Bull. Org. mond. Santé
Bull. Wld Hlth Org. } 1966’ 34’ 379'393
Population Genetics of Haemoglobin Variants,
medium falciparum parasite counts than do children
without this trait; (2) the finding of a low incidence
of sickling...
FIG. 1
the coastal regions for grain, as is proved by the very
localized areas along the south-west coast in which
F| G.2
Gone frequency (Z)
o~: .o. mSSZ3.'¢Ta'r33L‘. ’§'...
values of gene frequencies today are an obvious con-
sequence. Carcassi (1962) has shown that the same
situation pre...
tion-screening for 13-thalassaemia is performed by red-
cell fragility and blood-filrn studies alone (Table 4).
The ...
thesis that the Sardinians, unlike their neighbours,
must hav...
reasonable to suppose that it might, for instance, have
been the susceptibility of the enzyme-deficient red
cells of...
392 M.
The authors wish to express their gratitude to
Professor G. Montalenti,...
Bernini, L. , Latte, 13., Siniscalco, M. , Piomelli, S. ,
of 15

Population Genetics of Haemoglobin Variants, Thalassaemia and ...

Published on: Mar 4, 2016

Transcripts - Population Genetics of Haemoglobin Variants, Thalassaemia and ...

  • 1. Bull. Org. mond. Santé Bull. Wld Hlth Org. } 1966’ 34’ 379'393 Population Genetics of Haemoglobin Variants, Thalassaemia and Glucose—6-Phosphate Dehydrogenase Deficiency, with Particular Reference to the Malaria Hypothesis M. SINISCALCO} L. BERNINL1 G. FILIPPI, ’ B. LATTER P. MEERA KHAN} S. PIOMELLI‘ & M. RATTAZZI1 The authors report data on the genetic distribution of thalassaemia and of glucose- 6-phosphate dehydrogenase deficiency in the populations of certain Sardinian villages, many of wlzich are not only of great antiquity but have maintained isolation for very long periods and therefore possess the following three requirements for suitability for investiga- tion of the possible interrelationships among malaria, thalassaemia and G-6-PD deficiency: a reasonable degree of ethnic homogeneity, availability of reliable demographic data, and availability of malaria—free populations of adequate size and of ethnic background and genetic isolation similar to those of the malarial populations. Investigations including more than 6000 observations in 52 villages demonstrated a positive correlation between the incidences of t/ ralassaemia and G-6-PD deficiency. It is suggested that the genotype that carries thalassaemia and/ or the enzyme deficiency may have a high adaptive value in a malarial emit-onment. It is concluded that there is a need further to investigate human genetic structure and the biological fitness of the principal genotype combinations in both existing environments and those that will result from continued cultural evolution. INTRODUCTION The present status of knowledge of haemoglobin variants, thalassaemia and glucose-6-phosphate dehydrogenase (G-6-PD) deficiency has been summa- rized on several occasions in recent years. Excellent reviews of all aspects of these questions are now available (Allison, 1965; Baglioni, 1962; Fessas, l965; Ingram, 1963; Itano, 1965; Motulsky, 1965; Ruck- nagel, 1964; Silvestroni & Bianco, I963). Consideration of the recent exhaustive accounts of Silvestroni & Bianco (1963) on the world distribu- tion of haemoglobin variants and of thalassaemia, that of Motulsky (1965) on G-6-PD deficiency, to- gether with the elegant monograph of Rucknagel & Neel (1961) on the dynamics, at a population level, ‘Department of Human Genetics, University of Leiden, Leiden, Netherlands. 181’ Department of Genetics, University of Rome, Rome, 1 y. ' Town Hospital, Nuoro, Italy. ‘ Department of Genetics, University of Rome. Present address: Department of Pediatrics, New York University- Bellevue Medical Center, New York, N. Y., USA. 1728 of the genes controlling these conditions, and the discussions of the same subject by Livingstone (1964) and by Allison (1965) may serve as ideal introductions to the subject of the present report—that is, the hypothesis that malaria may have been the common ecological factor that was responsible for the selec- tion of these three groups of inherited abnormalities of the red cells. The idea that all individuals might not be equally liable to malarial infection was first proposed by Haldane (1949) to explain the preponderance of thalassaemia in the Mediterranean basin. A few years later, Allison (1954) reported the interesting finding that persons who carry the sickle-cell trait are indeed more resistant to subtertian malaria than those who do not. Although this claim has been disputed in some instances, it is now supported by a most impressive body of evidence, including data of the three following kinds (Allison, 1965): (1) direct demonstration, in areas where malignant malaria is still endemic, that young children who are hetero- zygous for the sickle-cell gene have lower Plas- —379—
  • 2. 380 medium falciparum parasite counts than do children without this trait; (2) the finding of a low incidence of sickling carriers among the cases of fatal malarial infections observed in the same areas; and (3) the overlap between the world distribution of the sickle- cell trait and that of malignant malaria. Motulsky (1960) was the first to report population data that showed that the distribution of G-6-PD deficiency in the Eastern Hemisphere also overlaps that of malignant malaria, and suggested that even this well-known example of sex-linked polymorphism owes its establishment to the higher fitness of enzyme- deficient genotypes in a malarial environment. On the other hand, studies of the malaria parasite counts in groups of normal and enzyme-deficient children yielded contradictory results. Thus, while Allison & Clyde (1961), in East Africa, and Harris & Gilles (1961) in West Africa found significantly lower P. falciparum counts in young enzyme-deficient children, Kruatzachue and his co-workers (1962), Motulsky (1965), and Edington & Watson-Williams (1965) failed to do so. Since, however, it is known that, for instance, a protein—deficient diet (Pérez et al. , 1964) and hypo- thyroidism (personal unpublished data) can lower G-6-PD activity, it is uncertain how much of this disagreement may be due to misclassification of the G-6-PD phenotypes, which reasonably may be expected, especially when dealing with indigenous African populations, in which enzyme deficiency is not as complete as in Caucasians, and when only screening tests have been used for the diagnosis. Equally contradictory have been the conclusions of studies on malarial parasite counts in the hetero- zygous carriers of haemoglobin C and haemoglobin E genes (Edington & Laing, 1957; Thompson, 1962; Edington & Watson—Williarns, 1965; Brumpt & Brumpt, 1958; Kruatzachue et al. , 1961). Investi- gations of these types have been impossible to perform on other haemoglobin variants because of their rarity, or on the different forms of thalassaemia because their correct diagnosis, which requires elabo- rate laboratory studies, is unreliable under field—work conditions. Moreover, the common occurrence, in the primitive areas Where malaria is still prevalent, of environmental and biological stress factors may dis- tort the haematological picture of thalassaemia. Thus, it is not surprising that the only evidence for the relationship between malaria, thalassaemia and G-6-PD deficiency should come exclusively from population studies showing a positive correlation between the incidence of these genes and malaria M. SINTSCALCO AND OTHERS morbidity or, rather, past malaria morbidity; since, for the reasons outlined above, the correct classifica- tion of these abnormalities of the red cells is difficult in primitive areas where malaria is still endemic. Nevertheless, even these types of studies present difficulties, since they can have very little value unless the following three essential requirements are met: (1) a reasonable degree of ethnic homo- geneity among the “ genetic isolates ” chosen for the study; (2) the availability of accurate historical data as well as general vital statistics, data on the malaria morbidity, mating patterns, and rates of consan- guinity in order to be able to estimate the relative importance of drift, migration and natural selection as potential causes of any genetic heterogeneity that might be found among neighbouring isolates; and (3) the availability of control populations within the malarial areas under consideration; that is, the ex- istence of “ malaria-free islands ” or human settle- ments of appreciable size, inhabited by persons of the same ethnic group, who have been living in genetic isolation for a very long time. Populations of nomadic habits are clearly useless for investigations of this type. Since these conditions are met only very rarely, it is impossible to attempt an over-all evaluation of the “malaria hypothesis” from the data available on the world distribution of the genes under discussion here. Instead, we prefer to report here, for the first time in full, the data that we collected in Sardinia and have so far published only in part (Siniscalco et al. , 1961; Siniscalco, 1964; Adinolfi et al. , 1960), since we feel that they represent one of the few sets of data among the population studies so far published on the subject that meet the requirements outlined above. HISTORICAL BACKGROUND The island of Sardinia is the ideal place for such population studies. The degree of isolation to be found there is still very high for most of its villages and, as shown in Fig. 1, its geography is such that, within a few hundred square miles, one can easily find isolated settlements with very high malaria morbidity in the past and others that have been prac- tically always free from this disease. Furthermore, it can reasonably be assumed that the populations of many Sardinian villages must have remained free from external admixtures over a very long period. Ancient Greek colonization was, in fact, limited to Olbia (the northern portion of the island), while the Romans and Carthaginians only exploited
  • 3. POPULATION GENETICS or TI-IALASSAEMIA AND G-6-PD DEFICIENCY 381 FIG. 1 LOCATIONS AND ALTITUDES ABOVE SEA-LEVEL OF THE VILLAGES IN SARDINIA CITED IN THE TEXT“ '5 Q '0 rm: .7; snnrigé, San Ieura 6‘a/ / / ~ ~ . ,}£’an. » 0” Earsacut ma : - . Irma/ ~a A . ruxvzzlm - . . If lrrmci cf, /M" , ls/ /mlva Ralf ahyfll. - / =W'¢""° 5’u/ I‘ M‘ ferry/ um: ‘(s I Ala/ It: Santa Imam‘ / Jal. Mal di V: /retro ‘(I til: //arim: 0/‘ta/ i . . at! salt at‘ 3 ‘ 0/‘ism/ n7 3 tuna at/ ion A . a/ -/my V ,9-gay”, . arflwi . .9 9 Fuss/ .1 . _ 9 -,6‘? -3. on/ 'nw. né. a.m Iarroscuxa ‘ . ' _( E can/ lira/ -2 ' / as-gvlus I , _ ° . - A 4-, ? I 6'41 mam . 3}" . - I ' '4.’ '5:; e cm. -.: m E . $‘. AIIt/ '41: - » i . 1.2;: MED/7'5]? /IA/ VEA/ V SEA 5AI}. ‘.2.'. .!"" »~ Land above 1000 '- In I: I0 29 3: IIANIJFD, Kl’/ VDLTN '1 Reproduced from Logan. .1. A. (1953) The Sardinian project. with the permis- sion of the Johns Hopkins Press, Baltimore, Md. , USA.
  • 4. 382 the coastal regions for grain, as is proved by the very localized areas along the south-west coast in which archaeological remains of their towns can now be found, as at the excavations at Nora. Later, the Vandals and Goths simply overran the island, which subsequently became the scene of struggles among the Pisans, Gcnovese and Saracens, none of whom, however, were there in sufficient numbers or for a long enough time to alter, significantly, the genetic structure of the autochthonous population. Even the Spanish, who ruled the island from 1297, when Pope Boniface VIII awarded it to James II of Aragon, never really cared to penetrate the rocky paths leading to the interior of the island, which remained half- forgotten, with its primitive villages consisting of huts clustered around the ancient nuraghi, and where little social or economic change took place until the accession of the House of Savoy in 1720. The nuraghi, those beautiful and impressive Bronze Age stone constructions, are scattered in thousands all over the island, showing how well organized the Sardinians were, even in prehistoric times, and how concerned with defence against invaders. It can hardly be denied, however, that as a result of these invasions, a few sets of “ external ” genes must have entered the Sardinian genetic pool from time to time. Malaria has probably been endemic in Sardinia since prehistoric times, according to the Roman historian Livy, and has remained one of the most important causes of infant mortality until the begin- ning of the present century. This disease was not eradicated completely until after the Second World War, when the Rockefeller Foundation, by a massive and well-planned anti-malaria campaign (Logan, 1953), successfully completed the pioneering work of Fermi and Missiroli, the distinguished Italian malariologists who had struggled against the disease in Sardinia for decades and made available to pos- terity the most accurate and complete information that could be desired on the malaria morbidity of every Sardinian village. This information, together with the very ancient church records and the first- class vital statistics on the island that have been available since the Savoy accession in 1720, enabled us to draw valid conclusions about the population structure and dynamics of the Sardinian isolates chosen for our studies. When we began our investigations, the existence of thalassaemia and G-6-PD deficiency in Sardinia already had been well established (Carcassi, Ceppel- lini & Pitzus, 1957; Larizza et al. , 1958), and the preliminary studies of Ceppellini (1955) had shown M. SINTSCALCO AND OTHERS that the incidence of thalassaemia in two non malarial villages in the Gennargentu Mountains was strikingly low as compared with that found in two lowland villages in formerly very malarial parts of the eastern coast of the island. We repeated these studies in a total of 19 villages and extended them to G-6-PD deficiency (Siniscalco et al. , 1961) and were able to demonstrate that there was, indeed, a very close positive correlation between the present-day frequencies of thalassaemia and G-6- PD deficiency and former malaria morbidity as reported by Fermi (1938). Additional studies of the same kind have been made in the past few years; new data are now available and are discussed in detail below. SUMMARY OF THE SARDINIAN POPULATION DATA Table 1 shows the frequency of the gene for G-6-PD deficiency, Gd(—), and that for /3-thalassaemia, Th(+), in 52 Sardinian villages, based on more than 6000 observations of unrelated individuals. The esti- mated frequencies of the two traits refer to the youngest generation, since the data for each village were obtained from a random sample of its school- boys. These estimates are better summarized in Fig. 2, in which a positive correlation between the frequency of thalassaemia and that of G-6-PD deficiency is evident. This correlation seems to be described reasonably well by a straight regression line up to a maximum frequency level of about 24 %. The correlation fades off above this level, since the gene for thalassaemia, being lethal in the homozygous condition, can never reach equilibrium levels as high as those that are possible for the much less unfavourable gene for the enzyme deficiency. Basic considerations of population genetics (Livingstone, 1964) make it evident that: (1) Levels of gene frequency such as those reported for the villages in the Sardinian plains can be ex- plained only by assuming a higher fitness of the heterozygous genotype for thalassaemia and of the heterozygous and, perhaps, the hemizygous and homozygous genotypes for enzyme deficiency. (2) A “ migration” hypothesis is grossly inade- quate to account for the very high frequencies of these traits found in the plains, since they could not be expected, even if one were to assume, against all historical evidence, that the autochthonous popu- lations had been totally replaced by equally numerous
  • 5. POPULATION GENETICS OF THALASSAEMIA AND G-6-PD DEFICIENCY TABLE 1 INCIDENCE OF GLUCOSE-6-PHOSPHATE DEHYDROGENASE DEFICIENCY AND THE 8-THALASSAEMIA TRAIT IN 52 SARDINIAN VILLAGES ' Gd(—) '7 TH(+) "' village a, Altitude . (metres) Number Percentage Number Percentage tested positive I tested positive i 1. Assemlni 5 108 E 20.4 108 11.0 2. Marrublu 7 as i 32.5 98 1 28.6 3. Cabras 9 200 35.0 100 28.0 4. Terralba 9 100 30.0 — — 5. Carloforte 10 99 5.0 99 5.0 6. Decimomannu 10 100 26.0 ' 100 25.0 7. S. Glusta 10 42 30.9 — — 8. Pula V 15 100 15.0 100 16.0 9. Tortoll 15 50 16.0 — — 10. Orosei 19 130 13.0 308 18.8 11. Torpe 24 100 22.0 100 38.0 12. Irgoll 26 100 15.0 100 32.0 13. Gaflelli 40 175 12.0 235 21.2 14. Siniscola 42 195 11.3 97 24.4 15. Barisardo 50 98 15.3 95 18.4 16. Teuladn 50 101 16.9 100 25.0 17. S. Gavino 53 100 26.0 -— - 18. Capoterra 54 02 16.3 92 20.6 19. Siliqua i 66 100 26.0 100 22.0 20. Vallermosa 70 86 20.9 86 21.0 21. Monastir 88 94 23.4 94 21.2 22. Nuraminis 86 100 25.0 . 100 17.0 23. Villamar 108 100 23.0 100 18.0 24. Guspini E 137 99 28.2 99 34.4 25. Domusnuvas ' 152 100 22.0 100 19.0 26. Gonnasfanadiga . 156 49 24.5 — — ‘27. Usinl i 190 99 6.1 99 14.2 23. Omaha 1 195 72 14.0 72 ' 23.0 29. Senorbi i 204 101 24.7 101 13.8 30. Tresnuraghes 251 86 12.7 85 31.4 31. Sedilo 288 100 22.0 96 13.8 32. Serrenti 307 100 21.0 ‘ 100 25.0 33. Arbus 311 95 35.7 '3 95 32.8 34. Abbasanta 315 97 18.5 1 92 1 22.8 35. Dualchi 321 75 21.3 J 100 1 18.0 36. Sunl . 333 9a 14.3 3 mo ' 25.0 37. Lode I 335 820 28.2 5 820 27.5 as. Gergel 3 374 92 13.5 ; 92 j 13.2 39. Borore 1' 399 100 ; 9.0 ' 99 33.2 40. Benetutti 1 406 100 ' 9.0 ‘ 100 12.0 41. Bolotana 1 412 93 I 11.8 . 93 I 21.4 42. Luras I see 100 ‘ 7.0 5 93 g 23.0 43. Lula ‘ 521 1oo 7.0 « 100 1 19.0 44. lsili 523 100 9.0 ‘ 100 j 17.0 45. Bitti 549 3 193 5.1 1 193 12.2 46. Lanusei 595 100 4.0 —— - 47. Ala dei Sardi 663 80 L 22.5 30 g 20.0 48. Orune 745 97 ‘ 6.1 ' 97 I 14.4 49. Gave! 777 93 1 3.0 1 9a 1 10.2 50. Desulo 3 391 313 ; 3.0 § 320 l 3.3 51. Tonara ‘ 935 148 I 4.0 , 102 1 4.8 52. Fonni 1 ooo , 1oo . 3.0 | — 1 — “ These villages are arranged and numbered in order 01 their increasing altitude. The data were obtained from random samples of the schoolboy: of each village. 1’ Glucose-6-phosphate dehydrouenase deficiency. ‘ Thalassaemia.
  • 6. 384 M. SINISCALCO AND OTHERS F| G.2 DISTRIBUTION OF G-6-PD DEFICIENCY AND OF THE THALASSAEMIA TRAIT IN SARDINIA“ Thalassnemia (7a) 0 5 l0 l5 G-6-PD deficiency (73) 20 25 30 35 40 vmo 60065 “ Estimates of the gene frequencies obtained from random samples of schooiboys. The numbers Identity the villages listed In Table 1. groups of immigrants, all of whom were carriers of thalassaemia and/ or G-6-PD deficiency. (3) If the genetic heterogeneity between the low- land villages and those in the high mountains is attributable to different adaptive values of the carrier genotypes, malignant malaria is the obvious ultimate factor, since the two environments are known to have differed from each other for centuries and, until only about twenty years ago, almost exclusively in respect to mortality and morbidity from malaria. The negative correlation with altitude above the level of 400 m is clearly demonstrated in Fig. 3, where the average frequencies (i 3 times their sampling errors) of the two traits are reported for each group of villages of similar altitude. This correlation is therefore positive when gene frequencies are com- pared with the relative incidence of malaria, as we demonstrated in a series of villages for which direct estimates of past malaria morbidity were available (Siniscalco et al. , 1961), and as is made clear by the data presented in Table 2. A few villages included in Fig. 2 and 3 (Carloforte, Usini, Lode and Ala dei Sardi) require special mention. Carloforte is the only village on the beautiful little island of San Pietro, which is adjacent to the very malarial plains of the south-western Sardinian coast. This island was first settled, about AD 1700, by a small group of Genovese fishermen who had come from the island of Tabarca, near the North African coast and, when expelled by the Bey of Tunis, had requested and received the hospitality of the King of Sardinia. This group, now numbering about 7000, kept itself in close isolation from the rest of Sardinia until very recently. It is thus not surprising that, despite the heavy malaria morbidity that was reported in the area until a decade ago, only a few genes for G-6-PD deficiency and thalassaemia can be found among them, and they had clearly been derived from Sardinian ancestors, as could be proved by genealo- gical studies. Usini is a small village not far from the north-west coast of Sardinia, where settlements of Genovese and Spanish origin are found. It is in this part of the island that the influence of the Catalan language on the local dialect can be observed most easily. Here, intermixture with the inland population has been more massive and frequent, and the intermediate
  • 7. INCIDENCE OF G-6-PD DEFICIENCY AND OF THE THALASSAEMIA TRAIT IN RELATION Gone frequency (Z) o~: .o. mSSZ3.'¢Ta'r33L‘. ’§'a‘.38$"‘aL": ‘.. ’ Usini and Ali: dei Sardl, which are considered separately, as explained in the text. POPULATION GENETICS OF TI-IALASSAEMIA AND G-6-PD DEFICIENCY Carlolorte @ O 0-50 _. _I 5i-100 @ G-6-PD deficiency O Thulassoemic 4 The figures In each of the large circles are the averages of the gene frequencies found in the villages that fall within the corresponding altitude groupings (0-50 metres, 51-100 metres, etc. ). These villages may be found easily in Table 1, in which they are arranged in sequence according to their altitude above sea-level. The smaller circles reter to the three villages of Carloforte, FIG. 3 TO ALTITUDE ABOVE SEA-LEVEL“ 101 - 200 1.. ... .” -201 - 400 401 Altitude above sea level (m) - 600 Q Ala‘ dei Sordi 601 — 800 41/ 44:; 801 - l000= ii. -.9 bcuéi 385 TABLE 2. SPLEEN AND PARASITE SURVEYS OF SCHOOLCHILDREN IN 66 SARDINIAN VILLAGES BETWEEN NOVEMBER 1947 AND MARCH 1948, BY VILLAGE ALTITUDE ABOVE SEA-LEVEL“ Spleen examinations Parasite examinations Altitude above N b 1 N b -t. b ‘ -l el "'3' 9' Spleen um er . p°5' “'5'. V T t l Parasite figaeuig) Number with raw Number Plasmodium species hugger {am examined palpable (V) examined 1 . t. (y) spleen I '’ E ‘ vivax ; falciparum l malariae W5‘ we ° I I a I 0-50 2 858 995 34.3 2 779 59 55 1 116 4.2 51 -100 2 015 595 29.5 1 861 39 25 1 69 3.7 101-200 1 282 238 22.5 1 239 58 29 1 83 5.8 201-300 1 215 225 13.5 1 201 I 25 11 I 0 37 3.1 301-400 1 094 212 19.4 1 144 12 I 2 I 0 > 14 1.2 401-500 1 540 215 14.0 1 539 12 10 D i 22 1.4 601 -300 1 431 217 in 1 431 18 9 I o 1 27 1.3 801-1000 1 430 106 7.4 1 an a 1 ‘ o l 9 0.7 . I ‘l I l . All altitudes 12 915 2 915 22.5 12 665 232 147 3 382 3.0 I “ Reproduced from Logan, J. A. (1953) The Sardinian project, with the permission of the Johns Hopkins Press, Baltimore, Md. . USA.
  • 8. 386 values of gene frequencies today are an obvious con- sequence. Carcassi (1962) has shown that the same situation prevails for Alghcro and its neighbouring villages. Lode and Ala dei Sardi are two villages that are especially useful to demonstrate the relationship between high gene frequencies and the former pre- valence of malaria. Beyond any doubt, these villages originated in the remote past; both of them are noted as sizable settlements in a 16th-century map of Sardinia drawn by Ignazio Donati and now kept in the Vatican Museum. Until a few years ago, their isolation was very strict, since a high mountain, difficult to cross even today, separates them from the coastal villages. Fermi (1938) reported a very high malaria morbidity for both of these villages, despite the relatively high altitude of one of them (Ala dei Sardi); the gene frequencies for both thalassaemia and G-6-PD deficiency are particularly high, unlike those observed in a neighbouring village (Bitti) located on the very summit of the mountain and reported by Fermi (1938) as being relatively free of malaria. All of these villages, together with others in the interior plains, such as Abbasanta and Guspini (see Table 1), again indicate that the suggestion that ethnic heterogeneity is a possible main cause of differences in gene frequencies is certainly not a likely one. Moreover, the blood-group distributions in the interior, coastal and mountain regions are remarkably similar, all showing an unusually high incidence of the M gene and a very low frequency of Rh-negative individuals, which seems to differentiate the Sardinian population from the general European population (Ceppellini, 1955). INTERACTION BETWEEN G-6-PD DEFICIENCY AND THALASSAEMIA AT THE INDIVIDUAL AND POPULATION LEVELS As expected, some individuals were found in Sardinia who carry both thalassaemia and G-6-PD deficiency. The association of both of these con- ditions in the same individual does not appear to involve a more serious red—cell defect, as has been demonstrated directly by chromium-51 studies that have showed that the reduction of red-cell survival time in these individuals is of the same order as that reported for carriers of G-6-PD deficiency alone iliernini et al. , 1964). Indeed, there are reasons to believe that the association of these two defects in the same person may, rather, produce higher biological fitness in him. M. SINISCALCO AND OTHERS For example, we have reported (Siniscalco et al. , 1961) that the frequency of severe haemolytic crises from exposure to fava beans (clinical favism) is less among carriers of both the enzyme deficiency and thalassaemia than among carriers of the enzyme deficiency alone. Since G~6-PD activity is always increased in carriers of thalassaemia, ‘ it was thought that some kind of compensation for the enzyme deficiency exists in the presence of tha- lassaemia. The probability that there is a higher fitness in the carriers of both of these genes is suggested by the finding that the number of such persons in the general population appears to exceed that which would be expected by calculation from the estimated gene frequencies for these traits in each village. While this excess is not significant within any village, it clearly becomes so when the data of the 21 villages studied for that purpose are pooled (Table 3). OTHER HAEMOGLOBIN VARIANTS AND DIFFERENT FORMS OF THALASSAI-IMIA IN SARDINIA To date, there has been no systematic search for haemoglobin variants or for other forms of thalassaemia in Sardinia, although the rarity of haemoglobin variants can be inferred from the scarcity of case reports during the last ten years. On the other hand, the presence of the so-called a-thalassaemia at an appreciable frequency appears probable, from our findings of a certain number of thalassaemic families without elevated haemoglobin A2 (Carcassi, Ceppellini & Siniscalco, 1957), from the report of Silvestroni & Bianco (1963) of two cases of Bart’s haemoglobin among Sardinians living in Rome and from the report of Fiaschi, Campanacci & Naccarato (1964) of a case of thalassaemia—haemoglobin H disease in a haemato- logical patient in the medical clinic of the University of Cagliari. We have performed an extensive study of nearly 1200 random blood samples collected in villages of the southern plains of Sardinia already known for their high frequency of thalassaemia and G-6-PD deficiency, and not a single instance of variant haemoglobin or of high foetal haemoglobin was found among them. Recently, however, the occur- ’Thalassaemic red cells, although smaller than normal ones, have the same enzymatic activity; thus, carriers of the Th (+) gene, who are polycythaemic, have relatively higher G-6-PD activity per blood-volume unit (Piomelli & Sinis- calco, 1966, to be published).
  • 9. POPULATION GENETICS OF THALASSAEMIA AND G-6-PD DEFICIENCY 387 TABLE 3 CALCULATION OF THE EXCESS OF DOUBLE CARRIERS IN 21 SARDINIAN VILLAGES) Double carriers: Gd(—) & Th(+) '1 Vmage H Double ; , i ‘°1°e's“t‘33“ °: .:a‘: :: , ; “:: ‘:; L°.2.: ;°. "" 1%. 2 99 a 7.00 +0.20 +0072 3 100 11 9.42 +2.59 +0399 11 100 5 0.79 -1.70 -0.577 14 97 2 2.44 ' -0.44 -0.231 15 98 i 3 2.49 1 +0.51 1 +0323 15 100 3 3.53 i -0.68 -0.354 24 99 12 7.90 +4.10 +1459 211 72 4 1.33 i +2.52 ; +2.23o 31 95 3 2.03 i +0.37 +0223 33 95 9 9.31 l -0.35 1 -0.100 34 92 3 3.44 -0.44 1 -0.237 as 93 4 3.13 +0.97 ‘ -0.492 37 020 65 55.00 E +10.00 +1350 as 92 2 2.10 -0.10 -0.071 39 99 2 0.24 +1.70 +3592 40 100 2 1.00 +1.00 +1.0oo 42 93 2 , 1.42 +0.53 +0487 43 100 ‘ 2 : 1.54 +0.45 ' +0.371 44 ‘ 100 1 1.39 i -0.33 -0.323 45 I 193 1 3 5 0.96 , +2.0 +2082 47 i 130 3 3.24 -0.24 -0.133 x i , ’ Z(£)= i‘3;i’i§ l E , I +12.399 Z / ’ , = 13% = 2.70 ; P<0.01 I “ Gd(—) = glucose-6-phosphate dehydrogenase deficiency; TH(+) = thalassaemia. 5 See Table 1 for numbering of villages. rence has been reported of a fast-moving haemo- globin variant that appears to be due to a mutation on the a-haemoglobin chain, probably similar to one that has been described for haemoglobin “Mexico” (Baglioni & Suljs, personal communica- tion). Moreover, an investigation of the distribution of haemoglobin A2 levels among a random sample of apparent carriers of thalassaemia in an area where this condition is present in about 3 3.3 of the population revealed the occurrence of a normal level of haemoglobin A. _. in about 4% of the cases. Further examination of these individuals and of their families led us to the conclusion that they had to be con- sidered instances of a-thalassaemia, although the presence of minor quantities of haemoglobin H could be established in only two of these individuals. If this conclusion is correct, it follows that the incidence of a-thalassaemia in the given area is of the order of 13-1; (O.30><0.04=0.0l2); thus only a minor classification error is involved when popula-
  • 10. 388 tion-screening for 13-thalassaemia is performed by red- cell fragility and blood-filrn studies alone (Table 4). The absence of the sickling trait in Sardinia is particularly noteworthy in view of its appreciable frequency in the neighbouring Mediterranean areas (Greece, North Africa, the Middle East, Sicily and, in general, southern Italy) and of its well-established adaptive value in a malarial environment (Allison, I965). It has been reported, however, that the incidence of the haemoglobin S gene tends to be correlated inversely with that of /3-thalassaemia in those populations in which both of these genes occur with appreciable frequency (Barnicot et al. , 1963). This phenomenon has been interpreted as a consequence of the frequently poor adaptability of the genotype that combines haemoglobin S and thalassaemia, thus leading to the elimination of the gene that M. SINISCALCO AND OTHERS happened to be the rarest when selective mechanisms of the present type became operative. Consequently, it may be postulated that, when thalassaemia and G-6-PD deficiency were introduced into Sardinia, the haemoglobin S gene may still have been uncommon in the Mediterranean basin and that it was therefore entirely eliminated from the Sardinian gene pool in the long run, while the other two genes successfully established themselves among the populations of the malarial plains. An alternative explanation is the assumption that both G-6-PD deficiency and thalassaemia are much older muta- tions than the haemoglobin S gene, and that the enzyme deficiency appeared in the Mediterranean basin when Sardinia had already split off from the mainland and isolated its population. At any rate, the absence of haemoglobin S from Sardinia is a good piece of evidence for the hypo- TABLE 4 DISTRIBUTION OF HAEMATOLOGICAL PARAMETERS IN A RANDOM GROUP OF 235 ADULT SARDINIAN MALES (157 NORMALS, 78 WITH THALASSAEMIA)“ D'scrimlna- Classifica- 1233' means 1 33:. -”. “r? ..'. ‘f 5:52:33‘ ram 1 ‘ 1 1 threshold error . 1 d Haemoglobin N 157 I 13.33 1.513 0.121 12.43 1 27% T 73 § 11.55 1.450 0.155 - I 1 Red cell number 9 N I 157 4.543 I 0.545 0.043 4.77 34% T , 79 1 5.019 1 0.501 0.059 ! l 1 1 Hematocrit N 156 42.56 3.571 0.286 1‘ 40.96 32 % T 79 39.17 3.972 0.450 ‘ 1 1 Mean corpuscular volume N 156 93.33 9.828 1 0.787 85.19 1 21 % T 79 77.05 9.923 1 1.112 : 1 1 . I Mean corpuscular haemoglobin N 157 29.53 3.854 0.308 25.70 16 % T 73 22.97 2.954 1 9.323 1 1 Mean corpuscular haemoglobin N 156 31.45 2.559 0.229 30.37 I 35 % °°"°°"“""“°" T 79 29.35 2.529 0.299 1 l Red cell fragility " N 127 0.413 0.024 1 0.002 0.379 7 % T 51 0.349 0.021 1 0.003 I 1 Haemoglobin Ar 1 N 157 2.329 0.392 5 0.031 3.37 " 0.7 % 1 T 1 79 5.195 0.590 0.077 “ Key: N = normal individuals: T = persons with thalassaemia (i. e., parents of persons with Cooley's disease). Data from Carcassi, Ceppellini at Slniscalco (1957). 5 Red cell fragility is expressed here as the concentration of NaCl required to produce 50 % haemolysis.
  • 11. POPULATION GENETICS OF THALASSAEMIA AND G-6-PD DEFICIENCY thesis that the Sardinians, unlike their neighbours, must have been genetically isolated for a very long time after the original arrival among them of the genetic raw material upon which natural selection must have been acting for at least 2000 years. An inverse correlation similar to that described for thalassaemia and sickling appears to exist between the genes for haemoglobins S and C in populations in which both of these variants are present (Allison, 1965). On the other hand, in Greece, a positive correlation has been reported between G-6-PD deficiency and B-thalassaemia (Allison et al. , 1963) and, in Greece as well as in Africa, between G-6-PD deficiency and the sickling trait (Motulsky, 1960; Allison, 1965). These interactions between different genes at a population level are an obvious illustration of the important role that natural selection must evidently play in the maintenance of the genic load of human populations. POPULATION DYNAMICS OF G-6-PD DEFICIENCY AND THALASSAEAHA In an attempt to provide a unitary explanation of the world distribution of genes known to involve genetic adaptability in a malarial environment, Zaino (1964) considered the possibility that they may have begun to have an adaptive value more than 50 000 years ago, when Europe was still bridged by land to Africa and malaria was probably already a strong factor in natural selection. When the glaciers receded, the Mediterranean basin was flooded, and its inhabitants either remained isolated along the newly formed seacoasts and islands or migrated towards the Middle and Far East into India, China, and, eventually, the Americas. In their new ecolo- gical niches, these genes were exposed to different selective pressures and, under the combined in- fluences of selective pressures such as consanguinity, mutation, interaction between genes, migration, drift, and differential survival, independently reached the diverse equilibrium frequencies that we observe today. There is naturally no hope of collecting evidence for or against these fascinating arguments. However, it seems to us to be irrelevant to establish whether a set of genes in a given area arose by mutation or migration, as long as it is made clear that their maintenance over an appreciably long time would not be possible without the intervention of selective mechanisms such as those proposed in the classic 389 works of Fisher, Haldane and Wright to explain the occurrence of stable genetic polymorphism. Livingstone (1964) has recently discussed, in detail, the population dynamics of the genes for thalassaemia, haemoglobin variants and G-6-PD deficiency and has presented the general equations for calculating the time required to attain genetic equilibrium in different selective models. In order to do so, he had to hazard some estimates of the fitness of the various genotypes for each of the genetic sys- tems under consideration. These estimates were pro- bably not far from reality for the haemoglobin S and B-thalassaemia genes, but we cannot agree with those proposed for the G-6-PD deficiency gene, which Livingstone supposed to have a positive adaptive value during malarial periods only in the female heterozygotes. This is, indeed, the obvious con- clusion if the gene frequencies observed today are considered “ equilibrium frequencies ”, but there are no strong reasons to assume that this is necessarily the case. Unlike the situation with the thalassaemia gene, whose lethality in homozygous conditions acts as a strong buffering factor to maintain the system under stable equilibrium at maximum levels of gene frequencies between 0.10 and 0.15, the G-6-PD deficiency gene, which certainly does not produce serious handicaps in its carriers, might well have been in a condition of transient equilibrium in malarial areas. Indeed, we feel that, at least in Sardinia, a situation similar to that presented in Table 5 is closer to reality. In other words, we assume that, when malaria was killing about one-half of the Sardinian popula- tion before reproductive age, the fittest genotypes may well have been the male hemizygote and the female homozygote, despite the slight risk of disease caused by the enzyme deficiency itself, since it is TABLE 5 THEORETICAL FITNESS VALUES OF THE DIFFERENT GENOTYPES FOR THE THREE SELECTIVE MODELS HYPOTHESIZED IN FIG. 4 E Males 1 Females . . . .v ‘ G-6-PD ' , I G‘6'FD E Normals I G‘s‘. PD I inter- Normals 1 oeficlent J 5 def: lentli mediate I I 1 I Model l 1.04 0.96 j 1.04 ‘ 1.00 l 0.96 l l I 1 Model II 1.01 . 0.95 I 1.01 1.00 0.95 —l—‘——- -—-— Model III 0.93 I 0.94 ! 0.98 1.00 0.94 . 1
  • 12. 390 reasonable to suppose that it might, for instance, have been the susceptibility of the enzyme-deficient red cells of these abnormal individuals to haemolysis that made the growth and multiplication of malarial parasites in their blood more difficult than in the blood of normal individuals. In such a situation, the greater the enzyme deficiency, the greater the protection against malaria. Under such a selective model, the ultimate fate of the enzyme-deficiency gene would, in the long run, have been its complete fixation, had the ecological factor that was responsible for the differential fitness not been removed (Fig. 4). This assumption does not appear to be unduly absurd when one considers the extremely high frequency of this gene in some other populations such as the Iraqi Jews, among whom frequencies of this enzyme deficiency in males range from 25% to 52 %, and among the Kurdish Jews, among whom frequencies as high as 70% have been recorded (Szeinberg, 1963). Nevertheless, it must be borne in mind that such attempts to express, in quantitative terms, the effects of natural selection in human populations are usually over-simplifications, since they can be done only by considering each genetic system separately. There are obvious difficulties in treating, in mathe- matical terms, more complicated models that would take into account the interactions between various gene systems. This is particularly true when different M. SINISCALCO AND OTHERS genes have been selected for by a common ecological factor. In the present instance, it is obviously possible that the greater fitness of the combined genotype Gd(—)Th(—i-), to which reference was made above, and the total absence of the haemo- globin S gene from Sardinia undoubtedly must have had their weight in influencing the gene-frequency distributions shown in Fig. 2 and 3 and in Table 1. Moreover, although there are adequate grounds for believing that malaria was the principal ecological factor responsible for the selection of these genes, the possible existence of other genetic and environmental selection factors should not be disregarded. Con- sanguinity, for example, must have been an impor- tant counteracting selective agent for the accumula- tion of thalassaemia genes in Sardinian isolates, since the increased homozygosity that follows close inbreeding would undoubtedly help in the elimina- tion of the lethal genes. The effects of consanguinity on G-6-PD deficiency must, instead, have been quite the opposite in the presence of malaria, if the fitness estimates shown in Table 5 are correct. To avoid the disturbing effects of consanguinity, we deliber- ately avoided the inclusion of villages that were of very different sizes and therefore likely to involve significant differences in inbreeding coeflicients that were otherwise known to be quite constant in all Sardinian Villages of ancient formation and long- standing genetic isolation (A. Moroni, personal communication). FIG. 4 THEORETICAL CURVES DESCRIBING THE POPULATION DYNAMICS OF THE G-6-PD DEFICIENCY GENE IN SARDINIA UNDER THREE SELECTIVE MODELS“ 2: . ° ~o .0 on 9 I Gene frequency (%) Z bx l40I30I20IIOI9080 70 60 50 40 30 20 I0 0 I0 20 30 A0 50 60 70 80 90 I00 ll0l20l30I40 Time (generations) I010 6006b “ The Roman numerals refer to the three selective models presented in Table 5. The curves were calculated by W. S. Volkers, Department of Human Genetics, Universlty of Leiden. according to the equations proposed by Llvingstone (1964).
  • 13. POPULATION GENETICS or THALASSAEMIA AND G-6-PD DEFICIENCY 391 POSSIBLE GENETIC HETEROGENEITY BETWEEN G-6-PD DEFICIENCY AND THALASSAEMIA PERTINENT TO DIFFERENT AREAS Unlike the case with haemoglobin variants, for which it can always be established with certainty whether or not one is dealing with a specific mutation (that is, with a specific amino-acid substitution in the haemoglobin molecule), one can never be sure of the genetic homogeneity of the several forms of thalas- saemia and G-6-PD deficiency that are reported from different parts of the world. The numerous and not always concordant attempts to classify different forms of thalassaemia (Fessas, 1965) and G-6-PD deficiency (Motulsky, I965) demonstrate the limitations of our knowledge in this matter, evidently because of the very fact that the diagnoses of tha- lassaemia and G-6-PD deficiency are made at a level far removed from the primary product of the genes. For the foregoing reasons, it appeared unwise, at present, to attempt an over-all evaluation of all of the population data so far published. It might well be, for example, that the ,3-thalassaemia observed in Sardinia is quite different from those reported in Greece or in the Far East, or even from that observed in the Ferrara district. The same is true for G-6-PD deficiency even within the so-called subclasses of the G-6-PD deficiency of the Caucasian type, which involves a complete red-ceH enzyme deficiency and no obvious electrophoretic differences, and the G-6-PD deficiency of the African type, which involves a partial red-cell enzyme deficiency asso- ciated with not yet well understood changes in the electrophoretic enzyme pattern. In view of the contrasting conclusions that have been drawn from linkage studies performed in Sardinia and elsewhere to establish the linear sequence on the X chromosome of the G-6-PD gene and other X-borne loci (Siniscalco, 1964; Siniscalco, Filippi & Latte, 1964), we have stressed the possibility that the G-6-PD deficiency may be genetically heterogeneous in different populations. We have also reported some population data that suggest a non-random distribution of colour blindness of the deutan type among enzyme-deficient and normal individuals in Sardinia, suggesting that the neutral or slightly detrimental gene that is responsible for this defect in colour vision has also enjoyed protec- tion in a malarial environment because of its very close linkage with the highly adaptive gene for G-6-PD deficiency (Siniscalco, 1963). Similar obser- vations on the non-random distribution of colour blindness and G-6'-PD deficiency have been reported in Israel (Adam, 1963). If these findings are confirmed, it could mean that, in fact, it is not the G-6-PD deficiency gene alone that confers a higher fitness upon its carriers, but a “ successful gene complex ” (such as, for example, the G-6-PD deficiency mutant plus a sex-linked modifier that is capable of reducing the red-cell enzyme deficiency to a minimum), which natural selection would have maintained because of its strong adaptive value in a malarial environment. If this were the case, the occurrence in some popu- lations of a genetic mechanism, such as chromosomal inversion, that is capable of reducing crossing-over, could be suspected. This would help to explain how the selective advantage of the G-6-PD deficiency gene may differ from one population to another, if not from one family to another. CONCLUSIONS The data that have been reported to date offer striking examples of the possibilities and lirnitations of studies on human population genetics. On the one hand, we find the availability of true natural populations living in all sorts of ecological situations and therefore exposed to all degrees of natural selection, a detailed knowledge of historical and pre- historical migrations, mating systems, inbreeding, public health and vital statistics, and the facilities for obtaining direct estimates of different genetic fitnesses and exhaustive information concerning their interaction at the individual and population levels. On the other hand, we must consider the difliculty in establishing the relative importance of the numerous identifiable factors of evolution, the time and expense required to collect an amount of data sufficient to justify significant conclusions, the technical difiiculties in performing critical studies in areas in which certain forces of natural selection, such as malaria, are still active, and the rapidly increasing impact of modern civilization on human ecology and therefore upon the relative fitness of a given genotype within the time interval of even a single human generation. However, from all of the foregoing, a conclusion clearly emerges; namely, that it is necessary to learn a great deal more about the genetic structure of our own species and on the biological fitnesses of the principal genotype combinations in the existing " natural environments ” as well as in the new environ- ments that our civilization is likely to produce.
  • 14. 392 M. SINISCALCO AND OTHERS ACKNOWLEDGEMENTS The authors wish to express their gratitude to Professor G. Montalenti, whose encouragement, help and advice were the main factors that made possible the fulfilment of the present research proj act. The collection of the blood samples would not have been possible without the generous and unlimited help of the health and administrative authorities of the districts of Cagliari and Nuoro and of the medical officers and school headmasters of the numerous Sardin- ian villages visited from time to time. Our gratitude goes also to the members of all of the families investigated, to the Directors of the Ophthahno- logical Clinic, the Paediatric Clinic, the Institute of Medical Pathology and the Institute of Genetics of the University of Cagliari for their hospitality to the members of the team during the collection of data. Special thanks go to Mr R. Palmarino, whose thorough and excellent technical assistance was of substantial significance throughout the entire investigation. The investigations reported in the present paper were supported by grants allocated to the Department of Genetics, University of Rome, by the Rockefeller Foundation (RF 62009, 1962-64), the National Science Foundation (G 19710, 1961-63) and the Italian National Research Council (Commissione di Genetica, 1960-64). RESUME Les conclusions de nombreux travaux ont appuyé Phypothese que le paludisme a représcnté un facteur écologique commun responsable de la selection de trois groupes d’anomalies héréditaires des hématies: hemo- globines atypiques, thalassemie et carence en glucose- 6-phosphate deshydrogénase (G-6-PD). Les auteurs ont étudié la distribution génétique de la thalassémie et de la carence en G-6-PD chez les popula- tions de certains villages sardes dont beaucoup sont trés anciens et sont restés isolés pendant de tres longues périodes. Ces collectivités offrent les trois conditions permettant une etude des relations possibles entre le paludisme, la thalassemie et la carence en G-6-PD: degré suffisant d’h0mogénéité ethnique; existence de statistiques démographiques et sanitaires dignes de foi; presence de groupes suffisamment importants exempts de paludisme, dont les caractéristiques ethniques et l’isolement génétique sont semblables a ceux des populations impaludées. Ces situations respectives sont restées inchangées jusqu"a l’achévement de l’éradication du paludlsme en Sardaigne. Plus de 6000 observations faites dans 52 villages out mis en evidence une correlation positive entre Pincidence de la thalassémie et eelle de la carence en G-6-PD et la morbidité par le paludisme observée dans le passe. Au cours de l’examen de pres de 1200 échantillons de sang, d’autres types d’hémoglobinose (dont l’anémie a hématies falciformes) n’ont pu étre décelés. Les auteurs passent en revue diverses hypotheses con- cernant le role de certaines combinaisons de genes dans Padaptation de 1’individu a un milieu d’endémie palu- déenne. Les porteurs du gene responsable de la déficience enzymatique auraient joui d’une immunité relative vis-it- vis du paludisme; cette action protectrice de la carence en G-6-PD pourrait avoir été influencée diversement par d’autres genes localjsés aux hétérochromosomes. Les auteurs concluent qu’il est nécessaire de mieux connaitre la structure génétique de l’espece humaine et d’étudier l’aptitude biologique des porteurs des princi- pales combinaisons génotypiques a vivre dans les milieux naturels aetuels et futurs. REFERENCES Adam, A. (1963) In: Goldschmidt, E. , ed. , The genetics of migrant and isolate populations, Baltimore, Williams & Wilkins, p. 111 Adinolfi, M. , Bernini, L. , Carcassi, U. , Latte, B. , Motulsky, A. G. & Siniscalco, M. (1960) Am‘ Accad. noz. Lincei R. C. , 28, Cl. Sc. , 716 Allison, A. C. (1954) Brit. med. J. , 1, 290 Allison, A. C. (1965) Population genetics of abnormal haemoglobin: and glucose-6-phosphate de/ tydrogenase deficiency. In: Jonxis, J. H. P. , ed. , Abnormal haemo- globins in Africa, Oxford, Blackwell, p. 365 Allison, A. C. & Clyde, D. F. (1961) Brit. med. 1., 1, 1346 Allison, A. , Askonas, B. A. , Bamicot, N. A. , Blumberg, B. S. & Krimbas, C. (1963) Ann. hum. Genet. , 26, 237 Baglioni, C. (1962) Correlation between genetic: and chemistry of human haemoglabins. In: Taylor, J. H. , ed. , Molecular genetics, Academic Press, New York. p. 405 Bamicot, N. A. , Allison, A. C. , Blumberg, B. S. , Deliyan- nis, G. , Krimbas, C. & Ballas, A. (1963) Ann. hum. Genet, 26, 229
  • 15. POPULATION GENETICS OF THALASSAEMIA AND G-6-PD DEFICIENCY Bernini, L. , Latte, 13., Siniscalco, M. , Piomelli, S. , Spada, U. , Adinolfi, M. & Mollison, P. L. (1964) Brit. J. Haemat. , 171, 10 Brumpt, L. C. & Brumpt, V. (1958) Bull. Soc. Path. exot. , 51, 217 Carcassi, U. (1962) Giornate di studio sulla microcitemia, Rome, Istituto Italiano di Medicina Sociale, pp. 78, 182 Carcassi, U. , Ceppellini, R. & Pitzus, F. (1957) Ball. Ist. sierater. milan. , 36, 206 Carcassi, U. , Ceppellini, R. & Siniscalco, M. (1957) Haematologica, 42, 1635 Ceppellini, R. (1955) Cold Spr. Harb. Symp. quant. Biol. , 20, 252 Edington, G. M. & Laing, W. N. (1957) Brit. med. J. , 2, 143 Edington, G. M. & Watson-Williams, E. J. (1965) Sickling, haemoglobin C, glucose-6-phosphate dehydro- genase deficiency and malaria in western Nigeria. In: Jcnxis, J. H. P. , ed. , Abnormal lzaemoglobins in Africa, Oxford, Blackwell, p. 393 Fermi, E. (1938) la malaria in Sardegna, Sassari, Stamperia Libreria italiana e straniera, vol. 2 Fessas, P. (1965) Forms of thalassaemia. In: Jonxis, J. H. P. , ed. , Abnormal haemoglobin: in Africa, Oxford, Blackwell, p. 71 Fiaschi, E. , Campanacei, L. & Naccarato, R. (1964) Minerva med. , 55, 1717 Haldane, J. B. S. (1949) Ricerca scient. , suppl. 2, p. 12 Harris, R. & Gilles, H. M. (1961) Ann. hum. Genet. , 25, 199 Ingram, V. I (1963) The haenzoglobiris in genetics and evolution, New York, Columbia University Press Itano, H. A. (1965) The synthesis and structure of normal and abnormal hemoglobins. In: Jonxis, J. H. P. , ed. , Abnormal haemoglobins in Africa, Oxford, Blackwell, p. 3 Kruatzachue, M. , Charoeularp, P. , Choughsuphajaisid- dhim, T. & Harinasuta, C. (1962) Lancet, 2, 1183 Kruatzachue, M. , Na-Nakoru, S. , Charoeularp, P. & Suwomakul, L. (1961) Ann. trap. Med. Parasit. , 5, 468 393 Larizza, P. , Ventura, S. , Mattioli, G. , Sulis, E. & Aresu, G. (1958) Haematologica, 43, 517 Livingstone, F. B. (1964) Am. J. hum. Genet. , 16, 435 Logan, J. A. (1953) The Sardinian project, Baltimore, Johns Hopkins Press Motulsky, A. G. (1960) Hum. Biol. , 32, 28 Motulsky, A. G. (1965) Theoretical and clinical problems of glucose-6-phosphate deltydrogenase deficiency : its occurrence in Africans and its combination with haemo- globinopathy. In: Jonxis, J. H. P. , ed. , Abnormal haemoglobin: in Africa, Oxford, Blackwell, p. 143 Pérez, N. , Clark-Turri, L. , Rabjille, E. & Niemeyer, H. (1964) J. biol. Chem. , 239, 2420 Rucknagel, D. L. (1964) Ann. N. Y. Acad. Sci. , 119, 436 Rucknagel, D. L. & Neel, J . V. (1961) The lzemoglobino- pathies. In: Steinberg, A. G. , ed. , Progress in medical genetics, New York, Grune & Stratton, p. 158 Silvestroni, E. & Bianco, I. (1963) Le emoglobiue umane, In: Gedda, L. , ed. De genetica medica, Rome, Istituto Mendel, vol. 4, pp. 197-336 Siniscalco, M. (1963) Linkage data for G6PD deficiency in Sardinian villages. In: Goldschmidt, E. , ed. , The genetics of migrant and isolate populations, Baltimore, Williams & Wilkins, p. 106 Siniscalco, M. (1964) In: Geerts, S. , ed. , Genetics today. Proceedings of the Xlth International Congress of Genetics, The Hague, 1963, London, Pergamon Press, vol. 2, p. 851 Siniscalco, M. , Bernini, L. , Latte, B. & Motulsky, A. G. (1961) Nature (Lond. j, 190, 1179 Siniscalco, M. , Filippi, G. & Latte, B. (1964) Nature (Lond. ), 204, I062 Sneinberg, A. (1963) G6PD deficiency among Jews- genetic and anthropological considerations. In: Gold- schmidt, E. , ed. , The genetics of migrant and isolate populations, Baltimore, Williams & Wilkins, p. 69 Thompson, G. R. (1962) Brit. med. J. , 1, 682 Zaino, E. C. (1964) Ann, N. Y. Acad. Sci. , 119, 404

Related Documents