Download Paraoxonase-2 Gene (PON2) G148 Variant Associated with

0021-972X/97/$03.00/0
Journal of Clinical Endocrinology and Metabolism
Copyright © 1997 by The Endocrine Society
Vol. 82, No. 10
Printed in U.S.A.
Paraoxonase-2 Gene (PON2) G148 Variant Associated
with Elevated Fasting Plasma Glucose in NoninsulinDependent Diabetes Mellitus*
ROBERT A. HEGELE†, PHILIP W. CONNELLY, STEPHEN W. SCHERER,
ANTHONY J. G. HANLEY‡, STEWART B. HARRIS, LAP-CHEE TSUI, AND
BERNARD ZINMAN
Department of Medicine, St. Michael’s Hospital (R.A.H., P.W.C.); Department of Genetics, The Hospital
for Sick Children (S.W.S., L.-C.T.); Samuel Lunenfeld Research Institute and Mount Sinai Hospital
(A.J.G.H., B.Z.), University of Toronto, Toronto; and Thames Valley Family Practice Research Unit,
University of Western Ontario (S.B.H.), London, Canada
ABSTRACT
Defining the genetic determinants of NIDDM requires evidence
from several complementary approaches, including both linkage and
association analyses using both discrete phenotypes and intermediate
quantitative traits. We tested for association between common
genomic variation in three genes that map to chromosome 7q21-q22
and quantitative traits related to NIDDM in a sample of Oji-Cree. We
found that a common genomic variation in codon 148 (alanine or
glycine) of the paraoxonase-2 gene (PON2) demonstrated a significant
association with a variation in fasting plasma glucose (P , 0.0001).
Furthermore, we found a significant association between a variation
in fasting plasma glucose and the interaction term comprised of a
PON2 codon 148 genetic variation and the presence of noninsulindependent diabetes mellitus (NIDDM; P , 0.0001). We then analyzed
subjects according to PON2 genotype and NIDDM status. In subjects
with NIDDM, the PON2 codon 148 G/G homozygotes had significantly
higher mean fasting plasma glucose than subjects with the other two
genotypes (P , 0.0001). However, in non-NIDDM subjects, there was
no difference in mean fasting plasma glucose among any of the genotypes. There was no association of the PON2 genotype with NIDDM
itself, with impaired glucose tolerance, or with other quantitative
traits related to NIDDM in this sample. These findings suggest that
1) the PON2 G148 gene variant worsens glycemia in subjects with
NIDDM; 2) defining the physiological role of the PON2 gene product
would be worthwhile; and 3) genetic factors can modify the severity
of clinical phenotypes in subjects with NIDDM. (J Clin Endocrinol
Metab 82: 3373–3377, 1997)
A
plasma glucose, may be useful in identifying NIDDM susceptibility genes (2).
Genome-wide scanning to identify NIDDM susceptibility
genes in the Pima Indians found linkages of DNA markers
at 7q21.3-q22.1 with both quantitative traits related to glucose uptake and storage and possibly with NIDDM itself (6).
Within this region are the genes encoding paraoxonase
(PON1) (7), a paraoxonase-like protein (PON2) (8), and an
isoenzyme of pyruvate dehydrogenase kinase (PDK4) (9).
Serum paraoxonase is associated with high density lipoprotein and hydrolyzes a number of organophosphates (10). In
contrast to PON1, which is mainly expressed in the liver,
PON2 is expressed in a variety of tissues, including the pancreas (8). Although the physiological role of the PON2 gene
product is unknown, its tissue distribution suggests a role
that unique from that of paraoxonase (8). The PDK4 product
modulates the activity of the mitochondrial pyruvate dehydrogenase complex, which catalyzes the first step in mitochondrial glucose oxidation; thus, PDK4 is also a reasonable
candidate gene for metabolic phenotypes (9).
We wished to determine whether common genomic
changes affecting the amino acid sequence at codon 148
(A3 G) predicted from the PON2 gene and within the PDK4
gene promoter were associated with NIDDM-related phenotypes in Oji-Cree. We tested for association between the
common genomic variation of PON2 and PDK4 and quantitative intermediate phenotypes of NIDDM. We also eval-
LTHOUGH noninsulin-dependent diabetes mellitus
(NIDDM) is acknowledged to have a genetic component, the identification of susceptibility genes to date has
proven to be difficult (1, 2). Various complementary strategies, including both linkage and association analysis, are
probably needed to dissect a genetic component for NIDDM
(2). To date, linkage studies have provided most of the promising leads for causative genes in NIDDM (3, 4). However,
common NIDDM is genetically heterogeneous, and association analysis might also be useful for identifying susceptibility or modifier genes (2, 5). Furthermore, Ghosh and
Schork have argued that analysis of the genetic determinants
of intermediate traits related to NIDDM, such as fasting
Received February 5, 1997. Revision received March 24, 1997.
Rerevision received June 3, 1997. Accepted June 17, 1997.
Address all correspondence and requests for reprints to: Robert A.
Hegele, M.D., DNA Research Laboratory, St. Michael’s Hospital, 30
Bond Street, Toronto, Ontario, Canada M5B 1W8. E-mail: robert.
[email protected].
* This work was supported by grants from the NIH (91-DK-01), the
Ontario Ministry of Health (04307), the Heart and Stroke Foundation of
Ontario (T2978), and the St. Michael’s Hospital Foundation and by a
Canadian Genome Analysis and Technology award (to L.-C.T. and
S.W.S.).
† Career Investigator with the Heart and Stroke Foundation of
Ontario.
‡ Supported by Health Canada through a National Health Research
and Development Program Research Training Award.
3373
3374
HEGELE ET AL.
uated the association between these quantitative traits and a
common sequence variant in another gene on chromosome
7q, namely the codon 192 R3 Q variant of paraoxonase (encoded by PON1) (7, 10), and a common sequence variant in
a gene on chromosome 4q, namely the codon 54 A3 T variant
in codon 54 of intestinal fatty acid binding protein (encoded
by FABP2) (11).
Subjects and Methods
Study subjects
The community of Sandy Lake, Ontario, is located about 2000 km
northwest of Toronto, in the subarctic boreal forest of central Canada.
The community is isolated and is accessible only by air during most of
the year. Most members of the community speak both English and
Oji-Cree, a member of the Algonkian family of languages (12). Historically, the ancestors of the contemporary residents of this region lived
a nomadic, hunting-gathering subsistence typical of other Algonkianspeaking peoples of the northeastern subarctic. Since the development
of the reservation and residential school systems, the lifestyle has
changed radically from physically active to sedentary. The primary
source of food has changed from wildlife with supplementation by roots
and berries to processed foods high in animal fats.
Seven hundred and twenty-eight members (72% of the total population) of this community aged 10 yr and above participated in the Sandy
Lake Health and Diabetes Project (12). Assessments included a questionnaire to assess medical history, including a previous diagnosis of
NIDDM. Body mass index (BMI) was defined as weight (kilograms)/
height (meters)2. The project was approved by the University of Toronto
ethics review committee.
Biochemical analyses
Plasma samples were obtained with informed consent. Exclusion
criteria included an inadequate blood sample available for all biochemical and/or genetic determinations. Volunteers provided plasma samples after fasting overnight for 8 –12 h. Blood was centrifuged at 2000
rpm for 30 min, and the plasma was stored at 270 C. Concentrations of
fasting glucose were determined as previously described (12). Concentrations of fasting plasma insulin were determined by RIA (Pharmacia,
Piscataway, NJ). A standard 75-g oral glucose tolerance test (OGTT;
Glucodex, Rougier, Chambly, Canada) was administered, and a second
blood sample was collected after 120 min for plasma glucose determination. Volunteers were excluded from the OGTT if they had physiciandiagnosed diabetes and were currently receiving treatment with insulin
and/or oral hypoglycemic agents or if they had a fasting blood glucose
level exceeding 11.1 mmol/L. Volunteers who were pregnant at the time
of recruitment had their OGTT deferred until 3 months postpartum.
NIDDM and impaired glucose tolerance (IGT) were diagnosed using
established criteria (13, 14).
The genotypes for FABP2 codon 54 and PON1 codon 192 were determined as previously described (15, 16). The genotype for PON2 codon
148 was determined using 0.5 mg genomic DNA amplified in a mismatch
PCR reaction using primers PON2–148-59 (59-AGT GGA AAT TTT TAA
ATT TGA AGC AG-39) and PON2–148-39 (59-TTG TTT GCA AAT GCT
GGG GAT-39). The resulting 130-bp fragment was digested with BsoFI,
which created two smaller fragments for an amplimer derived from a
PON2 allele encoding A148 and a single 130-bp fragment for an amplimer derived from a PON2 allele encoding G148. These fragments
were resolved on 4% agarose gels. The genotype for the PDK4 promoter
sequence was determined using primers PDK4-pro-59 (59-CCT CCG
AGT TGT AAA CAA GG-39) and PDK4-pro-39 (59-AAC GCG TCC TGA
ACT CCA G-39). Two single base substitutions, a C3 T and a T3 C,
respectively, at positions 2208 and 2153 nucleotides relative to the
transcription start site of PDK4, were included in this fragment and
could be simultaneously detected after digestion with MspI. There were
three combinations of MspI fragments in the Sandy Lake Oji-Cree; these
genotype combinations were treated as effective haplotypes. Haplotype
A contained the 2208 nucleotide MspI site but lacked the 2153 nucleotide MspI site; haplotype B contained both MspI sites; and haplotype
C lacked both MspI sites. We did not observe any alleles that contained
JCE & M • 1997
Vol 82 • No 10
the fourth possible haplotype, namely the absence and presence, respectively, of the 2208 and 2153 nucleotide MspI sites. The fragments
were resolved on 10% polyacrylamide gels.
Statistical analysis
SAS (version 6.11) was used for all statistical comparisons (17). Quantitative variables were log transformed and subjected to analysis of
normality as previously described (18). ANOVAs were performed using
the general linear models procedure to determine the sources of variation for log fasting plasma glucose and insulin and log plasma glucose
2 h after a standard glucose load. Adult subjects, aged 18 yr and older,
were analyzed. F tests were computed from the type III sums of squares
(17). This form of sums of squares is applicable to unbalanced study
designs and adjusts the level of significance to account for other independent variables included in the model. Independent variables for each
ANOVA were sex, age, the log of BMI, the presence of previously
diagnosed or newly diagnosed NIDDM, and genotypes of FABP2 codon
54, PON1 codon 192, PON2 codon 148, and the PDK4 promoter. We had
previously tested for association between the FABP2 codon 54 polymorphism with fasting plasma insulin and NIDDM status and had
found no association. However, we had not previously tested for association between the FABP2 codon 54 polymorphism and fasting
plasma glucose or plasma glucose 2 h after a standard glucose load.
Interaction terms composed of the presence of the NIDDM variable and
the significantly associated genotype variables were included in the
ANOVA. As recommended by others (2, 19), the nominal P value for a
significant association was taken to be less than 0.01. When a significant
genotype-phenotype was identified, the mean values for the trait were
compared between genotypic classes using pairwise comparisons of
least squares means (17).
Results
Clinical attributes of sample
Sufficient DNA and phenotypic information were obtained for analysis from 523 subjects, aged 18 yr and older,
of whom 298 (57%) were women. The mean 6 sd for the age
and BMI were, respectively, 35.8 6 14.6 yr and 28.1 6 5.3
kg/m2. One hundred and twenty-four subjects were classified as having NIDDM; 54 subjects were classified as having
newly diagnosed NIDDM based upon the OGTT, and 70
subjects had been previously diagnosed with NIDDM. Of the
subjects with NIDDM, 6 were taking insulin, 30 were taking
oral hypoglycemic agents, and the remainder were controlled by diet alone. Normal subjects and those with IGT
were classified as not having NIDDM (non-NIDDM).
Allele and genotype frequencies
The frequencies of the FABP2 T54, PON1 Q192, and PON2
G148 alleles were 0.14, 0.22, and 0.27, respectively. The frequency of the FABP2 T54 allele was the same in this population as we had reported previously. The frequencies of the
A, B, and C haplotypes of PDK4 were 0.732, 0.266, and 0.002,
respectively. Genotype frequencies of FABP2, PON1, and
PON2 and diploid haplotype frequencies of PDK4 did not
deviate from those predicted by the Hardy-Weinberg law in
this study sample (all P . 0.10).
Phenotype-genotype associations
Transformation using the natural logarithm for each variable resulted in a distribution that was not significantly different from normal. One ANOVA was performed for each of
log fasting plasma glucose, log 2 h postglucose challenge
plasma glucose, and log fasting plasma insulin level. There
PON2 G148 AND GLYCEMIA IN NIDDM
were no significant genotype-phenotype associations for log
2 h postglucose challenge plasma glucose and log fasting
plasma insulin levels. However, there were two significant
genotype-phenotype associations for log fasting plasma glucose. The first was between PON2 codon 148 genotype and
log fasting plasma glucose, and the second was between
FABP2 codon 54 genotype and log fasting plasma glucose
(Table 1). There was also a highly significant association
between log fasting plasma glucose and the interaction term
comprised of NIDDM status and PON2 genotype, and there
was a suggestion of an association between log fasting
plasma glucose and the interaction term comprised of
NIDDM status and FABP2 genotype (Table 1). Log fasting
plasma glucose was also significantly associated with both
NIDDM status and log BMI in adults (Table 1). There were
no significant phenotype-genotype associations for any
quantitative trait when normal adults and adults with IGT
were analyzed separately (data not shown). There were no
significant genotype-phenotype associations with the presence of NIDDM or IGT (data not shown). All significant
associations were unaffected by including medication use as
a covariate (data not shown).
Pairwise comparisons
Because of the highly significant association between log
fasting plasma glucose and the interaction term comprised of
NIDDM status and PON2 genotype, we examined NIDDM
and non-NIDDM subjects separately. The means 6 sds of log
fasting plasma glucose for NIDDM and non-NIDDM subjects are shown in Fig. 1. As expected, the mean log fasting
plasma glucose was significantly higher in NIDDM than in
non-NIDDM subjects (P , 0.0001). Pairwise comparisons in
the NIDDM subjects indicated that the mean log fasting
plasma glucose was significantly higher in the PON2 codon
148 G/G homozygotes than in subjects with the other two
genotypes (P , 0.0001). Pairwise comparisons in the nonNIDDM subjects indicated that the mean log fasting plasma
glucose level was not different between genotype classes.
There were no differences in log plasma glucose between
TABLE 1. Sources of variation of log fasting plasma glucose in
Sandy Lake adults (ANOVA)
Source of variation
df
F
Pr . F
Sex
Age
Log BMI
FABP2 genotype
PDK4 genotype
PON1 genotype
PON2 genotype
Diabetes status
FABP2* diabetes
PON2* diabetes
1
1
1
2
2
2
2
1
2
2
6.05
0.16
11.5
5.75
0.20
0.57
9.73
188.7
4.44
10.7
NS (0.014)
NS (0.69)
0.0007
0.0034
NS (0.93)
NS (0.57)
,0.0001
,0.0001
NS (0.012)
,0.0001
BMI, Body mass index; FABP2, intestinal fatty acid binding protein gene codon 54 genotype; PON1 genotype, paraoxonase gene codon
192 genotype; PON2 genotype, PON2 gene codon 148 genotype; PDK4
genotype, pyruvate dehydrogenase kinase isoform 4 gene promoter
genotype; diabetes status, presence or absence of NIDDM; FABP2*
diabetes, interaction term composed of FABP2 gene codon 54 genotype and diabetes status; PON2* diabetes, interaction term composed
of PON2 gene codon 148 genotype and diabetes status; Pr . F, probability of a greater F test; NS 5 not significant (nominal P , 0.01).
3375
FIG. 1. Fasting glucose and PON2 genotype in Sandy Lake adults.
Subjects are divided into those with NIDDM and those without
NIDDM (non-NIDDM). The number of subjects with each genotype is
indicated above each bar. Means and SDs for log fasting plasma glucose are shown. NIDDM subjects had significantly higher mean log
fasting plasma glucose than non-NIDDM subjects (P , 0.0001).
NIDDM subjects with the PON2 codon 148G/G genotype had significantly higher log fasting plasma glucose levels than all other subjects
(P , 0.0001), as indicated by the double asterisk.
genotype classes when IGT and normal subjects were analyzed separately (data not shown).
Because of the indication of an association between log
fasting plasma glucose and the interaction term comprised of
NIDDM status and FABP2 genotype, we examined NIDDM
and non-NIDDM subjects separately. The least squares
means of log fasting plasma glucose for NIDDM subjects
with FABP2 codon 54 genotypes T/T, T/A, and A/A were,
respectively, 2.91, 2.43, and 2.38, with T/T being significantly
different from the other two genotypes (P , 0.003). The least
squares means of log fasting plasma glucose for non-NIDDM
subjects with FABP2 codon 54 genotypes T/T, T/A, and A/A
were, respectively, 1.72, 1.68, and 1.69, with no significant
difference between the groups. However, it should be noted
that there were only three T/T subjects with NIDDM and five
T/T subjects without NIDDM; thus, there were only small
numbers of subjects to compare. Nevertheless, these observations are consistent with an influence of the FABP2 T54
allele on fasting plasma glucose only in NIDDM.
Discussion
The principal novel finding in this study of Oji-Cree from
Sandy Lake was that a conservative A3 G variant of codon
148 of PON2 gene was highly significantly associated with
variation in the fasting plasma glucose concentration. In particular, homozygosity for the PON2 G148 allele was highly
significantly associated with elevated plasma glucose in
adult NIDDM subjects. There was no association of the PON2
genotype with NIDDM itself, with IGT, or with other quantitative traits related to NIDDM in any subset of this sample.
There were no significant phenotype-genotype associations
3376
JCE & M • 1997
Vol 82 • No 10
HEGELE ET AL.
of any quantitative phenotype with variations in two other
genes on 7q21.3-q22.1, namely PON1 and PDK4.
The absence of associations with the other closely linked
genetic markers suggests that the association of a variation
in plasma glucose in NIDDM with the PON2 genotype is
specific, although the codon 148 change may be in linkage
disequilibrium with the actual functional DNA variant
within this gene. The association of homozygosity for PON2
G148 with worsened hyperglycemia in NIDDM suggests that
the variation in PON2 modulates a quantitative NIDDMrelated phenotype, but may not itself predispose to NIDDM.
The paraoxonase-related gene PON2 is located near PON1
and is part of a multigene family that includes PON1 and
PON3 (8). This proximity and the similarity in genomic structure of PON1 and PON2 suggests that the two genes might
have arisen by a tandem duplication event. PON1 was discovered after biochemical study of paraoxonase (7). Paraoxonase is associated with a subclass of antiatherogenic, high
density lipoprotein, and its physiological substrate is unknown (7, 8, 10). Paraoxonase might modify the oxidation of
atherogenic lipoproteins (16). The codon 192 polymorphism
in PON1 probably underlies interindividual variation in
plasma paraoxonase activity and has been associated with
variations in metabolic and cardiovascular phenotypes (16,
20, 21). In contrast, PON2 was discovered by a genetic approach (8), and there is presently no information regarding
its possible physiological role.
The structural similarities between PON1 and PON2 suggest that the PON2 gene product may have some function in
common with PON1 (8). However, the more widespread
tissue expression and multiple forms of the PON2 transcript
also suggest an independent function for the PON2 gene
product (8). Of special relevance to NIDDM is the expression
of PON2 in the pancreas (8). Furthermore, PON2 is expressed
in cardiac and skeletal muscle (8), suggesting that it could be
involved in the peripheral utilization of glucose. The unique
structural aspects and tissue distribution of PON2 taken
together with our observed genetic association with hyperglycemia suggest that defining its function would be
worthwhile.
There was also a less significant association between a
variation in codon 54 of the FABP2 gene and the fasting
plasma glucose concentration. This association appeared to
be due to elevated plasma glucose levels in NIDDM homozygotes for the FABP2 T54 allele; however, there were
only three such subjects. This is consistent with previous
observations of worsened diabetic phenotypes in subjects
with the FABP2 T54 allele (11). Furthermore, we previously
reported an association between the FABP2 T54 allele and
increased body mass, but not with NIDDM or fasting plasma
insulin (15). Taken together, our results are consistent with
a deleterious metabolic impact of the FABP2 T54 allele.
At least one genomic scan, performed in Mexican-Americans, showed no linkage between any chromosome 7q
marker and NIDDM (3). However, genomic scanning in
Pima Indian sibships with NIDDM (6) has revealed convincing linkages of D7S527, which is close to PON2, with
such quantitative phenotypes as glucose uptake, oxidation, and storage rates at physiological and maximally
stimulating insulin concentrations. There was less con-
vincing evidence for possible linkages with the plasma
insulin concentration and with NIDDM itself in the Pima
Indians (6). These results are consistent with our findings
that NIDDM per se is not related to genetic variation in
PON2. However, there appears to be a relationship between variation in this region and complex quantitative
phenotypes related to glucose uptake and/or storage. This
may become manifest as a variation in plasma glucose.
Also, the metabolic stress of overt NIDDM may force a
more extreme deviation in plasma glucose for Oji-Cree
individuals homozygous for PON2 G148 or the linked
variant for which G148 is a marker. The specific association with plasma glucose and not with insulin in our
sample suggests that these intermediate traits are themselves complex and have distinct genetic determinants (2).
In summary, we have observed that homozygosity for the
PON2 G148 allele was associated with worsened fasting hyperglycemia in Oji-Cree with NIDDM. This indicates that an
association analysis using quantitative intermediate traits
provides results complementary to those obtained from linkage studies (6). The PON2 variation was not associated with
the presence of NIDDM and thus cannot be interpreted as
being causative. However, PON2 may be a modifier gene for
a complex quantitative phenotype related to NIDDM. If genetic factors can worsen clinical phenotypes in NIDDM and
thus contribute to the spectrum of disease expression, then
determination of such genotypes may have an impact on
diagnosis and treatment in a particular subject even after the
onset of NIDDM.
Acknowledgments
We acknowledge the chief and council of the community of Sandy
Lake, the Sandy Lake community surveyors, the Sandy Lake nurses, the
staff of the University of Toronto Sioux Lookout Program, the Department of Clinical Epidemiology of the Samuel Lunenfeld Research Institute, Dr. Alexander Logan, Annette Barnie, Fang Sun, Teresa Lippingwell, Joel Napenas, Stefan Sadikian, and Cheri Tully. Dr. Michal
Prochazka provided sequence information for the PON2 and PDK4
amplification primers.
References
1. Elbein SC, Hoffman MD, Bragg KL, Mayorga RA. 1994 The genetics of
NIDDM. Diabetes Care. 17:1523–1533.
2. Ghosh S, Schork NJ. 1996 Genetic analysis of NIDDM: the study of quantitative traits. Diabetes. 45:1–14.
3. Hanis CL, Boerwinkle E, Chakraborty R, et al. 1996 A genome-wide search
for human non-insulin-dependent (type 2) diabetes genes reveals a major
susceptibility locus on chromosome 2. Nat Genet. 13:161–166.
4. Mahtani MM, Widen E, Lehto M, et al. 1996 Mapping of a gene for type 2
diabetes associated with an insulin secretion defect by a genome scan in
Finnish families. Nat Genet. 14:90 –94.
5. Lander ES, Schork NJ. 1994 Genetics of complex traits. Science. 265:2037–2048.
6. Prochazka M, Thompson B, Scherer SW, et al. 1995 Linkage and association
of insulin resistance and NIDDM with markers at 7q21.3-q22.1 in the Pima
Indians. Diabetes. 45:42A.
7. Humbert R, Adler DA, Disteche CM, Hassett C, Omiecinski CJ, Furlong, CE.
1993 The molecular basis of the human serum plasma paraoxonase activity
polymorphism. Nat Genet. 3:73–76.
8. Primo-Parmo SL, Sorenson RC, Teiber J, La Du BN. 1996 The human serum
paraoxonase/arylesterase gene (PON1) is one member of a multigene family.
Genomics. 33:498 –507.
9. Rowles J, Scherer SW, Xi T, et al. 1996 Cloning and characterization of PDK4
on 7q21.3 encoding a fourth pyruvate dehydrogenase kinase isoenzyme in
human. J Biol Chem. 271:22376 –22382.
10. Davies HG, Richter RJ, Keifer M, Broomfield CA, Sowalla J, Furlong CE.
1996 The effect of the human serum paraoxonase polymorphism is reversed
with diazoxon, soman and sarin. Nat Genet. 14:334 –336.
11. Baier LJ, Sacchettini JC, Knowler WC, et al. 1995 An amino acid substitution
PON2 G148 AND GLYCEMIA IN NIDDM
12.
13.
14.
15.
16.
in the human intestinal fatty acid binding protein is associated with increased
fatty acid binding, increased fat oxidation and insulin resistance. J Clin Invest.
95:1281–1287.
Harris SB, Gittelsohn J, Hanley AJG, et al. 1997 The prevalence of NIDDM,
and associated risk factors in native Canadians. Diabetes Care. 20:185–197.
WHO Expert Committee on Diabetes Mellitus. 1980 Second report. Geneva:
WHO; Technical Report Series 646.
National Diabetes Data Group. 1979 Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes.
28:1039 –1057.
Hegele RA, Harris SB, Hanley AJG, Sadikian S, Connelly PW, Zinman B.
1996 Genetic variation in intestinal fatty acid binding protein associated with
variation in body mass in aboriginal Canadians. J Clin Endocrinol Metab.
81:4334 – 4337.
Hegele RA, Brunt JH, Connelly PW. 1995 A polymorphism of the paraoxonase
17.
18.
19.
20.
21.
3377
gene associated with variation in blood pressure in a genetic isolate. Arterioscler Thromb Vasc Biol. 15:89 –95.
SAS Institute Inc. 1987 SAS/STAT guide for personal computers, version, 6th
ed. Cary: SAS Institute.
Hegele RA, Brunt JH, Connelly PW. 1995 Multiple genetic determinants of
variation of plasma lipoproteins in a genetic isolate. Arterioscler Thromb Vasc
Biol. 15:861– 871.
Kidd KK. 1993 Associations of disease with genetic markers: deja vu all over
again. Am J Med Genet. 48:71–73.
Ruiz J, Blanche H, James RW, et al. 1995 Gln-Arg192 polymorphism of
paraoxonase and coronary heart disease in type 2 diabetes. Lancet.
346:869 – 872.
Serrato M, Marrian AJ. 1995 A variant of the human paraoxonase/arylesterase
(HUMPONA) gene is a risk factor for coronary artery disease. J Clin Invest.
96:3005–3008.