Apolipoprotein E allele frequencies in non-insulin-dependent diabetes mellitus with hypertriglyceridemia (type IIb, III, IV, and V hyperlipoproteinemia)

Apolipoprotein Mellitus

E Allele> Frequencies in Non-Insulin-Dependent Diabetes With Hypertriglyceridemia (Type IIb, III, IV, and V Hyperlipoproteinemia) Masaaki Eto, Kiyoshi Watanabe, lsao Makino, and Kaneo lshiit

We examined apolipoprotein E (apo E) allele frequencies in non-insulin-dependent diabetes mellitus (NIDDM) patients with normolipidemia or various types of hypertriglyceridemia to elucidate the association of the apo E alleles with hypertriglyceridemia in NIDDM. NIDDM patients with normolipidemia (N = 134) or hypertriglyceridemia [type Ilb hyperlipoproteinemia (HLP) (N = 42). Ill HLP (N = 71, IV HLP (N = 96), and V HLP (N = 811 were randomly selected from our Diabetic Clinics. Apo E phenotypes (genotypes) were determined by our rapid flat-gel isoelectric focusing method. The frequency of the E4 allele was significantly higher in the type Ilb (20.2%, P < .Ol) and V (25.0%. P < .05) HLP patients than in the normolipidemic patients (8.9%). whereas the frequency of the ~3 allele was significantly (P < .025) lower in the type Ilb HLP patients (78.6%) than in the normolipidemic patients (89.2%). The frequency of the ~2 allele was significantly higher in the type Ill (64.3%. P < ,001) and IV (5.2%, P < .05) HLP patients than in the normolipidemic patients (1.9%). whereas the frequency of the l3 allele was significantly lower in the type Ill (28.6%. P < .OOl) and IV (82.8%, P < .05) HLP patients than in the normolipidemic patients. Thus, it has proven that the 4 allele is related to type Ill and IV HLP in NIDDM, whereas the ~4 allele is related to type Ilb and V HLP in NIDDM. It is concluded that the ~4 allele, as well as the a allele, may, at least in part, contribute to hypertriglyceridemia and probably atherosclerosis in NIDDM. Copyright 0 1991 by W.B. Saunders Company

A

POLIPOPROTEIN E (apo E) alleles (~2, ~3, and ~4) are located at a single genetic locus and code for the three major apo E isoproteins, E2, E3, and E4, respectively, and determine six apo E phenotypes (genotypes): E2/2(~2/ ~2), E3/2(~3/~2). E3/3(~3/~3), E4/2(~4/~2). E4/3(~4/~3), and E4/4(~4/~4), which are distinguished by isoelectric focusing.‘,’ Apo E, a protein constituent of very-low-density lipoprotein (VLDL) and high-density lipoprotein (HDL), plays an important role in lipoprotein metabolism through its ability to bind to the receptors.’ Since Utermann et al’.“ first found that the homozygote apo E2/2 is uniquely related to type III HLP. there is growing evidence that the apo E alleles (~2 and ~4) may be associated with hyperlipoproteinemia (HLP). Even the heterozygotes apo E3/2 and E4/2 have been reported to be related to type III HLP’ ’ or dyslipoproteinemia similar to type III HLP.‘.“’ In addition, it has been shown that apo E2 (~2) contributes to type IV HLP.‘,” On the other hand, apo E4 (~4) has been found to be related to type IIa and IIb HLP’ or type V hypercholesterolemia,“~” HLP.‘,’ Hypertriglyceridemia, which is thought to be a risk factor for coronary heart disease,‘” is a common feature in patients with non-insulin-dependent diabetes mellitus (NIDDM).” We first found that NIDDM patients with the apo E3/2 phenotype have a higher frequency of hypertriglyceridemia and higher levels of plasma triglyceride (TG). VLDL-TG, and apo E, and a higher VLDL-TG to VLDL-cholesterol

From the Second Department of Internal Medicine. Asahikawa Medical College. Asahikawa, Japan. Supported in part by a Grant for Specific Diseases from the Primar) Hyperlipidemia Research Committee, the Ministry of Health and Welfare, Japan; and by a grantfrom Sumitomo Pharmaceutical Co. fDeceased. Address reprint requests to Masaaki Eto, MD, Second Depatiment of Internal Medicine, Asahikawa Medical College, .4sahikawa 078, Japan. Copyright 0 1991 by WB. Saunders Cornpan) 00260495/91/4008-0002$03.00l0 776

(chol) ratio than NIDDM patients with the most common apo E313 phenotype.‘” Recently, Imari et al’- also have reported that the ~2 allele is significantly more frequent in NIDDM patients with hypertrigiyceridemia. These data suggest a close association of apo E2 (~2) with hypcrtriglyceridemia in NIDDM. However, hypertriglyceridemia includes type IIb, III, IV. and V HLP. It remains unclear which type of HLP is associated with apo E2 (~2) in NIDDM. It may be etiologically important to elucidate the association of the apo E alleles with various types of hypertriglyceridemia in NIDDM. In this study, we thercfore compared the apo E allele frequencies between NIDDM patients with normolipidemia and hypertriglyceridemia (type IIb, III, IV, and V HLP). MATERIALS

AND METHODS

Subjects Two hundred

eighty-seven

NIDDM patients aged 35 to 64 years who had normolipidemia (N = 134) or hypertriglyceridemia (plasma TG level 2 150 mg/dL. type IIb HLP [N = 411, III HLP [N = 71, IV HLP [N = 961. and V HLP [N = 81). were randomly selected from our Diabetic Clinics at Asahikawa Medical College Hospital and the affiliated hospitals. which were representative in the Asahikawa district. Japan. All the NIDDM patients including the 60 patients treated with insulin were not insulin-dependent and not ketotic, and NIDDM was diagnosed based on the criteria proposed by National Diabetes Data Group.‘” Clinical characteristics of the patients are summarized in Table I. No patients in this study were related. All NIDDM patients had normal hepatic, endocrine and renal (plasma creatinine < I.6 mg/dL, urine protein <(I.5 g/d) functions. In addition, our patients did not have familial hypercholesterolemia. which was diagnosed by clinical symptoms and family history because it is caused by another genetic factor of the defect of low-density lipoprotein (LDL) receptors independent of apo E alleles. NIDDM patients except type III and V HLP were relatively well controlled (HbA, < IO%) and below 12.5% ideal body weight (IBW), but HbA, and IBW were not taken into consideration when selecting the patients with type III and V HLP because of their small sample sizes. During the 2 weeks before blood sampling. no (181 men and 106 women),

Mefabolrsm,

Vol 40, No 8 (August),

1991:

pp 776.780

APO E ALLELES AND HYPERTRlGLYCERlDEMlA

777

Table 1. Clinical Characteristics of NIDDM Patients With Normolipidemia

and Hypertriglyceridemia Hypertriglyceridemia

Normolipidemia

N(M/F)

Ilb HLP

134 (82/52)

42 (21/21)

Ill HLP

IV HLP

V HLP

7 l4/3)

96 (66/30)

8 (8/O)

Mode of therapy Diet

55 (35/20)

14 (8/6)

1 (l/O)

44 (32/12)

5 (5/O)

Oral agent

51 (32/19)

20 (8/l 2)

3 (l/2)

31 (20/l 1)

3 (3/O)

Insulin

28 (15/13)

8 (5/3)

3 (2/l)

21 (14/7)

0

Age (vr)

53 + 2

53 * 1

50 + 3

54 2 2

51 +4

IBW (%)

107 + 1

115 + 2*

118 f 3t

112 * 1*

120 + 4t

Fasting plasma glucose (mg/dL)

131 24

141 +6

162 + 139

137 -t 3

HbA, (%)

9.0 ‘- 0.1

9.2 2 0.2

Duration (yr)

5.3 ‘- 0.1

6.9 * 1.2

Triglyceride (mg/dL)

98 ? 2

207 2 109

370 + 41

227 ? 90

863 + 1050

Total chol (mg/dL)

178 t 2

266 f 51

295 + 159

118 + 2

195 2 4§

256 + 215 -

193 + 46

LDL-chol (mg/dL)

119r3

106 + 13

HDL-chol (mg/dL)

46 f 1

41 + 1*

36 + 21

11.8 2 0.70 8.2 + 2.6

18Q? 29f

9.2 f 0.2

11.1 + l.O§

5.9 + 0.6

382

3.3 + 0.9*

l§

35 + 31

NOTE. Values are means ? SEM. Abbreviation: IBW, ideal body weight.

lP < .05, tP < .Ol, SP < ,005, §P < ,001 versus normolipidemia.

medication

known to influence

and antidiabetic

the lipid state, oral agent. was administered.

except

for insulin

Measurement of Plasma Lipids, Lipoproteins, and Other Parameters Blood samples were collected in 1 mg/mL EDTA from an antecubital vein after an overnight fast, and plasma was separated by centrifugation. Plasma glucose, TG and total chol were measured enzymatically. HbA, was measured by a microcolumn method or a high-performance liquid chromatography (HLC-723GHb, Toyo Soda, Tokyo, Japan). The VLDL fraction (d < 1.006) was separated by ultracentrifugation.“’ Plasma samples were centrifuged in a Hitachi Model 65P-7 ultracentrifuge (Hitachi Koki, Tokyo, Japan) with an RP-65T rotor (Hitachi koki) at 43,000 rpm and 10°C for 20 hours. The top VLDL fraction was collected, and TG (VLDL-TG) and chol (VLDL-chol) were measured enzymatitally. HDL-chol was determined by the phosphotungstic acid and magnesium precipitation method or the microliter scale ultracentrifugal method, as described previously.‘” LDL-chol was calculated by subtracting VLDL-chol and HDL-chol from total chol. The intraassay and interassay coefficients of variation for the above parameters were all less than 6%.

Tris-HCl buffer (pH 8.6) with 8 mol/L urea and 10 mmol/L dithiothreitol. Apo E phenotypes were then determined by our rapid flat-gel isoelectric focusing method, as described in detail elsewhere.” When it was difficult to distinguish the homozygote and the heterozygote (ie, apo E3/3 and E312, or apo E4/4 and E4/3) in our focusing system (approximately 5% of total samples), we could distinguish it using the methods of sialidase treatment and/or two-dimensional electrophoresis of apo VLDL, as reported previously.” Briefly, after mixing equal volumes of VLDL samples and 100 mmol/L acetate buffer (pH 5.0) and the addition of 25 uL of 3.3 UimL sialidase in acetate buffer, incubation was carried out at 37°C for 1 hour. The samples were then delipidated and subjected to isoelectric focusing. Two-dimensional isoelectric focusingsodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed. Apo E allele frequencies were calculated by the gene counting method.

StatisticalAnalysis All values are expressed as the mean 2 SEM. Comparisons of apo E phenotype and allele frequencies between the patients with

Determination of HLP HLP and normolipidemia were diagnosed according to the following criteria, which are common in Japan and described in detail previously’: type IIb HLP: TG t 150 mg/dL and LDLchol 2 165 mg/dL; type III HLP (which was diagnosed by the modified method of Mabuchi et al19): TG > 150 mgidL, VLDLcho1 2 55 mg/dL, VLDL-chol/VLDL-TG ratio 2 0.36, and the presence of a broad P-band on polyacrylamide disc gel electrophoresis’” (Fig 1); type IV HLP: TG 2 150 mg/dL, LDL-chol < 165 mg/dL, VLDL-chol < 55 mg/dL, VLDL-chol/VLDL-TG ratio < 0.36, and no chylomicron band; type V HLP: TG t 700 mg/dL, the presence of chylomicron band, and no evidence of postheparin lipoprotein lipase deficiency and apo CII deficiency; normolipidemia: TG < 150 mg/dL and LDL < 165 mg/dL.

Determination of Apo E Phenovpes The isolated VLDL was delipidated by the method of Warnick et al.” After drying, apo VLDL was solubilized in 10 mmol/L

VLDL LDL

Fig 1. Polyacrylamide disc gel electrophoretic patterns” of type Ilb, ill, and IV HLP. Type Ill HLP is characterized by the presence of a broad p-band between VLDL and LDL bands. VLDL and LDL bands are strong in type Ilb HLP, and VLDL band is strong in type IV HLP.

HDL

(+I IV

Ill

Ilb

778

ET0 ET AL

creased frequencies of apo EZ-present phenotypes. The frequency of the ~4 allele was significantly higher in the type IIb (x’ = 7.92, P < .Ol) and V (x’ = 4.24, P < .05) HLP patients than in the normolipidemic patients, whereas the frequency of the ~3 allele was significantly (xl = 6.25, P < .025) lower in the type IIb HLP patients. The frequency of the ~2 allele was significantly higher in the type III (xl = 109.00. P < .OOl) and IV (x’ = 3.96. P < .05) HLP patients than in the normolipidemic patients. whereas the frequency of the ~3 allele was significantly lower in the type III (x’ = 42.00, P < .OOl) and IV (x’ = 3.89. P < .05) HLP patients.

normolipidemia and various types of hypertriglyceridemia in NIDDM were made by the chi-square test. Comparisons of the mean values of parameters were made by ANOVA. Multiple comparisons were performed. A value of P < .05 was considered significant. RESULTS

As shown in Table 1, no difference was noted in age between the NIDDM patients with normolipidemia and different types of hypertriglyceridemia in NIDDM. The type IIb, III, IV, and V HLP patients showed significantly increased IBW as compared with the normolipidemic patients. There was no significant difference in fasting plasma glucose and HbA, levels between the normolipidemic and the type IIb and IV HLP patients, but the type III and V HLP patients showed significantly increased levels of fasting plasma glucose and HbA, as compared with the normolipidemic patients. Sex ratio, mode of therapy, and duration of diabetes are also shown in Table 1. It is reasonable that all NIDDM patients with hypertriglyceridemia had significantly higher levels of plasma TG than the normolipidemic NIDDM patients, and that the type IIb HLP patients had significantly higher levels of plasma LDL-chol. In addition, total chol levels were significantly higher, and HDL-chol levels were significantly lower in the patients with hypertriglyceridemia than those with normolipidemia. Table 2 shows apo E phenotype and allele frequencies. Apo E allele and phenotype frequencies were comparable in NIDDM patients with normolipidemia and general Japanese population (N = 576), as we previously reported.‘3 There were significant differences in apo E phenotype frequencies between the normolipidemic and the type IIb (x’ = 12.0, P < .005) and III (x’ = 93.94, P < .OOl) HLP patients. Type IIb (38.1%) and V (37.5%) HLP in NIDDM were characterized by increased frequencies of apo E4-present phenotypes, whereas type III (100%) and IV (10.4%) HLP in NIDDM were characterized by in-

DISCUSSION

In addition to the previous observation that the t2 allele contributes to hypertriglyceridemia in NIDDM.‘“,” we examined the apo E allele frequencies in various types of hypertriglyceridemia in NIDDM. It has proven that the t? allele is closely associated with type IV HLP, as well as type III HLP in NIDDM, whereas the ~4 allele is also closely associated with type IIb and V HLP in NIDDM, although our sample size is not large (N = 287). This finding suggests that the association of the apo E alleles with hypertriglyceridemia in NIDDM differs among various types of HLP, and therefore these types should be taken into consideration to elucidate the association of the apo E alleles with hypertriglyceridemia in NIDDM. In nondiabetic subjects the ~4 allele has been reported to have an LDL-elevating effect in normolipidemia.” ” to be associated with mixed hyperlipidemia (increases of TG and chol),” and with hypercholesterolemia alone,‘” although most investigators did not examine this association by the type of HLP. We previously reported that the frequency of the ~4 allele is increased in type IIa and IIb HLP in nondiabetics’ and obese subjects,” and that heterozygous familial hypercholesterolemic patients with the ~4 allele showed higher frequency of type IIb HLP.” Thus, this study

Table 2. Apo E Phenotype and Allele Frequencies in NIDDM Patients With Normolipidemia

and Hypertriglyceridemia,

and in the

General Japanese Population NIDDM NIDDM

N

0

E312

5 (Z/l/Z)

E313

107 (45/41/21)

E412

0

Japanese Ilb HLP

III HLP

%

N

%

N

%

0

0

0

2 (l/l/O)

28.6

3.7

0

0

4 (O/2/2)

57.1

79.9

26 (7/14/5)

61.9

0

1 (l/O/O)

2.4

N

0 9 (212151

0

0

2

0.3

0

0

35

6.1

5 (2/3/O)

62.5

414

71.9

0

0

0

1 (0/1101

1.0 20.9

65 (30/24/l 1)

33.3

0

0

20 (I Z/3/5)

E414

2 (l/l/O)

1.5

1 (O/l/O)

2.4

0

0

1 (O/l/O)

p < ,005

0 9.4

14.3

14 (6/5/3)

Significance v normolipidemia

N

0

14.9

E4

67.7

1.0

Q/o

2 (Z/O/O) 25.0 1 (l/O/O)

12.5

N

%

4

0.7

111

19.3

10

1.7

p < ,001

1.9

1.2

64.35

5.2”

89.2

78.6t

28.61

82.8”

75.0

84.6

8.9

20.2*

7.1

12.0

25.0”

11.7

NOTE. Parentheses give number of patients by mode of therapy (diet/oral agent/insulin). *P < .05, tP < ,025, tP < .Ol, §P < ,001 versus corresponding apo E allele of NIDDM patients with normolipidemia. llQuoted from our previous repwIZ3

PopulationD

%

1 (O/O/l)

20 (7/8/5)

t3

V HLP

IV HLP

-

E4/3

Apo E alleles E2

General

Patients

With Normolipidemia

Apo E phenotype E2/2

-___

Patients With Hypertriglycerldemla

0

3.7

APO E ALLELES AND HYPERTRIGLYCERIDEMIA

779

also suggests that hypertriglyceridemia in NIDDM that is accompanied by increased LDL-chol, (ie, type IIb HLP) is closely associated with the ~4 allele. This association of the ~4 allele with type IIb HLP in NIDDM seems not only through LDL-chol, because some reports showed a close relationship between the ~4 allele and type V HLP5,’ and hypertriglyceridemia.“~” In addition, it is unlikely that the ~2 allele contributes to type IIb HLP, because of the finding that the ~2 allele has a chol-lowering effect.“.24 We have already reported that NIDDM makes the l4-carrying subjects more susceptible to type IIa HLP, as well as type IIb HLP.‘* Seven NIDDM patients with type III HLP consisted of two with apo E2/2, four with apo E3/2, and one with apo E4/2. None of these subjects had evidence of apo E deficiency according to the isoelectric focusing method. All had the ~2 allele. Thus, the ~2 allele is closely related to type III HLP also in NIDDM. Of type III HLP NIDDM patients, 71.4% showed the heterozygotes, apo E3/2 or E4/2, and this observation supports our previous data29 that diabetes makes even the heterozygotes more susceptible to type III HLP. Janus et a13’ proposed the nomenclature homozygous (classical apo E2/2) and heterozygous (apo E3/2 or E4/2) type III HLP, and found that the former had a higher ratio of VLDL-chol to TG than the latter. This study showed that this nomenclature may be available in NIDDM and that plasma TG and apo E levels and the VLDL-chol to VLDL-TG ratio were higher in homozygous type III NIDDM than in heterozygous type III NIDDM (data not shown). Furthermore, four of the seven type III HLP patients were poorly controlled (fasting plasma glucose level > 150 mgidL). We have recently reported that poor glycemic control predisposes the e2-carrying NIDDM patients to type III HLP.29 An abnormal lipoprotein receptor-binding activity of apo E2 causes impaired TG-rich lipoprotein remnant catabolism and accumulation of remnant particles (type III HLP).” More increased remnants in plasma in NIDDM with the ~2 allele may be the result of overloading of TG-rich lipoproteins due to overproduction and/or decreased removal in diabetic state (particularly poor metabolic control).‘5 On the other hand, the frequency

of the ~2 allele (apo E3/2 and E4/2) was also increased in NIDDM with type IV HLP. That type IV HLP occurred in NIDDM patients with the heterozygotes for the ~2 may be explained by an intermediate increase of plasma remnants and TG-rich lipoproteins, as reported by Utermann et al.’ There are some reports’.” that the ~2 allele is associated with type IV HLP in nondiabetes. In this study, the ~4 allele was more frequent in type V HLP (increases of chylomicron and VLDL) in NIDDM. The development of type V HLP involves metabolic derangements, including the overloading of TG aggravated by diabetes and obesity, as described in type III HLP. Previous including us,’ reported that the ~4 allele is investigators, 5~32.33 related to type V HLP and that type V HLP patients had a higher frequency of diabetes (40%; 37%,’ 43%,32 and 52%33). In this study, the NIDDM patients with type V HLP showed poor control and increased IBW (Table l), but all three type V HLP patients with the ~4 allele showed good control (fasting plasma glucose < 122 mg/dL) and not increased IBM (mean, 111%). Thus, although the sample size is small, it is possible that the ~4 allele is related to type V HLP in NIDDM, independently of poor control and obesity. Stuyt et a134denied the association of the ~4 allele with type V HLP. Further studies are needed in this respect. We could not find any NIDDM patient with type I HLP, which shows severe hypertriglyceridemia caused by lipoprotein lipase deficiency. Only relatively well-controlled patients were studied, except type III and V HLP, to avoid the effects of diabetic control on plasma lipid levels.‘” The ~2 and ~4 alleles were also related to HLP in obesity,‘5,27and thus it is possible that the slightly increased IBW observed in type IIb, III, IV, and V HLP may be related to HLP, but the details are not known. It is concluded that the ~4 allele, as well as the ~2 allele, may, at least in part, contribute to hypertriglyceridemia in NIDDM, and that the ~2 allele is related to type III and IV HLP, whereas the ~4 allele is related to type IIb and V HLP. Both ~2 and ~4 alleles should be, therefore, considered as the etiology of hypertriglyceridemia and probably atherosclerosis’4.35 in NIDDM.

REFERENCES

1. Utermann G, Hees M, Steinmetz A: Polymorphism of apolipoprotein E and occurrence of dysbetalipoproteinemia in man. Nature 269:604-607, 1977 2. Zannis VI, Breslow JL: Human very low density lipoprotein E isoprotein polymorphism is explained by genetic variation and posttranslational modification. Biochemistry 20:1033-1041, 1981 3. Mahley RW, Innerarity TL: Lipoprotein receptors and cholesterol homeostasis. Biochim Biophys Acta. 737:197-222, 1983 4. Utermann GM, Jaeschke M, Menzel HJ: Familial hyperlipoproteinemia type III: Deficiency of a specific apolipoprotein (apo E-III) in the very-low-density lipoproteins. FEBS Lett 56:352355,1975 5. Ghiselli G, Gregg RE, Zech LA, et al: Phenotype study of apolipoprotein E isoforms in hyperlipoproteinemic patients. Lancet 2:405-407,1982 6. Breslow JL, Zannis VI, SanGiacomo TR, et al: Studies of familial type III hyperlipoproteinemia using as a genetic marker the apoE phenotype E212. J Lipid Res 23:1224-1235,1982

7. Eto M, Watanabe K, Ishii K: Apolipoprotein E alleles and hyperlipoproteinemia in Japan. Clin Genet 34:246-251,1988 8. Utermann G, Pruin N, Steinmetz A: Polymorphism of apolipoprotein E. III. Effect of a single polymorphic gene locus on plasma lipid levels in man. Clin Genet 15:63-72,1979 9. Kameda K, Matsuzawa Y, Kubo M, et al: Increased frequency of lipoprotein disorders similar to type III hyperlipoproteinemia in survivors of myocardial infarction in Japan. Atherosclerosis 51:241249,1984 10. Eto M, Watanabe K, Chonan H, et al: Apolipoprotein E4/2 phenotype and hyperlipoproteinemia-A report of four cases. J Jpn Atheroscler Sot 14:737742,1986 11. Sing CF, Davignon J: Role of the apolipoprotein E polymorphism in determining normal plasma lipid and lipoprotein variation. Am J Hum Genet 37:268-285, 1985 12. Utermann G, Kindermann I, Kaffarnik H, et al: Apolipoprotein E phenotypes and hyperlipidemia. Hum Genet 65:232-236, 1984

780

ET0 ET AL

13. Assmann G, Schmitz G. Menzel HJ, et al: Apolipoprotein E polymorphism and hyperlipidemia. Clin Chem 30:641-643, 1984 14. Fontbonne A. Eschwege E. Cambien F. et al: Hypertriglyceridemia as a risk factor of coronary heart disease mortality in subjects with impaired glucose tolerance or diabetes. Diabetologia 32:300-304, 1989 15. Howard BV: Lipid metabolism Res 28:613-620, 1987

in diabetes

mellitus.

J Lipid

16. Eto M. Watanabe K, Iwashima Y, et al: Apolipoprotein E polymorphism and hyperlipemia in type II diabetics. Diabetes 35:1374-1382.1986 17. Imari Y, Koga S, Ibayashi H: Phenotypes of apolipoprotein E and abnormalities in lipid metabolism in patients with noninsulin dependent diabetes mellitus. Metabolism 12:1134-1138. 1988 18. National Diabetes Data Group: Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 28:1039-1957. 1979 19. Mabuchi H. Ueda K, Ueda R, et al: Studies on the diagnosis of type III hyperlipoproteinemia. A suggestion of chemical criteria based on VLDL-cholesterol and VLDL-cholesterol/VLDL-triglyceride. Jpn J Geriatr 14:50-59, 1977 20. Narayan KA, Creinin HL, Kummerow FA: Disc electrophoreis of rat plasma lipoproteins. J Lipid Res 7:150-157, 1966 21. Warnick GR, Mayfield C, Albers JJ. et al: Gel isoelectric focusing method for specific diagnosis of familial hyperlipoproteinemia type 3. Clin Chem 25:279-284, 1979 22. Eto M, Watanabe K, Ishii K: A rapid flat gel isoelectric focusing method for the determination of apolipoprotein E phenotypes and its application. Clin Chim Acta 149:21-28, 1985 23. Eto M, Watanabe K, Ishii K: A racial difference in apolipoprotein E allele frequencies between the Japanese and Caucasian populations. Clin Genet 30:422-427, 1986 24. Eto M, Watanabe

K, Ishii K: Reciprocal

effects of apolipopro-

tein E alleles (t2 and ~4) on plasma lipid levels in normolipidemic subjects. Clin Genet ?9:477-484, 1986 25. Eto M, Watanabe K. Ishii K: Apolipoprotein E polymorphism and hyperlipoproteinemia in obesity. Int J Obesity 13:433440.1989 26. Eto M, Watanabe K, Chonan N, et al: Familial hypercholesterolemia and apolipoprotein E4. Atherosclerosis 72:123-128. 198X 27. Fumeron F. Rigaud D, Bertiere MC, et al: Association of apolipoprotein ~4 allele with hypertriglyceridemia in obesity. Clin Genet 34258-264. 1988 28. Eto M. Watanabe K. Iwashima Y, et al: Increased frequency of apolipoprotein ~4 allele in type II diabetes with hypercholesterolemia. Diabetes 36:1301-1306, 1987 29. Eto M. Watanabe K, Sato T. Makino I: Apolipoprotein E-7 and hyperlipoproteinemia in noninsulin-dependent diabetes mellitus. J Clin Endocrinol Metab 69:1207-1212. 19X9 30. Janus ED. Grant S. Lintott CJ. et al: Apolipoprotein E phenotypes in hyperlipidemic patients and their implications fol treatment. Atherosclerosis 57:249-266. 19X5 31. Schneider WJ. Kovanen PT. Brown MS. rt al: Familial dysbetalipoproteinemia: Abnormal binding of mutant apoprotein E to low density lipoprotein receptors of human fihroblasts and membranes from liver and adrenal of rats, rabbits and cows. J Clin Invest 6X:1975-1985. 1981 32. Sano R. Abe R. Oikawa S, et al: Apolipoprotein E phenotypes of normo- and hyperlipoproteinemia in Japanese. Tohoku J Exp Med 154:297-303. 1988 33. Kuusi T. Taskinen MR, Solakivi T. et al: Role of apolipoproteins E and C in type V hyperlipoproteinemia. J Lipid Res 29:293-79x. 198X 34. Stuyt PM, Stalenhoef AFM, Demacker PNM. et al: Hyperlipoproteinemia type V and apolipoprotein E4. Lancet 2:934, 1982 35. Eto M. Watanabe K. Makino I: Increased frequencies of apolipoprotein t2 and ~4 alleles in patients with ischrmic heart disease. Clin Genet 36:1X3-1X8. 1989