James V. Neel

FOR THE POPULATION CENETICIST, diabetes mellitus has long presented an enigma. Here is a relatively frequent disease, often interfering with reproduction by virtue of an onset during the reproductive or even pre-reproductive years, with a well-defined genetic basis, perhaps as simple in many families as a single recessive or incompletely recessive gene (cf. Allan, 1933; Pincus and White, 1933, 1934; Harris, 1950; Steinberg and Wilder, 1952; Lamy, Frezal and de Grouchy, 1957; Steinberg, 1959; Post, 1962a). If the considerable frequency of the disease is of relatively long duration in the history of our species, how can this be accounted for in the face of the obvious and strong genetic selection against the condition? If, on the other hand, this frequency is a relatively recent phenomenon, what changes in the environment are responsible for the increase? Current developments in the study of this disease suggest an explanation with important biological ramifications. THRIFTINESS OF DIABETIC GENOTYPE PRIOR TO ONSET OF DIABETES MELLITUS There is now much evidence to indicate that the individual predisposed to diabetes differs metabolically from the non-predisposed from birth onward. The frequency of over-sized infants among the offspring of women with overt or subclinical diabetes is well known (e.g., Bowcock and Greene, 1928; Bix, 1933; Allen, 1939; Miller, 1945, 1956; Kriss and Futcher, 1948; Gonce, 1949; Gilbert, 1949; Brosset and Werko, 1950; Moss and Mulholland, 1951; Jackson, 1952; Pirart, 1955; Hsia and Gellis, 1957, etc.). This phenomenon, which may antedate the development of clinical diabetes by 30),ears, has customarily been regarded as an expression of the mother's diabetes. The difficulties in duplicating this phenomenon in experimental animals rendered diabetic by alloxan (Miller, 1947, but see Lazarow, Kim and Wells, 1960) leaves some room for doubt as to whether this concept offers a complete explanation. Maternal genotype and phenotype may be only one factor in the etiology of the phenomenon. Inasmuch as from the genetic standpoint these children constitute a high risk group, and in view of the report that the juvenile diabetic at birth averages some 100 grams heavier than his non-diabetic siblings (Nilsson, 1962, but see White, 1960), the possibility must be considered that this phenomenon is also in part an expression of the infant's predisposition. Some evidence to this effect may be drawn from the observations of Sheldon (1949) on rapidly developing maternal obesity, in that women who presented with this problem, of whom a significant proportion had overt or subclinical diabetes and to whom large infants were often born, themselves often had been large infants. What is lacking is evidence concerning the relative risk of the development of diabetes in the normal-birthweight siblings of these individuals. Somewhat more cogent evidence regarding the role of infant genotype in birthweight comes from the reports that the infants born to normal mothers with diabetic husbands also have an increased birthweight (Jackson, 1952; Siegeler and Siegeler, 1960; Pyke, 1961; Nilsson, 1962). Children who develop diabetes tend to be somatically advanced for their age (White, 1939; Wagner, White and Bogan, 1942; Nilsson, 1962). The menarche has occurred, on the average, about a half to three-quarters of a year earlier in the diabetic woman who developed her disease after the age of 20 than in her siblings who did not later develop diabetes (Arduino and Ferreira, 1958; Post and White, 1958; Post, 1962). Theoretically (cf. Post and White, 1958) this might lead to earlier onset of childbearing and even to an increase in average fertility on the part of those with late-onset diabetes, capable of offsetting, at ]east in part, the impaired reproductive performance of those with early onset diabetes. The actual data on this point are consistent with a fertility differential in favor of the diabetic, but, as pointed out by Post and White (1958), are inadequate in several respects. …

The mtDNA variation of 321 individuals from 17 Native American populations was examined by high-resolution restriction endonuclease analysis. All mtDNAs were amplified from a variety of sources by using PCR. The mtDNA of a subset of 38 of these individuals was also analyzed by D-loop sequencing. The resulting data were combined with previous mtDNA data from five other Native American tribes, as well as with data from a variety of Asian populations, and were used to deduce the phylogenetic relationships between mtDNAs and to estimate sequence divergences. This analysis revealed the presence of four haplotype groups (haplogroups A, B, C, and D) in the Amerind, but only one haplogroup (A) in the Na-Dene, and confirmed the independent origins of the Amerinds and the Na-Dene. Further, each haplogroup appeared to have been founded by a single mtDNA haplotype, a result which is consistent with a hypothesized founder effect. Most of the variation within haplogroups was tribal specific, that is, it occurred as tribal private polymorphisms. These observations suggest that the process of tribalization began early in the history of the Amerinds, with relatively little intertribal genetic exchange occurring subsequently. The sequencing of 341 nucleotides in the mtDNA D-loop revealed that the D-loop sequence variation correlated strongly with the four haplogroups defined by restriction analysis, and it indicated that the D-loop variation, like the haplotype variation, arose predominantly after the migration of the ancestral Amerinds across the Bering land bridge.

Acatalasemia, a rare congenital abnormality characterized by an apparent lack of the enzyme catalase, was first discovered by Takahara and Miyamoto in 1947 (1, 2). The lack of catalase activity in this disease was initially noted in whole blood, but subsequent studies of the tissues of other organs such as the nasal and oral cavities, the pharynx, bone marrow and liver have revealed a similar absence of activity for this enzyme (3-5). Since the initial description of this disease, 38 cases in 17 families have been found and reported in Japan up to April 1959 (6-17); at this writing no reports of acatalasemia have appeared from any other countries of the world.

IF A DROP OF BLOOD is collected from each member of a randomly assembled series of American Negroes and sealed under a cover slip with vaseline, to be observed at intervals up to 72 hours, in the case of about 8 percent of the individuals composing the series a high proportion of the erythrocytes will be observed to assume various bizarre oat, or holly leaf shapes. This ability of the erythrocytes to sickle, as the phenomenon is commonly described, appears to be attended by no pathological consequences in the majority of these individuals, and they are spoken of as having sicklemia, or the sickle cell trait. However, a certain proportion of the individuals who sickle are the vic4ims of a severe, chronic, hemolytic type of anemia known as sickle cell anemia. This proportion has been variously estimated at between 1: 1.4 (8) and 1: 40 (4). The essential difference between sicklemia and sickle cell anemia appears at present to depend at least in part upon the relative ease with which sickling takes place. In sickle cell anemia the erythrocytes may frequently sickle under the conditions encountered in the circulating blood, whereas in sicklemia sickling does not usually occur under these conditions (12). This difference has been attributed to a greater tendency of the erythrocytes of sickle cell anemia to sickle when the 02-tension is reduced, although recently this viewpoint has been challenged (13). Perhaps because of this differencealthough there may be other factors involved, such as the anisoand poikilocytosis to be observed in some individuals with the disease, and a greater resistance to hemolysis of trait cells when sickled than sickle cell anemia cells when sickled-the erythrocytes of a patient with sickle cell anemia have a greatly shortened life span, both in the individuals with the disease and in normal persons who have been transfused with the cells of sickle cell anemia patients, whereas sicklemia erythrocytes have an normal life span (3, 14). The ability of the red cells to sickle was observed to have a genetic basis not long after sickle cell anemia 1 This investigation was supported in part by a grant from the U. S. Public Health Service. The study has been possible only through the generous cooperation of the Anemia Clinic of the Children's Hospital of Michigan, Detroit, Michigan, The University Hospital of the University of Michigan, Ann Arbor, Michigan, and the Wayne County General Hospital and Infirmary, Eloise, Michigan; all three institutions have made their case records of sickle cell anemia freely available. It is a pleasure to acknowledge my indebtedness to Mrs. Marion Weyrauch for technical assistance, and to Mrs. Laura Williams for case work. was recognized as a clinical entity (5). On the basis of a study of one large family, Taliaferro and Huck (15) postulated that the ability to sickle was due to a single dominant gene. At that time the clinical distinction between sicklemia and sickle cell anemia had not been clearly drawn, and the inference was that this gene was more strongly expressed in some individuals (sickle cell anemia) than in others (sicklemia). This has remained the accepted hypothesis up to the present time. Several years ago the author, in a review on the clinical detection of the genetic carriers of inherited disease (9), was led to suggest an alternative hypothesis-namely, that there existed in Negro populations a gene which in heterozygous condition results in sicklemlia, and in homozygous condition in sickle cell anemia. This hypothesis has a counterpart in the relationship which has been demonstrated to exist between thalassemia major and minor (10, 16). Recently the opportunity has arisen to give this hypothesis a thorough test. There exist a number of arguments permitting a critical decision between the two hypotheses. The present preliminary note will consider only one of these arguments. If the homozygous-heterozygotus hypothesis is correct, then both the parents of any patient with sickle cell anemia should always sickle (barring the occasional role of mutation; see below). If, on the other hand, the disease is due to a dominant gene with variable expression, only one parent need although occasionally, due to the chance marriage of two sicklers, both parents may sickle. In calculating the exact proportion of sicklemia to be expected among the parents of individuals with sickle cell anemia according to the dominant hypothesis, certain assumptions must be made. To the best of the author's knowledge, the question of the phenotype of the homozygote has never been raised by those who have accepted the variable dominant hypothesis of sickle cell anemia. For purposes of calculation we shall assume that under the variable dominant hypothesis all homozygotes have sickle cell anemiaalternative assumptions, such as intra-uterine lethality, are possible. We shall further assume that one in fifty heterozygotes also develops sickle cell anemia. Finally, we shall assume on the basis of the clinical data that the fertility of those with sickle cell anemnia approximates 20 percent of normal, with the result that only a few individuals with this disease-so few SCIENCE July 15, 1949, Vol. 1 10 64

Peer Reviewed