(1991) - Control of Expression of Insulin Resistance and Hyperglycemia by Different Genetic Factors in Diabetic C57BL-6J Mice. Surwit R. S. Diabetes 40, 82-87 PDF
Inheritance of the tendency to develop diet-induced non-insulin-dependent (type II) diabetes was analyzed in crosses between diabetes-prone C57BL / 6J (BL / 6) mice and diabetes-resistant a / J mice. Results suggest that diet-induced hyperglycemia is largely determined by a recessive gene.
Original Description:
Original Title
(1991). Control of expression of insulin resistance and hyperglycemia by different genetic factors in diabetic C57BL-6J mice. Surwit R. S. Diabetes 40, 82-87..pdf
Inheritance of the tendency to develop diet-induced non-insulin-dependent (type II) diabetes was analyzed in crosses between diabetes-prone C57BL / 6J (BL / 6) mice and diabetes-resistant a / J mice. Results suggest that diet-induced hyperglycemia is largely determined by a recessive gene.
(1991) - Control of Expression of Insulin Resistance and Hyperglycemia by Different Genetic Factors in Diabetic C57BL-6J Mice. Surwit R. S. Diabetes 40, 82-87 PDF
Inheritance of the tendency to develop diet-induced non-insulin-dependent (type II) diabetes was analyzed in crosses between diabetes-prone C57BL / 6J (BL / 6) mice and diabetes-resistant a / J mice. Results suggest that diet-induced hyperglycemia is largely determined by a recessive gene.
by Different Genetic Factors in Diabetic C57BL/6J Mice RICHARD S. SURWIT, MICHAEL F. SELDIN, CYNTHIA M. KUHN, CHRISTINA COCHRANE, AND MARK N. FEINGLOS The inheritance of the tendency to develop diet- induced non-insulin-dependent (type II) diabetes was analyzed in crosses between diabetes-prone C57BL/6J (BL/6) mice and diabetes-resistant A/J mice. The effects of a diabetogenic diet on blood glucose and insulin levels, insulin sensitivity, and weight were evaluated in F, and both (BL/6 x A/J) F, x BL/6 and (BL/6 x A/J) F, x A/J backcross mice. These results suggest that diet-induced hyperglycemia is largely determined by a recessive gene and diet-induced insulin resistance by a dominant gene. Analyses of both backcrosses indicated that insulin sensitivity and blood glucose levels were unrelated, suggesting that they are controlled by different genetic factors. This conclusion was supported by data from nine recombinant inbred BXA strains in which no correlation was observed between these variables. Furthermore, insulin sensitivity and body weight correlated differently in the two backcross groups, suggesting that insulin resistance is not simply a function of obesity. The number of genes that predominantly influence diabetic traits was estimated by comparing the variance observed in (BL/6 x A/J) F t x BL/6 backcross mice with that observed in parental mice. The data suggest that relatively few genes predominantly affect the diabetic phenotype in this murine model. Diabetes 40:82-87,1991 U nderlying genetic factors are probably critical in the pathogenesis of non-insulin-dependent (type II) diabetes mellitus. Although studies of identical twins have shown 95% concordance of disease (1) and abnormal glucose tolerance in as many as 30% of relatives of patients with type II diabetes (2), little is known From the Departments of Psychiatry, Medicine, and Pharmacology, Duke University Medical Center, Durham, North Carolina. Address correspondence and reprint requests to Dr. Richard S. Surwit, Box 3842, Duke University Medical Center, Durham, NC 27710. Received for publication 19 February 1990 and accepted in revised form 8 August 1990. about the specific inherited defects. In certain populations, a form of type II diabetes appears to be inherited in a com- plex fashion, with insulin resistance and impaired pancreatic activity segregating independently (3,4). Linkage studies have failed to demonstrate a clear association of these traits with any genetic locus, including those of the gene encoding either insulin or insulin receptors (5). Animal models of type II diabetes are amenable to genetic analysis. Inbred mouse strains are of particular interest be- cause several forms of type II diabetes exist in inbred strains, which provide large numbers of genetically identical indi- viduals (6). Finally, the mouse genome has been extensively mapped, and much of it is conserved in humans, which suggests that identification of a genetic locus or region in the mouse genome may facilitate human genetic analyses (7). Several mutations in C57BL/6J (BL/6) mice have resulted in type II diabetes-like syndromes (8), suggesting to us that this strain may carry many of the genes necessary for this disease (9). Consistent with this hypothesis, studies in our laboratory demonstrated that the BL/6 mouse but not the A/J mouse will develop an analogue of type II diabetes if weaned onto a high-fat high-simple-carbohydrate diet and allowed to become obese (9). In this article, we describe the metabolic characteristics of an FT cross of diabetes-prone BL/6 mice and diabetes- resistant A/J mice raised on the diabetogenic diet and those of the backcross of the F, onto the BL/6 and A/J strains and a few recombinant inbred (Rl) strains. The distribution of these values supports the hypothesis that different genetic factors influence the expression of diet-induced insulin re- sistance and hyperglycemia. RESEARCH DESIGN AND METHODS Twenty BL/6 mice, 20 A/J mice, 20 F, crosses (BXA), and 55 backcrosses on the BL/6 parent strain (BL/6 x A/J fe- male) F, x BL/6 were obtained at 1 mo of age from Jackson (Bar Harbor, ME). A second group of 44 (BL/6 x A/J) F, x A/J backcross mice were bred at Duke University from parental stock obtained from Jackson. Finally, 5 animals from 82 DIABETES, VOL. 40, JANUARY 1991
R.S. SURWIT AND ASSOCIATES each of Rl BXA strains 8, 10, 11, 13, 19, 24, and 25 and 4 each from strains 7 and 17 were obtained from the laboratory of M. Nesbitt (University of California, San Diego). All animals studied were male. Note that Rl strains are derived from crosses between two inbred progenitor strains in which the mice from the F 2 generation and each subsequent genera- tion are mated in accordance with a strict inbreeding pro- gram. After 20 generations, each of the resultant Rl strains has a unique contribution from each original progenitor in which the alleles from either progenitor strain have become fixed at each locus. Thus, observations may be made on different individuals, each of which has an identical geno- type. On their arrival at the laboratory, half of the animals of each strain were placed on a high-fat high-simple-carbo- hydrate diet previously shown to induce type II diabetes in BL/6 mice (9). The diet contained 20.5% protein, 35.8% fat, 0.4% fiber, 3.6% ash, 3.1% moisture, and 36.8% carbohy- drate (primarily disaccharides; diet 1850, Bio-Serve, French- town, NJ). Typically, BL/6 animals fed this diet for 3 mo show fasting plasma glucose >11.11 mM and plasma insulin >898.2 pM. The rest of the animals were maintained on standard Purina Rodent Chow (Ralston-Purina, Richmond, IN). This diet contained 23% protein, 4.5% fat, 6% fiber, 8% ash, and 56% complex carbohydrate. All animals were housed in the vivarium at Duke University in group cages, with five animals to each cage. They were given food and water ad libitum and maintained on a 12-h light-dark diurnal cycle. All animals were weighed weekly. Once per month, beginning with mo 2, all animals were fasted for 8 h, after which blood was sampled for serum glucose (assayed by Beckman autoanalyzer) and serum in- sulin (assayed by a Cambridge Medical Diagnostics ra- dioimmunoassay). After 4.5 mo, five animals of each parental strain and all backcrosses were given an insulin-sensitivity test. Animals were fasted for 8 h before testing and then given an injection of regular pork insulin (0.5 lU/kg i.p.; Lilly, Indianapolis, IN). Blood was sampled at 15 and 30 min. Blood samples were analyzed for serum glucose as described above. Insulin sen- sitivity (K|) was measured as the slope of the fall in glucose over time after the injection of insulin. This technique, when applied to humans, has been shown to give results com- parable to those obtained with glucose-insulin-clamp tech- niques (10). All blood samples were collected by retro-orbital sinus puncture. Data from the 5th mo were used to calculate estimates of genetic dominance for fasting glucose, insulin, insulin sen- sitivity (mo 4.5), and body weight. These data were also used to calculate the degree to which these traits segregated together and the number of genetic loci that determined each trait. Month five was selected because previous studies have shown that diabetes is fully developed in these mice at this time (9). To analyze the genetics of diet-induced diabetes, diabe- tes-prone BL/6 mice were crossed with diabetes-resistant A/J mice. The mice were given either a control or diabeto- genic diet as previously described (9). We then compared the parental strains to the F, on both control and experimental diets to determine the dominance of each of the above meta- bolic characteristics. Dominance {h) of the BL/6 genotype was estimated by the formula (11) mean BL/6 x A/J F, - mean A/J h = mean BL/6 - mean A/J Variance (var) of h was estimated with the following formula (kindly provided by B. Weir, Dept. of Statistics, North Carolina State Univ., Raleigh). For the quantity h = X] X2 AQ AD where X, is the mean BL/6 x A/J F,, X 2 is the mean A/J, and X 3 is the mean BL/6, an approximate expression for the variance is var(h) = var(X,)+ var(X2)+ ^ - - x- var(X,) X, - X 3 Y {A3 A2) / var(X2) x, - x 2 Y \A$ A2J / var(X3) where d represents the derivative of the function. Note that there are no covariances, because the three Xs are found from different samples. To further characterize the relation- ship of genetic factors that result in the diabetic phenotype, the F, mice were backcrossed with the diabetes-prone BL/6 parent and diabetes-resistant A/J parent. The analysis of correlation in genetically heterogeneous populations (backcross mice) allows us to assess the degree to which the same genetic factors control phenotypic expression of the various traits, provided that environmental factors are not as important as genetic factors in determining disease. The method of Wright (11), after modifications suggested by Lande (12) and Cockerham (13), was used to estimate the number of genes responsible for each biological property measured between mo 4 and 5. The following formulas were applied to data obtained from each parent, F,, and (BL/6 x A/J) F, x BL/6 backcross to determine the mean SD number of effective loci (S,) in which the F, differs from the BL/6 parent. Note that these methods are imprecise and serve as estimates of the number of loci that predominantly affect each trait, because many additional genes may make minor contributions that are quantitatively important. For as- sumptions and limitations of these procedures, see refs. 11- 13. S, = (mean BL/6 - mean F,) 2 2as oy 2 is the additive genetic variance stemming from differences among gene frequencies of the two parental populations us 2 = 2a 2 BC, - 1 /6(7aBL/6 2 + crA/J 2 4<rF, 2 ) where BC, = (BL/ 6 x A/ J) F, x BL/ 6 backcross mice. The SD of S, was cal cul ated as , , /4(a 2 BL/6/n BL , 6 + q Fl 2 /n Fl ) var q s CTs ^' ' \ (mean BL/6 - mean F,) 2 a s 4 DIABETES, VOL. 40, JANUARY 1991 83
GENETIC FACTORS IN DIABETIC C57BL/6J MICE where var a s 2 = ( 3a BL/ 6 2 qA/ J 2 2a s 8n B and n is the number of mice in each group. Similar analysis with data from the reciprocal backcross (BL/6 x A/J) F, x A/J to determine S 2 was not performed because parental strains were not run simultaneously. RESULTS Effect of diabetogenic diet on cross between diabetes- prone and diabetes-resistant strains. The mo-5 values for glucose, insulin, and weight and mo-4.5 values for /<, are shown in Fig. 1. BL/6 mice fed the control diet had slightly higher glucose values than the A/J or F, mice given this diet, but insulin levels, K u and body weight were the same as those of A/J and F, animals. The experimental diet raised glucose and insulin levels and body weight of all animals, but this effect was much more pronounced for the diabetes- prone BL/6 mice. Although the effect of the diabetogenic diet on insulin levels and body weight became more pro- nounced during the 5 mo of analyses, hyperglycemia was evident after 1 mo (data not shown). The FT mice on the diabetogenic diet showed glucose and insulin levels and body weights that were intermediate be- tween the parental strains; fasting insulin and body weights were significantly lower than those of BL/6 mice and signif- icantly higher than those of A/J mice (Fig. 1). The diabeto- genic diet greatly reduced the K, of BL/6 mice but did not affect ft, of A/J mice. In contrast, the K x values for F, animals on the experimental diet were less than those of A/J mice but not significantly greater than those of BL/6 mice. At mo 5, the mean SD h of the BL/6 genotype for blood glucose was estimated to be 0.19 0.09, suggesting that this characteristic is recessive. Mean SD h for insulin- sensitivity was measured at 0.71 0.21, suggesting that this trait is controlled mostly by dominant genes. Thus, hy- perglycemia and insulin sensitivity appear to be inherited differently. Correlation analysis of genetic and nongenetic factors influencing different diabetic characteristics. The regres- sion values for four dependent variables are listed for both backcross groups in Table 1. In only one case was the cor- relation in one backcross significantly different from that in the other backcross. Although K, was significantly correlated with body weight in the (BL/6 x A/J) F, x A/J backcross (r = 0.44, P < 0.01), it was not correlated at all in the 1000- BL/ 6 FIG. 1. Metabolic parameters of C57BL/6J (BL/6), A/J, and (BL/6 x A/J) F,, mice on diabetogenic and control diets. Fasting serum glucose (F,) fasting serum insulin, and body weight after 5 mo and insulin sensitivity (K,) after 4.5 mo in BL/6, A/J, and F, mice on control (open bars) or diabetogenic (hatched bars) diet. Results are expressed as means SE (n = 8-10/group for fasting serum glucose, fasting serum insulin, and body weight; n = 4-5/group for insulin sensitivity). There was significant effect of diet, strain, and diet x strain interaction by analysis of variance allowing following pairwise comparisons. Diet had significant effect on glucose for BL/6 and A/J (P < 0.02 or better), on insulin and body weight for all strains (P < 0.001), and on insulin sensitivity for BL/6 and F, mice only (P < 0.05 or better). On control diet, F, animals differed from BL/6 mice for body weight only and from A/J mice for glucose and body weight. On diabetogenic diet, F, animals differed from BL/6 mice in glucose, insulin, and body weight and from A/J mice in all except glucose. P < 0.05 or better for latter probabilities. 84 DIABETES, VOL. 40, JANUARY 1991
R.S. SURWIT AND ASSOCIATES TABLE 1 Correlation coefficients of diabetic parameters at mo 5 in backcross C57BL/6J (BL/6) mice on diabetogenic diet BG and BW IRI and BW BG and IRI K, and BW K, and BG K, and IRI (BL/6 x A/J) F, x BL/6 55 (BL/6 x A/J) F, x A/J 44 (BL/6 x A/J) F, x BL/6 55 (BL/6 x A/J) F, x A/J 41 0.10 0.28 0.26 0.45 0.43 0.14 0.00 0.44* 0.17 0.10 0.26 0.47 BG, blood glucose; BW, body weight; IRI, immunoreactive insulin; K u insulin sensitivity. *P < 0.01 such that r, ^ r 2 for crosses (BL/6 x A/J) F, x BL/6 vs. (BL/6 x A/J) F, x A/J. (BL/6 x A/J) F, x BL/6 backcross (r = 0). These corre- lations were significantly different (P < 0.01) from each other (14). This suggests that the presence of animals homozy- gous for the BL/6 genotype in the (BL/6 x A/J) F, x BL/6 backcross adds significant variability to the relationship be- tween insulin resistance and body weight above that attrib- utable to the environment. Thus, insulin resistance is not simply a result of obesity; the degree of insulin resistance depends on genetic background. Another notable finding is the lack of correlation between K, values and blood glucose. In both backcross groups, the relationship between these variables appears to be random. This is in contrast to a strong correlation observed in the BL/6 parental group (r = 0.96). The probability that the correlation observed in the BL/6 parental group is different from the observed correlations in either backcross group is P< 0.05 (14). Analyses of hyperglycemia and insulin resistance in Rl strains. To further investigate the inheritance of genes pre- disposing to the diabetic phenotype, nine BXA Rl strains were examined. Because of the few animals per strain, glu- cose values were averaged over mo 3, 4, and 5. K, values were determined at mo 4.5. Mice from each of these strains are homozygous at each locus, but each strain contains a unique contribution of the two progenitor strains. There is no relationship between serum glucose values and K, values (Fig. 2). Rank-sum correlations between K, values and serum 11 8 17 25 7 13 10 19 (BXA) Rl Strains 24 FIG. 2. Serum glucose and insulin sensitivity (K,) of recombinant inbred (Rl) strains on diabetogenic diet. Mean SE serum glucose (open bars) and K, {hatched bars) values in 9 C57BL/6J x A/J (BXA) Rl strains. Blood glucose values are averaged over mo 3, 4, and 5. Glucose data are based on 15 observations for strains 8, 10, 11, 13, 24, and 25, with 14 observations for strain 19,12 for strain 7, and 11 for strain 17. Insulin-sensitivity determinations were made at mo 4.5. n = 5 mice/group except for strains 11 and 17 (n = 4). glucose or insulin levels were not significant (data not shown). These data support the conclusion that insulin re- sistance and hyperglycemia are determined by different ge- netic factors in this mouse model. Analyses of genetic factors influencing diabetic pheno- type with backcross mice. Glucose, insulin, and weight data of all strains measured at mo 5 and K, values at mo 4.5 are presented in Table 2, and the individual glucose values are depicted in Fig. 3. On the diabetogenic diet, diabetes-prone BL/6 and dia- betes-resistant A/J animals were >3SD apart for each pa- rameter except AC,. Furthermore, differences in variance between parental and F, mice were small, indicating the genetic homogeneity of parental populations and the rela- tively small environmental variance. Therefore, the large vari- ance of backcross values relative to the parental and F, mice indicates genetic heterogeneity of the backcross mice. A clear bimodal or trimodal distribution of glucose (Fig. 3), insulin (data not shown), and K, (data not shown) was not observed. However, with the method of Wright (11) as adapted and modified by Lande (12) and Cockerham (13), the number of major genetic factors (S,) for which the dia- betes-prone BL/6 mouse differed from the relatively dia- betes-resistant F, mouse was calculated. S, was <1 for glucose, log glucose, and insulin; 4.1 for weight; and 1.10 for log insulin. Similarly, S, for /<, was 0.06 (calculated at mo 4.5). These data suggest that relatively few genes are prob- ably responsible for differences in glucose, insulin, insulin sensitivity, and weight observed between F, mice and (BL/6 x A/J) F, x BL/6 backcross mice maintained on the diabetogenic diet. DISCUSSION These data allow us to draw several conclusions about the inheritance of the tendency to develop type II diabetes in the BL/6 mouse. Of primary importance is that diet-induced insulin resistance and hyperglycemia are controlled by dif- ferent genetic factors. Three lines of evidence allow us to reach this conclusion. First, in both backcross groups, fast- ing glucose and insulin sensitivity do not correlate, indicating that insulin resistance and hyperglycemia segregate inde- pendently. Second, the differential dominance of blood glu- cose levels and insulin sensitivity in the F, crosses show that, although insulin resistance is inherited in a largely dominant fashion, hyperglycemia appears to be a recessive trait. Third, and most conclusively, hyperglycemia and insulin sensitivity segregated independently in Rl strains. Thus, at least two distinct genetic factors are necessary for the appearance of type II diabetes characteristic of the BL/6 mouse. Unfortu- nately, the physiological basis for hyperglycemia or insulin DIABETES, VOL. 40, JANUARY 1991 85
GENETIC FACTORS IN DIABETIC C57BL/6J MICE TABLE 2 Serum glucose, log glucose, insulin, log insulin, insulin sensitivity, and body weight for experimental animals at mo 5 Glucose (mM) BL/6 A/J (BL/6 x A/J) F, (BL/6 x A/J) F, x BL/6 Log glucose BL/6 A/J (BL/6 x A/J) F, (BL/6 x A/J) F, x BL/6 Insulin (pM) BL/6 A/J (BL/6 x A/J) F, (BL/6 x A/J) F, x BL/6 Log insulin BL/6 A/J (BL/6 x A/J) F, (BL/6 x A/J) F, x BL/6 Insulin sensitivity (/<,) BL/6 A/J (BL/6 x A/J) F, (BL/6 x A/J) F, x BL/6 Body weight (g) BL/6 A/J (BL/6 x A/J) F, (BL/6 x A/J) F, x BL/6 Mean SD 12.5 1.3 8.3 1.1 9.1 0.8 12.3 2.8 0.130 0.130 0.120 0.120 0.122 0.002 0.128 0.005 893.6 128.3 187.9 59.6 401.4 188.3 985.7 377.1 14.286 0.487 9.989 0.924 12.164 1.417 15.000 1.900 -2. 10 1.09 -5. 17 1.17 -3. 04 1.15 -0. 60 2.18 49.5 4.0 35.2 4.3 41.1 3.8 50.5 4.5 Variance 1.7 1.3 0.7 6.1 6.98 x 10" 6 1.0 x 10" 5 3.4 x 10" 6 2.3 x 10" 5 16556.8 3520.3 34789.8 150844.2 0.227 0.869 2.045 2.759 1.18 1.38 1.33 4.79 16.1 18.2 14.6 20.0 n 10 9 10 55 10 9 10 55 10 8 10 55 10 8 10 55 5 4 5 55 10 10 10 55 S, SD 0.65 0. 18 0.95 0.36 0.49 0. 10 1.10 0.37 0.06 0.06 4.10 2.49 BL/6 = C57BL/6J. S, represents an estimate of the number of major genetic factors for which the diabetes-prone BL/6 mouse differs from the relatively diabetes-resistant F, mouse. The formula for calculation of S, is given in METHODS. resistance in the BL/6 mouse remains uncertain. As in hu- man type II diabetes, hyperglycemia is accompanied by hyperinsulinemia in our mouse model (15). However, subtle defects in pancreatic function, such as impaired first-phase insulin release, are characteristic of BL/6 mice (16). Although we cannot immediately generalize these findings to the phenomenon of human diabetes, several studies have suggested that insulin sensitivity and impaired pancreatic activity may be separate phenomena, which do not have to occur together in human type II diabetes (17-20). Lillioja et al. (17) found that impaired glucose tolerance in Pima In- 1 0 ' 20 30 animal number 50 FIG. 3. Individual serum glucose values for (BL/6 x A/J) F, x BL/6 backcross mice after 5 mo on diabetogenic diet. Note, BL/6 is our abbreviation for C57BL/6J. dians was due to insulin resistance, whereas the develop- ment of frank hyperglycemia was associated with a defect in pancreatic function. Furthermore, in the Pima population, insulin resistance appears to be determined by one codom- inant gene, whereas an additional recessive gene is thought to be responsible for hyperglycemia (18). Eriksson et al. (19) showed that even insulin sensitivity alone could not explain impaired glucose tolerance in a Scandanavian population; a defect in first-phase insulin release had to be present as well. Finally, Banerji and Lebovitz (20) described two forms of type II diabetes, one in which insulin sensitivity is normal and a significant pancreatic defect exists and another in which there is a mild impairment of glucose-stimulated in- sulin secretion accompanied by significant insulin resist- ance. Reaven (21) has suggested that insulin resistance is a very common problem, and therefore cannot be by itself the cause of type II diabetes. Our data support the conclu- sion that impaired insulin sensitivity does not cause hyper- glycemia in the BL/6 mouse and that the development of diabetes in this animal model is primarily dependent on some other genetic factor. Another conclusion that can be reached from our data is that insulin sensitivity is not simply a function of body weight. The different correlations between insulin sensitivity and weight observed in the two backcross groups suggest that insulin sensitivity is dependent in part on genetic back- ground. Thus, insulin resistance may be inherited indepen- dent of obesity. Studies in several human populations have suggested similar conclusions. Lillioja et al. (17) reported that the familial component of insulin action and hyperin- 86 DIABETES, VOL. 40, JANUARY 1991
R.S. SURWIT AND ASSOCIATES sulinemia occurs in addition to the effects of obesity in Pima Indians, and Haffner et al. (22) found a similar independence of genetic and environmental effects in Mexican Americans. Finally, estimation of the number of genes that differentiate the F, from the BL/6 strain suggests that relatively few genes are primarily responsible for the manifestation of diet-in- duced insulin resistance and diet-induced hyperglycemia. These estimates suggest that the diabetes of the BL/6 mouse is amenable to genetic analysis. Future studies of Fa- generation crosses of BL/6 and A/J mice will provide further precision in estimating the number of genes responsible for hyperglycemia and insulin resistance. Furthermore, gene- mapping studies with available Rl lines may allow identifi- cation of one or more genetic loci that are important in the pathogenesis of this murine model of type II diabetes and perhaps eventual identification of the human homologues. ACKNOWLEDGMENTS This research was supported by a grant from the American Diabetes Association, the John D. and Catherine T. Mac- Arthur Foundation Research Network on the Consequences and Determinants of Health Promoting and Health Damaging Behavior, and by Research Scientist Development Award K01-MH-00303 from the National Institute of Mental Health. M.F.S. is a Charles E. Culpeper Foundation Scholar; the work was supported in part by this foundation. We are indebted to Dr. Muriel Nesbitt for providing BXA strains, Dr. Bruce Weir for valuable discussions on estimating dominance and the number of genes for a quantitative char- acter, and Dr. Marc K. Drezner for critical review of the manu- script. Some of the research contained in this article was pre- sented at the scientific sessions of the 50th annual meeting of the American Diabetes Association, Atlanta, Georgia, 16- 19 June 1990. REFERENCES 1. Barnett AH, Eff C, Leslie RDG, Pyke DA: Diabetes in identical twins: a study of 200 pairs. Diabetologia 17:333-43, 1979 2. Kobberling J: Studies on the genetic heterogeneity of diabetes mellitus. Diabetologia 7:46-49, 1971 3. Lillioja S, Mott DM, Howard V, Bennett PH, Hannele YJ, Freymond D, Nyomba BL, Zurlo F, Swinburn B, Bogardus C: Impaired glucose toler- ance as a disorder of insulin action: longitudinal and cross-sectional studies in Pima Indians. N Engl J Med 318:1217-24, 1988 4. O'Rahilly S, Turner RC, Matthews DR: Impaired pulsatile secretion of insulin in relatives of patients with non-insulin-dependent diabetes. N Engl J Med 318:1225-30, 1988 5. O'Rahilly S, Wainscoat JS, Turner RC: Type 2 (non-insulin-dependent) diabetes mellitus: new genetics for old nightmares. Diabetologia 31:407- 14,1988 6. Hedrich HJ: Genetic monitoring. In The Mouse in Biomedical Research. Vol. 1. Foster HL, Small JD, Fox JG, Eds. New York, Academic, 1981, p. 159-76 7. Searle AG, Peters J, Lyon MF, Evans EP, Edwards JH, Buckle VW: Chro- mosome maps of man and mouse. Genomics 1:3-18, 1987 8. Coleman DL: Obese and diabetes: two mutant genes causing diabetes- obesity syndromes in mice. Diabetologia 14:141-48, 1978 9. Surwit RS, Kuhn CM, Cochrane C, McCubbin JA, Feinglos MN: Diet- induced type II diabetes in C57BL/6J mice. Diabetes 37:1163-67, 1988 10. Bonora E, Mogehetti P, Zancanaro C, Cigolini M, Querena M, Cacciatori M, Corgnati V, Muggeo A: Estimates of in vivo insulin action in man: comparison of insulin tolerance tests with euglycemic and hyperglycemic glucose clamp studies. J Clin Endocrinol Metab 68:374-78, 1989 11. Wright S: The genetics of quantitative variability. Evolution and the Ge- netics of Populations: A Treatise in Three Volumes. Genetic and Biometric Foundations. Vol. 1, chapt. 15. Chicago, IL, Univ. of Chicago Press, 1968, p. 373-420 12. Lande R: The minimum number of genes contributing to quantitative variation between and within populations. Genetics 99:541-53, 1981 13. Cockerham CC: Modifications in estimating the number of genes for a quantitative character. Genetics 114:659-66, 1986 14. Cohen J: Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ, Erlbaum, 1988 15. Niskanen LK, Uusitupa Ml, Sarlund H, Siitonen O, Pyorala K: Five-year follow-up study on plasma insulin levels in newly diagnosed NIDDM pa- tients and nondiabetic subjects. Diabetes Care 13:41-48, 1990 16. Kaku K, Fiedorek FT Jr, Province M, Permutt MA: Genetic analysis of glucose tolerance in inbred mouse strains: evidence for polygenic control. Diabetes 37:707-13, 1988 17. Lillioja S, Mott DM, Zawadzki JK, Young AA, Abbott WGH, Knowler WC, Bennett PH, Moll P, Bogardus C: In vivo insulin action is familial char- acteristic in nondiabetic Pima Indians. Diabetes 36:1329-35, 1987 18. Bogardus C, Lillioja S, Nyomba BL, Swinburn B, Esposito-del Puente A, Knowler WC, Ravussin E, Mott DM, Bennett PH: Distribution of in vivo insulin action in Pima Indians as mixture of three normal distributions. Diabetes 38:1423-32, 1989 19. Erikkson J, Franssila-Killunki A, Estrand A, Saloranta C, Widen S, Schalin C, Groop L: Early metabolic defects in persons at increasing risk for non- insulin-dependent diabetes mellitus. N Engl J Med 321:339-43, 1989 20. Banerji MA, Lebovitz HE: Insulin-sensitive and insulin-resistant variants in NIDDM. Diabetes 38:784-92, 1989 21. Reaven GM: Banting lecture 1988: role of insulin resistance in human disease. Diabetes 37:1595-607, 1988 22. Haffner SM, Stern MP, Hazuda HP, Mitchell BD, Patterson JK: Increased insulin concentrations in nondiabetic offspring of diabetic parents. N Engl J Med 319:1297-301, 1988 DIABETES, VOL. 40, JANUARY 1991 87