Common DNA Polymorphisms at the Lipoprotein Lipase Gene
Association With Severity of Coronary Artery Disease and Diabetes
Background DNA variants of the lipoprotein lipase gene are associated with changes in lipid metabolism similar to those in diabetes and may relate to the development of atherosclerotic lesions, particularly premature lesions.
Methods and Results To determine whether lipoprotein lipase gene variants are relevant to ongoing atherogenesis, we explored relationships between two common lipoprotein lipase gene polymorphic markers, Pvu II at intron 6 and HindIII at intron 8; the severity of coronary artery disease (CAD); and lipid variables in 475 white patients 65 years of age or younger. We assessed CAD severity as the number of significantly stenosed (>50% luminal obstruction) major coronary arteries at angiography and by the Green Lane coronary score. We found a significant association between the Pvu II polymorphism and the number of significantly diseased vessels (P=.0099) and coronary score (P=.028), with the Pvu II(−) alleles associated with less severe disease. The HindIII polymorphism was not associated with severity but had an additive effect with the Pvu II polymorphism. There was a close relationship between the Pvu II(+/+) genotype and the presence of diabetes (P=.0025), with an OR of 3.12 (95% CI, 1.30 to 7.49) compared with the Pvu II(−/−) genotype. The interaction between these polymorphisms and CAD severity (rather than occurrence) was independent of the levels of triglycerides and HDL cholesterol and of other lipid variables. There was also a dosage-dependent relationship between the Pvu II polymorphism and levels of triglyceride. The Pvu II(−) allele was associated with low levels and variances of triglycerides.
Conclusions We conclude that the lipoprotein lipase Pvu II polymorphism is significantly associated with CAD severity and with type II diabetes in CAD patients, independent of changes in circulating lipid levels. These findings may be relevant to mechanisms mediating atherogenesis.
In a recent study, we examined the relationship between known atherogenic variables and the severity of premature CAD determined at angiography.1 We postulated that factors predicting the occurrence of CAD may not be quantitatively the same as those predicting CAD severity (ie, the initiating factors may not be exactly the same as the factors that cause progression). Our results identified a hierarchy of predictors of severity that was somewhat different from that thought to be associated with the occurrence of CAD. But we could account for only ≈50% of the variance in CAD severity. Clearly, factors other than the standard risk factors, which in our study also included Lp(a) and apo A-I and B, also were contributing. In the present study, we have attempted to identify unrecognized contributors to CAD severity by using a candidate gene approach in a further series of consecutively referred patients. We have explored a possible contribution from the DNA polymorphisms of the gene for LpL because these could result in both local and circulating atherogenic effects.
LpL is a glycoprotein synthesized in the parenchymal cells of several tissues. It hydrolyzes circulating triglycerides of exogenous (chylomicron) and endogenous (VLDL) origins to provide free fatty acids for oxidation in the heart and other tissues and for storage in adipose tissue.2 It affects circulating triglyceride levels by generating lipoprotein remnants. These are then processed by hepatic lipase. After secretion, LpL is attached to the luminal surface of endothelial cells, where it not only has a major role in the metabolism of lipoproteins in the circulation but also interacts with lipoproteins locally.2 It has been demonstrated that LpL increases the retention of LDL and VLDL particles by the subendothelial matrix of the artery wall and that it potentiates their conversion to more atherogenic forms.3 These localized effects could not be assessed by measurement of either circulating LpL or triglyceride levels but could be explored by identification of genetic variants of LpL and relation of these to the presence and extent of atherosclerotic lesions.
We have now investigated first the possibility that the DNA polymorphisms of LpL Pvu II and HindIII are associated with CAD severity determined at angiography. Polymorphisms of LpL Pvu II are caused by the presence or absence of a C→T transition in intron 6 and of LpL HindIII by the presence or absence of a T→G transversion in intron 8, and they have been reported to be associated with circulating triglyceride levels in some studies4 5 6 7 but not in others.7 8 These variations are common in white populations.4 5 6 7 8 Because diabetes has an important role in the development of CAD and hypertriglyceridemia occurs frequently in diabetics, we also explored the relationship between the LpL polymorphisms and the presence of diabetes in this patient population. We have not investigated other mutations that induce severe LpL deficiency because they are too rare to make a significant overall contribution in the coronary population.
We studied men and women 65 years of age or younger who were referred consecutively to the Eastern Heart Clinic at Prince Henry Hospital (Sydney, Australia) for coronary angiographies over a 10-month period in 1994. We excluded only patients shown to have significant left main disease (>50% luminal obstruction) because it was difficult to categorize this small proportion of the total (4%) within the classification system we used (see below). Written consent was obtained from every patient after a full explanation of the study, which was approved by the Ethics Committee of the University of New South Wales.
Among the 505 patients studied, 479 (94%) were of European origin and were classified as white. There were 15 Arabic patients (3.0%), 4 Indians (0.8%), 4 Pacific Islanders (0.8%), and 3 Asian patients (0.6%). White patients only were used in the statistical analysis because the numbers in other ethnic groups were too small for meaningful analyses. Table 1⇓ gives demographic information about the patients. Table 2⇓ gives the number of significantly diseased vessels in relation to sex. Ninety-nine patients reported no angina, 173 reported stable angina, and 207 reported unstable angina.
A 4-mL venous blood sample was drawn into an EDTA sample tube before the angiogram after at least a 6-hour fast. The blood sample was centrifuged within 2 hours, and plasma and cellular components were stored separately at −70°C in aliquots until analysis.
DNA Analysis for the Detection of LpL Pvu II and HindIII Polymorphisms
DNA was extracted from the frozen cellular blood component by a salting-out method adapted from that described by Miller et al9 for whole frozen blood. The extracted DNA was stored at 4°C until analysis. The polymerase chain reaction was used to detect the LpL Pvu II and HindIII RFLPs. Primers were designed to amplify the intron 6 region of chromosome 8 for the detection of the Pvu II polymorphism as described by Peacock et al.5 A C→T transition within a Pvu II site in the middle of intron 6 is responsible for the Pvu II polymorphism.10 11 Digests were subsequently run on an 8% polyacrylamide gel at 100 V for 1 hour, and two fragments of 330 and 110 bp were produced when the restriction site was present. The HindIII polymorphism is due to a replacement of T by G at the HindIII site as described by Gotoda et al,12 and primers were designed to amplify the region of intron 8 for its detection. The digests were subsequently run in a 1.5% agarose gel, and two fragments of 600 and 115 bp were produced when the restriction site was present.
The identified genotypes were named according to the presence or absence of the enzyme restriction sites, so Pvu II(+/+), (+/−), and (−/−) and HindIII(+/+), (+/−), and (−/−) are homozygotes for the presence of the site, heterozygotes for the presence and absence of the site, and homozygotes for the absence of the site, respectively. In 4 patients, blood samples were inappropriately stored, and no DNA samples were available for study.
Total cholesterol, HDL cholesterol, and triglyceride levels were measured by the hospital’s Clinical Chemistry Department with standard enzymatic methods. The LDL cholesterol levels were calculated with the Friedewald formula. Because this formula is inaccurate when triglyceride levels are >4.52 mmol/L, we did not estimate LDL cholesterol in the 19 patients of the series with triglyceride levels >4.52 mmol/L (5.93±0.23 mmol/L; range, 4.6 to 8.4 mmol/L). We measured levels of apoA-I, apoB, and Lp(a) using ELISA methods developed in our laboratory and reported previously.13 14 15
We obtained each patient’s medical history using a questionnaire with standardized choices of answers to be ticked during the interview. We recorded the presence or absence (yes/no) of a history of hypertension requiring treatment, diabetes requiring treatment, angina pectoris, and previous myocardial infarction. A “don’t know” box was included for those patients who were not clear about aspects of their medical history that were unavailable from the patient’s file. Current medications were recorded, particularly the usage of lipid-lowering drugs and β-adrenergic blocking agents. The presence or absence of CAD among first-degree relatives (parents and siblings) and the age of first onset were recorded for quantitative assessment of family history of premature CAD. We recorded the presence and severity of angina according to whether each patient was experiencing no angina, stable angina, or unstable angina before and during the current hospitalization. All those classified as having unstable angina had an increase in pain frequency and pain while at rest. The lifetime smoking dose in pack-years and whether the patient was a current smoker or an exsmoker were recorded as described previously,1 and BMI was measured. The patient’s country of origin also was obtained to document ethnicity.
Documentation of CAD Severity
The severity of CAD was determined as follows. The angiograms were assessed by two cardiologists who were unaware that the patients were to be included in the study. Each angiogram was classified either as revealing no coronary lesion with >50% luminal stenosis or as having one, two, or three major epicardial coronary arteries with >50% luminal obstructions. In a second approach, we used the Green Lane coronary scoring system,16 which provides a numerical value for lesion severity and takes into account the amount of myocardium supplied by an affected vessel; the maximum score is 15.
Levels of the quantitative variables are presented as mean±SEM. We used a full factorial design of ANOVA to assess relationships between the quantitative and categorical variables. The latter included the number of significantly diseased vessels (>50% luminal obstruction) and the LpL Pvu II and HindIII polymorphisms. The distributions of the quantitative variables were assessed before the ANOVA to establish that they were normally distributed to meet the conditions required for the analysis. Lp(a) levels were transformed to logarithmic levels (log10) to achieve a close-to-normal distribution. We compared the means of each subgroup and reported F values for levels of significance using two-tailed probability values. The equality of variances among different subgroups of diseased vessels and genotypes was evaluated by the Bartlett-Box F test for homogeneity of variance as appropriate.
We used log linear analysis for associations between the number of diseased vessels and LpL Pvu II and HindIII genotypes. The number of diseased vessels was regarded as an ordinal variable (0, 1, 2, or 3 significantly diseased vessels). We also explored the possible dosage effects of the Pvu II and the HindIII polymorphisms on the number of significantly diseased vessels by treating them as ordinal variables (+/+=1, +/−=2, and −/−=3) using a linear-and-linear association model in the log linear analysis. The log linear model also was used to evaluate the relationship between genotypes and the medical history, which included categorical variables (eg, history of hypertension requiring treatment and diabetes). Likelihood ratio χ2 values were used to assess significance. Because there is a sex difference in CAD risk profiles, we analyzed results not only among all patients but also in male and female patients separately.
We used a logistic regression model to evaluate the interaction between the genotypes and quantitative variables in relation to the number of diseased vessels. The number of diseased vessels was regarded as the dependent ordinal variable, and all other variables were regarded as independent variables appropriately classified as quantitative or categorical. Fewer than 2% of the responses to questions were “don’t know,” and they were treated as missing values. The statistical analyses were executed by the SPSS statistical program.
Genetic Characteristics of the Patient Population
There was no difference in frequencies of the rare Pvu II(−) allele between male (0.45) and female (0.45) patients. The HindIII(−) allele tended to be less frequent among women (0.26) compared with men (0.31), although it was not statistically significant. Hardy-Weinberg equilibrium was evident for the Pvu II genotypes (P=.82) but not for the HindIII genotypes (P<.01) in this patient population. The polymorphisms of Pvu II and HindIII were in strong linkage disequilibrium (P<.0001) as shown in Table 3⇓.
LpL Pvu II and HindIII Polymorphisms and Number of Significantly Diseased Vessels
There was a marked association between the Pvu II polymorphism and the number of significantly diseased vessels. As Table 4⇓ shows, patients with Pvu II(−/−) genotypes were significantly less likely to have triple-vessel disease and more likely to have no significantly diseased vessels. The OR for Pvu II(+/+) to have triple-vessel disease compared with Pvu II(−/−) was 1.73 (95% CI, 1.034 to 2.895). When Pvu II was treated as an ordinal variable, there was no allelic dosage effect. The significant association between CAD severity and LpL Pvu II polymorphism also was confirmed with Green Lane coronary scores, which were highly correlated with the number of significantly diseased vessels (r=.88, P=.0001). There was a significant decrease in the severity scores with the Pvu II(−) alleles (6.35±0.41, 5.61±0.30, and 4.72±0.40 for +/+, +/−, and −/− genotypes, respectively; F=3.59, P=.028 by ANOVA). We also conducted analyses among male and female patients separately. The association remained statistically significant in the 350 male patients (likelihood ratio χ2=14.02, P=.02). The same trend was observed among the 125 female patients but did not reach statistical significance (P=.21). The small number of female patients could be the reason. Only 1 female patient with triple-vessel disease had Pvu II(−/−) genotype, whereas the expected number of patients in this cell calculated by the model was 3.6. In contrast, the LpL HindIII polymorphism did not show any association with either the number of diseased vessels (P=.456) or the Green Lane scores [5.63±0.29, 5.26±0.35, and 6.38±0.60 for HindIII(+/+), (+/−), and (−/−), respectively; P=.281].
Because there was a strong linkage disequilibrium between Pvu II and HindIII, we included both polymorphisms in the log linear model. This analysis also assessed the relationship between the diseased vessels and the haplotypes defined by both LpL Pvu II and HindIII polymorphisms. After the HindIII polymorphism was controlled for, the relationship between Pvu II and the number of diseased vessels remained highly significant (χ2=16.336, P=.0026). When we controlled for the Pvu II polymorphism, HindIII also showed a weak relationship with CAD severity (χ2=9.124, P=.0581). The HindIII and Pvu II polymorphisms were still in disequilibrium after we controlled for the number of significantly diseased vessels (χ2=42.67, P=.000001). The three-way interaction among these three variables also was evident (χ2=15.32, P=.0531). However, the association is weak, and combining the genotypes did not predict the number of diseased vessels.
Using logistic regression analysis, we also included in the model the quantitative variables—levels of lipoproteins, apolipoproteins, BMI, smoking dose, and age—and the categorical variables—sex, LpL Pvu II, HindIII. The factors found to be predictive of severity in our earlier study1 —sex, diabetes, smoking dose, ratio of total cholesterol to HDL cholesterol, log Lp(a), age, positive family history, and hypertension—remained significant predictors in this study, whereas the LpL Pvu II polymorphism was still independently predictive (P<.01) of the number of significantly diseased vessels. In particular, triglyceride level was not a significant predictor of the number of significantly diseased vessels. This was the case in the analysis of all patients and when the analysis was confined to those patients not receiving lipid-lowering drugs.
LpL Pvu II and HindIII RFLPs and Lipoprotein Variables
There was no direct association between the Pvu II or HindIII genotype and triglyceride levels. Mean triglyceride levels (±SEM) were 2.27±0.1 and 2.07±0.06 mmol/L for the 125 patients receiving lipid-lowering drugs and the 354 nonusers (P=.10). Because the usage of lipid-lowering drugs could attenuate the effect of the polymorphisms on triglyceride levels, we assessed the effects of both Pvu II and HindIII on triglyceride levels only in those patients not receiving lipid-lowering drugs. As Table 5⇓ shows, there were statistically significant differences among different genotypes. There was clearly a dose-dependent relationship between the presence of Pvu II(−) alleles and low triglyceride levels but no consistent relationship between triglyceride and HDL cholesterol and the HindIII genotypes (Table 5⇓). The relationship between the Pvu II polymorphism and triglyceride levels was independent of BMI in a multivariate ANOVA even though BMI was associated with triglyceride levels as expected (r=.158, P=.001). BMI was not different among patients with different Pvu II genotypes.
We also evaluated the effects of the Pvu II and HindIII polymorphisms on triglyceride and HDL cholesterol levels in patients with and without significantly diseased vessels. As Table 6⇓ shows, the effects were far stronger in patients with no significantly diseased vessels. There was a linear decrease in triglyceride levels with the presence of Pvu II(−) alleles (P=.0044) and a linear increase in HDL cholesterol levels with the presence of HindIII(−) alleles (P=.023) in patients without significantly diseased vessels. The relationship between the Pvu II polymorphism and triglyceride levels was more highly significant than that between the HindIII polymorphism and HDL cholesterol levels.
As expected, there was a significant negative correlation between levels of HDL cholesterol and triglycerides among all patients (r=−.446, P=.0001), which also was seen in patients with Pvu II(−/−) (r=−.345, P=.009), Pvu II(+/−) (r=−.505, P=.0001), and Pvu II(+/+) genotypes (r=−.430, P=.0001).
LpL Pvu II and HindIII Polymorphisms and Diabetes and Other Medical Conditions
As Table 7⇓ shows, the coexistence of diabetes with the Pvu II(+/+) genotype was far more than that by chance alone (r=.152, P=.0025 by logistic regression analysis). The OR was 3.12 (95% CI, 1.30 to 7.49) for the association of Pvu II(+/+) genotype with diabetes compared with the Pvu II(−/−) genotype (Table 7⇓). But there was no dosage effect in relation to the (+) allele (Table 7⇓). The inclusion of triglyceride levels and BMI in the predictive model did not alter the relationship between the Pvu II polymorphism and diabetes (r=.140, P=.007); the triglyceride level itself, which was elevated in the patient population as a whole, also was not significantly associated with diabetes in our patients (P=.715), whereas BMI was strongly associated with diabetes (r=.177, P=.0006). The findings for diabetes were not altered after exclusion of patients using lipid-lowering drugs or when the analysis was among male patients only. The BMI was, as expected, larger in the diabetic than the nondiabetic patients (30.1±0.7 [n=55] versus 27.9±0.2 kg/m2 [n=420], P=.003).
Neither the Pvu II nor the HindIII genotype was related to a positive family history of premature CAD categorically (either negative or positive history) or quantitatively (the number and age of first-degree relatives with CAD). There were also no associations with hypertension (likelihood ratio χ2=4.01, P=.13), or severity of angina pectoris (χ2=0.95, P=.91), or history of myocardial infarction (χ2=3.25, P=.196). Patients with unstable angina had lower apoA-I levels (P=.012) and higher Lp(a) levels (P=.05) than patients with no or stable angina, but other lipid variables were not different among patients with angina of differing severity. Lipoprotein profiles also were not different among patients with different family histories of premature CAD.
Our present study explored the association between DNA polymorphisms of the LpL gene and CAD severity. We determined severity from the number of significantly stenosed (>50%) major coronary arteries and a commonly used scoring system.16 Our findings relate to a representative group of consecutively investigated white patients judged by cardiologists to require coronary angiography. This selected patient population is, of course, different from the general population but is representative of patients currently being assessed for the diagnosis and severity of CAD. Nevertheless, the allele frequencies of Pvu II(−) and HindIII(−) in our study were similar to those found in other white populations. Ahn et al17 reported allele frequencies of 0.45 and 0.26 for Pvu II(−) and HindIII(−), respectively, in a control population (n=539). Peacock et al5 also reported frequencies of 0.435 and 0.228 for Pvu II(−) and HindIII(−), respectively, among 92 healthy control subjects. There was a strong linkage disequilibrium between the two Pvu II and HindIII polymorphisms we measured; this disequilibrium also was found in several other studies.5 6 7 8 18 19
LpL Polymorphisms, CAD Severity, and Lipids
The strong association between the Pvu II polymorphism and the number of significantly diseased vessels described above was independent of lipoprotein and apolipoprotein levels, particularly triglyceride and HDL cholesterol levels. Patients with the Pvu II(−) allele were less likely to have severe triple-vessel disease, whereas the +/+ genotype was associated with more triple-vessel disease although this association was less strong. Because the base substitution for this polymorphism occurs in the middle of intron 6, it may not have a direct functional role in this association. More likely, it is in linkage disequilibrium with changes at other sites that alter the systemic and local expression of LpL in response to environmental change. Alternatively, it could also be in linkage disequilibrium with some mutations that produce defective LpL with partial activity, which is expressed with unfavorable environmental influences. For example, Ma et al20 reported that a ser172-cys mutation in exon 5 caused hypertriglyceridemia during pregnancy. Hayden et al21 recently reviewed gene and environment interaction in relation to defective LpL expression.
We have no information on mechanisms that could be mediating the association between the Pvu II polymorphism and CAD severity, but it is known that LpL may enhance the retention of LDL and VLDL particles by the subendothelial matrix.3 The polymorphism may be relevant to that atherogenic effect. Although the Pvu II and HindIII polymorphic markers were in strong linkage disequilibrium in our study, we failed to show an independent association between the HindIII polymorphism and severity. However, there was a weak combined effect of these two polymorphisms on severity (P=.056).
Reports about the relationship between LpL polymorphisms and the severity of CAD have been controversial. Mattu et al18 and Thorn et al6 found a significant association between HindIII(+) and an increase in severity of CAD, whereas Peacock et al5 reported that HindIII(+) was associated with decreased severity. Neither group found a relationship with the Pvu II polymorphism. It is difficult to reconcile these observations except to note that the patient numbers in these studies were considerably smaller than in our own and that the other studies used only the Green Lane scoring system to quantify severity. The Green Lane score is an assessment of the extent and site of the arterial stenosis and takes into account the amount of myocardium supplied by the vessel. Although it is a good indicator of the severity of ischemic heart disease, which in our study correlated strongly with the number of significantly stenosed arteries, it represents a different end point in the evaluation of atherosclerosis, particularly in the assessment of a lipid-related atherogenic factor. Our finding that Pvu II had a stronger association with the number of diseased vessels than with the Green Lane scores could offer some support for this view. Nevertheless, the Green Lane scores in our study were also significantly related to Pvu II.
Reports about the associations between lipid variables and LpL Pvu II and HindIII polymorphism have been just as controversial. Several studies have reported significant associations between HindIII(+) and higher triglyceride and lower HDL cholesterol levels,4 5 6 7 findings not confirmed by Heinzmann et al.8 Chamberlain et al7 reported significant associations between Pvu II and triglyceride and HDL cholesterol levels, another finding not confirmed by others.4 5 6 Our findings are in agreement with those of Chamberlain et al7 in that there was a significant dosage-dependent relationship between elevated triglycerides and the Pvu II(+) allele. We also found a trend for higher triglyceride levels in patients with the HindIII(+) allele. Although no association between HindIII and HDL cholesterol levels was found in all patients, a statistically significant dosage-dependent relationship between high HDL cholesterol levels and HindIII(−) alleles was observed in patients with no significantly diseased vessels. In our study, the relationships between triglycerides and Pvu II genotypes were also stronger among subgroups of patients who were not using lipid-lowering drugs or had no significantly stenosed vessels. These findings indicate that confounding factors such as the presence of CAD or the use of lipid-lowering drugs could affect the relationship between LpL polymorphism and circulating lipid levels. Nonetheless, our study clearly identifies an important contribution by the LpL gene to the severity of coronary atherosclerosis, which may not be related to circulating levels of triglyceride and HDL cholesterol.
Pvu II Polymorphism and Diabetes
Our patients with the Pvu II(+/+) genotype were significantly more likely to have diabetes (OR, 3.12; 95% CI, 1.30 to 7.49; P=.001). As far as we are aware, this has not been reported previously. Hypertriglyceridemia and decreased adipose tissue LpL activity occur commonly in diabetic subjects22 ; therefore, it is relevant that the association we found was independent of triglyceride levels. Our diabetic patients as a group were considerably overweight (BMI, 30.1±0.7 kg/m2), significantly more so than the overweight nondiabetic patients (BMI, 28.2±0.2 kg/m2), and virtually all had non–insulin-dependent diabetes. Whether or not obese patients homozygous for the Pvu II(+/+) genotype are at increased risk of developing non–insulin-dependent diabetes requires further exploration.
In summary, our study describes associations between polymorphisms of the LpL gene and CAD severity, maturity-onset diabetes, and some lipid variables. These observations may contribute to an understanding of the determinants of CAD severity.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|CAD||=||coronary artery disease|
|RFLP||=||restriction fragment length polymorphism|
This work was supported by grants from the National Health and Medical Research Council of Australia and the Rebecca Cooper Medical Research Foundation. We wish to thank Dr Bridget Wilcken for reviewing the manuscript; L. Fenech, S. Brown, S. Brouwer, Dr Greg Cranny, and all the nurses in the Eastern Heart Clinic for their assistance in clinical data collection; A.S. Sim for laboratory assistance; and J. Kessy for data entry. We are also grateful to all the cardiologists in the department for allowing us to study their patients.
- Received August 7, 1995.
- Revision received October 25, 1995.
- Accepted October 30, 1995.
- Copyright © 1996 by American Heart Association
Wang XL, Tam C, McCredie RM, Wilcken DEL. Determinants of severity of coronary artery disease in Australian men and women. Circulation. 1994;89:1974-1981.
Saxena U, Klein MG, Vanni TM, Goldberg IJ. Lipoprotein lipase increases low density lipoprotein retention by subendothelial cell matrix. J Clin Invest. 1992;89:373-380.
Peacock RE, Hamsten A, Nilsson-Ehle P, Humphries SE. Associations between lipoprotein lipase gene polymorphisms and plasma correlations of lipids, lipoproteins and lipase activities in young myocardial infarction survivors and age-matched healthy individuals from Sweden. Atherosclerosis. 1992;97:171-185.
Miller SA, Dykes DD, Polesky HF. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 1988;16:1215.
Oka K, Tkalcevic GT, Stocks J, Galton DJ, Brown WV. Nucleotide sequence of Pvu II polymorphic site at the human lipoprotein lipase gene locus. Nucleic Acids Res. 1989;17:6752.
Gotoda T, Yamada N, Murase T, Shimano H, Shimada M, Harada K, Kawamure M, Kozaki K, Yazaki Y. Detection of three separate DNA polymorphisms in the human lipoprotein lipase gene by gene amplification and restriction endonuclease digestion. J Lipid Res. 1992;33:1067-1072.
Wang XL, Dudman NPB, Wilcken DEL. Enzyme-linked immunosorbent assay for apolipoprotein B in blood spotted onto filter paper, suitable for neonatal screening. Clin Chem. 1989;35:1000-1004.
Ahn YI, Kamboh MI, Hamman RF, Cole SA, Ferrell RE. Two DNA polymorphisms in the lipoprotein lipase gene and their associations with factors related to cardiovascular disease. J Lipid Res. 1993;34:421-428.
Mattu RK, Needham EW, Morgan R, Rees A, Hackshaw AK, Stocks J, Elwood PC, Galton DJ. DNA variants at the LpL gene locus associate with angiographically defined severity of atherosclerosis and serum lipoprotein levels in a Welsh population. Arterioscler Thromb. 1994;14:1090-1097.
Ma Y, Liu M-S, Ginzinger D, Frohlich J, Brunzell JD, Hayden MR. Gene-environment interaction in the conversion of a mild-to-severe phenotype in a patient homozygous for a Ser172→Cys mutation in the lipoprotein lipase gene. J Clin Invest. 1993;91:1953-1958.
Simsolo RB, Ong JM, Saffari B, Kern PA. Effect of improved diabetes control on the expression of lipoprotein lipase in human adipose tissue. J Lipid Res. 1992;33:89-95.