Diabetologia
Diabetologia 15,447-451 (1978)
© by Springer-Verlag 1978
HLA-Typing in Juvenile Diabetics with and without Positive Family History and in Families with One and Two Diabetic Siblings N. Bengsch a, J. K6bberling 1, G. Eckert 2, and B. Willms3 1 Department of Medicine, University of G6ttingen, 2 Medizinische Hochschule Hannover, Abteilung ffir Klinische Immunologie und Transfusionsmedizin, and 3 Spezialklinik fiir Diabetes und Stoffwechselkrankheiten, Bad Lauterberg, FRG
Summary. HLA-typing was performed in two groups of juvenile-onset diabetics, one with (n = 58) and one without (n = 109) a family history of the disease. The association of this type of diabetes with certain H L A antigens (excess of B8 and B 15, shortage of B7) was confirmed. No heterogeneities could be established between the two groups. This suggests that the aetiologic basis in single and familial cases of juvenile diabetes is the same. The hypothesis, that the B 8 associated gene is more penetrant than the B 15 associated gene, cannot be confirmed. Haplotypes were determined in families with one and two diabetic siblings. The findings of high haplotype concordance among diabetic siblings was confirmed: concordance of 2, 1 and 0 haplotypes in 7, 5 and 3 pairs respectively. There was a low degree of haplotype concordance between diabetics and nonaffected siblings in the families with two diabetics: 2, 1, and 0 haplotypes in 2, 8 and 6 pairs respectively. This led to the hypothesis of negative selection against these HLA-linked "diabetogenic" genes. This tendency was not, however, observed in families with only one diabetic. The report of a high recombination rate in families with juvenile diabetics could not be confirmed. Key words: Juvenile diabetes, HLA, family history, heterogeneity, genetics, haplotype concordance, negative selection, recombination.
Several positive and negative associations between juvenile-onset diabetes and genes of the HLA-region have been established. Some results have suggested that heterogeneities exist within juvenile diabetics, e. g. according to autoimmunity [3], to seasonal patterns of incidence [4], or to various other factors [12].
In order to clarify whether a subgroup of genetic diabetes exists independent of the H L A system, two groups of juvenile diabetics with and without a family history of diabetes have been studied with respect to H L A antigens. Cudworth and Woodrow [6] have reported a high concordance of haplotypes among siblings with juvenile-onset diabetes. By HLA-typing of families with two affected children this observations has been reexamined. In addition, the HLA-typing of all available non-diabetic siblings should indicate possible genetic influence of the H L A system in these families besides diabetes. Such family data also allow reexamination of the claim by Rubinstein et al. [13] that recombination frequency is increased in families with one or more juvenile-onset diabetics.
Materials and Methods Juvenile-onset diabetes was defined as diabetes diagnosed at age 30 or younger and insulin-dependent, or at least insulin treated from the time of diagnosis. Patients with at least one first-degree relative or two seconddegree relatives with insulin-dependent diabetes were classified as "positive family history". "Negative family history" was assumed in patients with no insulin-dependent diabetic among first- and second-degree relatives and no more than one second-degree relative with maturity-onset diabetes. 58 diabetics with positive and 109 with negative family history have been investigated. The family studies with HLA-typing of all available parents and siblings were performed in 15 families with two diabetic children (47 children in total) and in 15 families with only one diabetic child (54 children). HLA antigens were determined with the lymphocytotoxicity test, using 15 antisera of the A-series, 19 of the B-series, and, with some exceptions, 2 to 4 antisera of the C-series. Only the antigens B 7, B 8 and B 15 were analysed statistically according to preformulated hypotheses. Only in the case of the haplotype concordance of diabetic and non-diabetic siblings had no hypothesis been formulated in advance. For statistical analysis binomial and chisquare tests were used.
0012-186X/78/0015/0447/$01.00
N. Bengsch et al.: HLA in Juvenile Diabetics
448 Table 1. Frequencies of HLA antigens in juvenile-onset diabetics
Percent frequency HLAantigen
Group with negative family history (n = 109)
Group with positive family history (n = 58)
Juvenile diabetics from the literature a
B B B B B B
12.8 42.4 16.5 4.6 1.8 4.6
12.1 44.8 24.1 5.2 0.0 6.9
14.6 (n = 634) 44.6 (n = 848) 23.0 (n = 848)
7 8 15 7 and B 8 7 and B 15 8 and B 15
a Combined data [5, 6, 7, 8, 9, 11, 14, 15]
Table 2. Haplotype concordance between diabetic siblings
2 haplotypes 1 haplotype in common in common
No haplotype in common
Sum
Expected occurrence for any given siblings 25%
50%
25%
100%
Diabetic siblings
5 ( = 33.3%)
3 ( = 20.0%)
15
7 ( = 47.7%)
Table 3. Haplotype concordance between the first diabetic and his
non-affected siblings in families with two diabetic siblings 2 haplotypes in common
I haplotype in common
No haplotype in common
1 ( = 7% )
7 ( = 50%)
6 ( = 43%)
Sum 14
Table 4. Haplotype concordance between the diabetic and his nonaffected siblings in families with only one diabetic 2 haplotypes
1 haplotype
No haplotype
in common
in common
in common
Sum
12 (= 31.5%)
14 (= 37%)
12 (= 31.5%)
38
Results
H L A frequencies for the groups with and without family history, together with combined data from the literature are shown in Table 1. A X2-test after separation in disjunct classes did not reveal significant differences between the groups. Also, the other tested antigens of the A-, B- and Cloci did not indicate heterogeneity between subjects with and without family history.
In Table 2 the degrees of haplotype concordance of the diabetic siblings are shown. Using the binomial test, the increased frequency of two common haplotypes is not significant (0.05 < P < 0.1). Two of the three pairs without a common haplotype had distinct aetiologies. In one pair both siblings became diabetic within one week of a mumps infection. In the second pair, one sibling as well as his monozygotic twin (not included in this study) has Addison's disease, thus belonging more properly in the category of autoimmune-polyendocrinopathy. If the data of table 2 are corrected for "idiopathic" diabetes, leaving out the two pairs, the high haplotype identity is more impressive and significant (P < 0.05). The haplotype concordances between the first diabetic child (selected randomly) and all nonaffected siblings are shown in Table 3. The two families mentioned above have been excluded. The deviation from a random distribution in the opposite direction was suggestive, but not significant. None of the six siblings of those diabetic pairs with identical haplotypes share the same hapiotypes. In Table 4 the corresponding data are given from families with only one diabetic child. No significant deviation from a random distribution was found. One recombination between A- and B-locus was found among 112 cases examined for this study. Detailed data are listed in appendices 1 and 2.
Discussion
No normal control population was H L A typed for this study. However, the H L A frequencies of non selected inhabitants of various parts of Germany are about equal and correspond well with those found by other authors in E u r o p e and North America [1]. Thus, the association of juvenile diabetes with special H L A antigens is confirmed in view of the good
N. Bengsch et al.: HLA in Juvenile Diabetics
accordance of the H L A frequencies with those reported in the literature [5, 6, 7, 8, 9, 11, 14, 15]. The hypothesis of an HLA-independent subgroup with a strong family history could not be verified. The similarity of H L A frequencies in the groups with and without family history suggests, on the contrary, that identical aetiologies are responsible for the diabetes in both groups. The familial incidence of juvenile diabetes is, therefore, when compared with the single cases, a random occurrence. It is possible, however, that there are various alleles at the locus of the "diabetogenic gene". Thus, genes of different penetrance could be in the same linkage disequilibrium with certain H L A antigens. Nelson et al. [10], in their studies on identical twin pairs, have observed an increased frequency of HLA-B 8 only among concordant juvenile diabetic pairs (18/31), but not among discordant pairs [10/ 31]. HLA-B15 had been found to be equally increased among concordant and discordant pairs [11/31]. Although these differences are not significant the authors have speculated that the allele linked with B 8 would be more penetrant. This speculation has been quoted by Rotter and Rimoin [12] to support their hypothesis of a heterogeneity among type I diabetes. The data of the present study do not allow such conclusions, since HLA-B 8 was equally frequent among the groups with and without diabetic family history. On the contrary, HLA-B 15 was more frequent among the familial group, but the difference is not significant. This possible difference should be retested with further data. The fact that two of the three pairs of diabetic siblings without common haplotypes had probably distinct causes for their diabetes suggests that in some cases of juvenile diabetes HLA-associated genes are probably not involved. A cross-over between the "diabetes-gene" and the HLA-B-locus cannot be ruled out, but seems unlikely considering the known recombination rates [13]. By analysing the family data, the findings of a high degree of haplotype identity among diabetic siblings [2, 6] could be confirmed. This assures the importance of HLA-associated factors for the development of juvenile-onset diabetes. The low haplotype concordance between diabetic and nondiabetic siblings (table 3) was remarkable, though not significant. Provided that this observation represents a systematic effect that could not be established due only to the small number of cases, one has to look for an explanation. The fact that a carrier of the same H L A antigens as a diabetic has a slightly increased chance of developping diabetes himself does not suffice to
449
explain such a negative association. One has to postulate negative selection against the HLA-associated diabetogenic factors independent of overt diabetes, e. g. by an increased number of early abortions. If it is assumed that the aetiology is not different between the diabetics with or without positive family history, the negative association of haplotypes between diabetic and non-diabetic siblings should also be observed in families with only one diabetic child. For this reason the hypothesis of negative selection was reexamined in such families, which are far more frequent. By doing so the above hypothesis could not be confirmed. Also the data of Barbosa et al. [2] can be interpreted in this sense, although these authors have incorrectly used highly dependent data more than once for the same analysis. We could not confirm the high recombination frequency in families with juvenile diabetics that Rubinstein et al [13] reported. The frequency of 1:112 is in good accordance with that usually assumed [16].
Appendix 1 H L A antigens of diabetics with and without family history of diabetes - given only antigens of the A- and B-series. A l w a y s determined: 1, 2, 3, 9, 10, 11, 28, 29, 30, 31, 32, 5, 7, 8, 12, 13, 14, 18, 27, 15, 17, 21, 22, 35. Not always determined: 33, 34, 36, 43, 16, 37, 40, 41, 42. With family history No. Antigens
No.
FM
FW
1 2 3 4 5 6 7 8 9 10
1,8 1,2, 7,27 2, 3, 14, 15 2, 9, 12, 39 1, 24, 8, 18 3, 24, 8, 18 2, 40 2,9,5,15 1,8,15 1, 8
11 12 13 14 15 16 17 18 19 20
2, 2, 3, 1, 1, 2, 1, 2, 2, 2,
21 22 23 24 25 26 27 28 29 30
3, 9, 12, 27 1, 2, 8, 38 30, 5, 15 1,2, 8 1, 8 2, 35 2, 28, 8, I2 1, 2, 15 1, 2, 8, 13 1, 2, 8, 18
31 32 33
2, 11, 15, 35 2, 40 1,2,5,8
3, 12, 41 3, 13, 39 25, 18, 39 2, 8, 40 24, 8, 15 24, 8, 22 11, 15 24, 5, 22 29, 13 30, 18, 35
Antigens 1 2 3 4 5 6 7 8 9 10
3, 9, 8, 35 1,8, 40 1,3,8,35 25, 12, 18 23, 24, 7, 35 2, 24, 8, 15 1,3,7,8 2, 18, 21 2, 40 2, 11, 14, 27
11 12 13 14 15 16 17 18 19 20
1, 2, 15, 35 2, 3, 15, 35 2, 3, 7, 39 2, 12, 35 1,3,7,8 1,2,7,8 2, 11, 5, 15 1, 11, 8, 35 2,32, 5 , 2 2 1, 2, 8, 15
21 22 23 24 25
2, 2, 1, 1, 2,
3, 7, I3 25, 12, 15 26, 8, 12 8 24, 12, 39
450
N. Bengsch et al.: H L A in Juvenile Diabetics
Without family history No. Antigens
No.
KM
KM
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
2, 24, 16, 35 1, 24, 12 1, 8 1, 35 2, 9, 17, 37 2, 15 3, 10, 8 2, 3, 7, 15 1, 2, 8 1, 8 1, 2, 8, 15 2, 3, 16, 17 2, 3, 17, 35 9, 28, 12 2, 9 1, 9, 8, 39 2, 30, 18, 27 2, 25 29, 31, 12, 15 l, 2, 15, 39 2, 25, 40 1, 3, 8, 27 2, 8, 12 3, 25, 7, 18 2, 3, 15 2, 17
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68
Antigens
No.
3, 28, 12, 18 1, 9, 8, 12 1, 28, 8, 15 2, 29, 8, 35 1, 32, 8, 41 2, 9, 15, 39 2, 15 1, 9, 7, 8 2, 3, 35, 40 2, 7, 15 2, 11, 21, 35 2, 7, 27 2, 25, 18, 27 1, 2, 8, 40 2, 8, 12 11, 8, 35 2, 3, 7, 13 2, 12, 35
KW
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 2l 22 23 24 25 26
Antigens
No.
2, 3, 12, 40 3, 26, 12, 27 1, 2, 8, 13 3, 29, 5 1, 32, 8 3, 9, 35 1, 8 2, 9, 12 2, 3, 7 2, 3, 21, 22 2, 9, 15, 18 2, 5, 12 1, 26, 8, 35 1, 2, 8 28, 18 30, 13, 35 25, 28, 5, 35 1, 2, 8 1, 8, 27 9, 27, 40 l, 2, 7, 8 2, 24, 27, 39 2, 39, 40 28, 12 3, 9, 12, 15 24, 18, 27
KM
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
Antigens
No.
2, 3, 15, 18 1, 3, 8, 15 25, 14, 18 1, 2, 7, 8 1, 3, 7, 8 2, 3, 35 24, 30, 18, 2 7 2, 26, 18, 40 l, 11, 8, 35 1, 26, 8, 22 2, 9, 7, 40 2, 25, 8, 15 2, 25, 8, 18 3, 9, 7, 16 1, 2, 8, 21 2, 9, 7 1, 8 2, 3, 18, 39 3, 26, 7, 8 24, 13, 22 1, 24, 8 1, 3, 8, 18 2, 3, 8 2, 3, 15, 27
KW
Antigens 27 28 29 30
1,2,8,40 1, 29, 12 2, 3, 15, 35 1, 2, 8, 13
31 32 33 34 35 36 37 38 39 40
2, 8 2, 29, 8, 12 1, 9 , 8 1, 9, 8 , 2 7 2, 24, 8, 22 2, 3, 12, 18 1,2,8 1, 28, 8, 15 2, 5, 40 11, 25, 12, 35
41
11, 25, 18, 35
Appendix 2 F a m i l y data: Given only antigens of the A- and B-series. a, b: paternal haplotypes; c, d: maternal haplotypes. Haplotypes which could not be determined by H L A typing but were deduced are given in brackets. In family K 13, both parents could not be HLA-typed. The naming "paternal" and "maternal" is arbitrary. Sibling 4 in family F6 shows a recombination: 24, 5, 12
Diabetics Family
a
b
F 1 F 2 F 3 F 4 F 5 F 6 F 7 F 8 F 9 F10 Fll F12 F13 F14 F15
3, 2, (2, 11, 29, (2, (2, 23, 1, 2, 2, 1, 23, 11, 2,
K 1 K 2 K 3 K 4 K 5 K 6 K 7 K 8 K 9 K10 Kll K12 K13 K14 K15 K16
2, 8 29, 12 (30, 27) 2, 8 2, 15 2, 15 32, 41 2, 12 24, 12 2, 27 (3, 15) 2, 40 (2, 15) 24, 35 1, 8 1, 8
35 15 15) 5 41 22) 13) 7 37 12 35 8 8 35 38
c
24, 12 28, 12 (1, 8) 2, 15 3, 18 (24, 12) • (24, 12) 2, 40 2, 5 1, 8 ?, 40 28, 5 1, 13 1, 8 1, 8 ?, 2, (2, 2, ?, 3, 24, 1, 1, 28, (11, 1, (11, 2, 32, 3,
35 12 41) 15 7 7 18 8 8 12 ?) 17 35) 12 22 27
Non-affected siblings
d
1
2
1
2
ac ac ad ad bd ad bd ac ac ac ac ad ad bd ac
ad
ad
bc ad ac bd
bd
(2, 3, 11, 2, 2, 24, 1, 24, (11, 25, 30, 2, 3, 24, 1,
15) 14 35 15 13 8 8 35 15) 15 18 12 35 22 8
(2, 40) 2, 21 1, 8 24, 40 1, 8 11, 5 2, 15 2, 27 (2, 12) 1, 8 24, 27 1, 8 1, 13 1, 8 3, 15
ac ac ac ac ac ac ac ac ac ac ac ac ac ac ac
24, 31, 24, 2, 24, 1, 1, ?, 28, 24, 2, 24, (?, 2, 1, 2,
22 15 18 12 18 8 8 5 12 39 35 7 7) 39 8 8
29, 2, 3, 3, 1, 25, 3, 28, 24, 1, 2, 3, (1, 26, 3, 1,
ac ac ac ac ac ac ac ac ac ac ac ac ac ac ac ac
12 7 15 35 8 18 40 40 12 15 40 40 7) 27 7 12
3
4
bd ac ad bd
bc
b c/d
bd
ad
bd ad
ac bc ac bd ac bc ad ad bd ac bd ad ad bc bd ae
ad bd bc ac ac bd bc ad bd bc bd ac bd ac bc bd
ad
bd ad ac bd
ac
N. Bengsch et al.: HLA in Juvenile Diabetics References
1. Albert, E.D., Scholz, S., Bertrams, J., Ewald, R.W., Westphal, E., Ratschko, K.-W., Spielmann, W., Seidl, S.: Representative HL-A phenotype and haplotype frequencies of the German population. Z. Immunitaetsforsch. 148, 367-413 (1975) 2. Barbosa, J., King, R., Noreen, H., Yunis, E. J.: The histocompatibility system in juvenile, insulindependent diabetic multiplex kindreds. J. Clin. Invest 60, 989-998 (1977) 3. Christy, M., Nerup, J., Bottazzo, G. F., Doniach, D., Platz, P., Svejgaard, A., Ryder, L.P., Thomsen, M.: Association between HLA-B 8 and autoimmunity in juvenile diabetes mellitus. Lancet 1976 II, 142-143 4. Cudworth, A.G., White, G. B. B., Woodrow, J. C., Gamble, D. R., Lendrum, R., Bloom, A.: Aetiology of juvenile-onset diabetes. A prospective study. Lancet 1977 I, 385-388 5. Cudworth, A. G., Woodrow, J. C.: HL-A system and diabetes mellitus. Diabetes 24, 345-349 (1975) 6. Cudworth, A. G., Woodrow, J. C.: Evidence for HL-A-linked genes in "juvenile" diabetes mellitus. Br. Med. J. 1975 III, 133-135 7. Cudworth, A.G., Woodrow, J.C.: Genetic susceptibility in diabetes mellitus: analysis of the HLA association. Br. Med. J. 1976 II, 846-848 8. Ludvigsson, L., S~ifwenberg, J., Heding, L.G.: HLA-types, C-peptide and insulin antibodies in juvenile diabetes. Diabetologia 13, 13-17 (1977) 9. Mirouze, J., Seignalet, J., Selam, J.-L., Lapinski, H., Jaffiol, C.: Antig6nes HLA chez les diab6tiques insulino-d6pendants. Nouv. Presse Med. 5, 1628-1630 (1976) 10. Nelson, P. G., Pyke, D. A., Cudworth, A. G., Woodrow, J. C., and Batchelor, J.R.: Histocompatibility antigens in diabetic identical twins. Lancet 1975 II, 193-194
451 11. Nerup, J., Ortved Andersen, O., Christy, M., Platz, P., Ryder, L., Thomsen, M., Svejgaard, A.: HLA, autoimmunity, virus and the pathogenesis of juvenile diabetes mellitus. Acta Endoerinol. [Suppl.] (Kbh.) 205, 167-175 (1976) 12. Rotter, J. I., Rimoin, D. L.: Heterogeneity in diabetes mellitus - update, 1978. Evidence for further genetic heterogeneity within juvenile-onset insuline-dependent diabetes mellitus. Diabetes 27, 599-608 (1978) 13. Rubinstein, O., Suciu-Foca, N., Nicholson, J. F., Fontino, M., Molinatio, A., Harisiadis, K., Hardy, M.A., Reemtsma, K., Allen, F.H. jr.: The HLA system in the families of patients with juvenile diabetes mellitus. J. Exp. Med. 143, 1277-1282 (1976) 14. Schernthaner, G., Ludwig, H., Mayr, W. R.: Juvenile diabetes mellitus: HLA-antigen frequencies dependent on the age of onset of the disease. J. Immunogenet. 3, 117-121 (1976) 15. Singal, D.P., Blajchmann, M. A.: Histocompatibility (IlL-A) antigens, lymphocytotoxic antibodies and tissue antibodies in patients with diabetes mellitus. Diabetes 22, 429-432 (1973) 16. Svejgaard, A., Hauge, M., Jersild, C., Platz, P., Ryder, L. P., Staub Nielsen, L., Thomsen, M.: The HLA system. An introductory survey. Monogr. Hum. Genet. 7, 25 (1975) Received: June 15, 1978, and in revised form: August 9, 1978
Prof. Dr. med. J. K6bberling Department of Medicine University of G6ttingen Robert-Koch-StraBe 40 D-3400 G6ttingen Federal Republic of Germany
Diabetologia 15,447-451 (1978)
© by Springer-Verlag 1978
HLA-Typing in Juvenile Diabetics with and without Positive Family History and in Families with One and Two Diabetic Siblings N. Bengsch a, J. K6bberling 1, G. Eckert 2, and B. Willms3 1 Department of Medicine, University of G6ttingen, 2 Medizinische Hochschule Hannover, Abteilung ffir Klinische Immunologie und Transfusionsmedizin, and 3 Spezialklinik fiir Diabetes und Stoffwechselkrankheiten, Bad Lauterberg, FRG
Summary. HLA-typing was performed in two groups of juvenile-onset diabetics, one with (n = 58) and one without (n = 109) a family history of the disease. The association of this type of diabetes with certain H L A antigens (excess of B8 and B 15, shortage of B7) was confirmed. No heterogeneities could be established between the two groups. This suggests that the aetiologic basis in single and familial cases of juvenile diabetes is the same. The hypothesis, that the B 8 associated gene is more penetrant than the B 15 associated gene, cannot be confirmed. Haplotypes were determined in families with one and two diabetic siblings. The findings of high haplotype concordance among diabetic siblings was confirmed: concordance of 2, 1 and 0 haplotypes in 7, 5 and 3 pairs respectively. There was a low degree of haplotype concordance between diabetics and nonaffected siblings in the families with two diabetics: 2, 1, and 0 haplotypes in 2, 8 and 6 pairs respectively. This led to the hypothesis of negative selection against these HLA-linked "diabetogenic" genes. This tendency was not, however, observed in families with only one diabetic. The report of a high recombination rate in families with juvenile diabetics could not be confirmed. Key words: Juvenile diabetes, HLA, family history, heterogeneity, genetics, haplotype concordance, negative selection, recombination.
Several positive and negative associations between juvenile-onset diabetes and genes of the HLA-region have been established. Some results have suggested that heterogeneities exist within juvenile diabetics, e. g. according to autoimmunity [3], to seasonal patterns of incidence [4], or to various other factors [12].
In order to clarify whether a subgroup of genetic diabetes exists independent of the H L A system, two groups of juvenile diabetics with and without a family history of diabetes have been studied with respect to H L A antigens. Cudworth and Woodrow [6] have reported a high concordance of haplotypes among siblings with juvenile-onset diabetes. By HLA-typing of families with two affected children this observations has been reexamined. In addition, the HLA-typing of all available non-diabetic siblings should indicate possible genetic influence of the H L A system in these families besides diabetes. Such family data also allow reexamination of the claim by Rubinstein et al. [13] that recombination frequency is increased in families with one or more juvenile-onset diabetics.
Materials and Methods Juvenile-onset diabetes was defined as diabetes diagnosed at age 30 or younger and insulin-dependent, or at least insulin treated from the time of diagnosis. Patients with at least one first-degree relative or two seconddegree relatives with insulin-dependent diabetes were classified as "positive family history". "Negative family history" was assumed in patients with no insulin-dependent diabetic among first- and second-degree relatives and no more than one second-degree relative with maturity-onset diabetes. 58 diabetics with positive and 109 with negative family history have been investigated. The family studies with HLA-typing of all available parents and siblings were performed in 15 families with two diabetic children (47 children in total) and in 15 families with only one diabetic child (54 children). HLA antigens were determined with the lymphocytotoxicity test, using 15 antisera of the A-series, 19 of the B-series, and, with some exceptions, 2 to 4 antisera of the C-series. Only the antigens B 7, B 8 and B 15 were analysed statistically according to preformulated hypotheses. Only in the case of the haplotype concordance of diabetic and non-diabetic siblings had no hypothesis been formulated in advance. For statistical analysis binomial and chisquare tests were used.
0012-186X/78/0015/0447/$01.00
N. Bengsch et al.: HLA in Juvenile Diabetics
448 Table 1. Frequencies of HLA antigens in juvenile-onset diabetics
Percent frequency HLAantigen
Group with negative family history (n = 109)
Group with positive family history (n = 58)
Juvenile diabetics from the literature a
B B B B B B
12.8 42.4 16.5 4.6 1.8 4.6
12.1 44.8 24.1 5.2 0.0 6.9
14.6 (n = 634) 44.6 (n = 848) 23.0 (n = 848)
7 8 15 7 and B 8 7 and B 15 8 and B 15
a Combined data [5, 6, 7, 8, 9, 11, 14, 15]
Table 2. Haplotype concordance between diabetic siblings
2 haplotypes 1 haplotype in common in common
No haplotype in common
Sum
Expected occurrence for any given siblings 25%
50%
25%
100%
Diabetic siblings
5 ( = 33.3%)
3 ( = 20.0%)
15
7 ( = 47.7%)
Table 3. Haplotype concordance between the first diabetic and his
non-affected siblings in families with two diabetic siblings 2 haplotypes in common
I haplotype in common
No haplotype in common
1 ( = 7% )
7 ( = 50%)
6 ( = 43%)
Sum 14
Table 4. Haplotype concordance between the diabetic and his nonaffected siblings in families with only one diabetic 2 haplotypes
1 haplotype
No haplotype
in common
in common
in common
Sum
12 (= 31.5%)
14 (= 37%)
12 (= 31.5%)
38
Results
H L A frequencies for the groups with and without family history, together with combined data from the literature are shown in Table 1. A X2-test after separation in disjunct classes did not reveal significant differences between the groups. Also, the other tested antigens of the A-, B- and Cloci did not indicate heterogeneity between subjects with and without family history.
In Table 2 the degrees of haplotype concordance of the diabetic siblings are shown. Using the binomial test, the increased frequency of two common haplotypes is not significant (0.05 < P < 0.1). Two of the three pairs without a common haplotype had distinct aetiologies. In one pair both siblings became diabetic within one week of a mumps infection. In the second pair, one sibling as well as his monozygotic twin (not included in this study) has Addison's disease, thus belonging more properly in the category of autoimmune-polyendocrinopathy. If the data of table 2 are corrected for "idiopathic" diabetes, leaving out the two pairs, the high haplotype identity is more impressive and significant (P < 0.05). The haplotype concordances between the first diabetic child (selected randomly) and all nonaffected siblings are shown in Table 3. The two families mentioned above have been excluded. The deviation from a random distribution in the opposite direction was suggestive, but not significant. None of the six siblings of those diabetic pairs with identical haplotypes share the same hapiotypes. In Table 4 the corresponding data are given from families with only one diabetic child. No significant deviation from a random distribution was found. One recombination between A- and B-locus was found among 112 cases examined for this study. Detailed data are listed in appendices 1 and 2.
Discussion
No normal control population was H L A typed for this study. However, the H L A frequencies of non selected inhabitants of various parts of Germany are about equal and correspond well with those found by other authors in E u r o p e and North America [1]. Thus, the association of juvenile diabetes with special H L A antigens is confirmed in view of the good
N. Bengsch et al.: HLA in Juvenile Diabetics
accordance of the H L A frequencies with those reported in the literature [5, 6, 7, 8, 9, 11, 14, 15]. The hypothesis of an HLA-independent subgroup with a strong family history could not be verified. The similarity of H L A frequencies in the groups with and without family history suggests, on the contrary, that identical aetiologies are responsible for the diabetes in both groups. The familial incidence of juvenile diabetes is, therefore, when compared with the single cases, a random occurrence. It is possible, however, that there are various alleles at the locus of the "diabetogenic gene". Thus, genes of different penetrance could be in the same linkage disequilibrium with certain H L A antigens. Nelson et al. [10], in their studies on identical twin pairs, have observed an increased frequency of HLA-B 8 only among concordant juvenile diabetic pairs (18/31), but not among discordant pairs [10/ 31]. HLA-B15 had been found to be equally increased among concordant and discordant pairs [11/31]. Although these differences are not significant the authors have speculated that the allele linked with B 8 would be more penetrant. This speculation has been quoted by Rotter and Rimoin [12] to support their hypothesis of a heterogeneity among type I diabetes. The data of the present study do not allow such conclusions, since HLA-B 8 was equally frequent among the groups with and without diabetic family history. On the contrary, HLA-B 15 was more frequent among the familial group, but the difference is not significant. This possible difference should be retested with further data. The fact that two of the three pairs of diabetic siblings without common haplotypes had probably distinct causes for their diabetes suggests that in some cases of juvenile diabetes HLA-associated genes are probably not involved. A cross-over between the "diabetes-gene" and the HLA-B-locus cannot be ruled out, but seems unlikely considering the known recombination rates [13]. By analysing the family data, the findings of a high degree of haplotype identity among diabetic siblings [2, 6] could be confirmed. This assures the importance of HLA-associated factors for the development of juvenile-onset diabetes. The low haplotype concordance between diabetic and nondiabetic siblings (table 3) was remarkable, though not significant. Provided that this observation represents a systematic effect that could not be established due only to the small number of cases, one has to look for an explanation. The fact that a carrier of the same H L A antigens as a diabetic has a slightly increased chance of developping diabetes himself does not suffice to
449
explain such a negative association. One has to postulate negative selection against the HLA-associated diabetogenic factors independent of overt diabetes, e. g. by an increased number of early abortions. If it is assumed that the aetiology is not different between the diabetics with or without positive family history, the negative association of haplotypes between diabetic and non-diabetic siblings should also be observed in families with only one diabetic child. For this reason the hypothesis of negative selection was reexamined in such families, which are far more frequent. By doing so the above hypothesis could not be confirmed. Also the data of Barbosa et al. [2] can be interpreted in this sense, although these authors have incorrectly used highly dependent data more than once for the same analysis. We could not confirm the high recombination frequency in families with juvenile diabetics that Rubinstein et al [13] reported. The frequency of 1:112 is in good accordance with that usually assumed [16].
Appendix 1 H L A antigens of diabetics with and without family history of diabetes - given only antigens of the A- and B-series. A l w a y s determined: 1, 2, 3, 9, 10, 11, 28, 29, 30, 31, 32, 5, 7, 8, 12, 13, 14, 18, 27, 15, 17, 21, 22, 35. Not always determined: 33, 34, 36, 43, 16, 37, 40, 41, 42. With family history No. Antigens
No.
FM
FW
1 2 3 4 5 6 7 8 9 10
1,8 1,2, 7,27 2, 3, 14, 15 2, 9, 12, 39 1, 24, 8, 18 3, 24, 8, 18 2, 40 2,9,5,15 1,8,15 1, 8
11 12 13 14 15 16 17 18 19 20
2, 2, 3, 1, 1, 2, 1, 2, 2, 2,
21 22 23 24 25 26 27 28 29 30
3, 9, 12, 27 1, 2, 8, 38 30, 5, 15 1,2, 8 1, 8 2, 35 2, 28, 8, I2 1, 2, 15 1, 2, 8, 13 1, 2, 8, 18
31 32 33
2, 11, 15, 35 2, 40 1,2,5,8
3, 12, 41 3, 13, 39 25, 18, 39 2, 8, 40 24, 8, 15 24, 8, 22 11, 15 24, 5, 22 29, 13 30, 18, 35
Antigens 1 2 3 4 5 6 7 8 9 10
3, 9, 8, 35 1,8, 40 1,3,8,35 25, 12, 18 23, 24, 7, 35 2, 24, 8, 15 1,3,7,8 2, 18, 21 2, 40 2, 11, 14, 27
11 12 13 14 15 16 17 18 19 20
1, 2, 15, 35 2, 3, 15, 35 2, 3, 7, 39 2, 12, 35 1,3,7,8 1,2,7,8 2, 11, 5, 15 1, 11, 8, 35 2,32, 5 , 2 2 1, 2, 8, 15
21 22 23 24 25
2, 2, 1, 1, 2,
3, 7, I3 25, 12, 15 26, 8, 12 8 24, 12, 39
450
N. Bengsch et al.: H L A in Juvenile Diabetics
Without family history No. Antigens
No.
KM
KM
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
2, 24, 16, 35 1, 24, 12 1, 8 1, 35 2, 9, 17, 37 2, 15 3, 10, 8 2, 3, 7, 15 1, 2, 8 1, 8 1, 2, 8, 15 2, 3, 16, 17 2, 3, 17, 35 9, 28, 12 2, 9 1, 9, 8, 39 2, 30, 18, 27 2, 25 29, 31, 12, 15 l, 2, 15, 39 2, 25, 40 1, 3, 8, 27 2, 8, 12 3, 25, 7, 18 2, 3, 15 2, 17
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68
Antigens
No.
3, 28, 12, 18 1, 9, 8, 12 1, 28, 8, 15 2, 29, 8, 35 1, 32, 8, 41 2, 9, 15, 39 2, 15 1, 9, 7, 8 2, 3, 35, 40 2, 7, 15 2, 11, 21, 35 2, 7, 27 2, 25, 18, 27 1, 2, 8, 40 2, 8, 12 11, 8, 35 2, 3, 7, 13 2, 12, 35
KW
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 2l 22 23 24 25 26
Antigens
No.
2, 3, 12, 40 3, 26, 12, 27 1, 2, 8, 13 3, 29, 5 1, 32, 8 3, 9, 35 1, 8 2, 9, 12 2, 3, 7 2, 3, 21, 22 2, 9, 15, 18 2, 5, 12 1, 26, 8, 35 1, 2, 8 28, 18 30, 13, 35 25, 28, 5, 35 1, 2, 8 1, 8, 27 9, 27, 40 l, 2, 7, 8 2, 24, 27, 39 2, 39, 40 28, 12 3, 9, 12, 15 24, 18, 27
KM
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
Antigens
No.
2, 3, 15, 18 1, 3, 8, 15 25, 14, 18 1, 2, 7, 8 1, 3, 7, 8 2, 3, 35 24, 30, 18, 2 7 2, 26, 18, 40 l, 11, 8, 35 1, 26, 8, 22 2, 9, 7, 40 2, 25, 8, 15 2, 25, 8, 18 3, 9, 7, 16 1, 2, 8, 21 2, 9, 7 1, 8 2, 3, 18, 39 3, 26, 7, 8 24, 13, 22 1, 24, 8 1, 3, 8, 18 2, 3, 8 2, 3, 15, 27
KW
Antigens 27 28 29 30
1,2,8,40 1, 29, 12 2, 3, 15, 35 1, 2, 8, 13
31 32 33 34 35 36 37 38 39 40
2, 8 2, 29, 8, 12 1, 9 , 8 1, 9, 8 , 2 7 2, 24, 8, 22 2, 3, 12, 18 1,2,8 1, 28, 8, 15 2, 5, 40 11, 25, 12, 35
41
11, 25, 18, 35
Appendix 2 F a m i l y data: Given only antigens of the A- and B-series. a, b: paternal haplotypes; c, d: maternal haplotypes. Haplotypes which could not be determined by H L A typing but were deduced are given in brackets. In family K 13, both parents could not be HLA-typed. The naming "paternal" and "maternal" is arbitrary. Sibling 4 in family F6 shows a recombination: 24, 5, 12
Diabetics Family
a
b
F 1 F 2 F 3 F 4 F 5 F 6 F 7 F 8 F 9 F10 Fll F12 F13 F14 F15
3, 2, (2, 11, 29, (2, (2, 23, 1, 2, 2, 1, 23, 11, 2,
K 1 K 2 K 3 K 4 K 5 K 6 K 7 K 8 K 9 K10 Kll K12 K13 K14 K15 K16
2, 8 29, 12 (30, 27) 2, 8 2, 15 2, 15 32, 41 2, 12 24, 12 2, 27 (3, 15) 2, 40 (2, 15) 24, 35 1, 8 1, 8
35 15 15) 5 41 22) 13) 7 37 12 35 8 8 35 38
c
24, 12 28, 12 (1, 8) 2, 15 3, 18 (24, 12) • (24, 12) 2, 40 2, 5 1, 8 ?, 40 28, 5 1, 13 1, 8 1, 8 ?, 2, (2, 2, ?, 3, 24, 1, 1, 28, (11, 1, (11, 2, 32, 3,
35 12 41) 15 7 7 18 8 8 12 ?) 17 35) 12 22 27
Non-affected siblings
d
1
2
1
2
ac ac ad ad bd ad bd ac ac ac ac ad ad bd ac
ad
ad
bc ad ac bd
bd
(2, 3, 11, 2, 2, 24, 1, 24, (11, 25, 30, 2, 3, 24, 1,
15) 14 35 15 13 8 8 35 15) 15 18 12 35 22 8
(2, 40) 2, 21 1, 8 24, 40 1, 8 11, 5 2, 15 2, 27 (2, 12) 1, 8 24, 27 1, 8 1, 13 1, 8 3, 15
ac ac ac ac ac ac ac ac ac ac ac ac ac ac ac
24, 31, 24, 2, 24, 1, 1, ?, 28, 24, 2, 24, (?, 2, 1, 2,
22 15 18 12 18 8 8 5 12 39 35 7 7) 39 8 8
29, 2, 3, 3, 1, 25, 3, 28, 24, 1, 2, 3, (1, 26, 3, 1,
ac ac ac ac ac ac ac ac ac ac ac ac ac ac ac ac
12 7 15 35 8 18 40 40 12 15 40 40 7) 27 7 12
3
4
bd ac ad bd
bc
b c/d
bd
ad
bd ad
ac bc ac bd ac bc ad ad bd ac bd ad ad bc bd ae
ad bd bc ac ac bd bc ad bd bc bd ac bd ac bc bd
ad
bd ad ac bd
ac
N. Bengsch et al.: HLA in Juvenile Diabetics References
1. Albert, E.D., Scholz, S., Bertrams, J., Ewald, R.W., Westphal, E., Ratschko, K.-W., Spielmann, W., Seidl, S.: Representative HL-A phenotype and haplotype frequencies of the German population. Z. Immunitaetsforsch. 148, 367-413 (1975) 2. Barbosa, J., King, R., Noreen, H., Yunis, E. J.: The histocompatibility system in juvenile, insulindependent diabetic multiplex kindreds. J. Clin. Invest 60, 989-998 (1977) 3. Christy, M., Nerup, J., Bottazzo, G. F., Doniach, D., Platz, P., Svejgaard, A., Ryder, L.P., Thomsen, M.: Association between HLA-B 8 and autoimmunity in juvenile diabetes mellitus. Lancet 1976 II, 142-143 4. Cudworth, A.G., White, G. B. B., Woodrow, J. C., Gamble, D. R., Lendrum, R., Bloom, A.: Aetiology of juvenile-onset diabetes. A prospective study. Lancet 1977 I, 385-388 5. Cudworth, A. G., Woodrow, J. C.: HL-A system and diabetes mellitus. Diabetes 24, 345-349 (1975) 6. Cudworth, A. G., Woodrow, J. C.: Evidence for HL-A-linked genes in "juvenile" diabetes mellitus. Br. Med. J. 1975 III, 133-135 7. Cudworth, A.G., Woodrow, J.C.: Genetic susceptibility in diabetes mellitus: analysis of the HLA association. Br. Med. J. 1976 II, 846-848 8. Ludvigsson, L., S~ifwenberg, J., Heding, L.G.: HLA-types, C-peptide and insulin antibodies in juvenile diabetes. Diabetologia 13, 13-17 (1977) 9. Mirouze, J., Seignalet, J., Selam, J.-L., Lapinski, H., Jaffiol, C.: Antig6nes HLA chez les diab6tiques insulino-d6pendants. Nouv. Presse Med. 5, 1628-1630 (1976) 10. Nelson, P. G., Pyke, D. A., Cudworth, A. G., Woodrow, J. C., and Batchelor, J.R.: Histocompatibility antigens in diabetic identical twins. Lancet 1975 II, 193-194
451 11. Nerup, J., Ortved Andersen, O., Christy, M., Platz, P., Ryder, L., Thomsen, M., Svejgaard, A.: HLA, autoimmunity, virus and the pathogenesis of juvenile diabetes mellitus. Acta Endoerinol. [Suppl.] (Kbh.) 205, 167-175 (1976) 12. Rotter, J. I., Rimoin, D. L.: Heterogeneity in diabetes mellitus - update, 1978. Evidence for further genetic heterogeneity within juvenile-onset insuline-dependent diabetes mellitus. Diabetes 27, 599-608 (1978) 13. Rubinstein, O., Suciu-Foca, N., Nicholson, J. F., Fontino, M., Molinatio, A., Harisiadis, K., Hardy, M.A., Reemtsma, K., Allen, F.H. jr.: The HLA system in the families of patients with juvenile diabetes mellitus. J. Exp. Med. 143, 1277-1282 (1976) 14. Schernthaner, G., Ludwig, H., Mayr, W. R.: Juvenile diabetes mellitus: HLA-antigen frequencies dependent on the age of onset of the disease. J. Immunogenet. 3, 117-121 (1976) 15. Singal, D.P., Blajchmann, M. A.: Histocompatibility (IlL-A) antigens, lymphocytotoxic antibodies and tissue antibodies in patients with diabetes mellitus. Diabetes 22, 429-432 (1973) 16. Svejgaard, A., Hauge, M., Jersild, C., Platz, P., Ryder, L. P., Staub Nielsen, L., Thomsen, M.: The HLA system. An introductory survey. Monogr. Hum. Genet. 7, 25 (1975) Received: June 15, 1978, and in revised form: August 9, 1978
Prof. Dr. med. J. K6bberling Department of Medicine University of G6ttingen Robert-Koch-StraBe 40 D-3400 G6ttingen Federal Republic of Germany
