Immunoglobulin (Gm) Allotypes in a Sample of Canadian Ashkenazic Jews L. LEIGH FIELD,' J. ALEXANDER LOWDEN ' AND AJIT K. RAY ' Department ofdnthropology, UriiLNersityof Toronto. Toronto, Ontario, Canada, M5S 1AI and Research Institute, The Hospital f o r s i c k Children, Toronto, Ontario, Canada, M5G 1x8 i K EY WORDS Gm variation . Haplotypes a Ashkenazi Jews - Genetic ABSTRACT Gm typing on t h e serum specimens of 507 Ashkenazic Jews (predominantly of Polish-Russian ancestry! from Toronto, Canada has established andGm 1,17;5, and the absence of the presence of h a p l o t y p e ~ G m ~Gm : ~ ,'z21, Gm haplotypes Gm1:13J5.16, and Gm1;5.6,24which have been found in other Jewish peoples. It is suggested t h a t Ashkenazic populations have lower frequencies of haplotype Gm 1 ~ 1 7 : 5than non-European Jewish populations, and t h a t some eastern European Jewish populations have acquired the Gm 1;13,15,16 h a PlotYPe through gene flow from Central Asia. Thus Jewish populations show differences in the Gm system; many of t h e differences may be in the direction of similarities to neighbouring non-Jewish populations. It would be interesting to know t h e degree of genetic differentiation among modern Jewish populations, and t h e degree of similarity of each population to local non-Jewish peoples inhabiting t h e same geographic region. Such information, accumulated for many genetic markers, would provide the beginnings for a study of drift, selection and gene flow within and across a cultural boundary. The Gm genetic system is particularly appropriate for describing interpopulational differences, since the kind as well as the frequency of Gm haplotypes often varies from population to population (for review, see Johnson e t al., "77; Schanfield and Fudenberg, '75). Table 1 presents the Gm haplotypes which commonly occur in populations relevant to this paper; i t also gives the abbreviated haplotype notation t h a t will be used throughout the paper. Steinberg ('73) reported t h e presence 0fGrn','~3~ and Gm1,13J5J6 in a sample of American Ashkenazic Jews. I t was suggested t h a t these haplotypes indicate admixture with Black Africans and San (Bushmen), respectively. The results of Gm typing in another sample of Ashkenazic Jews a r e reported here, followed by a discussion of how the data relates to the more extensive problem mentioned above. AM. J. PHYS. ANTHROP. (1978)48: 159-164. MATERIALS AND METHODS Over 6,000 serum specimens of self-identified Ashkenazic Jews were available from a Tay-Sachs carrier screening programme conducted in Toronto, Canada by J. A. Lowden. The 507 individuals randomly selected for Gm testing were primarily of the same generation and included only two pairs of known relatives. Table 2 shows the origins of the parents of t h e test individuals. Eighty percent of the parents originated from Poland and Russia, and many of the remainder came from other eastern European countries. About one-half (263/497 = 53%)of the test individuals were of homogeneous parentage - t h a t is, both parents originated from the same country. Eighty-seven percent (2291263) of these had parents from Poland or Russia; only 341263 (13%)had parents from some other country. Of the 234 individuals of heterogeneous parentage, 2171234 (93%)had one parent from Poland and/or Russia, while 17/234 (7%)had neither parent from Poland or Russia. Homogeneous parentage occurred in 64% of those born before 1940, but in only 46%of those born during or after 1940 (x' = 12.89, df = 1, P < 0.001). 159 160 L. LEIGH FIELD, J. ALEXANDER LOWDEN, AND AJIT K. RAY TABLE 1 Gm haplotypes common in certain population groups Gm haplotype Extended notation Abbreviated notation Gm3,5.10,11,13,14,26,27 Gm 1,17.21.26,27 Gm 1,2.17;21,26.27 Gm 1,17,5,10,11,19,14,26,27 Gm 1,17,6,6,10.11,14.26.27 Gm 1,220.127.116.11,24,26 Gm1,17;10,11,13.15,16,27 ~~1,3,18.104.22.168.14.26,27 TABLE 2 Country of origin of CoUntTv No Darents Poland Russia ' Austria Roumania Hungary Germany Czechoslovakia England North Africa Other Europe 442 355 45 43 35 20 15 13 10 16 994 parents %, of total 44.5 35.7 4.6 4.3 3.5 2.0 1.5 1.3 1.0 1.6 100.0 I The high percentage of Polish ancestors IS characteristic of the Toronto Jewish population and 1s not seen in other Canadian cities, where the percentage Poliah varies from 24-26 and the percentage Russian varies from 40-57 (Lowden, '75). The exact composition wa3: 312 "Russia," 2 "White Russia," 29 "Lithuania," 3 "Latvia," 9 "Ukraine." 497:507 responded to the question abuut origins. TABLE 3 Reagents usrd for testing Grn and A m antigens Antigen Numeric Alphameric Gl m G3m A2m 1 a 2 X 3 f 17 z 5 6 13 b' 14 15 16 21 b' C 1 b3 S t g 1 1 2 2 Population p r o w Caucasian Caucasian, Oriental, Bushmen Caucasian, Oriental Negro, Bushmen Negro Negro Oriental, Bushmen Oriental Ayylutlnator Mor Mor TY 1 How Haw R-15 Anti-D coat Bra Bar Bar Dul Ell Bar Pay Cur Ing Ska Goe Gai Ros Cli Cam Kre Far Tay Her For Dur Dur Dur Puh Puh Bar L The nomenclature is based on recommendations of a W.H.O. meeting held in J u l y 1974 (W.H.O. Committee on Human Immunoglobulin Allotypes, '761. The serum specimens were tested, using the standard agglutination-inhibition technique described by Steinberg ('621, for Glm(1,2,3) and G3m(5,6,13,14,21), with the reagents listed in table 3. Selected samples, indicated in table 4, were also tested for Glm(171, G3m(15,16), and A2m(1,2) to search for certain rare haplotypes. Appropriate controls were employed at all times. Twenty-five of the unknown sera exhibited agglutinating activity; these were heat inactivated a t 63°C and re-typed. Haplotype frequencies were estimated by the maximum likelihood programme MAXIM (Kurczynski and Steinberg, '67). RESULTS The Gm phenotypes of the 507 individuals are given in table 4. In the absence of family data, decisions about the probable genotypes of the observed phenotypes must be based on knowledge of the commonly occurring haplotypes in relevant populations. The first five phenotypes listed provide evidence for the presence of the haplotypes Gm3i5,Gm and Gm1~2;21, which are common in Europe. The last two phenotypes, Gm(1;5,13,14,21) and Gm(1,2;5,13,14,21), suggest the presence of the haplotype Gm1.17i5,common in Africa but rare in Europe. Phenotype Gm(1,3;5,13,14) without further testing could be explained by many genotypes, the most probable being heterozygous combination of Gm3i5with Gm l-17i5, Gm1;13915316,or Gm1.3;5. The latter two haplotypes are common in Asia. Testing for Glm(17) and G3m(15,16) helps to distinguish between these possibilities, since genotype Gm3.5/Gm1,'7;5would be (17, - 15, - 1 6 ) , Gm3;5/Gm1:13J5J6would be (17,15,16), and Gm3;51Gm1~3;5 would be ( - 17,- 15,-16). When the 12 samples with phenotype Gm (1,3;5,13,14) were tested for Glm(17) and 161 GM ALLOTYPES IN CANADIAN ASHKENAZIC JEWS TABLE 4 Gm phenotypes of the total sample and the subgroups Phenotype Total sample ' 3;5,13,14 1,3;5,13,14,21 1,2,3;5,13,14,21 1;21 1,2;21 1,3,17;5,13,14 1;5,13,14,21 1,2;5,13,14,21 No. Y, 291 134 41 20 7 12 57.4 26.4 8.1 3.9 1.4 2.4 0.2 0.2 100.0 1 1 507 Polish subgroup No. 79 33 11 5 % 2 59.4 24.8 8.3 3.8 1.5 3 0 0 133 2.2 0.0 0.0 100.0 Russian subgroup xo. 52 29 8 3 1 2 0 1 96 54.2 30.2 8.3 3.1 1.0 2.1 0.0 1.0 100.0 ' All samples were tested for Glm(1.2,31 and G3rni5,6,13,14,21i. 'These samples were also tested for Glm(17i and C3m(15,161. In addition, they were tested for A2m(1,2) TABLE 5 Estimated haplotype frequenczes of Total sample Haplotype Gm3,5 Gm1:PI Gm1.23 Gm1,17;5 the total sample and the subgroups Polish subgroup Russian subgroup Frequency * S.E. Frequency t S.E. Fresuencv 0.758 0.179 0.049 0.014 0.013 0.012 0.007 0.004 0.771 0.168 0.050 0.011 0.026 0.023 0.013 0.006 0.745 0.187 0.053 0.015 1 S.E. 0.031 0.028 0.016 0.009 The fit to Hardy~Weinbergexpectations using these estmated frequencies is satlsfactorg for each of the three samplesTotal sample: x' = 2.91. df = 3. 0.5 > P > 0.3. Polish subgroup: x1 = 0.58, df :1. 0.5 > P 1 0 3. Russian subgroup: x' - 0.54, df = 1. 0 5 > P > 0.3. G3m(15,16), they were all (17,-15, -161, thus providing corroborative evidence for the presence of Gm l.17i5. Finally, phenotypes Gm (1,3;5,13,14,2 1) and Gm (1,2,3;5,13,14,21) could possess t h e haplotype Gm1*3;5, but in this case no differentiating tests are available: Glm(17) would always be positive in these phenotypes because it is carried by the Gm';21 andGm',2;21haplotypes. Thus, in the 507 individuals tested, there was evidence only for t h e haplotypes Gm33, Gm Gm1,2;21,and Gm'.'7;5. The estimated frequencies of these haplotypes are presented in table 5. The 12 individuals with t h e phenotype Gm (1,3,17;5,13,14), probable genotype Gm 3:5/ Gmi.17;5,were also tested for A2m(1,2). The A2m markers, only recently discovered, a r e closely linked to Gm markers. A2m(2) is rare in "European" haplotypes; in Africa, Gm Am2 is much more common than Gm1J7;5Am1 (Schanfield and Fudenberg, '75). Only 2/12 (17%) of the individuals tested were A2m(2) positive, indicating t h a t Gm 1~17;5 Am' predominates in our population. DISCUSSION In order to search for differences in Gm fre- quencies between Jewish populations, two large subgroups were isolated from the Toronto sample: (1) people with Polish ancestry only (2) people with Russian ancestry only. Table 4 shows t h e phenotype distributions of these two groups, and table 5 gives the estimated haplotype frequencies. The difference between the phenotype distributions is not significant (xz = 0.87, df = 3, 0.9 > P > 0.8). I t is not surprising t h a t Jewish people with Polish origins and Russian origins have similar Gm frequencies: many of their ancestors probably came from the region surrounding what is now the Polish-Russian border, the northern part of a n area called the Pale of Settlement in the Nineteenth Century when i t was contained within the Russian Empire (Fraikor, '77). In other words, these ancestors came from a relatively restricted geographical area, which is not reflected in t h e political terms chosen by their descendents t o describe ancestral origins. The one other detailed study of Gm variation in Ashkenazic Jews is t h a t of Steinberg ('731, who typed 248 American Jews for the same array of Gm factors tested in the present analysis. The population was not identified as 162 L. LEIGH FIELD, J. ALEXANDER LOWDEN, A N D AJIT K. RAY TABLE 6 Cumparison of the phenotype distributions of Polish Jewish and non-Jewish samples Phenotype 1 IT 111 Polish Jewish (this study) Polish non-Jewish (Podliachouk e t a1 , '63) Polish non-Jewish (Schlesinger e t al., ' i l l No. 3;5 1.3;5 1,2,3;5 1 1.2 %> 79 36 I1 59.4 27.0 8.3 5 3.8 2 133 1.5 100.0 No. %, No. x. 193 77 24 6 5 305 63.3 182 66 30 14 8 300 60.7 22.0 10.0 4.7 2.7 100.0 25.2 7.9 2.0 1.6 100.0 Tests of significance I YS. 11. ,y' = I 56, df = 4, 0.9 > P > 0.8. I VB. 111: x' = 2.03. df - 4, 0.8 > I' > 0.7. ' Glm!3l or G3m161 or both were detected, depending on the reagents used in the study to origin, but was (we may presume) predomi- population comparisons should be considered nantly from eastern Europe. Chi-square test- tentative. However, several of the non-Euroing on the phenotype distributions showed no pean groups (e.g., populations No. 37 and No. significant difference between the American 43 from Morocco and Iran in Ropartz e t al., and the Toronto samples (x' = 6.71, df = 5, '65; Kurdish Jews from northern Iraq in 0.3 > P > 0.2).3 It may be more appropriate to Steinberg et al., '70; Yemenite Jews and Kurperform a x 2 test using the estimated hap- dish Jews from Iran in Godber e t al., '73) aplotype frequencies (Elandt-Johnson, '71); this pear to possess higher frequencies of "Grn135" is done by generating a set of expected pheno- haplotypes than Ashkenazic Jews. An exceptypic proportions from the pooled haplotype tion is the Karaite isolate from Iraq, which frequencies, and then comparing observed dis- Goldschmidt et al. ('76) showed to have only But evidence from blood tributions to those predicted from the ex- Grn395 and Grn1%21. pected proportions. Again, the x 2 results were group and isoenzyme systems suggests that not significant ( x 2 = 7.66, df = 7, 0.5 > P > this community has experienced considerable 0.3). The only interesting difference between genetic drift due to small population size. Furthese two Ashkenazic populations was the ther studies are needed to determine the geopresence in the American group of the hap- graphic variation of Grn1.17.5frequency in nonlotype Grn';'33'5~16(carried by 1.6%of the peo- European and European Jews. In the present ple tested; estimated frequency = 0.008) and study, the parental origins of the 14 individits absence in the Toronto group. The hap- uals carrying Grn 1 ~ 1 7 , 5were distributed in the lotype seems to exist in the Polish-Jewish same manner as the origins of the entire gene pool, since it has been recorded in a Toronto group. Podliachouk and Eyquem ('63) tested 305 Polish Jewish family (mentioned in Steinberg et al., '70). However, it may have a frequency Poles for Glm(1,2) and G3m(5); Schlesinger lower than 0.008 in Poland and adjoining and Luczkiewicz-Mulczykowa ('71) tested 300 Russia, and a higher frequency in some area of Poles for G l m (1,2,3). Since it is unlikely that eastern Europe t h a t is better represented in either sample contained significant numbers the American sample. Amongst non-Jews in of Jews, they may be considered non-Jewish eastern Europe, Czechoslovakians have a for purposes of comparison. It is not possible to G m 1 ; 1 3 J 5 . 1 6 frequency of about 0.004 (Schandetermine the presence or absence of the field e t al., '75b), while further south, Hun- Gm1J7;5haplotype from their testing, but in garians have a frequency of about 0.014 terms of the phenotypes observed, there are no (Schanfield et al., '75a). Two samples of significant differences between the Polish disRoumanians (quoted in Johnson et al., '77) tributions and the Toronto Polish-Jewish disshow frequencies of 0.003 (north Roumania) The rare haplotypes Gm L17;5 and Gm 1;13,15,16were considered and 0.013 (Bucharest). as one for purposes of xi testing, sinceGm1;13,15.16 did not exist in Although some Gm data has been published the Toronto sample. "Gm1;5" may include Grn1;5,6 and Gm1;5,6,24 as well as on non-European Jews, it is less extensive, in All three haplotypes were detected by Steinberg et al. ('701 sample size or in number of factors tested, Gm1.17;5 in Iraqi Kurdish Jews, but the other studies of non-EuropeanJews did than the Ashkenazic material and therefore not test for G3m (6,241, which would have permitted differentiation. GM ALLOTYPES IN CANADIAN ASHKENAZIC JEWS tribution (table 6 ) . More detailed studies of Gm variation in Poland, and in Russia, have not yet been conducted. I t is suggested t h a t Gm differences exist between geographically separated Jewish populations and t h a t Jews may resemble their nonJewish neighbours. However, this does not imply how the differences and similarities have arisen. Admixture may be postulated, but i t should be supported by the history of population movements and contacts. For example, we consider i t more probable t h a t the haplotype Gm1:13.153detected by Steinberg ('73) in Ashkenazic Jews is the result of gene flow from peoples of Central Asia who invaded Europe around 500-1500 A.D., rather t h a n gene flow from S a n (Bushmen) of southern Africa (in the past, people related to the San inhabited eastern Africa). Evidence from non-Jewish Hungarians (Schanfield e t al., '75a) suggests t h a t these westward moving Asians had higher frequencies of Gm 1 ; 1 3 ~ 1 5 ~ 1 6 t h a n Gm1x3i5;this may explain the failure to find the latter haplotype in Jewish populations t h a t possess t h e former. The Gm 'J~;' haplotype exists in non-polymorphic frequencies (-- 0.003) in several western a n d e a s t e r n European countries (Schanfield et al., '75b). Non-European Jewish populations appear to have higher frequencies - for example, 0.033 in Kurdish Jews from northern Iraq (Steinberg e t al., '70). Middle Easterners generally may have higher frequencies - for example, 0.037 in Lebanese (Lefranc e t al., '76). Thus the frequencies observed in Ashkenazic Jews - 0.014 (this study) and 0.010 (Steinberg, '73) - a r e intermediate. Again, i t is possible to postulate admixture. The haplotype was first acquired by J e w s through gene flow from Black Africans, then during the Diaspora, i t was reduced in frequency by cryptic gene flow from Europeans. However, perhaps admixture is being given too prominent a role in explaining differences and similarities in Gm frequencies. The presence of Gm1J7,5in Jews does not require a n African origin, since this haplotype could arise through recombination between Gm3;5and Gm1iZ1(Schanfield et al., '75a). The results of A2m typing in this study may be interpreted as supporting a nonAfrican origin: t h e Gm1J7:5Am' haplotype predominates in t h e Jewish sample whereas Gm1J7;5Am2is the more common haplotype in Africa. Furthermore, any decrease in t h e frequency of Gm1J7;5during t h e time Jews have 163 inhabited Europe could be the result of evolutionary convergence. The frequencies of genetic variants in the structure of immunoglobulin molecules may be strongly affected by environmental factors, even if the operation of selection is not at present demonstrable. ACKNOWLEDGMENTS We would like to thank Doctor A. G. Steinberg for generously providing many of the anti-sera used, training in Gm typing, and access to the MAXIM programme. Thanks also go to Doctor H. Gershowitz for donation of anti-D for Glm(3) and to Doctor M. S. Schanfield for kindly typing G3m(15,16) and A2m (1,2).We have benefited from the comments of Doctors H. Gershowitz, T. E. Reed, M. S. Schanfield, A. G. Steinberg, and E. J. E. Szathmary. This research was supported in part by a n Ontario Graduate Scholarship and by Postgraduate Scholarship 1560 from the National Research Council of Canada. The Office of Research Administration, University of Toronto, also gave financial assistance. LITERATURE CITED Elandt-Johnson, R. 1971 Probability Models and Statistical Methods in Genetics. Wiley & Son, New York. Fraikor, A. L. 1977 An anthropological analysis of TaySachs disease: genetic drift among the Ashkenazic Jews. SOC.Biol., 24: 117.134. Godber, M. J., A. C. Kopec, A. E. Mourant, D. Tills and E. E. Lehmann 1973 The hereditary blood factors of the Yemenite and Kurdish Jews. Phil. Trans. R. Soc. Lond. B, 266: 169-184. Goldschmidt, E., K. Fried, A. G. Steinberg and T. Cohen 1976 The Karaite community of Iraq in Israel: a genetic study. Am. J. Human Genet., 28: 243.252. Johnson, W. E., P. H. Kohn and A. G. Steinberg 1977 Population genetics of the human allotypes Gm, Inv, and A2m, a n analytical review. Clinical Immunol. and Immunopath., 7: 97-113. Kurczynski, T. W., and A. G. Steinberg 1967 A general program for maximum likelihood estimation of gene frequencies. Am. J. Human Genet., 19: 178-179. Lefranc, G., L. Rivat, C. Rivat, J. Loiselet and C. Ropartz 1976 Evidence for "deleted' or "silent" gene homozygous a t the locus coding for the constant regionof the y 3 chain. Am. J. Human Genet., 28: 51-61. Lowden, J. A. 1975 Tay-Sachs screening and prevention: the Canadian experience. Paper presented a t the First Internat. Conf. on Tay-Sachs Disease, Palm Springs, California. Podliachouk, L., and A. Eyquem 1963 Les facteurs seriques Gm (a), Gm (b), Gm (XI, e t Gm-like dans la race blanche. Comptes Rendus Societe de Biologie, 157: 732-736. Ropartz, C., P:Y. Rousseau and L. Rivat 1965 Hypotheses sur la genetique formelle du systeme Gm chez les Caucasiens. Humangenetik, 1: 483-496. Schanfield, M. S., and H. H. Fudenberg 1975 The a n t h r o p logical usefulness of the IgA allotypic markers. In: Biosocial Interrelationships in Population Adaptation. Mouton, Amsterdam, pp. 105-114. 164 L. LEIGH FIELD, J. ALEXANDER LOWDEN, AND AJIT K. RAY Schanfield, M. S., J. Gergely and H. H. Fudenberg 1975a Immunoglobulin allotypes of European populations I. Gm and Km (Inv) allotypic markers in Hungarians. Hum. Hered., 25: 370-377. Schanfield, M. S.. P. Herzog andH. H. Fudenberg 1975b Immunoglobulin allotypes of European populations 11. Gm, Am, and Km (Inv)allotypic markers in Czechoslovakians. Hum. Hered., 25: 382-392. Schlesmger, D., and A. Luczkiewicz-Mulczykowa 1971 The Gm (l),Gm (21, and Gm (4) factors in the Polish population. Archs. Immunol. Ther. Exp., 19; 703-708. Steinberg, A. G. 1962 Progress in the study of genetically determined human gamma globulin types !the Gm and Inv types). In: Progress in Medical Genetics. Vol. 11. A. G. Steinberg and A. G. Bearn, eds. Grune and Stratton, New York, pp. 1-33. - 1973 The Gm and Inv allotypes of some Ashkenazic Jews living in northern U.S.A. Am. J. Phys. Anthrop., 39: 409-412. Steinberg, A. G., C. Levene, E. Goldschmidt and T. Cohen 1970 The Gm and Inv allotypes in kindreds of Kurdish Jews. Am. J. Human Genet., 22: 652-661. W.H.O. Committee on Human Immunoglobulin Allotypes 1976 Review of the notation for the allotypic and related markers of human immunoglobulins. The J. Immunol., I 1 7: 1056-1058.