- Email: [email protected]
Diabetes and cows' milk
THE LANCET
Letters to the Editor
Diabetes and cows’ milk SIR—Cavallo and colleagues (Oct 5, p 926)1 report that more diabetics show a lymphocyte proliferative response to bovine casein than do healthy controls or patients with autoimmune thyroid diseases.1 They note sequence identity between bovine -casein and the B-cell glucose transporter GLUT-2 and conclude that a specific immune reaction to milk protein may lead to B-cell damage by a process involving molecular mimicry. T cells respond to short peptide sequences, and sometimes such sequences from different proteins may be similar or identical, so it is logical that molecular mimicry should occur. However, these workers’ observations shed no light on which way round the mimicry is operating. As they mention, many patients with recent-onset type-1 diabetes have antibodies to GLUT-2,2 which suggests immune reactivity to a component of this endogenous protein. Why could not the T cells from the diabetic patients in Cavallo and colleagues’ study be initially sensitised, through an autoimmune mechanism, to a GLUT-2 peptide and the response to bovine casein be an in vitro artifact due to molecular mimicry with one or more bovine casein peptides? The other-wayround argument must always be remembered in studies that purport to show molecular mimicry in immune reactions. Derek W R Gray Nuffield Department of Surgery, John Radcliffe Hospital, Headington, Ox ford OX3 9DU, UK
1
2
Cavallo MG, Fava D, Monetini L, Barone F, Pozzilli P. Cell-mediated immune response to  casein in recent-onset insulin-dependent diabetes: implications for disease pathogenesis. Lancet 1996; 348: 926–28. Inman LR, McAllister CT, Chen L, et al. Autoantibodies to the GLUT-2 glucose transporter of  cells in insulin-dependent diabetes mellitus of recent onset. Proc Nat Acad Sci USA 1993; 90: 1281–84.
Authors’ reply SIR—In reply to Gray, at present there is no evidence that βcasein-responsive T cells are crossreactive with GLUT-2, although a region of sequence identity exists between the two molecules. Therefore, the hypothesis of molecular mimicry between bovine β-casein and GLUT-2 or other β-cell antigens is yet to be proven. However, in the event that such mimicry is shown, we agree with Gray that GLUT-2 could indeed be the primary antigen to induce an autoimmune response and that T-cell clones could crossreact with βcasein. Nevertheless, whichever comes first—GLUT-2 or βcasein—environmental proteins could be the factors responsible for sustaining the activation of autoreactive T-cell clones, through continual intake of substances that contain β-
β-casein BSA OVA
IDDM patients positive/total
Controls positive/total
20/54 (37%)* 6/53 (11·3%) 2/40 (5·0%)
3/53 (5·6%) 2/48 (4·2%) 2/40 (5·0%)
*p<0·0001 vs controls (χ2 test).
Table: β-casein antibodies in IDDM patients and controls
Vol 348 • December 14, 1996
casein, such as milk. Finally, antibodies in recent-onset IDDM patients are present not only against GLUT-2 but also against β-casein, as we report here. In our report we postulated that T cells activated against βcasein may crossreact, via a mechanism of molecular mimicry, with protein expressed on pancreatic β-cells. We have now tested the antibody response to β-casein and to bovine serum albumin (BSA), another interesting although controversial milk protein candidate for triggering IDDM.1,2 We obtained serum samples from 54 recent-onset IDDM patients (mean age 16·8 years, SD 9), all living in the Lazio region (central Italy), where the incidence rate of IDDM is 7·9/105 per year in people aged under 15 years,3 and from 53 healthy controls matched for age and sex. Antibodies against bovine β-casein and BSA were measured by ELISA. Ovalbumin (OVA) was also tested as a negative control. Antibodies to these proteins were detected, after reaction of diluted serum samples with β-casein-coated plates, with an antihuman IgG peroxidase-conjugated antibody and 3,3',5,5',-tetramethylbenzidine as developer of the final reaction. A selected pool of ten samples from IDDM patients was used as a standard reference for each experiment. The value of optical density of each sample was divided by the value of the pool and the antibody titre expressed in units (U). Since an association between antibodies against milk proteins and IDDM is controversial, we attempted to test the specific reactivity to β-casein by immunoblotting.4 As shown in the table, antibodies to β-casein (2 SD above the mean units for healthy individuals) were detected by ELISA in 37% of recent-onset IDDM patients and in 5·6% of controls (p<0·0001, χ2 test). Reactivity to BSA was higher in IDDM patients (0·67 U, 0·2 SD) compared with controls (0·59 U, 0·1; p=0·07). Finally, reactivity to OVA did not differ between IDDM patients (0·74 U, 0·3) and controls (0·67 U, 0·3) (p=0·36). Immunoblotting with both whole milk and purified β-casein, confirmed the presence of antibodies reacting with a single band of 25 kDa corresponding to bovine β-casein.* Thus, antibodies against β-casein are present in over a third of IDDM patients but in few healthy individuals. These results, in addition to those we reported earlier, further support the notion that bovine β-casein may be implicated in the pathogenesis of IDDM, although the importance of antiβ-casein immunity remains to be elucidated. *Immunoblot available from the authors on request.
Maria Gisella Cavallo, Laura Monetini, Barry Walker, Robin Thorpe, *Paolo Pozzilli *Istituto II Clinica Medica and Department of Experimental Medicine, University of Rome “La Sapienza”, 00161 Rome, Italy; National Institute for Biological Standards and Control, South Mimms, UK; and St Bartholomew’s Hospital, London, UK.
1
2
3
4
Karjalainen J, Martin JM, Knip M, et al. A bovine albumin peptide as a possible trigger of insulin-dependent diabetes mellitus. N Engl J Med 1992; 327: 302–07. Atkinson MA, Bowman MA, Kao KJ, et al. Lack of immune responsiveness to bovine serum albumin in insulin-dependent diabetes. N Engl J Med 1993; 329: 1853–58. Sebastiani L, Visalli N, Adorisio E, et al. A 5-year (1989–1993) prospective study of the incidence of IDDM in Rome and the Lazio region in the age-group 0–14 years. Diabetes Care 1996; 19: 70–74. Johnstone A, Thorpe R. Immunochemistry in practice. 3rd edn. Oxford: Blackwell Science, 1996.
1655
Letters to the Editor
Diabetes and cows’ milk SIR—Cavallo and colleagues (Oct 5, p 926)1 report that more diabetics show a lymphocyte proliferative response to bovine casein than do healthy controls or patients with autoimmune thyroid diseases.1 They note sequence identity between bovine -casein and the B-cell glucose transporter GLUT-2 and conclude that a specific immune reaction to milk protein may lead to B-cell damage by a process involving molecular mimicry. T cells respond to short peptide sequences, and sometimes such sequences from different proteins may be similar or identical, so it is logical that molecular mimicry should occur. However, these workers’ observations shed no light on which way round the mimicry is operating. As they mention, many patients with recent-onset type-1 diabetes have antibodies to GLUT-2,2 which suggests immune reactivity to a component of this endogenous protein. Why could not the T cells from the diabetic patients in Cavallo and colleagues’ study be initially sensitised, through an autoimmune mechanism, to a GLUT-2 peptide and the response to bovine casein be an in vitro artifact due to molecular mimicry with one or more bovine casein peptides? The other-wayround argument must always be remembered in studies that purport to show molecular mimicry in immune reactions. Derek W R Gray Nuffield Department of Surgery, John Radcliffe Hospital, Headington, Ox ford OX3 9DU, UK
1
2
Cavallo MG, Fava D, Monetini L, Barone F, Pozzilli P. Cell-mediated immune response to  casein in recent-onset insulin-dependent diabetes: implications for disease pathogenesis. Lancet 1996; 348: 926–28. Inman LR, McAllister CT, Chen L, et al. Autoantibodies to the GLUT-2 glucose transporter of  cells in insulin-dependent diabetes mellitus of recent onset. Proc Nat Acad Sci USA 1993; 90: 1281–84.
Authors’ reply SIR—In reply to Gray, at present there is no evidence that βcasein-responsive T cells are crossreactive with GLUT-2, although a region of sequence identity exists between the two molecules. Therefore, the hypothesis of molecular mimicry between bovine β-casein and GLUT-2 or other β-cell antigens is yet to be proven. However, in the event that such mimicry is shown, we agree with Gray that GLUT-2 could indeed be the primary antigen to induce an autoimmune response and that T-cell clones could crossreact with βcasein. Nevertheless, whichever comes first—GLUT-2 or βcasein—environmental proteins could be the factors responsible for sustaining the activation of autoreactive T-cell clones, through continual intake of substances that contain β-
β-casein BSA OVA
IDDM patients positive/total
Controls positive/total
20/54 (37%)* 6/53 (11·3%) 2/40 (5·0%)
3/53 (5·6%) 2/48 (4·2%) 2/40 (5·0%)
*p<0·0001 vs controls (χ2 test).
Table: β-casein antibodies in IDDM patients and controls
Vol 348 • December 14, 1996
casein, such as milk. Finally, antibodies in recent-onset IDDM patients are present not only against GLUT-2 but also against β-casein, as we report here. In our report we postulated that T cells activated against βcasein may crossreact, via a mechanism of molecular mimicry, with protein expressed on pancreatic β-cells. We have now tested the antibody response to β-casein and to bovine serum albumin (BSA), another interesting although controversial milk protein candidate for triggering IDDM.1,2 We obtained serum samples from 54 recent-onset IDDM patients (mean age 16·8 years, SD 9), all living in the Lazio region (central Italy), where the incidence rate of IDDM is 7·9/105 per year in people aged under 15 years,3 and from 53 healthy controls matched for age and sex. Antibodies against bovine β-casein and BSA were measured by ELISA. Ovalbumin (OVA) was also tested as a negative control. Antibodies to these proteins were detected, after reaction of diluted serum samples with β-casein-coated plates, with an antihuman IgG peroxidase-conjugated antibody and 3,3',5,5',-tetramethylbenzidine as developer of the final reaction. A selected pool of ten samples from IDDM patients was used as a standard reference for each experiment. The value of optical density of each sample was divided by the value of the pool and the antibody titre expressed in units (U). Since an association between antibodies against milk proteins and IDDM is controversial, we attempted to test the specific reactivity to β-casein by immunoblotting.4 As shown in the table, antibodies to β-casein (2 SD above the mean units for healthy individuals) were detected by ELISA in 37% of recent-onset IDDM patients and in 5·6% of controls (p<0·0001, χ2 test). Reactivity to BSA was higher in IDDM patients (0·67 U, 0·2 SD) compared with controls (0·59 U, 0·1; p=0·07). Finally, reactivity to OVA did not differ between IDDM patients (0·74 U, 0·3) and controls (0·67 U, 0·3) (p=0·36). Immunoblotting with both whole milk and purified β-casein, confirmed the presence of antibodies reacting with a single band of 25 kDa corresponding to bovine β-casein.* Thus, antibodies against β-casein are present in over a third of IDDM patients but in few healthy individuals. These results, in addition to those we reported earlier, further support the notion that bovine β-casein may be implicated in the pathogenesis of IDDM, although the importance of antiβ-casein immunity remains to be elucidated. *Immunoblot available from the authors on request.
Maria Gisella Cavallo, Laura Monetini, Barry Walker, Robin Thorpe, *Paolo Pozzilli *Istituto II Clinica Medica and Department of Experimental Medicine, University of Rome “La Sapienza”, 00161 Rome, Italy; National Institute for Biological Standards and Control, South Mimms, UK; and St Bartholomew’s Hospital, London, UK.
1
2
3
4
Karjalainen J, Martin JM, Knip M, et al. A bovine albumin peptide as a possible trigger of insulin-dependent diabetes mellitus. N Engl J Med 1992; 327: 302–07. Atkinson MA, Bowman MA, Kao KJ, et al. Lack of immune responsiveness to bovine serum albumin in insulin-dependent diabetes. N Engl J Med 1993; 329: 1853–58. Sebastiani L, Visalli N, Adorisio E, et al. A 5-year (1989–1993) prospective study of the incidence of IDDM in Rome and the Lazio region in the age-group 0–14 years. Diabetes Care 1996; 19: 70–74. Johnstone A, Thorpe R. Immunochemistry in practice. 3rd edn. Oxford: Blackwell Science, 1996.
1655