journal of Internal Medicine 1991 : 2 3 0 : 3 9 7 4 0 5
Effect of apolipoprotein E polymorphism and Xbal polymorphism of apolipoprotein B on response to lovastatin treatment in familial and non-familial hypercholesterolaemia J.-P. O J A L A t , E. H E L V E t , C. EHNHOLMS, K. AALTO-SETALAS, K. K. KONTULA*§ & M. J. TIKKANEN From the First. Second and t Third Departments of Medicine, University of Helsinki, the *National Public Health Institute, and the 5 Institute of Biotechnology. University of Helsinki, Helsinki, Finland
Abstract. Ojala J-P, Helve E, Ehnholm C, Aalto-Setala K, Kontula KK, Tikkanen MJ (First, Second and Third Departments of Medicine, University of Helsinki, National Public Health Institute, and Institute of Biotechnology, University of Helsinki, Helsinki, Finland). Effect of apolipoprotein E polymorphism and Xbal polymorphism of apolipoprotein B on response to lovastatin treatment in familial and non-familial hypercholesterolaemia. journal of lnternal Medicine 1991: 2 3 0 : 3 9 7 4 0 5 . Despite the well-documented efficacy of lovastatin, a wide inter-individual variation in treatment responses has been observed. The aim of the present study was to investigate the possible roles of apolipoprotein E (apo E) phenotype and apolipoprotein B (apo B) Xbal genotype on this variation. The apo E phenotype was determined in 232 subjects (78 cases of familial hypercholesterolaemia [FH] and 154 cases of non-familial hypercholesterolaemia [non-FH]) and the apo B Xbal genotype was determined in 21 1 subjects (67 cases of FH, 144 cases of non-FH). Depending on their baseline total serum cholesterol levels, these patients used a starting dose of lovastatin of either 20 or 40 mg nightly. After 6 weeks of therapy, slightly but significantly smaller reductions in LDLcholesterol were observed in patients with the E4/3 phenotype compared with those with the E3/3 phenotype in non-FH with lovastatin 20 mg (-20 vs.-28%; P = 0.043) and in total cholesterol in FH with lovastatin 40 mg (-23 vs.-27%; P = 0.023). No significant differences were found in non-FH patients starting with lovastatin, 40 mg. After doubling of the lovastatin doses, all treatment responses became similar among apo E phenotypes. Moreover, when all patients using lovastatin 40 mg either at 6 or 12 weeks were pooled ( n = 224), no differences in treatment responses were observed between the E3/2, E3/3, E4/3 and E4/4 phenotypes. The apo B Xbal genotype did not affect the hypocholesterolaemic efficacy of lovastatin in any of the patient groups. Thus our results indicate that inter-individual variation in the treatment response to lovastatin in both familial and non-familial hypercholesterolaemia is mainly due to factors other than the apo E phenotype or apo B Xbal genotype. K e y w o r d s : apolipoprotein B, apolipoprotein E, genetic polymorphism, lovastatin, primary hypercholesterolaemia, treatment response.
Introduction Lovastatin, one of the hydroxymethylglutaryl coenzyme A (HMG CoA) reductase inhibitors, is an effectivehypocholesterolaemic agent. Its mechanism of action is to reduce the synthesis of cholesterol,
mainly in the liver, resulting in up-regulation of the low-density-lipoprotein (LDL) receptors. This in turn enhances the clearance of LDL particles and verylow-density lipoprotein (VLDL) and VLDL remnant particles (containing both apolipoprotein B [apo B] and apolipoprotein E [apo El). This results in the
397
398
J.-P. OJALA et al.
lowering of plasma LDL-cholesterol and apo B levels, and may also contribute to the less consistent fall in plasma triglycerides [l]. Activation of the LDLreceptors is apparently the primary mechanism underlying the lowering of cholesterol levels by lovastatin in subjects with heterozygous familial hypercholesterolaemia (FH) [2]. Turnover studies suggest that lovastatin can also decrease the production of apo B-containing lipoproteins in the liver in other forms of hypercholesterolaemia [3. 41. A wide inter-individual variation in treatment responses to lovastatin in terms of LDL cholesterollowering has been observed in a number of studies [S-81. As with many other drugs [9], this variation can be partially explained by various environmental and genetic factors that affect the disposition of lovastatin (i.e. absorption, distribution, biotransformation, excretion, or a combination of these) in each individual. However, environmental and genetic factors may also exert their effects by modulating substrates and/or structures mandatory for the action of the drug. Thus, bearing in mind the mechanism of action of lovastatin, all factors that alter the function of the LDL-receptor or the ligands for the receptor (i.e. apo B or apo E) might be expected to contribute to the variation in treatment response to lovastatin. Two genetic polymorphisms, the apo E polymorphism and the Xbal polymorphism of apo B, have been shown to be associated with an altered in-vitro or in-vivo function of the apolipoprotein which they are coding [lo-1 21. Interestingly, both polymorphisms may affect the response of serum cholesterol to dietary intervention [13, 141, and the apo E polymorphism may influence the serum cholesterol response to procubol [15]. The present study was designed to establish whether these two genetic polymorphisms also affect the treatment response to lovastatin, and whether such an influence would differ in familial and non-familial (non-FH) hypercholesterolaemia.
Methods Patients
The present study is a re-analysis (according to the apo E phenotype and the apo B Xbal genotype) of the treatment efficacy data for patients who received lovastatin either in a previously reported, randomized, double-blind multicentre comparison
study between lovastatin and gemfibrozil (‘ parallel study ’, n = 1 67) [161, or in its open-label extension study, where the patients treated with gemfibrozil were switched to lovastatin therapy (‘switch study ’, n = 141) [17]. Sera or fresh blood samples for the retrospective apo E phenotyping or Xbal genotyping of apo B could not be obtained from all centres. Thus determination of the apo E phenotype was possible in 232 patients, the apo B Xbal genotype was determined in 2 l l patients, and both were determined in 155 of the total of 308 patients who participated in these two studies. A total of 8 6 patients were diagnosed as having FH on the basis of clinical assessment, the criteria being hypercholesterolaemia associated with tendon xanthomas and/or a family history indicative of FH [18]. However, since it is difficult to diagnose FH solely on clinical grounds, it is possible that a small proportion of the patients were wrongly classified. The remaining subjects (including the patients with possible familial combined hyperlipidaemia) were classified as having non-familial hypercholesterolaemia. Study design
The study designs of the ‘parallel’ and ‘switch’ studies have been described in detail previously [16, 171. Briefly, 334 patients whose total serum cholesterol and triglyceride levels on a standard cholesterol-lowering diet (American Heart Association Phase 1 diet) were 2 6.2 mmol I-’ and < 4.00 mmol I-’, respectively, entered the ‘parallel’ study [16]. After a 4-week placebo run-in period, the patients were stratified into stratum 1 (cholesterol 6.2-7.79 mmol I-’) or stratum 2 (cholesterol 2 7.8 mmol I-’), and in both strata were randomly assigned to either lovastatin or gemfibrozil treatment. The patients in the lovastatin arm of the ‘parallel’ study commenced the study on lovastatin, 2 0 mg nightly (stratum 1) or 4 0 mg nightly (stratum 2). After 6 weeks the dose of lovastatin was doubled if the target of 5 mmol I-’ for total serum cholesterol had not been reached. After the ‘parallel’ study, most of the gemfibrozil-treated subjects entered the ‘switch’ study, in which they were switched to lovastatin therapy [171. The original stratification was maintained, and the study design was identical to that of the lovastatin arm of the ‘parallel’ study. Other medication known to influence lipids and the lipid-lowering diet remained unchanged throughout both studies.
APO E A N D B POLYMORPHISMS
399
Table 1 . Distribution and clinical characteristics of the patients, according to the apolipoprotein E phenotypes Apolipoprotein E phenotype
E3/2
n (no. of women) Frequency (%)
E3/3
E4/2
141 (77)
7 (3) 3.0
E4/3
E4/4
72 (35)
1 1 (7) 4.7
31.0
60.8
+ 10
Age (years)
54k11
52
50k9
54*7
Body mass index (kg m-')
27.1 k 4 . 3
25.4k3.0
26.0 k 4 . 0
26.0k3.5
Concomitant drugs Beta-blockers (yes/no) Diuretics (yes/no) Oestrogens (yes/no)
116 116 017
45/96 241117 51136
32/40 11/61 5/67
318 011 1
FH/non-FH
215
45/96
28/44
318
215 611
42/99
14/58
318
123118
6616
912
Stratum l/Stratum 2" Doubled dose of lovastatin at week 12 (yes/no)
* The baseline total cholesterol levels were 6.2-7.79 and
516
7.8 mmol I-' in strata 1 and 2, respectively.
All differences between the phenotypes were non-significant. Mean values _+ SD are shown.
Our previous report indicated that the treatment response to lovastatin was similar in the 'parallel' and 'switch ' studies [17], and therefore in the present re-evaluation the data for the patients who received lovastatin in either of these studies were combined. Laboratory met hods Lipid, lipoprotein and apolipoprotein analyses. Total serum cholesterol, triglycerides and HDL-cholesterol levels were determined in a central laboratory (Yhtyneet Laboratoriot, Helsinki, Finland). Cholesterol and triglycerides were determined by an enzymatic colorimetric method (SMAC-analyzer, Technicon, Tarrytown, New York, USA). The HDLfraction was obtained by the Mg '+/dextran sulphate precipitation method [191. LDL-cholesterol levels were calculated using the Friedewald formula [20]. Apo E phenotyping. Apo E phenotyping was performed using a modification [21] of the method of Havekes et al. [22], which is based on isoelectric focusing of delipidated plasma followed by immunoblotting using apoE antiserum. Cysteamine treatment of plasma prior to isoelectric focusing was performed as described by Weisgraber et al. [23].
Apo B Xbal genotyping. DNA was isolated from frozen whole blood samples and digested with the restriction enzyme Xbal (Promega; 2-5 U pg-' DNA), frac-
tionated by gel electrophoresis on 0.6% agarose, and transferred to nitrocellulose filters. The conditions for prehybrization, hybridization, washing and autoradiography of the filters were as described previously [24]. Labelling of apo B cDNA probe pB23 [25] was accomplished by random oligonucleotide priming. The allele resulting in the formation of an 8.6-kb Xbal restriction fragment is designated X1, and the allele which generates a 5-kb fragment is designated x2. Statistical analyses
The drug efficacy within and between groups with different genotypes was assessed by analysing the percentage changes from baseline after 6 and 12 weeks of treatment with lovastatin. The values obtained at the end of the placebo run-in period of the 'parallel' study were used as baseline values. Statistical analyses were performed using the BMDP statistical software package (University of California Press, 1988). The Wilcoxon signed-rank test was used for pairwise comparison of means within groups. Comparisons between groups were made by parametric analysis of variance where appropriate, according to Shapiro and Wilk's W statistic : otherwise, the non-parametric Kruskal-Wallis test was used. In addition, the baseline lipid and lipoprotein levels were adjusted for age and body mass index, using analysis of covariance. Because this adjustment did
400
J.-P. OJALA et al.
6 4
? a
2 ?
3
tl
not alter the original results, these calculations have not been described separately. The differences in categorical variables (sex, use of beta-blocking agents, diuretics and oestrogens, as well as the number of patients with no dose titration at week 6) between groups were analysed by the Pearson Chisquare test.
d!
T 27
.*. m
w
n v
I
2tl -?
n
I
xj +I d!
m
k Y
m
I
w
'9 n +I \ N
x
m
I
w
Y
-s
t
m
a?
Apolipoprotein E polymorphism and baseline parameters
-4
\
..
Results
m Z
pz r(
+I
o! Do
y! 3
+I p? a\
P-
+ +I F
m
-?
+ +I
2
The distribution of apo E phenotypes and the clinical characteristics of the patients are shown in Table 1. Approximately 90% of the subjects had the E3/3 or E4/3 phenotype, while the remaining 10%had the E4/4 or E3/2 phenotype. One patient with the E4/2 phenotype was excluded from subsequent analyses. Thus the following analyses concern the effects of the E3/2, E3/3, E4/3 and E4/4 phenotypes on the treatment response to lovastatin. There were no statistically significant differences between the phenotypes with regard to mean age, body mass index, sex distribution, or the number of patients who were using beta-blocking agents, diuretics or oestrogens, or who had familial hypercholesterolaemia. The dose of lovastatin was doubled in most patients at week 6 because the predefined target for serum total cholesterol had not been reached (see Study design section). However, the number of patients with dose titration did not differ significantly between the apo E phenotypes. The data for baseline lipid and lipoprotein levels are summarized in Table 2. Apart from the HDLcholesterol levels (P = 0.046), the baseline levels did not differ significantly between phenotypes. Moreover, a higher triglyceride level was observed in the E3/2 phenotype than in the other phenotypes, although the difference was not statistically significant.
Apo E phenotype and response to lovastatin. 40 mg nightly In the first set of analyses, the data for all patients who were treated with lovastatin, 40 mg nightly, after either 6 weeks of treatment in stratum 2 ( n = 170) or 12 weeks of treatment in stratum 1( n = 54) were combined. Seven patients who remained on the
APO E A N D B POLYMORPHISMS
E3/2 (n = 6)
401
E3/3
E4/3
€414
(n = 135)
(n = 72)
(n=ll)
-40
Fig. 1 . Percentage changes in LDL-cholesterol. according to the apo E phenotypes. in subjects treated with lovastatin. 40 mg nightly. Ears indicate mean values.
-50 L
-60
1
20-mg dose in stratum 1 and the one E4/2 patient were excluded. Thus 224 of the 232 patients shown in Table 1 were included in these analyses. All changes from baseline induced by lovastatin treatment were significant, except for the increases in HDL-cholesterol concentration in the E3/2 and E4/4 phenotypes (Table 2). Despite the large variation within each phenotype with regard to the LDLcholesterol response (Fig. l),significant differences between phenotypes were not observed. Moreover, in order to minimize the effects of possible noncompliers, the analyses were repeated after excluding individuals with a poor treatment response (defined as < 10% reduction in LDL-cholesterol levels), but this did not alter the results. Furthermore, there were no significant differences in changes in total cholesterol, HDL-cholesterol or triglycerides between phenotypes (Table 2).
Apo E phenotype and response to lovastatin in non-FH and FH Due to the small sample sizes in the E3/2 and E4/4 phenotypes (see Table 1)only the effects of the E3/3 and E4/3 phenotypes on treatment response were analysed separately in the non-FH and FH patient groups. The non-FH group consisted of strata 1 and 2, but the FH group consisted only of strata 2 (see Study design section). Apart from higher BMI values in E4/3 subjects than in E3/3 subjects in stratum 1, clinical characteristics (cf. Table 1) were similar among the apo E phenotypes at baseline. The baseline levels and percentage changes in total and LDLcholesterol in the E3/3 and E3/4 phenotypes after 6
and 12 weeks of treatment are shown in Table 3. The baseline levels of either these lipid parameters or HDL-cholesterol and triglycerides (data not shown) did not differ significantly between the phenotypes. Significant differences in LDL-cholesterol were observed in the non-FH patients who were receiving lovastatin 2 0 mg nightly (decreases of 27.9 vs. 20.1 %; P = 0.043) in E3/3 and E4/3, respectively, and in total cholesterol in the FH group using lovastatin 4 0 mg nightly (decreases of 27.4 vs. 22.6%; P = 0.023) (Table 3). In addition, significant differences in HDL-cholesterol increases ( + 10.5 vs. 1.0%;P = 0.024) in E3/3 and E4/3, respectively, were observed in non-FH patients taking 20 mg of lovastatin, and in triglyceride decreases (-26.8 vs.-ll.O%; P = 0.010)inE3/3 andE4/3, respectively, in FH patients taking 4 0 mg of lovastatin (data not shown). However, at week 12 with doubled doses of lovastatin, all responses of lipid fractions in both non-FH and FH groups became uniform among phenotypes. These results remained unchanged after inclusion of the data for the few patients (Table 1) whose lovastatin dose was not doubled after 6 weeks of treatment (data not shown).
+
Xbal polymorphism of the apolipoprotein B The effects of the Xbal genotypes (i.e. X l X 1 , X1X2 and X2X2) on treatment responses were analysed separately in the non-FH and FH patient groups. The non-FH groups consisted of strata 1 ( n = 68) and 2 ( n = 76), while the FH group only included stratum 2 ( n = 67) (see Study design section). Apart from the fact that a higher proportion of the non-FH patients
6.5k0.8 n = 23 10.8k1.5 8.7k1.6
6.5k0.8
n = 39 10.9kl.5
9.Ok1.6
n = 28
5.3k0.5
n = 13 7.4k0.6
8.9kO.8
= 49
5.3k0.6
34 7.5k0.6
=
8.7k0.8
n
n
E4/3
NS
NS
NS
NS
NS
NS
P-value'
E4/3
-20.1k9.4
-14.5k8.3
-32.1k9.6
-27.4f7.0
-35.9k11.1
-28.0f8.7
-29.4k9.9
-22.6k8.9
-36.5k9.3
-27.3k5.4
Lovastatin 40 mg nightly
-27.9k12.2
-19.0k6.9
Lovastatin 20 mg nightly
E3/3
NS
0.023
NS
NS
0.043
NS
P-value'
E4/3
-34.4511.8
-24.8k10.7
-39.0k9.4
-32.3k7.1
-45.0k10.1
-34.3f8.8
-38.0f8.6
-30.6k6.9
-43.8k7.6
-33.7f6.8
Lovastatin 80 mg nightly
-33.4k11.9
-24.0k8.4
Lovastatin 40 mg nightly
E3/3
Percentage change at week 12
NS
NS
NS
NS
NS
NS
=
P-value*
Mean values k SD are shown. All differences between baseline and treatment values were significant ( P < 0.05). * Level of significance of differences between the phenotypes estimated by analysis of variance or the Kruskal-Wallis test, depending on normalities of the distributions: NS significant. t The baseline total cholesterol levels were 6.2-7.79 and > 7.8 mmol I-' in strata I and 2, respectively.
Stratum 2 t Non-FH Total cholesterol LDLcholesterol FH Total cholesterol LDL cholesterol
cholesterol
LDL-
Stratum lt Non-FH Total cholesterol
E3/3
Percentage change at week 6
non-
?
P
4
phenotypes, in the non-FH and FH patients
Baseline level (mmol I-')
9
Table 3. Baseline levels and percentage changes in total and LDL-cholesterol levels after 6 and 12 weeks of treatment with lovastatin. according to the apolipoprotein E3/3 and E4/3
F
9
3
0
?
?
APO E AND B POLYMORPHISMS
with the X l X l genotype in stratum 2 used diuretics, the clinical characteristics (cf. Table 1)did not differ significantly between the genotypes. The lovastatin dose was not doubled at 6 weeks in 24 of the 211 subjects. The efficacy data at 6 weeks were analysed with and without inclusion of these patients, but the results obtained were similar. There were no significant differences between genotypes in baseline levels of any measured lipid parameter (data not shown). Uniform decreases in total and LDL-cholesterol were observed during treatment in all Xbal genotype groups. The average increase in HDL-cholesterol was somewhat greater in X l X l subjects than in the other genotypes after 12 weeks of treatment in the non-FH patients in stratum 2 (increases of 12.4% vs. 2.1% vs. 6.4% in X l X 1 , X1X2 and X2X2, respectively; P = 0.046, analysis of variance). Otherwise, the differences in HDLcholesterol and triglycerides between the genotypes in both non-FH and FH groups at 6 and 12 weeks (data not shown) were not statistically significant.
Discussion As we have also reported previously [7, 81, a wide inter-individual variation in treatment responses to lovastatin was observed in the present study. Although dietary and other environmental confounding factors were kept as constant as possible, they may still have influenced this variation. In addition, the treatment response could be dependent on variation at the apolipoprotein gene loci. In theory, the apo E and apo B Xbal polymorphisms could modulate the LDL-receptor/lipoprotein particle interaction and thereby interfere with the action of lovastatin. Both polymorphisms have been shown to affect the serum cholesterol response to dietary change in the Finnish population [13, 141. In the present study, combined data from two previous studies were re-analysed in order to determine whether these polymorphisms also affected the treatment response to lovastatin. Although the studies were not originally designed for this purpose, and blood samples were not available from all participating centres, it is unlikely that this would have biased our results. The loss of samples was probably random, and the number of analyses remained relatively large. The contribution of the apo E phenotype to serum total and LDL-cholesterol is well established in many populations [26-301. In Finland, the levels of these
403
lipid parameters in random samples of apparently healthy subjects have been shown to occur in the following order: E4/4 > E4/3 > E3/3 > E3/2 > E4/2 > E2/2 [13, 291. Such an association was not observed in our hypercholesterolaemic patients. One possible reason for this could be that the baseline lipid levels were determined during a low-fat, lowcholesterol diet, which has been shown to reduce the correlation between apo E phenotype and serum cholesterol levels [ 131. Theoretically, the response to lovastatin could be impaired in subjects who carry the e4 allele. This is based on the finding that the apo E4-containing postprandial lipoproteins are cleared more effectively by the liver than those with other apo E isoforms, probably resulting in hepatic cholesterol loading and suppression of HMG CoA reductase activity [31]. Accordingly, HMG CoA reductase could already be partially inhibited in these subjects, resulting in a blunted response to lovastatin. On the other hand, the efficiency of intestinal cholesterol absorption is reported to be greater in subjects who carry the e4 allele than in those who carry the e2 allele [32]. As there is preliminary evidence [ 3 31 to suggest that lovastatin inhibits intestinal cholesterol absorption, it could also be argued that lovastatin is more effective in patients who carry the e4 allele. The only significant differences between apo E phenotypes in terms of total or LDL-cholesterol reduction occurred after 6 weeks of treatment. In non-FH patients, LDL-cholesterol was reduced slightly, but significantly less in E4/3 subjects than in E3/3 subjects, and in FH patients a similar difference was observed between total cholesterol responses. After doubling of the lovastatin doses (from 2 0 to 4 0 mg in non-FH patients, and from 4 0 to 80 mg nightly in FH patients) even these slight differences disappeared completely. In theory, this could suggest that the e4 allele has a small opposing effect on cholesterol reduction with low doses of lovastatin. This appears unlikely, as individuals with two e4 alleles (apo E4/4 phenotype) did not exhibit any degree of diminished response to lovastatin (see Fig. 1, FH and non-FH subjects combined). In any case, the effects of apo E phenotypes on treatment response were clinically insignificant. When the data for all patients who were taking 40 mg of lovastatin were pooled and analysed together, the same result was obtained. While we are unaware of any previous report on the effects of apo E phenotype on the treatment
404
J.-P. OJALA et al.
response to HMG CoA reductase inhibitors in non-FH patients, two recent papers have addressed this question in patients with heterozygous FH. In these studies no significant difference was detected between apo E phenotypes in this respect during lovastatin [34] and simvastatin [35] therapy. This finding is confirmed by the present study and might be expected, as it is unlikely that the enhanced receptormediated removal of LDL by lovastatin would be influenced by apo E contained within VLDL and IDL. The slight reduction in LDL-cholesterol-lowering in E4/3 subjects vs. E3/3 subjects using the lowest lovastatin dose (20 mg d-l) in patients with non-FH may reflect some unknown mechanism of cholesterol elevation in these patients. De Knijff et al. [35] reported that a better response with regard to the LDL-cholesterol decrease was observed with simvastatin in women than in men with the E3/3 phenotype. We observed no sexrelated differences in response (data not shown). To our knowledge, this is the first report of a study of the possible effects of the apo B Xbal polymorphism on treatment response to an HMG CoA reductase inhibitor. This polymorphism is determined by two alleles: X 1 (absence of Xbal cutting site) and X2 (presence of Xbal cutting site), which produce three genotypes: X l X 1 , X1X2 and X2X2. The presence of the X2 allele has been associated with increased total and LDL-cholesterol levels in several [24, 36-40] but not all studies [25, 41, 421. In our study population, which consisted of high-risk individuals with severe hypercholesterolaemia, this association was not observed. The Xbal mutation does not result in an amino acid change [43]. However, recent in-vivo and in-vitro studies have indicated that, apparentiy due to impaired receptor binding properties of LDL, subjects with the X2X2 genotype catabolized LDL more slowly than X l X l individuals [ l l , 121, suggesting that the Xbal mutation is closely associated with an as yet unknown functionally significant mutation. Despite the possible role of the X2 allele as a marker of this functionally important DNA change, we did not observe any differences between the three Xbal genotypes within the non-FH groups or the FH group in terms of treatment response. This suggests that the proposed differences between LDL particles, if they exist, have no clinically significant effects on the lovastatin response. In conclusion, our results did not reveal any clinically significant effects of the apo E or apo B Xbal
polymorphisms on inter-individual variation in treatment response to lovastatin, in either familial or non-familial hypercholesterolaemia. Further studies are needed to investigate the effects of other genetic factors on this variation.
Acknowledgements Financial support for this study was provided by the Sigrid Juselius Foundation (to C.E. and M. J.T.), the Paavo Nurmi Foundation, the Academy of Finland and the University of Helsinki (to M. J.T.).
References 1 Grundy SM. HMG-CoA reductase inhibitors for treatment of hypercholesterolemia. N Engl ] Med 1 9 8 8 ; 3 1 9 : 24-33.
2 Bilheimer DW. Grundy SM. Brown MS. Goldstein JL. Mevinolin and colestipol stimulate receptor-mediated clearance of low density lipoprotein from plasma in familial hypercholesterolemia heterozygotes. Proc N a d Acad Sci USA 1 9 8 3 ; 80: 4124-8. 3 Grundy SM. Vega GL. Influence of mevinolin on metabolism of low density lipoproteins in primary moderate hypercholesterolemia. ] Lipid Res 1 9 8 5 ; 2 6 : 1464-75. 4 Arad Y, Ramakrishnan R. Ginsberg HN. Lovastatin therapy reduces low density lipoprotein apoB levels in subjects with combined hyperlipidemia by reducing the production of apoBcontaining lipoproteins : implications for the pathophysiology of apoB production. ] Lipid Res 1 9 9 0 ; 31 : 567-82. 5 Illingworth DR. Sexton GJ. Hypocholesterolemic effects of mevinolin in patients with heterozygous familial hypercholesterolemia. / Clin Invest 1 9 8 4 ; 74: 1972-8. 6 Hoeg JM. Maher MB, Zech LA el a/. Fffectiveness of mevinolin on plasma lipoprotein concentrations in type I1 hyperlipoproteinemia. A m J Cardiol 1 9 8 6 ; 5 7 : 933-9. 7 Ojala J-P, Helve E. Karjalainen K. Tarkkanen A, Tikkanen MJ. Long-term maintenance of therapeutic response to lovastatin in patients with familial and non-familial hypercholesterolemia: a 3-year follow-up. Atherosclerosis 1 9 9 0 : 82 : 85-95. 8 Ojala J-P, Helve E. Tikkanen MJ. Treatment of combined hyperlipidemia with lovastatin vs. gemfibrozil: a comparison study. Cardiology 1 9 9 0 ; 77 (Suppl. 4): 3 9 4 9 . 9 Vesell ES. Genetic and environmental factors affecting drug disposition in man. Clin Pharmacol Ther 1 9 7 9 ; 22: 659-79. 10 Davignon 1. Gregg RE, Sing CF. Apolipoprotein E polymorphism and atherosclerosis. Arteriosclerosis 1988 ; 8 : 1-21. 11 Demant T, Houlston RS, Caslake MJ et al. Catabolic rate of low density lipoprotein is influenced by variation in the apolipoprotein B gene. / Clin Invest 1988: 82: 797-802. 12 Series J, Cameron I. Caslake M. Gaffney D. Packard CJ. Shepherd J. The Xbal polymorphism of the apolipoprotein B gene influences the degradation of low density lipoprotein in vitro. Biochirn Biophys Acta 1989: 1003: 183-8. 13 Tikkanen MJ, Huttunen JK. Ehnholm C. Pietinen P. Apolipoprotein E, homozygosity predisposes to serum cholesterol elevation during high fat diet. Arteriosclerosis 1 9 9 0 : 10: 285-8.
A P O E A N D B POLYMORPHISMS
1 4 Tikkanen MJ, Xu C-F. Hamalainen Tet al. XbAl polymorphism of the apolipoprotein B gene influences plasma lipid response to diet intervention. CIin Genet 1 9 9 0 ; 37: 327-34. 1 5 Nestruck AC. Bouthillier D. Sing CF. Davignon J. Apolipoprotein E polymorphism and plasma cholesterol response to probucol. Metabolism 1987: 36: 743-7. 16 Tikkanen MJ, Helve E. Jaattela A et al. Comparison between lovastatin and gemfibrozil in the treatment of primary hypercholesterolemia : the Finnish multicenter study. Am j Cardiol 1 9 8 8 : 62: 3 5 J 4 3 J . 1 7 Ojala J-P. Helve E. Tikkanen MJ et al. Switch from gemfibrozil to lovastatin (mevinolin) therapy in patients with primary hypercholesterolaemia. A multicentre study. Drug Investigation 1990: 2 (Suppl. 2): 40-47. 18 Goldstein JL, Brown MS. Familial hypercholesterolemia. In: Stanbury JB. Wyngaarden JB, Fredrickson DS et a/.. eds. The Metabolic Basis of Inherited Disease. New York: McGraw Hill, 1983: 672-712. 1 9 Finley PR, Schifman RB. Williams RJ. Lichti DA. Cholesterol in high-density lipoprotein : use of Mg2+/dextran sulfate in its enzymic measurement. Clin Chem 1 9 7 8 ; 24: 931-3. 20 Friedewald WT. Levy RI. Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972 ; 18: 499-502. 21 Lukka M. Metso J, Ehnholm C. Apolipoprotein A-IV polymorphism in the Finnish population : gene frequencies and description of a rare allele. Hum Hered 1988; 38: 359-62. 22 Havekes LM.de KnijfTP. Beisiegel U. Havinga J. Smit M. Klasen E. A rapid micromethod for apolipoprotein E phenotyping directly in serum. I Lipid Res 1987: 28: 455-63. 23 Weisgraber KH. Innerarity TL. Mahley RW. Abnormal lipoprotein receptor-binding activity of the human E apoprotein due to cysteine-arginine interchange at a single site. I Biol Chem 1 9 8 2 ; 257: 2518-521. 24 Aalto-SetaII K. Tikkanen MJ, Taskinen M-R. Nieminen M, Holmberg P. Kontula K. Xbal and c/g polymorphisms of the apolipoprotein B gene locus are associated with serum cholesterol and LDL cholesterol levels in Finland. Arteriosclerosis 1988: 74: 47-54. 25 Hegele RA. Huang LS.Herbert PN et a/. Apolipoprotein Bgene DNA polymorphisms associated with myocardial infarction. N Englj Med 1986: 315: 1509-15. 26 Utermann G. Pruin N, Steinmetz A. Polymorphism of apolipoprotein E. 111. meet of a single polymorphic gene locus on plasma lipid levels in man. Clin Genet 1 9 7 9 ; 15: 63-72. 27 Menzel H-J. Kladetzky R-G. Assmann G. Apolipoprotein E polymorphism and coronary artery disease. Arteriosclerosis 1983: 3: 310-15. 28 Sing CF, Davignon J. Role of the apolipoprotein E polymorphism in determining normal plasma lipid and lipoprotein variation. Am I Hum Genet 1 9 8 5 ; 29: 477-84. 29 Ehnholm C. Lukka M, Kuusi T. Nikkila E. Utermann G. Apolipoprotein E polymorphism in the Finnish population : gene frequencies and relation to lipoprotein concentrations. j Lipid Res 1986: 27: 227-35.
405
3 0 Ordovas JM, Litwack-Klein L. Wilson PWF, Schaefer MM. Schaefer El. Apolipoprotein E isoform phenotyping methodology and population frequency with identification of apo E l and apo E5 isoforms. j Lipid Res 1987: 28: 371-80. 31 Weintraub MS, Eisenberg S. Breslow JL. Dietary fat clearance in normal subjects is regulated by genetic variation in apolipoprotein E. CIin Invest 1987: 80: 1571-7. 32 Kesaniemi YA. Ehnholm C. Miettinen TA. Intestinal cholesterol absorption efficiency in man is related to apoprotein E phenotype. I Clin Invest 1987: 80: 578-81. 33 Miettinen TA. Tikkanen MJ, Helve E, Ojala J-P. Inhibition of dietary cholesterol absorption during lovastatin (mevinolin) treatment. Drug Investigation 1990: 2 (Suppl. 2): 29-35. 34 O'Malley JP, Illingworth DR. The influence of apolipoprotein E phenotype on the response to lovastatin therapy in patients with heterozygous familial hypercholesterolemia. Metabolism 1990: 39: 150-54. 35 De Knijff P. Stalenhoef AFH, Mol MJTM et al. Influence of apo E polymorphism on the response to simvastatin treatment in patients with heterozygous familial hypercholesterolemia. Atherosclerosis 1990: 83: 89-97. 36 Berg K. DNA polymorphism at the apolipoprotein B locus is associated with lipoprotein level. Clin Genet 1986: 30: 515-20. 3 7 Law A, Powell LM. Brunt H et al. Common DNA polymorphism within coding sequence of apolipoprotein B gene associated with altered lipid levels. Lancet 1986: i: 1301-3. 38 Talmud PJ, Barni N. Kessling M et a/. Apolipoprotein B gene variants are involved in the determination of serum cholesterol levels : a study in normo- and hyperlipidaemic individuals. Atherosclerosis 1987: 67: 81-9. 39 Leren TP, Berg K. Hjermann I. Leren P. Further evidence for an association between the Xbal polymorphism at the apolipoprotein B locus and lipoprotein level. CIin Genet 7 988 : 34: 347-51. 4 0 Aalto-SetaIa K. Gylling H, Helve E et al. Genetic polymorphisms of the apolipoprotein B gene locus influence serum LDL cholesterol level in familial hypercholesterolemia. Hum Genet 1989: 82: 305-7. 4 1 Aburatani H. Matsumoto A, Itoh H et al. A study of DNA polymorphism in the apolipoprotein B gene in a Japanese population. Atherosclerosis 1988: 72: 71-6. 42 Darnfors C. Wiklund 0. Nilsson J et al. Lack of correlation between the apolipoprotein B XbAl polymorphism and blood lipid levels in a Swedish population. Atherosclerosis 1989: 75: 183-8. 4 3 Carlsson P. Darnfors C. OlotTson S, Bjursell G. Analysis of the human apoliprotein B gene: complete structure of the B74 region. Gene 1986: 49: 29-51. Received 18 January 1991, accepted 2 April 1991. Correspondence: Jukka-Pekka Ojala. MD. Third Department of Medicine, University of Helsinki, Haartmaninkatu 4, SF-00290 Helsinki, Finland.
Effect of apolipoprotein E polymorphism and Xbal polymorphism of apolipoprotein B on response to lovastatin treatment in familial and non-familial hypercholesterolaemia J.-P. O J A L A t , E. H E L V E t , C. EHNHOLMS, K. AALTO-SETALAS, K. K. KONTULA*§ & M. J. TIKKANEN From the First. Second and t Third Departments of Medicine, University of Helsinki, the *National Public Health Institute, and the 5 Institute of Biotechnology. University of Helsinki, Helsinki, Finland
Abstract. Ojala J-P, Helve E, Ehnholm C, Aalto-Setala K, Kontula KK, Tikkanen MJ (First, Second and Third Departments of Medicine, University of Helsinki, National Public Health Institute, and Institute of Biotechnology, University of Helsinki, Helsinki, Finland). Effect of apolipoprotein E polymorphism and Xbal polymorphism of apolipoprotein B on response to lovastatin treatment in familial and non-familial hypercholesterolaemia. journal of lnternal Medicine 1991: 2 3 0 : 3 9 7 4 0 5 . Despite the well-documented efficacy of lovastatin, a wide inter-individual variation in treatment responses has been observed. The aim of the present study was to investigate the possible roles of apolipoprotein E (apo E) phenotype and apolipoprotein B (apo B) Xbal genotype on this variation. The apo E phenotype was determined in 232 subjects (78 cases of familial hypercholesterolaemia [FH] and 154 cases of non-familial hypercholesterolaemia [non-FH]) and the apo B Xbal genotype was determined in 21 1 subjects (67 cases of FH, 144 cases of non-FH). Depending on their baseline total serum cholesterol levels, these patients used a starting dose of lovastatin of either 20 or 40 mg nightly. After 6 weeks of therapy, slightly but significantly smaller reductions in LDLcholesterol were observed in patients with the E4/3 phenotype compared with those with the E3/3 phenotype in non-FH with lovastatin 20 mg (-20 vs.-28%; P = 0.043) and in total cholesterol in FH with lovastatin 40 mg (-23 vs.-27%; P = 0.023). No significant differences were found in non-FH patients starting with lovastatin, 40 mg. After doubling of the lovastatin doses, all treatment responses became similar among apo E phenotypes. Moreover, when all patients using lovastatin 40 mg either at 6 or 12 weeks were pooled ( n = 224), no differences in treatment responses were observed between the E3/2, E3/3, E4/3 and E4/4 phenotypes. The apo B Xbal genotype did not affect the hypocholesterolaemic efficacy of lovastatin in any of the patient groups. Thus our results indicate that inter-individual variation in the treatment response to lovastatin in both familial and non-familial hypercholesterolaemia is mainly due to factors other than the apo E phenotype or apo B Xbal genotype. K e y w o r d s : apolipoprotein B, apolipoprotein E, genetic polymorphism, lovastatin, primary hypercholesterolaemia, treatment response.
Introduction Lovastatin, one of the hydroxymethylglutaryl coenzyme A (HMG CoA) reductase inhibitors, is an effectivehypocholesterolaemic agent. Its mechanism of action is to reduce the synthesis of cholesterol,
mainly in the liver, resulting in up-regulation of the low-density-lipoprotein (LDL) receptors. This in turn enhances the clearance of LDL particles and verylow-density lipoprotein (VLDL) and VLDL remnant particles (containing both apolipoprotein B [apo B] and apolipoprotein E [apo El). This results in the
397
398
J.-P. OJALA et al.
lowering of plasma LDL-cholesterol and apo B levels, and may also contribute to the less consistent fall in plasma triglycerides [l]. Activation of the LDLreceptors is apparently the primary mechanism underlying the lowering of cholesterol levels by lovastatin in subjects with heterozygous familial hypercholesterolaemia (FH) [2]. Turnover studies suggest that lovastatin can also decrease the production of apo B-containing lipoproteins in the liver in other forms of hypercholesterolaemia [3. 41. A wide inter-individual variation in treatment responses to lovastatin in terms of LDL cholesterollowering has been observed in a number of studies [S-81. As with many other drugs [9], this variation can be partially explained by various environmental and genetic factors that affect the disposition of lovastatin (i.e. absorption, distribution, biotransformation, excretion, or a combination of these) in each individual. However, environmental and genetic factors may also exert their effects by modulating substrates and/or structures mandatory for the action of the drug. Thus, bearing in mind the mechanism of action of lovastatin, all factors that alter the function of the LDL-receptor or the ligands for the receptor (i.e. apo B or apo E) might be expected to contribute to the variation in treatment response to lovastatin. Two genetic polymorphisms, the apo E polymorphism and the Xbal polymorphism of apo B, have been shown to be associated with an altered in-vitro or in-vivo function of the apolipoprotein which they are coding [lo-1 21. Interestingly, both polymorphisms may affect the response of serum cholesterol to dietary intervention [13, 141, and the apo E polymorphism may influence the serum cholesterol response to procubol [15]. The present study was designed to establish whether these two genetic polymorphisms also affect the treatment response to lovastatin, and whether such an influence would differ in familial and non-familial (non-FH) hypercholesterolaemia.
Methods Patients
The present study is a re-analysis (according to the apo E phenotype and the apo B Xbal genotype) of the treatment efficacy data for patients who received lovastatin either in a previously reported, randomized, double-blind multicentre comparison
study between lovastatin and gemfibrozil (‘ parallel study ’, n = 1 67) [161, or in its open-label extension study, where the patients treated with gemfibrozil were switched to lovastatin therapy (‘switch study ’, n = 141) [17]. Sera or fresh blood samples for the retrospective apo E phenotyping or Xbal genotyping of apo B could not be obtained from all centres. Thus determination of the apo E phenotype was possible in 232 patients, the apo B Xbal genotype was determined in 2 l l patients, and both were determined in 155 of the total of 308 patients who participated in these two studies. A total of 8 6 patients were diagnosed as having FH on the basis of clinical assessment, the criteria being hypercholesterolaemia associated with tendon xanthomas and/or a family history indicative of FH [18]. However, since it is difficult to diagnose FH solely on clinical grounds, it is possible that a small proportion of the patients were wrongly classified. The remaining subjects (including the patients with possible familial combined hyperlipidaemia) were classified as having non-familial hypercholesterolaemia. Study design
The study designs of the ‘parallel’ and ‘switch’ studies have been described in detail previously [16, 171. Briefly, 334 patients whose total serum cholesterol and triglyceride levels on a standard cholesterol-lowering diet (American Heart Association Phase 1 diet) were 2 6.2 mmol I-’ and < 4.00 mmol I-’, respectively, entered the ‘parallel’ study [16]. After a 4-week placebo run-in period, the patients were stratified into stratum 1 (cholesterol 6.2-7.79 mmol I-’) or stratum 2 (cholesterol 2 7.8 mmol I-’), and in both strata were randomly assigned to either lovastatin or gemfibrozil treatment. The patients in the lovastatin arm of the ‘parallel’ study commenced the study on lovastatin, 2 0 mg nightly (stratum 1) or 4 0 mg nightly (stratum 2). After 6 weeks the dose of lovastatin was doubled if the target of 5 mmol I-’ for total serum cholesterol had not been reached. After the ‘parallel’ study, most of the gemfibrozil-treated subjects entered the ‘switch’ study, in which they were switched to lovastatin therapy [171. The original stratification was maintained, and the study design was identical to that of the lovastatin arm of the ‘parallel’ study. Other medication known to influence lipids and the lipid-lowering diet remained unchanged throughout both studies.
APO E A N D B POLYMORPHISMS
399
Table 1 . Distribution and clinical characteristics of the patients, according to the apolipoprotein E phenotypes Apolipoprotein E phenotype
E3/2
n (no. of women) Frequency (%)
E3/3
E4/2
141 (77)
7 (3) 3.0
E4/3
E4/4
72 (35)
1 1 (7) 4.7
31.0
60.8
+ 10
Age (years)
54k11
52
50k9
54*7
Body mass index (kg m-')
27.1 k 4 . 3
25.4k3.0
26.0 k 4 . 0
26.0k3.5
Concomitant drugs Beta-blockers (yes/no) Diuretics (yes/no) Oestrogens (yes/no)
116 116 017
45/96 241117 51136
32/40 11/61 5/67
318 011 1
FH/non-FH
215
45/96
28/44
318
215 611
42/99
14/58
318
123118
6616
912
Stratum l/Stratum 2" Doubled dose of lovastatin at week 12 (yes/no)
* The baseline total cholesterol levels were 6.2-7.79 and
516
7.8 mmol I-' in strata 1 and 2, respectively.
All differences between the phenotypes were non-significant. Mean values _+ SD are shown.
Our previous report indicated that the treatment response to lovastatin was similar in the 'parallel' and 'switch ' studies [17], and therefore in the present re-evaluation the data for the patients who received lovastatin in either of these studies were combined. Laboratory met hods Lipid, lipoprotein and apolipoprotein analyses. Total serum cholesterol, triglycerides and HDL-cholesterol levels were determined in a central laboratory (Yhtyneet Laboratoriot, Helsinki, Finland). Cholesterol and triglycerides were determined by an enzymatic colorimetric method (SMAC-analyzer, Technicon, Tarrytown, New York, USA). The HDLfraction was obtained by the Mg '+/dextran sulphate precipitation method [191. LDL-cholesterol levels were calculated using the Friedewald formula [20]. Apo E phenotyping. Apo E phenotyping was performed using a modification [21] of the method of Havekes et al. [22], which is based on isoelectric focusing of delipidated plasma followed by immunoblotting using apoE antiserum. Cysteamine treatment of plasma prior to isoelectric focusing was performed as described by Weisgraber et al. [23].
Apo B Xbal genotyping. DNA was isolated from frozen whole blood samples and digested with the restriction enzyme Xbal (Promega; 2-5 U pg-' DNA), frac-
tionated by gel electrophoresis on 0.6% agarose, and transferred to nitrocellulose filters. The conditions for prehybrization, hybridization, washing and autoradiography of the filters were as described previously [24]. Labelling of apo B cDNA probe pB23 [25] was accomplished by random oligonucleotide priming. The allele resulting in the formation of an 8.6-kb Xbal restriction fragment is designated X1, and the allele which generates a 5-kb fragment is designated x2. Statistical analyses
The drug efficacy within and between groups with different genotypes was assessed by analysing the percentage changes from baseline after 6 and 12 weeks of treatment with lovastatin. The values obtained at the end of the placebo run-in period of the 'parallel' study were used as baseline values. Statistical analyses were performed using the BMDP statistical software package (University of California Press, 1988). The Wilcoxon signed-rank test was used for pairwise comparison of means within groups. Comparisons between groups were made by parametric analysis of variance where appropriate, according to Shapiro and Wilk's W statistic : otherwise, the non-parametric Kruskal-Wallis test was used. In addition, the baseline lipid and lipoprotein levels were adjusted for age and body mass index, using analysis of covariance. Because this adjustment did
400
J.-P. OJALA et al.
6 4
? a
2 ?
3
tl
not alter the original results, these calculations have not been described separately. The differences in categorical variables (sex, use of beta-blocking agents, diuretics and oestrogens, as well as the number of patients with no dose titration at week 6) between groups were analysed by the Pearson Chisquare test.
d!
T 27
.*. m
w
n v
I
2tl -?
n
I
xj +I d!
m
k Y
m
I
w
'9 n +I \ N
x
m
I
w
Y
-s
t
m
a?
Apolipoprotein E polymorphism and baseline parameters
-4
\
..
Results
m Z
pz r(
+I
o! Do
y! 3
+I p? a\
P-
+ +I F
m
-?
+ +I
2
The distribution of apo E phenotypes and the clinical characteristics of the patients are shown in Table 1. Approximately 90% of the subjects had the E3/3 or E4/3 phenotype, while the remaining 10%had the E4/4 or E3/2 phenotype. One patient with the E4/2 phenotype was excluded from subsequent analyses. Thus the following analyses concern the effects of the E3/2, E3/3, E4/3 and E4/4 phenotypes on the treatment response to lovastatin. There were no statistically significant differences between the phenotypes with regard to mean age, body mass index, sex distribution, or the number of patients who were using beta-blocking agents, diuretics or oestrogens, or who had familial hypercholesterolaemia. The dose of lovastatin was doubled in most patients at week 6 because the predefined target for serum total cholesterol had not been reached (see Study design section). However, the number of patients with dose titration did not differ significantly between the apo E phenotypes. The data for baseline lipid and lipoprotein levels are summarized in Table 2. Apart from the HDLcholesterol levels (P = 0.046), the baseline levels did not differ significantly between phenotypes. Moreover, a higher triglyceride level was observed in the E3/2 phenotype than in the other phenotypes, although the difference was not statistically significant.
Apo E phenotype and response to lovastatin. 40 mg nightly In the first set of analyses, the data for all patients who were treated with lovastatin, 40 mg nightly, after either 6 weeks of treatment in stratum 2 ( n = 170) or 12 weeks of treatment in stratum 1( n = 54) were combined. Seven patients who remained on the
APO E A N D B POLYMORPHISMS
E3/2 (n = 6)
401
E3/3
E4/3
€414
(n = 135)
(n = 72)
(n=ll)
-40
Fig. 1 . Percentage changes in LDL-cholesterol. according to the apo E phenotypes. in subjects treated with lovastatin. 40 mg nightly. Ears indicate mean values.
-50 L
-60
1
20-mg dose in stratum 1 and the one E4/2 patient were excluded. Thus 224 of the 232 patients shown in Table 1 were included in these analyses. All changes from baseline induced by lovastatin treatment were significant, except for the increases in HDL-cholesterol concentration in the E3/2 and E4/4 phenotypes (Table 2). Despite the large variation within each phenotype with regard to the LDLcholesterol response (Fig. l),significant differences between phenotypes were not observed. Moreover, in order to minimize the effects of possible noncompliers, the analyses were repeated after excluding individuals with a poor treatment response (defined as < 10% reduction in LDL-cholesterol levels), but this did not alter the results. Furthermore, there were no significant differences in changes in total cholesterol, HDL-cholesterol or triglycerides between phenotypes (Table 2).
Apo E phenotype and response to lovastatin in non-FH and FH Due to the small sample sizes in the E3/2 and E4/4 phenotypes (see Table 1)only the effects of the E3/3 and E4/3 phenotypes on treatment response were analysed separately in the non-FH and FH patient groups. The non-FH group consisted of strata 1 and 2, but the FH group consisted only of strata 2 (see Study design section). Apart from higher BMI values in E4/3 subjects than in E3/3 subjects in stratum 1, clinical characteristics (cf. Table 1) were similar among the apo E phenotypes at baseline. The baseline levels and percentage changes in total and LDLcholesterol in the E3/3 and E3/4 phenotypes after 6
and 12 weeks of treatment are shown in Table 3. The baseline levels of either these lipid parameters or HDL-cholesterol and triglycerides (data not shown) did not differ significantly between the phenotypes. Significant differences in LDL-cholesterol were observed in the non-FH patients who were receiving lovastatin 2 0 mg nightly (decreases of 27.9 vs. 20.1 %; P = 0.043) in E3/3 and E4/3, respectively, and in total cholesterol in the FH group using lovastatin 4 0 mg nightly (decreases of 27.4 vs. 22.6%; P = 0.023) (Table 3). In addition, significant differences in HDL-cholesterol increases ( + 10.5 vs. 1.0%;P = 0.024) in E3/3 and E4/3, respectively, were observed in non-FH patients taking 20 mg of lovastatin, and in triglyceride decreases (-26.8 vs.-ll.O%; P = 0.010)inE3/3 andE4/3, respectively, in FH patients taking 4 0 mg of lovastatin (data not shown). However, at week 12 with doubled doses of lovastatin, all responses of lipid fractions in both non-FH and FH groups became uniform among phenotypes. These results remained unchanged after inclusion of the data for the few patients (Table 1) whose lovastatin dose was not doubled after 6 weeks of treatment (data not shown).
+
Xbal polymorphism of the apolipoprotein B The effects of the Xbal genotypes (i.e. X l X 1 , X1X2 and X2X2) on treatment responses were analysed separately in the non-FH and FH patient groups. The non-FH groups consisted of strata 1 ( n = 68) and 2 ( n = 76), while the FH group only included stratum 2 ( n = 67) (see Study design section). Apart from the fact that a higher proportion of the non-FH patients
6.5k0.8 n = 23 10.8k1.5 8.7k1.6
6.5k0.8
n = 39 10.9kl.5
9.Ok1.6
n = 28
5.3k0.5
n = 13 7.4k0.6
8.9kO.8
= 49
5.3k0.6
34 7.5k0.6
=
8.7k0.8
n
n
E4/3
NS
NS
NS
NS
NS
NS
P-value'
E4/3
-20.1k9.4
-14.5k8.3
-32.1k9.6
-27.4f7.0
-35.9k11.1
-28.0f8.7
-29.4k9.9
-22.6k8.9
-36.5k9.3
-27.3k5.4
Lovastatin 40 mg nightly
-27.9k12.2
-19.0k6.9
Lovastatin 20 mg nightly
E3/3
NS
0.023
NS
NS
0.043
NS
P-value'
E4/3
-34.4511.8
-24.8k10.7
-39.0k9.4
-32.3k7.1
-45.0k10.1
-34.3f8.8
-38.0f8.6
-30.6k6.9
-43.8k7.6
-33.7f6.8
Lovastatin 80 mg nightly
-33.4k11.9
-24.0k8.4
Lovastatin 40 mg nightly
E3/3
Percentage change at week 12
NS
NS
NS
NS
NS
NS
=
P-value*
Mean values k SD are shown. All differences between baseline and treatment values were significant ( P < 0.05). * Level of significance of differences between the phenotypes estimated by analysis of variance or the Kruskal-Wallis test, depending on normalities of the distributions: NS significant. t The baseline total cholesterol levels were 6.2-7.79 and > 7.8 mmol I-' in strata I and 2, respectively.
Stratum 2 t Non-FH Total cholesterol LDLcholesterol FH Total cholesterol LDL cholesterol
cholesterol
LDL-
Stratum lt Non-FH Total cholesterol
E3/3
Percentage change at week 6
non-
?
P
4
phenotypes, in the non-FH and FH patients
Baseline level (mmol I-')
9
Table 3. Baseline levels and percentage changes in total and LDL-cholesterol levels after 6 and 12 weeks of treatment with lovastatin. according to the apolipoprotein E3/3 and E4/3
F
9
3
0
?
?
APO E AND B POLYMORPHISMS
with the X l X l genotype in stratum 2 used diuretics, the clinical characteristics (cf. Table 1)did not differ significantly between the genotypes. The lovastatin dose was not doubled at 6 weeks in 24 of the 211 subjects. The efficacy data at 6 weeks were analysed with and without inclusion of these patients, but the results obtained were similar. There were no significant differences between genotypes in baseline levels of any measured lipid parameter (data not shown). Uniform decreases in total and LDL-cholesterol were observed during treatment in all Xbal genotype groups. The average increase in HDL-cholesterol was somewhat greater in X l X l subjects than in the other genotypes after 12 weeks of treatment in the non-FH patients in stratum 2 (increases of 12.4% vs. 2.1% vs. 6.4% in X l X 1 , X1X2 and X2X2, respectively; P = 0.046, analysis of variance). Otherwise, the differences in HDLcholesterol and triglycerides between the genotypes in both non-FH and FH groups at 6 and 12 weeks (data not shown) were not statistically significant.
Discussion As we have also reported previously [7, 81, a wide inter-individual variation in treatment responses to lovastatin was observed in the present study. Although dietary and other environmental confounding factors were kept as constant as possible, they may still have influenced this variation. In addition, the treatment response could be dependent on variation at the apolipoprotein gene loci. In theory, the apo E and apo B Xbal polymorphisms could modulate the LDL-receptor/lipoprotein particle interaction and thereby interfere with the action of lovastatin. Both polymorphisms have been shown to affect the serum cholesterol response to dietary change in the Finnish population [13, 141. In the present study, combined data from two previous studies were re-analysed in order to determine whether these polymorphisms also affected the treatment response to lovastatin. Although the studies were not originally designed for this purpose, and blood samples were not available from all participating centres, it is unlikely that this would have biased our results. The loss of samples was probably random, and the number of analyses remained relatively large. The contribution of the apo E phenotype to serum total and LDL-cholesterol is well established in many populations [26-301. In Finland, the levels of these
403
lipid parameters in random samples of apparently healthy subjects have been shown to occur in the following order: E4/4 > E4/3 > E3/3 > E3/2 > E4/2 > E2/2 [13, 291. Such an association was not observed in our hypercholesterolaemic patients. One possible reason for this could be that the baseline lipid levels were determined during a low-fat, lowcholesterol diet, which has been shown to reduce the correlation between apo E phenotype and serum cholesterol levels [ 131. Theoretically, the response to lovastatin could be impaired in subjects who carry the e4 allele. This is based on the finding that the apo E4-containing postprandial lipoproteins are cleared more effectively by the liver than those with other apo E isoforms, probably resulting in hepatic cholesterol loading and suppression of HMG CoA reductase activity [31]. Accordingly, HMG CoA reductase could already be partially inhibited in these subjects, resulting in a blunted response to lovastatin. On the other hand, the efficiency of intestinal cholesterol absorption is reported to be greater in subjects who carry the e4 allele than in those who carry the e2 allele [32]. As there is preliminary evidence [ 3 31 to suggest that lovastatin inhibits intestinal cholesterol absorption, it could also be argued that lovastatin is more effective in patients who carry the e4 allele. The only significant differences between apo E phenotypes in terms of total or LDL-cholesterol reduction occurred after 6 weeks of treatment. In non-FH patients, LDL-cholesterol was reduced slightly, but significantly less in E4/3 subjects than in E3/3 subjects, and in FH patients a similar difference was observed between total cholesterol responses. After doubling of the lovastatin doses (from 2 0 to 4 0 mg in non-FH patients, and from 4 0 to 80 mg nightly in FH patients) even these slight differences disappeared completely. In theory, this could suggest that the e4 allele has a small opposing effect on cholesterol reduction with low doses of lovastatin. This appears unlikely, as individuals with two e4 alleles (apo E4/4 phenotype) did not exhibit any degree of diminished response to lovastatin (see Fig. 1, FH and non-FH subjects combined). In any case, the effects of apo E phenotypes on treatment response were clinically insignificant. When the data for all patients who were taking 40 mg of lovastatin were pooled and analysed together, the same result was obtained. While we are unaware of any previous report on the effects of apo E phenotype on the treatment
404
J.-P. OJALA et al.
response to HMG CoA reductase inhibitors in non-FH patients, two recent papers have addressed this question in patients with heterozygous FH. In these studies no significant difference was detected between apo E phenotypes in this respect during lovastatin [34] and simvastatin [35] therapy. This finding is confirmed by the present study and might be expected, as it is unlikely that the enhanced receptormediated removal of LDL by lovastatin would be influenced by apo E contained within VLDL and IDL. The slight reduction in LDL-cholesterol-lowering in E4/3 subjects vs. E3/3 subjects using the lowest lovastatin dose (20 mg d-l) in patients with non-FH may reflect some unknown mechanism of cholesterol elevation in these patients. De Knijff et al. [35] reported that a better response with regard to the LDL-cholesterol decrease was observed with simvastatin in women than in men with the E3/3 phenotype. We observed no sexrelated differences in response (data not shown). To our knowledge, this is the first report of a study of the possible effects of the apo B Xbal polymorphism on treatment response to an HMG CoA reductase inhibitor. This polymorphism is determined by two alleles: X 1 (absence of Xbal cutting site) and X2 (presence of Xbal cutting site), which produce three genotypes: X l X 1 , X1X2 and X2X2. The presence of the X2 allele has been associated with increased total and LDL-cholesterol levels in several [24, 36-40] but not all studies [25, 41, 421. In our study population, which consisted of high-risk individuals with severe hypercholesterolaemia, this association was not observed. The Xbal mutation does not result in an amino acid change [43]. However, recent in-vivo and in-vitro studies have indicated that, apparentiy due to impaired receptor binding properties of LDL, subjects with the X2X2 genotype catabolized LDL more slowly than X l X l individuals [ l l , 121, suggesting that the Xbal mutation is closely associated with an as yet unknown functionally significant mutation. Despite the possible role of the X2 allele as a marker of this functionally important DNA change, we did not observe any differences between the three Xbal genotypes within the non-FH groups or the FH group in terms of treatment response. This suggests that the proposed differences between LDL particles, if they exist, have no clinically significant effects on the lovastatin response. In conclusion, our results did not reveal any clinically significant effects of the apo E or apo B Xbal
polymorphisms on inter-individual variation in treatment response to lovastatin, in either familial or non-familial hypercholesterolaemia. Further studies are needed to investigate the effects of other genetic factors on this variation.
Acknowledgements Financial support for this study was provided by the Sigrid Juselius Foundation (to C.E. and M. J.T.), the Paavo Nurmi Foundation, the Academy of Finland and the University of Helsinki (to M. J.T.).
References 1 Grundy SM. HMG-CoA reductase inhibitors for treatment of hypercholesterolemia. N Engl ] Med 1 9 8 8 ; 3 1 9 : 24-33.
2 Bilheimer DW. Grundy SM. Brown MS. Goldstein JL. Mevinolin and colestipol stimulate receptor-mediated clearance of low density lipoprotein from plasma in familial hypercholesterolemia heterozygotes. Proc N a d Acad Sci USA 1 9 8 3 ; 80: 4124-8. 3 Grundy SM. Vega GL. Influence of mevinolin on metabolism of low density lipoproteins in primary moderate hypercholesterolemia. ] Lipid Res 1 9 8 5 ; 2 6 : 1464-75. 4 Arad Y, Ramakrishnan R. Ginsberg HN. Lovastatin therapy reduces low density lipoprotein apoB levels in subjects with combined hyperlipidemia by reducing the production of apoBcontaining lipoproteins : implications for the pathophysiology of apoB production. ] Lipid Res 1 9 9 0 ; 31 : 567-82. 5 Illingworth DR. Sexton GJ. Hypocholesterolemic effects of mevinolin in patients with heterozygous familial hypercholesterolemia. / Clin Invest 1 9 8 4 ; 74: 1972-8. 6 Hoeg JM. Maher MB, Zech LA el a/. Fffectiveness of mevinolin on plasma lipoprotein concentrations in type I1 hyperlipoproteinemia. A m J Cardiol 1 9 8 6 ; 5 7 : 933-9. 7 Ojala J-P, Helve E. Karjalainen K. Tarkkanen A, Tikkanen MJ. Long-term maintenance of therapeutic response to lovastatin in patients with familial and non-familial hypercholesterolemia: a 3-year follow-up. Atherosclerosis 1 9 9 0 : 82 : 85-95. 8 Ojala J-P, Helve E. Tikkanen MJ. Treatment of combined hyperlipidemia with lovastatin vs. gemfibrozil: a comparison study. Cardiology 1 9 9 0 ; 77 (Suppl. 4): 3 9 4 9 . 9 Vesell ES. Genetic and environmental factors affecting drug disposition in man. Clin Pharmacol Ther 1 9 7 9 ; 22: 659-79. 10 Davignon 1. Gregg RE, Sing CF. Apolipoprotein E polymorphism and atherosclerosis. Arteriosclerosis 1988 ; 8 : 1-21. 11 Demant T, Houlston RS, Caslake MJ et al. Catabolic rate of low density lipoprotein is influenced by variation in the apolipoprotein B gene. / Clin Invest 1988: 82: 797-802. 12 Series J, Cameron I. Caslake M. Gaffney D. Packard CJ. Shepherd J. The Xbal polymorphism of the apolipoprotein B gene influences the degradation of low density lipoprotein in vitro. Biochirn Biophys Acta 1989: 1003: 183-8. 13 Tikkanen MJ, Huttunen JK. Ehnholm C. Pietinen P. Apolipoprotein E, homozygosity predisposes to serum cholesterol elevation during high fat diet. Arteriosclerosis 1 9 9 0 : 10: 285-8.
A P O E A N D B POLYMORPHISMS
1 4 Tikkanen MJ, Xu C-F. Hamalainen Tet al. XbAl polymorphism of the apolipoprotein B gene influences plasma lipid response to diet intervention. CIin Genet 1 9 9 0 ; 37: 327-34. 1 5 Nestruck AC. Bouthillier D. Sing CF. Davignon J. Apolipoprotein E polymorphism and plasma cholesterol response to probucol. Metabolism 1987: 36: 743-7. 16 Tikkanen MJ, Helve E. Jaattela A et al. Comparison between lovastatin and gemfibrozil in the treatment of primary hypercholesterolemia : the Finnish multicenter study. Am j Cardiol 1 9 8 8 : 62: 3 5 J 4 3 J . 1 7 Ojala J-P. Helve E. Tikkanen MJ et al. Switch from gemfibrozil to lovastatin (mevinolin) therapy in patients with primary hypercholesterolaemia. A multicentre study. Drug Investigation 1990: 2 (Suppl. 2): 40-47. 18 Goldstein JL, Brown MS. Familial hypercholesterolemia. In: Stanbury JB. Wyngaarden JB, Fredrickson DS et a/.. eds. The Metabolic Basis of Inherited Disease. New York: McGraw Hill, 1983: 672-712. 1 9 Finley PR, Schifman RB. Williams RJ. Lichti DA. Cholesterol in high-density lipoprotein : use of Mg2+/dextran sulfate in its enzymic measurement. Clin Chem 1 9 7 8 ; 24: 931-3. 20 Friedewald WT. Levy RI. Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972 ; 18: 499-502. 21 Lukka M. Metso J, Ehnholm C. Apolipoprotein A-IV polymorphism in the Finnish population : gene frequencies and description of a rare allele. Hum Hered 1988; 38: 359-62. 22 Havekes LM.de KnijfTP. Beisiegel U. Havinga J. Smit M. Klasen E. A rapid micromethod for apolipoprotein E phenotyping directly in serum. I Lipid Res 1987: 28: 455-63. 23 Weisgraber KH. Innerarity TL. Mahley RW. Abnormal lipoprotein receptor-binding activity of the human E apoprotein due to cysteine-arginine interchange at a single site. I Biol Chem 1 9 8 2 ; 257: 2518-521. 24 Aalto-SetaII K. Tikkanen MJ, Taskinen M-R. Nieminen M, Holmberg P. Kontula K. Xbal and c/g polymorphisms of the apolipoprotein B gene locus are associated with serum cholesterol and LDL cholesterol levels in Finland. Arteriosclerosis 1988: 74: 47-54. 25 Hegele RA. Huang LS.Herbert PN et a/. Apolipoprotein Bgene DNA polymorphisms associated with myocardial infarction. N Englj Med 1986: 315: 1509-15. 26 Utermann G. Pruin N, Steinmetz A. Polymorphism of apolipoprotein E. 111. meet of a single polymorphic gene locus on plasma lipid levels in man. Clin Genet 1 9 7 9 ; 15: 63-72. 27 Menzel H-J. Kladetzky R-G. Assmann G. Apolipoprotein E polymorphism and coronary artery disease. Arteriosclerosis 1983: 3: 310-15. 28 Sing CF, Davignon J. Role of the apolipoprotein E polymorphism in determining normal plasma lipid and lipoprotein variation. Am I Hum Genet 1 9 8 5 ; 29: 477-84. 29 Ehnholm C. Lukka M, Kuusi T. Nikkila E. Utermann G. Apolipoprotein E polymorphism in the Finnish population : gene frequencies and relation to lipoprotein concentrations. j Lipid Res 1986: 27: 227-35.
405
3 0 Ordovas JM, Litwack-Klein L. Wilson PWF, Schaefer MM. Schaefer El. Apolipoprotein E isoform phenotyping methodology and population frequency with identification of apo E l and apo E5 isoforms. j Lipid Res 1987: 28: 371-80. 31 Weintraub MS, Eisenberg S. Breslow JL. Dietary fat clearance in normal subjects is regulated by genetic variation in apolipoprotein E. CIin Invest 1987: 80: 1571-7. 32 Kesaniemi YA. Ehnholm C. Miettinen TA. Intestinal cholesterol absorption efficiency in man is related to apoprotein E phenotype. I Clin Invest 1987: 80: 578-81. 33 Miettinen TA. Tikkanen MJ, Helve E, Ojala J-P. Inhibition of dietary cholesterol absorption during lovastatin (mevinolin) treatment. Drug Investigation 1990: 2 (Suppl. 2): 29-35. 34 O'Malley JP, Illingworth DR. The influence of apolipoprotein E phenotype on the response to lovastatin therapy in patients with heterozygous familial hypercholesterolemia. Metabolism 1990: 39: 150-54. 35 De Knijff P. Stalenhoef AFH, Mol MJTM et al. Influence of apo E polymorphism on the response to simvastatin treatment in patients with heterozygous familial hypercholesterolemia. Atherosclerosis 1990: 83: 89-97. 36 Berg K. DNA polymorphism at the apolipoprotein B locus is associated with lipoprotein level. Clin Genet 1986: 30: 515-20. 3 7 Law A, Powell LM. Brunt H et al. Common DNA polymorphism within coding sequence of apolipoprotein B gene associated with altered lipid levels. Lancet 1986: i: 1301-3. 38 Talmud PJ, Barni N. Kessling M et a/. Apolipoprotein B gene variants are involved in the determination of serum cholesterol levels : a study in normo- and hyperlipidaemic individuals. Atherosclerosis 1987: 67: 81-9. 39 Leren TP, Berg K. Hjermann I. Leren P. Further evidence for an association between the Xbal polymorphism at the apolipoprotein B locus and lipoprotein level. CIin Genet 7 988 : 34: 347-51. 4 0 Aalto-SetaIa K. Gylling H, Helve E et al. Genetic polymorphisms of the apolipoprotein B gene locus influence serum LDL cholesterol level in familial hypercholesterolemia. Hum Genet 1989: 82: 305-7. 4 1 Aburatani H. Matsumoto A, Itoh H et al. A study of DNA polymorphism in the apolipoprotein B gene in a Japanese population. Atherosclerosis 1988: 72: 71-6. 42 Darnfors C. Wiklund 0. Nilsson J et al. Lack of correlation between the apolipoprotein B XbAl polymorphism and blood lipid levels in a Swedish population. Atherosclerosis 1989: 75: 183-8. 4 3 Carlsson P. Darnfors C. OlotTson S, Bjursell G. Analysis of the human apoliprotein B gene: complete structure of the B74 region. Gene 1986: 49: 29-51. Received 18 January 1991, accepted 2 April 1991. Correspondence: Jukka-Pekka Ojala. MD. Third Department of Medicine, University of Helsinki, Haartmaninkatu 4, SF-00290 Helsinki, Finland.
