APOE Polymorphism and the Progression of Diabetic Nephropathy in Japanese Subjects With Type 2 Diabetes

Results of a prospective observational follow-up study

  1. Shin-ichi Araki, MD, PHD,
  2. Daisuke Koya, MD, PHD,
  3. Tetsuya Makiishi, MD,
  4. Toshiro Sugimoto, MD, PHD,
  5. Motohide Isono, MD, PHD,
  6. Ryuichi Kikkawa, MD, PHD,
  7. Atsunori Kashiwagi, MD, PHD and
  8. Masakazu Haneda, MD, PHD
  1. From the Department of Medicine, Shiga University of Medical Science, Otsu, Shiga, 520-2192, Japan
  1. Address correspondence and reprint requests to Masakazu Haneda, MD, Department of Medicine, Shiga University of Medical Science, Otsu, Shiga, 520-2192, Japan. E-mail: haneda{at}belle.shiga-med.ac.jp

Abstract

OBJECTIVE—The aim of this study is to clarify the conflicting results of the ε2/ε3/ε4 APOE polymorphism as a risk factor on diabetic nephropathy by a cohort study.

RESEARCH DESIGN AND METHODS—A total of 429 Japanese subjects with type 2 diabetes and with normoalbuminuria (n = 299) or with microalbuminuria (n = 130) were enrolled in a prospective observational follow-up study during 1995–1998 and followed until 2001 (for at least 3 years). The endpoint was the occurrence of a renal event defined as the progression to a higher stage of diabetic nephropathy.

RESULTS—During the study (the mean follow-up period: 4.4 ± 1.0 years), 31 of 429 subjects progressed: 21 from normoalbuminuria to microalbuminuria and 10 from microalbuminuria to overt proteinuria. The allele frequency of the APOE polymorphism was significantly different between the progressors and the nonprogressors. Eight of 42 ε2 carriers (19%) progressed, whereas 23 of 387 noncarriers (6%) progressed with a relative risk of 3.2 (95% CI 1.5–6.7). When subjects were stratified by renal status at baseline, each relative risk for the progression in the ε2 carriers was 2.7 (0.99–7.4) in those with normoalbuminuria and 4.2 (1.3–13.3) in those with microalbuminuria. Furthermore, when analyzed only in subjects with normoalbuminuria and short duration of diabetes (<15 years) at baseline, the risk in the ε2 carriers became higher to 3.2 (1.2–8.8).

CONCLUSIONS—Our follow-up study indicates that the ε2 allele of the APOE polymorphism is a prognostic risk factor for both the onset and the progression of diabetic nephropathy in Japanese type 2 diabetes.

Diabetic nephropathy (DN) associated with type 2 diabetes is a leading cause of end-stage renal disease (ESRD) in Japan and the U.S. (1,2). Although long-standing hyperglycemia is an important risk factor for the development of DN, epidemiologic and family studies suggest that genetic susceptibility to this complication is also required (3). The abnormalities of lipid metabolism have been proposed as one plausible mechanism in the pathogenesis of the development of DN (4). Thus, the genes encoding components of this pathway that regulate lipid metabolism can be considered as candidate genes susceptible to DN (5).

Apolipoprotein E (apoE) is a 299-amino acid glycoprotein that plays a central role in lipid metabolism. ApoE is well known to consist of three common isoforms (E2, E3, E4) encoded by three alleles (ε2, ε3, ε4) in exon 4 of the apoE gene (APOE) (6). The lipid profiles or atherogenic factors are, in part, determined by this genetic variation of APOE (6). Carriers of apoE4 have higher plasma levels of total cholesterol and low-density lipoprotein cholesterol than carriers of apoE2 or apoE3 homozygotes (6). On the other hand, the binding of apoE2 to lipoprotein receptors is defective in comparison with that of E3 or E4 and results in delayed clearance of triglyceride (TG)-rich lipoprotein.

Recently, this ε2/ε3/ε4 polymorphism of APOE has been investigated for the association with DN in subjects with both type 1 and type 2 diabetes using a case-control comparison, but the findings were inconclusive (715). Very recently, we reported that the ε2 allele of APOE increases the risk of DN in type 1 diabetes in a large number of case-control comparisons and a family-based study called the transmission/disequilibrium test (16). The latter method is bias free and is the most reliable study design for detecting an association between DNA sequence differences and a specific disease (17). Although our previous result provided stronger evidence of the association between the APOE polymorphism and DN than other case-control studies, a question whether the diabetic subjects with the ε2 allele of APOE actually develop DN remains to be solved, because all previous studies were cross-sectional association studies. Thus, the role of the ε2 allele as the prognostic factor for DN should be confirmed by a cohort study. Also, the role of this polymorphism on the onset of microalbuminuria, which is a powerful predictor of the late development of persistent overt proteinuria and ESRD in subjects with type 2 diabetes (18,19), has not been established in previous studies.

The aim of this study is to investigate the effect of the APOE polymorphism as the prognostic risk factor for the development of DN. For this purpose, we carried out a prospective observational follow-up study in Japanese subjects with type 2 diabetes and without overt proteinuria.

RESEARCH DESIGN AND METHODS

Study population and examination

For a prospective observational follow-up study, subjects were recruited from patients who were regularly cared for at the outpatient clinic of Third Department of Medicine, Shiga University of Medical Science during 1995–1998. A total of 759 subjects were clinically diagnosed with type 2 diabetes. On the basis of multiple measurements of the albumin excretion rate (AER) in a 24-h urine sample collection, patients were classified into the following categories regardless of duration of diabetes: normoalbuminuria, microalbuminuria, and overt proteinuria. Among them, 83 patients were excluded because their nephropathy stage could not be ascertained and 1 because of a nondiabetic renal disease. Also, 91 patients with overt proteinuria at entry were not included because this study was designed to analyze the progression of patients with normoalbuminuria or microalbuminuria. After consenting to participate in the study, each subject underwent a standardized physical examination and provided a diabetes history regarding its diagnosis, treatment, and the occurrence of complications. Each individual provided a blood sample for biochemical measurements and DNA extraction. A total of 135 patients were not enrolled at the time because they refused to participate, they had chronic disease unrelated to diabetes (such as malignancy, liver cirrhosis), or the APOE polymorphism was not genotyped. Finally, 449 patients (312 with normoalbuminuria and 137 with microalbuminuria) were enrolled in this follow-up study. The participants, during the follow-up periods, underwent the standardized physical examination, biochemical measurements, and a measurement of AER in a 24-h urine collection at least once a year. The follow-up lasted at least 3 years until the end of 2001 or death. All participants received treatment based on the standardized strategies for diabetes, hypertension, and hyperlipidemia during the follow-up periods.

The study protocol and informed consent procedure were approved by the Ethics Committee of Shiga University of Medical Science.

Diagnosis of DN and retinopathy

The DN status of each patient was determined based on the AER levels measured by immunoturbidimetry assay (HITACHI 7070E, Hitachi High-Technologies, Tokyo, Japan) in 24-h urine samples as follows after the elimination of urinary infection. Patients were classified as: normoalbuminuria if AER was <20 μg/min, microalbuminuria if AER was ≥20 μg/min and <200 μg/min, and overt proteinuria if AER was ≤200 μg/min in two consecutive measurements. In the follow-up study, the progressors on DN were defined as the subjects who shifted to a higher stage of DN from that at the baseline. An ophthalmologist made the diagnosis of diabetic retinopathy. Diabetic retinopathy was defined as background or more.

Genotyping of the APOE polymorphism

Genomic DNA was extracted from peripheral lymphocytes using a phenol/chloroform method. Fragments containing the polymorphic site were amplified by polymerase chain reaction methods in a total 25 μl of volume with 20 ng of genomic DNA according to a previous described method (16). The genotypes at a CfoI restriction fragment length polymorphism site in exon 4 of APOE (APOE ε2/ε3/ε4) were determined.

Statistical analysis

The study groups were compared by χ2 tests for frequencies or Fisher’s exact test if the frequencies were small (StatView; SAS Institute, Cary, NC). Student’s t test was used for continuous variables. Relative risks for the progression of DN and associated CI were calculated by comparing the risk in the ε2 carriers to risk in the noncarriers. Odds ratios and their CI were used to estimate relative risk in a multiple logistic regression analysis (20). Multiple logistic regression analysis was used to assess the prognostic effect of the ε2 carriers of the APOE polymorphism for the progression of DN. The following independent variables were used: APOE genotype (indicator variable for the ε2 allele carriers), total cholesterol (mg/dl), TG (mg/dl), BMI (kg/m2), and retinopathy (indicator variable for the presence of retinopathy). The ε2/ε3/ε4 polymorphism consists of the combination of two single nucleotide polymorphisms in exon 4 of APOE (6). Thus, Hardy-Weinberg tests were performed by a standard observed-expected χ2 test at each polymorphic site. A value of P < 0.05 was taken to be statistically significant.

RESULTS

Of 449 participants in the prospective observational follow-up study, 20 patients were lost during the study period: one with the ε4ε4 genotype died due to acute myocardial infarction, and 19 had less than 3 years for the follow-up period (ε2ε3/ε3ε3/ε3ε4: n = 2/14/3). Finally, the remaining 429 patients (299 with normoalbuminuria and 130 with microalbuminuria) were used for the analysis. Selected clinical characteristics of these 429 subjects are summarized in Table 1. Sex, duration of diabetes, the levels of HbA1c, systolic blood pressure (sBP), diastolic blood pressure (dBP), BMI, TG, and the proportion of diabetic retinopathy at the baseline in diabetic subjects with microalbuminuria were higher than in those with normoalbuminuria, but age and the levels of total cholesterol were not. The baseline clinical characteristics and the distribution of the APOE polymorphism were similar between the lost subjects and the remainder (data not shown).

The mean follow-up period was 4.4 ± 1.0 years (range, 3–6 years). During the study period, 31 of 429 subjects progressed to a higher stage of DN (the progressors, 7.2%): 21 of 299 with normoalbuminuria to microalbuminuria and 10 of 130 with microalbuminuria to overt proteinuria. No patients with normoalbuminuria at the baseline progressed to overt proteinuria. The clinical characteristics of the progressors and the nonprogressors at the baseline are shown in Table 2. Sex, age, duration of diabetes, the levels of HbA1c, sBP, dBP, total cholesterol, and TG at the baseline were similar between the two groups. The levels of BMI at the baseline were significantly higher in the progressors than in the nonprogressors. During the follow-up period, the mean levels of sBP, TG, and BMI were higher in the progressors than in the nonprogressors (Table 2). As expected, the proportion of diabetic retinopathy at the baseline and during the follow-up period was higher in the progressors than in the nonprogressors.

The genotype and allele frequencies of the APOE polymorphism are presented in Table 3. Genotype distributions of the APOE polymorphism in our population were in agreement with the Hardy-Weinberg equilibrium. Also, allele frequencies were similar to those that were previously reported in Japanese healthy control subjects (n = 1,090, ε2/ε3/ε4, 5%/86%/9%; 21). At the baseline, the genotype distribution of the APOE polymorphism was significantly different between the subjects with normoalbuminuria and those with microalbuminuria but not the allele frequency, indicating that the overall genotype distribution varies between two groups. In the follow-up study, the genotype and allele frequencies in the progressors were significantly different from those in the nonprogressors (Table 3). As shown in Table 4, 8 of 42 subjects with the ε2 allele (19%) progressed to a higher stage of DN, whereas only 23 of 387 without the ε2 allele (6%) progressed (χ2 = 9.71 with 1 degree of freedom, P = 0.0018). Relative risk for the progression in the ε2 carriers in comparison to the noncarriers was 3.2 (95% CI 1.5–6.7). When subjects were stratified by renal status at the baseline, each relative risk for the progression in the ε2 carriers was 2.7 (0.99–7.4) in those with normoalbuminuria and 4.2 (1.3–13.3) in those with microalbuminuria. In this study, subjects with normoalbuminuria were recruited regardless of duration of diabetes. Subjects with long duration of diabetes are considered to have a low risk at the onset of DN. Thus, only those with normoalbuminuria and short duration (<15 years) were analyzed. In 206 patients who fulfilled these criteria, the risk for the onset of microalbuminuria in the ε2 carriers significantly became higher to 3.2 (1.2–8.8). In a multiple logistic regression analysis, the risks of the progressors for the ε2 carriers were not changed by adjustment for the levels of total cholesterol, TG, BMI, and the presence of retinopathy at the baseline (Table 4). Also, stratification by sex, levels of HbA1c, and the presence of diabetic retinopathy did not change results (data not shown).

CONCLUSIONS

This prospective observational follow-up study provides evidence that the ε2 allele of the APOE polymorphism is a valuable prognostic factor for DN in Japanese subjects with type 2 diabetes. The risk in the ε2 carriers was threefold higher than the noncarriers in the mean 4-year follow-up period. Furthermore, this study shows that the ε2 allele is an independent risk factor for both the onset and the progression of DN.

Our findings regarding the APOE polymorphism as the genetic risk factor for DN are consistent with the results of previous case-control studies in Japanese subjects with type 2 diabetes (7,8). Eto et al. (8) reported that the carriers of the ε2 allele had an odds ratio of 3.0 (95% CI 1.2–7.7) for DN (135 control and 146 case subjects). The value was very similar to the relative risk obtained in the present study. The effect of the APOE polymorphism on the risk of DN has also been examined in Caucasian subjects with type 1 diabetes. Chowdhury et al. (10) reported that the presence of the ε2 allele was associated with increased risk of DN (197 control and 252 case subjects). In our previous case-control comparison (including 196 control and 223 case subjects), the risk of DN was 3.1 times higher in the ε2 carriers than in noncarriers (16). However, several studies, even in Japanese subjects with type 2 diabetes, did not confirm this association (1215). Kimura et al. (12) did not find the association between the ε2 carriers and ESRD in a total of 178 Japanese patients with type 2 diabetes. Also, Hadjadj et al. (15) examined 494 subjects with type 1 diabetes and various stages of DN and found that the distribution of the APOE genotypes did not differ according to the stage of DN.

Explanations for the discrepant results of the case-control studies are unknown. To confirm these conflicting results from the case-control comparison, one may consider performing other reliable methods, such as the family-based study or the cohort study. Recently, we have reported the positive association between the ε2 allele and DN in Caucasian subjects with type 1 diabetes in the family-based study as well as the case-control comparison (16). That family-based study demonstrated the preferential transmission of the ε2 allele from heterozygous parents for the ε2 allele to DN offspring (64% transmitted and 36% nontransmitted). The present follow-up study also indicates that the ε2 carriers more frequently progressed to the high stage of DN than the noncarriers, although the number of the progressors in the ε2 carriers was small. Therefore, we consider that the ε2 allele of the APOE polymorphism is one of the valuable risk factors for the development of DN, which account for a limited proportion of cases.

Previous case-control studies were not investigated regarding which stages of DN were influenced by the APOE polymorphism. This study showed that the ε2 carriers had high relative risk for the progression from microalbuminuria to overt proteinuria. In addition, the carriers of the ε2 allele in the subjects with normoalbuminuria and short duration of diabetes (<15 years) had a risk of 3.2 (95% CI 1.2–8.8), which was similar to that in the subjects with microalbuminuria. Recently, Werle et al. (11) reported that the ε2 allele was a positive predictor of the level of urinary albumin excretion in type 1 diabetic patients. Therefore, the ε2 allele of the APOE polymorphism is at increased risk for both the onset and the progression of DN.

The mechanism by which the APOE polymorphism influences the development of DN remains unclear at present. One possibility is the lipid abnormalities related to the APOE polymorphism. Recently, Eto et al. (22) reported that remnant lipoproteins associated with the ε2 allele may have an important role in the development of DN. Although we did not measure the remnant lipoproteins in our study, this hypothesis must be confirmed in a longitudinal study. A second possibility is a direct effect of the apoE protein to renal mesangial cells. ApoE2 isoform has been reported to have less protective effect on the growth factor-induced mesangial proliferation in comparison with apoE3 or E4 (23). Thus, the ε2 carriers may have less autocrine protective effects on renal function in diabetic conditions than other alleles.

In conclusion, the present follow-up study indicates that the ε2 allele of the APOE polymorphism is one of the prognostic risk factors involved in the development of DN in Japanese subjects with type 2 diabetes. Right now, no treatment strategy for the ε2 carriers of the APOE polymorphism to prevent the onset and the progression of DN has been established. Further studies should be required to clarify the practical implications such as therapeutic intervention.

Table 1—

Selected clinical characteristics at baseline according to DN status

Table 2—

Selected clinical characteristics at baseline and during the follow-up periods according to the progression of DN

Table 3—

Genotype distribution and allele frequency of the APOE polymorphism at baseline and during the follow-up periods according to the progression of DN

Table 4—

Relative risk and adjusted odds ratio (OR) of DN for carriers of the ε2 allele according to the progression of nephropathy

Acknowledgments

This work was supported by Grant-in-Aid for Scientific Research on Priority Areas (C) “Medical Genome Science” from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Footnotes

  • A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

    • Accepted April 25, 2003.
    • Received January 15, 2003.

References

| Table of Contents