.7) and the patients with established IDDM (21.4 ±16.1) was significantly higher than that of the normal subjects (9.1 ±16.0), relatives of patients with IDDM who were negative for islet-cell antibodies (14.4 ±18.6), and relatives of patients with IDDM who were positive for islet-cell antibodies (8.2 ±8.9) (P = 0.03). This observation may reflect the younger mean age of the two groups of patients with IDDM (see the Methods section). Overall, 15 percent of the patients with new-onset IDDM, 23 percent of those with established IDDM, 5 percent of first-degree relatives who were negative or positive for islet-cell antibodies, and no normal subjects had cellular responses to insulin, in accordance with the results of previous studies of insulin-specific T-cell responses22. The proliferative responses to insulin were significantly associated with new-onset and established IDDM (five-group analysis of variance, P = 0.04). In a randomly selected subgroup of eight patients with new-onset IDDM, the peripheral-blood mononuclear cells of four responded to glutamate decarboxylase, in accordance with our previous findings21.
Serum Anti-BSA Antibodies
The serum concentrations of IgG anti-BSA antibodies in the study groups are shown in Figure 2. The frequency of anti-BSA antibodies in patients with new-onset IDDM was not significantly higher than in the normal subjects (Table 1). However, as compared with the frequency in normal subjects, the frequency was increased in first-degree relatives of patients with IDDM regardless of their islet-cell antibody status and in patients with thyroiditis. Subjects at increased risk for IDDM (i.e., their relatives) had anti-BSA antibodies more often (9 of 42, 21 percent; P = 0.02) than did normal subjects. The presence of islet-cell antibodies in these relatives did not add an additional risk factor for the presence of anti-BSA antibodies (P = 1.0). As compared with the normal subjects, patients with autoimmune disease (new-onset IDDM, established IDDM, chronic autoimmune thyroiditis, systemic lupus erythematosus, or rheumatoid arthritis) also had an increased frequency of anti-BSA antibodies (14 of 82, 17 percent; P = 0.02).
|
|
Discussion
IDDM is an autoimmune disorder in which the etiologic role of environmental agents has been difficult to define. In a recent report, anti-BSA and anti-BSA peptide (ABBOS) antibodies were strongly associated with IDDM14. Since the pathogenic events in IDDM are more likely to be associated with cellular than serologic immunity, we analyzed the response of peripheral-blood mononuclear cells to BSA and ABBOS in patients with newly diagnosed IDDM, those with established disease, subjects at increased risk for IDDM, and subjects at low risk for IDDM.
We were unable to demonstrate any reactivity of peripheral-blood mononuclear cells to BSA or ABBOS in patients with IDDM or subjects at increased risk for the disease -- results that are in disagreement with a recent report by Karjalainen et al.23. Indeed, the difference between the two studies is striking, since we used two proliferation assays, the second of which was similar to that used by those investigators23. One major difference between the two studies is that Karjalainen et al.23 analyzed the results of antigen stimulation as a percentage of the seven-day response to phytohemagglutinin, which is variable and represents mitogenic stimulation, rather than the specific stimulation in response to antigens. The use of such values is problematic, since the response to phytohemagglutinin is often maximal before day 7. Analysis of our BSA- and ABBOS-induced stimulation results as a function of the four-day phytohemagglutinin response revealed no significant differences in stimulation between the study groups (data not shown). The studies by Karjalainen and coworkers23 were also limited in that no positive control antigens (e.g., tetanus) were tested.
Our findings of specific anti-BSA antibodies in patients with autoimmune diseases other than IDDM and in relatives of patients with IDDM raise the question of the specificity of anti-BSA immunity for IDDM. However, our results are in agreement with the results of other studies of anti-BSA immunity. Serum or synoviocytes from patients with rheumatoid arthritis bind soluble BSA as well as secrete anti-BSA antibodies24,25. Patients with chronic liver disease (e.g., chronic active hepatitis or cirrhosis) or proximal gastrointestinal tract disease (e.g., gastric cancer or peptic ulcer) have anti-BSA antibodies, whereas patients with distal gastrointestinal disease (e.g., Crohn's disease or ulcerative colitis) do not26,27. Patients with IgA nephropathy often have antibodies against multiple dietary antigens including BSA,28 and many plasma cells from the jejunal mucosa of children with intolerance to cow's milk protein produce anti-BSA antibodies29. Patients with familial dysautonomia have high titers of antihuman serum albumin and anti-BSA antibodies that may arise after the aspiration of bovine milk antigens during infancy30. Although many infants do not have anti-BSA antibodies at birth and acquire them only after the ingestion of BSA-containing nutritional formulas, others have anti-BSA antibodies at birth, presumably passively transferred from their mothers31. However, the titers of the anti-BSA antibodies in these two groups of infants are similar when the infants are compared at later ages. Furthermore, the gut of premature infants32 or intestinal epithelium damaged by retroviral or other infections33,34 may allow excessive passage of intact proteins. This situation alone may lead to the production of antibodies to BSA or other proteins in cow's milk, but diabetes has not been reported in such subjects.
Since our results support the observation that anti-BSA immunity is not restricted to IDDM, how might the reports of increased anti-BSA antibodies in IDDM be explained? One potential explanation comes from a large study measuring antibody responses to a number of autoantigens or natural self-proteins (e.g., double-stranded DNA, actin, tubulin, thyroglobulin, myosin, myoglobin, transferrin, interferon, and BSA) in patients with systemic lupus erythematosus and normal subjects35. Although not every subject had antibodies to every antigen, the patients with autoimmune disease had antibodies with higher affinities for all antigens. Therefore, antibodies to BSA in IDDM may reflect a nonspecific defect in the formation of immunologic tolerance (e.g., generally higher affinity of self-reactive antibodies) associated with a predisposition to autoimmunity, rather than immunity specific to beta cells. Other explanations include the possibility of polyclonal B-lymphocyte activation, the presence of antibodies with low antigenic specificity, or a generalized failure to develop tolerance to dietary antigens.
Antibodies from patients with IDDM have recently been identified to react with an islet-cell autoantigen with a relative molecular mass of 69,000 daltons (islet-cell p69)36. However, the hypothetical relation between anti-ABBOS, anti-BSA, and anti-islet-cell p69 immunity is unclear since islet-cell p69 only shares two small regions of sequence homology with BSA and these homologous regions are not related to the putative ABBOS region of BSA36. Furthermore, unlike the uniform recognition frequencies reported for anti-ABBOS antibodies in patients with IDDM,14 only a minority of patients with IDDM have antibodies that react with islet-cell p6936. Finally, there is a major discrepancy between the reported expression of ABBOS by tissue14 and the reported expression of p69 by tissue36. Therefore, although recognition of p69 as an autoantigen of IDDM has been firmly established, the relation between the islet-cell p69 and anti-BSA and anti-ABBOS immunity remains uncertain.
In view of previous reports of the stimulation of peripheral-blood mononuclear cells from patients with IDDM by insulin,22 an islet-cell antigen with a relative molecular mass of 38,000 daltons,37 islet cells,38 and glutamate decarboxylase,21 the results of this study call into question whether anti-BSA and anti-ABBOS immunity is of relevance to the pathogenesis of IDDM.
Supported in part by grants from the National Institutes of Health (R29DK45342 and R01HD19469) and by the Clinical Research Center at the University of Florida, the Diabetes Research and Education Foundation, the Juvenile Diabetes Foundation, and the American Diabetes Association.
We are indebted to Dr. Hans-Michael Dosch for his technical advice and contribution of ABBOS peptide used for our studies.
Source Information
From the Department of Pathology and Laboratory Medicine, College of Medicine, University of Florida, Gainesville (M.A.A., M.A.B., K.-J.K., L.C., P.J.D., N.K.M.); the Department of Pediatrics, University of South Florida, Tampa (S.C.S.); and the Department of Pediatrics, University of Turku, Turku, Finland (O.S.).
Address reprint requests to Dr. Atkinson at the Department of Pathology, University of Florida, Box 100275 JHMHC, Gainesville, FL 32610.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Lack of Immunity to Bovine Serum Albumin in Insulin-Dependent Diabetes Mellitus
Dosch H.-M., Karjalainen J., VanderMeulen J., Atkinson M. A., Kao K.J., Maclaren N. K.
Extract |
Full Text
N Engl J Med 1994;
330:1616-1617, Jun 2, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
