J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Review Article J. med. Genet. (I967). 4, 44. Human Polymorphism J. PRICE* F-rom the Nuffield Unit of Medical Genetics, Department of Medicine, University ofLLiverpool Ford (I940) has defined polymorphism as the If this were the case, any attempt to relate this occurrence together in the same habitat of two or polymorphism to differences in function of the more discontinuous forms or phases of a species various forms of acid phosphatase would plainly not in such proportions that the rarest of them cannot succeed. be maintained merely by recurrent mutation. So few selective mechanisms (similar to that In his Genetic Polymorphism (I965) he states relating sickle cell trait to falciparum malaria) (p. I4) that 'A unifactorial character must be poly- have been discovered, that the present position is morphic if found even in I% of a considerable that most polymorphisms described are, so to population, amounting perhaps to 5oo individuals speak, 'in search of a disease', which they seem on or more, when random genetic drift may reasonably present evidence unlikely to find. The selective be excluded as unimportant'. This figure of I% factors may, of course, not be concerned with has been used in this paper as a rough dividing disease, but be connected with fertility, ability to line between what needs to be mentioned and what gather food, and other forms of fitness. does not. This fruitless searching for selective mechanismscopyright. Ford's definition excludes continuous variation, which will give a clear-cut explanation of known as exemplified by human height and skin colour, polymorphisms seems to have deterred a great and it excludes rare disadvantageous recessive con- many investigators from making any attempt to ditions, such as albinism and haemophilia. determine how a polymorphic situation has come Two types of polymorphism are included in the about. Many publications give lists of data for the balanced and the transient. A definition: poly- many human polymorphic systems from every http://jmg.bmj.com/ morphism may be balanced, that is, with the ratio corner of the globe, without referring to possible between two (or more) different characters remain- selective factors that may be maintaining them. ing constant throughout many generations, through So often the data collected are wholly genetic- the operation of one of several mechanisms of which thus we may be given the haptoglobin and trans- heterozygous advantage is the best known (see the ferrin phenotypes, but we are seldom given the section on blood groups). Altematively, poly- serum iron or the haemoglobin level. morphisms may be transient; if two discontinuous The situation of polymorphisms in search of forms (A) and (B) exist, then in time one or other of disease has a corollary: disease in which neither on September 27, 2021 by guest. Protected the transitions (A) -* (A + B) -- (A) or (A) -* (A + B) the number of genes involved nor their precise -÷ (B) will take place. If a certain form has a io% effects are known, but in which various facets incidence in a population can we be sure whether suggest that a polymorphic situation may exist. it is increasing, decreasing, or stable in that popu- Two such diseases may be mentioned here- lation? The answer in many cases seems to be no, diabetes and schizophrenia: both are common; because of lack of knowledge of frequencies in the both are inherited conditions; both have interfered past and of the selective forces involved. markedly (and still do, to an extent which is greater Genes always have multiple effects. It therefore for schizophrenia) with biological fitness. Further- follows that we may discover a polymorphism re- more, in each condition mechanisms may exist to lating to one effect, for example red cell acid phos- maintain the frequency of the responsible gene or phatase, in which it is another property of the genes: in diabetes the increase in prediabetic genes responsible for acid phosphatase production matemal fertility; in schizophrenia, increased which is bringing about the polymorphic situation. resistance to stress (this does not seem to be estab- ished, but it remains a possibility). *Bates Research Fellow, Mental Health Research Fund. Particular attention has been paid in this review 44 J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 45 to those polymorphisms in which selective factors would rise steadily. One disadvantage could occur are known to exist, or in which hypotheses have in compatible heterozygote matings (e.g. AO x AO been constructed to explain a polymorphic situa- or AB x AB) if heterozygous offspring survived tion. Data on some of the newly-discovered poly- better than homozygous offspring. morphisms, not yet widely known, have also been Both Chung and Morton (I96I) and Matsunaga included. (i962b) found an excess of AB children in AB x AB mating. The same might be true of AO x AO A: BLOOD GROUP POLYMORPHISMS matings, but these are difficult to study because the Three aspects of these complex polymorphisms AO phenotype cannot be established by simple are discussed here. testing as can the AB phenotype. Chung and Morton (I96I) stress that 'the genetic evidence I: Selective Forces Acting on ABO and Rh strongly supports the assumption of Brues (1954) Blood Groups that the ABO polymorphism is maintained because The subject of blood groups and natural selection heterozygous advantage in compatible matings is discussed (principally with reference to ABO and more than compensates for selection against hetero- Rh) by Sheppard (1959), Reed (I96I), Clarke (I964), zygotes in incompatible mating'. They further hold Chung and Morton (I96I) among many others. that 'the overall effect of maternal/foetal incompati- Among the selective factors the following call for bility is virtually the same for 0, A, and B mothers, consideration. despite the fact that clinical haemolytic disease is almost restricted to children of 0 mothers'. This Prezygotic Selection. Three possible mechan- should, perhaps, bring home the importance of not isms exist: the first two mechanisms (which cannot overestimating clinical haemolytic disease in con- be distinguished by testing) are confined to the sidering selection and ABO blood groups. heterozygous male. Reed, Gershowitz, Soni, and Napier (I964) divide Meiotic drive-unequal numbers of sperms carry- the reproductive indicator into several independent cate- ing each type of allele. gories: the number of pregnancies, the proportion of copyright. Sperm competition-unequal opportunity for each pregnancies terminating in abortion or in stillbirth; type of sperm in the act of fertilization. and the proportion of liveborn children dying non- Thus with an AO father meiotic drive would accidentally under the age of 5. These data were mean more A sperms than 0 sperms (or vice versa) obtained from several hundred couples where the wife i.e. a was over 40, so that reproduction had virtually been instead of equal numbers, quantitative differ- completed. Six blood groups were studied in these ence. Sperm competition would mean some couples (ABO, Rh, MN, Kell, Duffy, P) as was the http://jmg.bmj.com/ intrinsic property favouring A sperms at the secretor status. expense of 0 sperms (or vice versa), i.e. a qualita- An effect found to be significant at the o-ooi level tive difference. was that K females w-re found to have m re pregnancies; Sperm incompatability-meaning, for example, at o oos level Group N females were more likely to be that a sperm carrying an A gene coming in contact sterile (taken from amended data: Reed, I965). Further with cervical secretions containing anti-A would, studies of this type will plainly be required, both to thereby, be eliminated. substantiate or refute associations already found, and to Matsunaga (i962b) provides evidence to support extend the work to other blood groups. The task of on September 27, 2021 by guest. Protected the determining the selective factors operating in the blood hypothesis that prezygotic selection operates group polymorphisms is formidable indeed, particu- throughout so as to favour group 0 sperms: Morton larly as interactions such as those between ABO and (I965) does not consider the case proven. Rh may occur. Zygotic Selection. By this is meant the elimina- The Rh System. Newcombe (I963) and (I965) tion of zygotes whose blood group is not that of the (amended data) has studied the ABO and Rh mother, by spontaneous abortion or by clinical blood groups ofmothers experiencing foetal death in haemolytic disease. That this occurs has been New York, and has found that, with the exception of established by Chung and Morton (I96I), and AB Rh- mothers (who fare worst, being most likely Matsunaga (i962b). The loss is always of hetero- to become Rh sensitized), the risk of foetal death zygotes. to mothers over 30 rises as the number of anti- genically active ABO and Rh alleles unrepresented ABO Heterozygotes. The tendency for group in the mother's genotype rises (this number is o 0 sperms to be favoured must be balanced by dis- for AB Rh +; I for A Rh + and B Rh +; 2 for advantages, otherwise blood group 0 frequency ARh -,BRh -,andORh +;and3forORh -). J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 46 J. Price In connexion with Rh incompatibility, a theoretical Livingstone (I960) and Vogel, Pettenkofer, and point that does not seem to be widely understood may Helmbold (I960) discuss this subject. be mentioned. If three genotypes appearing with The effect of differences in incidence of duodenal frequencies p2, 2 pq, and q2 (corresponding to genes p ulcer, gastric carcinoma, and pernicious anaemia and q) have relative selective advantages of a, b, and c, on the ABO polymorphism is infinitesimal compared respectively, then for equilibrium: P ap2 bpq with prenatal and early postnatal influences in main- q bpq +cq2 taining the polymorphism, because these diseases p _b-c whence: affect biological fitness hardly at all.