212

© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 212–215

Introduction

Paraoxonase is a glycoprotein, bound to high-density lipo-proteins (HDL) which prevents oxidative modification oflow-density lipoproteins (LDL)

in vitro

[1]. An association

with cardiovascular disease has been found for polymor-phisms of the

PON1

gene [2,3] and more recently for a poly-morphism in the

PON2

[4]. Diabetic microvascular diseaseshares many risk factors with macrovascular disease, includinglipid dis-orders [5,6] and oxidative stress [7].

We have previously found a strong association between theMet54Leu polymorphism of

PON1

and diabetic retinopathyin adolescents with Type 1 diabetes [8]. This study wasdesigned to investigate potential association of

PON1

and

PON2

polymorphisms with diabetic retinopathy and micro-albuminuria in a larger group.

Correspondence to

: Dr Kim Donaghue, Ray Williams Institute of Paediatric Endocrinology, Diabetes and Metabolism, The Children’s Hospital at Westmead, Locked Bag 4001, Westmead NSW 2145, Australia. E-mail: kimd@chw.edu.au

Abstract

Background

Paraoxonase is a serum enzyme, which prevents oxidation of low-density lipoprotein (LDL) by hydrolyzing lipid peroxides. Two polymorphismsin

PON1

gene have been associated with cardiovascular and microvascular dis-eases in both diabetic and non-diabetic patients.

Aims

The current project was designed to investigate the association betweenthe polymorphisms of two

PON

genes and diabetes microvascular diseases(retinopathy and microalbuminuria) and any potential linkage betweenMet54Leu of

PON1

and Cys311Ser of

PON2

gene.

Methods

Diabetic retinopathy and albumin excretion rate were assessedin 372 adolescents with Type 1 diabetes who were genotyped for the twopolymorphisms.

Results

We confirmed the increased susceptibility for diabetic retinopathy forthe Leu/Leu genotype (odds ratio (OR) 3.34 (confidence interval (CI) 1.95,5.75),

P

< 0.0001). The Ser/Ser genotype was significantly more common inthose patients with microalbuminuria (albumin excretion rate

≥

20

µ

g /min)compared with those with albumin excretion rate < 20

µ

g /min (OR 4.72 (CI2.65, 8.41),

P

< 0.0001). The Ser311 of

PON2

was in strong linkage disequilib-rium with Leu54 of

PON1

gene (

∆

= 23

×

10

4

,

P

< 0.001). The delta value washigher for those without complications (28

×

10

4

,

P

< 0.001) compared withthose with complications (15.5

×

10

4

,

P

< 0.001).

Conclusions

This study supports the hypothesis that diabetic microangiopathyis genetically heterogeneous.

PON1

Leu/Leu increases the risk for retinopathyand

PON2

Ser/Ser increases the risk for microalbuminuria.

Diabet. Med. 19, 212–215 (2002)

Keywords

paraoxonase, microalbuminuria, adolescents, Type 1 diabetes,retinopathy

Blackwell Science LtdOxford, UKDMEDiabetic Medicine0742-3071Blackwell Science Ltd, 2002October 200119000000Original ArticleOriginal articleParaoxonase and diabetes

Y. Kao et al.

Paraoxonase gene cluster is a genetic marker for early microvascular complications in Type 1 diabetes

Y. Kao*, K. C. Donaghue*§, A. Chan*, B. H. Bennetts†§, J. Knight‡ and M. Silink*§

*Ray Williams Institute of Paediatric Endocrinology, Diabetes and Metabolism, †Department of Molecular Genetics and ‡Centre for Kidney Research, The Children’s Hospital at Westmead, Westmead, and §University of Sydney, Sydney, Australia

Accepted 24 August 2001

DME_660.fm Page 212 Thursday, March 7, 2002 8:39 AM

Original article

213

© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 212–215

Patients and methods

Patients

In total 372 adolescents with Type 1 diabetes were genotypedwhen they presented for routine diabetes complications screen-ing. Ninety-three percent were Caucasian. All participants gaveinformed consent and the Ethics Committee of the hospitalapproved the study.

A single observer blinded to the patient’s genotype gradedstereoscopic fundal photography of seven fields (obtained in369). The presence of background retinopathy was defined asany microaneurysm or haemorrhage [9]. The mean of threetimed overnight urine collections was obtained in 316 adoles-cents. Albumin was measured using a polyclonal radioimmuno-assay (Pharmacia AB, Uppsala, Sweden). Elevated albuminexcretion rate (AER) was examined at two levels: mean AER

≥

20

µ

g /min (microalbuminuria) and mean AER

≥

7.5

µ

g /min(which is greater than the 95th centile for Australian school-children [10]). Glycated haemoglobin was measured by thediamat HPLC assay (BioRad Laboratories, Hercules, CA, USA).The non-diabetic range for HbA

1c

is 4–6%.

Molecular analysis

DNA was extracted from peripheral leucocytes by a conven-tional phenol–chloroform method. For Cys311Ser, polymerasechain reaction (PCR) primers as described by Sanghera

et al

.were used [4]. An aliquot of 100 ng DNA was denatured at94

°

C for 5 min and then amplified for 35 cycles. Each cycleconsisted of denaturation at 94

°

C for 1 min, annealing at 50

°

Cfor 1.5 min and extension at 72

°

C for 2 min with a final exten-sion time of 7 min. The 262-bp PCR products were digestedwith restriction enzyme

Dde

I. The digested products were sep-arated by electrophoresis on 3.5% agarose gel and identified byethidium bromide staining. Genotyping for Met54Leu has beenpreviously described [8].

Statistical analysis

The software package of SAS (Version 6.12; SAS Institute,Cary, NC, USA; 1996) was used to analyse the data.

χ

2

test wasused to assess association between genotypes and each com-plication. Wilcoxon rank sum test was used for comparingcontinuous variables between the retinopathy and elevatedAER groups. The effect of genotype and biological variables oneach complication was assessed by logistic regression with oddsratios (OR) and 95% confidence intervals (CI) shown.

Results

Diabetes complications screening showed 46% had retino-pathy, 25% had AER

≥

7.5

µ

g /min and 13% had both abnorm-alities. Sixteen individuals (5%) had microalbuminuria.Differences between those with and without complications areshown in Tables 1 and 2.

For the

PON1

Met54Leu polymorphism, the frequency ofLeu/Leu (L/L) genotype was higher in those with retinopathy

than without retinopathy (52% vs. 23%) (Table 1). The OR ofLeu/Leu genotype for increased susceptibility for retinopathywas 3.34 (CI 1.95, 5.75) (Table 3), independent of the effectsof age, duration and cholesterol. Heterozygosity was lesscommon in those with retinopathy than in those withoutretinopathy (44% vs. 56%) (Table 1), suggesting that theM allele may be more protective than the susceptibility of theL allele. The OR of Met/Met genotype for reduced risk ofretinopathy was 0.17 (CI 0.062, 0.44) (Table 3).

There was no significant difference in

PON1

genotypefrequencies between patients with AER

≥

7.5

µ

g /min and thosewith AER < 7.5

µ

g /min (

χ

2

= 3.23,

P

= 0.20) or microalbu-minuria (

χ

2

= 3.95,

P

= 0.14) (Table 2).The frequency of Ser/Ser genotype was higher in those with

AER

≥

7.5

µ

g /min than in those with AER < 7.5

µ

g/min (50%vs. 19%), and in those patients with microalbuminuria thanin those without (56% vs. 25%) (Table 2). The OR of Ser/Serfor AER

≥

7.5

µ

g/min was 4.72 (CI 2.65, 8.41), independent ofthe effect of age, duration and their interaction. The OR ofSer/Ser for microalbuminuria was 4.41 (CI 1.50, 13.01),independent of the effect of duration (Table 3).

The genotype frequencies at both positions were inHardy–Weinberg equilibrium. Of those with the

PON2

Ser/Ser genotype, 74% also had the

PON1

Leu/Leu genotype,21% had the

PON1

Met/Leu and 5% the

PON1

Met/Metgenotype. Linkage disequilibrium was confirmed between the

PON1

Leu54 and the

PON2

Ser311 genotypes (

∆

= 23

×

10

4

,

P

< 0.001). The delta value was higher for those withoutcomplications (28

×

10

4

,

P

< 0.001) compared with thosewith complications (15.5

×

10

4

,

P

< 0.001).

Discussion

This study confirmed the previous association of

PON1

Leu/Leu with retinopathy in a larger group, identifying a nearly

Table 1 Demographic variables and genotyping of adolescents with and without retinopathy

No retinopathy (n = 198)

Retinopathy (n = 171) P-value

Gender 99 M, 99 F 71 M, 100 F 0.10Age (years) 13.0 [11.8–14.7] 14.8 [13.2–16.5] < 0.0001Duration (years) 4.1 [2.8–6.1] 7.6 [5.8–10.3] < 0.0001HbA1c (%) 8.5 [7.9–9.4] 8.6 [7.7–9.6] 0.63Mean AER (µg/min) 4.3 [3.1–6.7] 5.5 [4.0–9.8] < 0.0001Cholesterol (mmol/ l) 4.2 [3.7–4.6] 4.4 [3.9–4.9] < 0.005

Met-Leu 54MM 42 (21%) 7 (4%)ML 111 (56%) 75 (44%) < 0.001LL 45 (23%) 89 (52%)

Cys-Ser 311CC 55 (28%) 33 (19%)CS 95 (48%) 84 (49%) 0.10SS 48 (24%) 54 (32%)

DME_660.fm Page 213 Thursday, March 7, 2002 8:39 AM

© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 212–215

214

Paraoxonase and diabetes •

Y. Kao et al.

four-fold increased risk due to this genotype after allowing forother confounding factors. The new finding is the associationof the

PON2

Ser/Ser genotype with a four-fold increased riskfor microalbuminuria in Type 1 diabetes.

A weakness of this study is the small number of adolescentswith microalbuminuria. When an earlier cut-off for abnormal-ity (AER

≥

7.5

µ

g /min) was used, the genotype also increasedthe risk four-fold. A continuous elevation of albumin excretionbelow traditional cut-offs for microalbuminuria has beenshown by others to predict microalbuminuria [11,12] and con-firmed in adolescents [13].

This potential association of

PON2

and microalbuminuriahas also recently been described in a group of Swiss patientswith Type 2 diabetes [14]. As in the current study, there was noassociation of microalbuminuria with the

PON1

genotype,nor was it found in another group of Caucasian patients withpersistent proteinuria [15].

The

PON1

Met54Leu polymorphism has been found tomodify serum concentration and activity of paraoxonase [16,17].

In vitro

, HDL from Met/Met gives greatest protection toLDL against lipid peroxides [1]. More recently the

PON2

Cys311Ser polymorphism has also been shown to modifyparaoxonase activity [18], but HDL from Ser/Ser does noteffect lipid peroxide

in vitro

activity. The physiological roleof the

PON2

gene is uncertain. In contrast to

PON1

, which ismainly expressed in liver,

PON2

is expressed in most tissues,suggesting a more extensive function [19] which may beindependent of paraoxonase activity.

Diabetic retinopathy was associated with Met54Leupolymorphism but not the Gln192Arg polymorphism inAustralian adolescents [8]. In contrast, no association of com-plications with

PON1

polymorphisms was found in a Japanesegroup of Type 2 diabetic patients [20] and no association with

PON1

or

PON2

polymorphisms and retinopathy was foundin a group from Manchester [18]. This difference may be dueto the differences in ethnic background of the patient groups.

The

PON1

genotype was not associated with microalbu-minuria and the

PON2

genotype was not associated withretinopathy. This study supports the hypothesis that diabeticmicroangiopathy is genetically heterogeneous. Intriguingly,despite linkage disequilibrium being found between the twopolymorphisms of

PON1

(Met54Leu) and

PON2

(Cys311Ser),the former was associated with early retinopathy and the latterwith microalbuminuria. This is consistent with the linkagedisequilibrium being greater for those without either compli-cation than with a complication, again indicating that retino-pathy and nephropathy may have distinct genetic susceptibility.

This heterogeneity between retinopathy and renal abnorm-alities does challenge the initial observations that all patientswith diabetic nephropathy will almost invariably have retino-pathy [21]. The clinical data in this study show that there is notcomplete concordance between retinopathy and early urine

Table 2 Demographic variables and genotyping of adolescents with albumin excretion rate (AER) < and ≥ 7.5 µg/min and for those with AER < and ≥ 20 µg/min

AER < 7.5 µg/min (n = 236)

AER ≥ 7.5 µg/min (n = 80) P-value

AER < 20 µg/min (n = 300)

AER ≥ 20 µg/min (n = 16) P-value

Gender 111 M, 125 F 37 M, 43 F 0.90 142 M, 158 F 6 M, 10 F 0.44Age (years) 13.2 [11.8–14.9] 14.4 [13.0–15.7] < 0.0001 13.4 [12.2–15.2] 14.1 [13.1–16.2] 0.12Duration (years) 5.5 [3.3–7.8] 6.2 [3.3–10.2] 0.038 5.6 [3.3–8.0] 8.3 [5.7–11.7] 0.011HbA1c (%) 8.5 [7.7–9.5] 8.6 [7.8–9.7] 0.36 8.5 [7.8–9.5] 9.0 [8.1–11.3] 0.10Mean AER (µg/min) 4.1 [2.9–5.1] 11.6 [9.2–16.8] < 0.0001 4.6 [3.4–6.8] 29.6 [25.1–41.1] < 0.0001Cholesterol (mmol/l) 4.2 [3.8–4.8] 4.4 [3.9–4.8] 0.39 4.3 [3.9–4.8] 4.4 [3.7–5.7] 0.49

Met-Leu 54MM 36 (15%) 8 (10%) 44 (14%) 0 (0%)ML 129 (55%) 40 (50%) 0.20 161 (53%) 8 (50%) 0.14LL 71 (30%) 32 (40%) 95 (31%) 8 (50%)

Cys-Ser 311CC 72 (31%) 7 (9%) 76 (25%) 3 (19%)CS 119 (50%) 33 (41%) < 0.001 148 (49%) 4 (25%) 0.023SS 45 (19%) 40 (50%) 76 (25%) 9 (56%)

Table 3 Adjusted odds ratios of PON1 and PON2 genotypes for various complications

GenotypeAdjusted odds ratios(95% confidence intervals) P-value

Retinopathy*M/M 0.17 (0.062, 0.44) < 0.0005L/L 3.34 (1.95, 5.75) < 0.0001

AER ≥ 7.5 µg/min†CC 0.19 (0.077, 0.46) < 0.0005SS 4.72 (2.65, 8.41) < 0.0001

Microalbuminuria‡SS 4.41 (1.50, 13.01) 0.0071

*Odds ratios adjusted for age, duration and cholesterol.†Odds ratios adjusted for age, duration and their interaction.‡Odds ratio adjusted for duration.

DME_660.fm Page 214 Thursday, March 7, 2002 8:39 AM

Original article

215

© 2002 Diabetes UK.

Diabetic Medicine

,

19

, 212–215

abnormalities: 13% had early urine abnormalities withoutretinopathy and 13% were concordant. There is alreadyevidence for genetic heterogeneity for retinopathy and micro-albuminuria. Given sufficient diabetes duration, retinopathy isalmost universal, but not all are at risk for renal disease. In theDiabetes Control and Complications Trial, familial clusteringwas found for retinopathy and for nephropathy, but no overlapin clustering for retinopathy in families positive and negativefor nephropathy was described [22]. Meta-analysis of ACEgene polymorphism studies showed uniformity in increasednephropathy risk between different ethnic groups, Type 1 and2 diabetes, microalbuminuria and more severe renal disease.However, no increased risk for retinopathy was detected [23].

It is possible that neither the

PON1

Leu54 nor the

PON2

Ser311 is the functional polymorphism responsible forincreased risk of microangiopathy, but that they are in linkagedisequilibrium with a yet to be determined functional mutationin the

PON

gene cluster on chromosome 7. Both functional andgenetic studies are needed in order to test the nature of theassociation described here and to detect new mutations in thegene cluster, acting as susceptible loci for vascular disorders.

Acknowledgements

The authors would like to thank Dr Jan Fairchild for patientrecruitment, Mandy Crocker, Janine Cusumano for expertretinal photography, Dr Stephen Hing for expert grading, andDarna Bradford and Mirijana Tepsa for laboratory assays.

References

1 Mackness B, Mackness MI, Arrol S, Turkie W, Durrington PN.Effect of the human serum paraoxonase 55 and 192 genetic poly-morphisms on the protection by high density lipoprotein against lowdensity lipoprotein oxidative modification.

FEBS Letters

1998; 423:57–60.

2 Odawara M, Tachi Y, Yamashita K. Paraoxonase polymorphism(Gln192-Arg) is associated with coronary heart disease in Japanesenoninsulin-dependent diabetes mellitus. J Clin Endocrinol Metab1997; 82: 2257–2260.

3 Ruiz J, Blanche H, James RW, Garin MC, Vaisse C, Charpentier Get al. Gln-Arg192 polymorphism of paraoxonase and coronary heartdisease in type 2 diabetes. Lancet 1995; 346: 869–872.

4 Sanghera DK, Aston CE, Saha N, Kamboh MI. DNA polymorphismsin two paraoxonase genes (PON1 and PON2) are associated with therisk of coronary heart disease. Am J Human Genet 1998; 62: 36–44.

5 Chew EY, Klein ML, Ferris FL, Remaley NA, Murphy RP, ChantryK et al. Association of elevated serum lipid levels with retinal hardexudate in diabetic retinopathy. Early Treatment Diabetic Retino-pathy Study (ETDRS) Report 22. Arch Ophthalmol 1996; 114:1079–1084.

6 Sibley SD, Hokanson JE, Steffes MW, Purnell JQ, Marcovina SM,Cleary PA et al. Increased small dense LDL and intermediate-densitylipoprotein with albuminuria in type 1 diabetes. Diabetes Care 1999;22: 1165–1170.

7 Baynes JW, Thorpe SR. Role of oxidative stress in diabetic complica-tions: a new perspective on an old paradigm. Diabetes 1999; 48:1–9.

8 Kao YL, Donaghue K, Chan A, Knight J, Silink M. A variant ofparaoxonase (PON1) gene is associated with diabetic retinopathy inIDDM. J Clin Endocrinol Metab 1998; 83: 2589–2592.

9 Klein R, Klein BE, Moss SE, Davis MD, DeMets DL. The WisconsinEpidemiologic Study of Diabetic Retinopathy. IX. Four-year incid-ence and progression of diabetic retinopathy when age at diagnosisis less than 30 years. Arch Ophthalmol 1989; 107: 237–243.

10 Couper JJ, Staples AJ, Cocciolone R, Nairn J, Badcock N, HenningP. Relationship of smoking and albuminuria in children with insulin-dependent diabetes. Diabet Med 1994; 11: 666–669.

11 Gilbert RE, Tsalamandris C, Bach LA, Panagiotopoulos S,O’Brien RC, Allen TJ et al. Long-term glycemic control and the rateof progression of early diabetic kidney disease. Kidney Int 1993; 44:855–859.

12 Chase HP, Marshall G, Garg SK, Harris S, Osberg I. Borderlineincreases in albumin excretion rate and the relation to glycemiccontrol in subjects with type I diabetes. Clin Chem 1991; 37: 2048–2052.

13 Couper JJ, Clarke CF, Byrne GC, Jones TW, Donaghue KC, Nairn Jet al. Progression of borderline increases in albuminuria in adoles-cents with insulin-dependent diabetes mellitus. Diabet Med 1997; 14:766–771.

14 Pinizzotto M, Castillo E, Fiaux M, Temler E, Gaillard R, Ruiz J.Praoxonase 2 polymorphisms are associated with nephropathy intype 11 diabetes. Diabetologia 2001; 44: 104–107.

15 Araki S, Makita Y, Canani L, Ng D, Warram JH, Krolewski AS.Polymorphisms of human paraoxonase 1 gene (PON1) andsusceptibility to diabetic nephropathy in Type 1 diabetes mellitus.Diabetologia 2000; 43: 1540–1543.

16 Humbert R, Adler DA, Disteche CM, Hassett C, Omiecinski CJ,Furlong CE et al. The molecular basis of the human serum paraoxo-nase activity polymorphism. Nature Genet 1993; 3: 73–76.

17 Garin MC, James RW, Dussoix P, Blanche H, Passa P, Froguel P et al.Paraoxonase polymorphism Met-Leu54 is associated with modifiedserum concentrations of the enzyme. A possible link between theparaoxonase gene and increased risk of cardiovascular disease indiabetes. J Clin Invest 1997; 99: 62–66.

18 Mackness B, Durrington PN, Abuashia B, Boulton AJ, Mackness MI.Low paraoxonase activity in type II diabetes mellitus complicated byretinopathy. Clin Sci 2000; 98: 355–363.

19 Primo-Parmo SL, Sorenson RC, Teiber J, La Du BN. The humanserum paraoxonase/arylesterase gene (PON1) is one member of amultigene family. Genomics 1996; 33: 498–507.

20 Ikeda Y, Suehiro T, Inoue M, Nakauchi Y, Morita T, Arii K et al.Serum paraoxonase activity and its relationship to diabetic complica-tions in patients with non-insulin-dependent diabetes mellitus. MetabClin Exp 1998; 47: 598–602.

21 Hanssen KF. The determinants of microvascular complications indiabetes: an overview. In Pickup JC, Williams G eds. Chronic Com-plications of Diabetes. Oxford: Blackwell Scientific Publications,1994: 3–4.

22 DCCT Research Group. Clustering of long-term complications infamilies with diabetes in the diabetes control and complications trial.Diabetes 1997; 46: 1829–1839.

23 Fujisawa T, Ikegami H, Kawaguchi Y, Hamada Y, Ueda H,Shintani M et al. Meta-analysis of association of insertion/deletionpolymorphism of angiotensin I-converting enzyme gene with diabeticnephropathy and retinopathy. Diabetologia 1998; 41: 47–53.

DME_660.fm Page 215 Thursday, March 7, 2002 8:39 AM