Triglyceride-rich lipoproteins of subjects heterozygous for apolipoprotein E2(Lysl46→Gln) are inefficiently converted to cholesterol-rich lipoproteins

ATHEROSCLEROSIS

ELSEVIER

Atherosclerosis 108 (1994) 183- 192

Triglyceride-rich lipoproteins of subjects heterozygous for apolipoprotein E2(Lys 146- Gln) are inefficiently converted to cholesterol-rich lipoproteins Monique Mulder”, Hans van der Booma, Peter de Knijffa Carine Braama, Am van den Maagdenbergb, Jan A. Gevers Leuvena, Louis M. Havekes*a “TN0 Institute of Prevention and Health Research, Gaubius Laboratory. P.O. Box 430, 2300 AK Leiden, The Netherlands bMGC-Dept. of Human Generics, University of Leiden, Leiden, The Netherlands

(Received 24 December 1993; revision received 24 March 1994; accepted 11April 1994)

Abstract

The APOE*2(Lys146-Gln) allele behaves like a dominant trait in the expression of familial dysbetalipoproteinemia (FD) (Smit et al., J. Lipid Res. 1990; 31: 45-53). FD patients carrying the APOE*2(Lys146-Gln) allele exhibit less elevated cholesterol to triglyceride ratios in the d < 1.019 g/ml lipoprotein density fraction as compared to classical FD patients displaying homozygosity for the APOE*2(Arg158-Cys) allele (0.8 vs. 1.4). Upon treatment of complete serum with lipoprotein lipase (LPL), the mean cholesterol to triglyceride molar ratio of the d < 1.019 g/ml lipoprotein fraction in these FD patients increased only ma&ally (from 0.8 to 1. l), as compared with that of classical FD subjects (from 1.4 to 2.6) and non-FD control subjects (from 0.7 to 1.5). In order to obtain further evidence for an inefticient lipolysis of the d < 1.019 g/ml lipoprotein fraction in APOE*2(Lys146-Gln) carriers, possibly in combination with a less efficient cholesteryl ester transfer protein (CETP) activity, blood samples of APOE*2(Lys146-Gln) allele carrying FD patients were analysed and compared with classical FD patients and controls. In the APOE*2(Lys146-Gln) FD patients, the increase in plasma cholesterol was mainly confined to the very low density lipoprotein (VLDL) fraction, whereas in classical FD patients, the levels of cholesterol in the intermediate density lipoprotein (IDL) fraction was also dramatically increased (ratios of VLDL to IDL cholesterol are 4.7 and 2.6, respectively). Family analyses of the APOE*2(Lys146-Gln) FD subjects showed that the apo E to apo B ratio in the d c 1.019 g/ml lipoprotein fraction of allele carriers is 3.5 times as high as that found in non-carriers (2.8 vs. 0.8, by wt.). Also, in the APOE*2(Lys146-Gln) allele carrying family members, the ratio of cholesterol to triglyceride of the d < 1.019 g/ml lipoprotein fraction is less markedly elevated upon addition of LPL when compared to that in non-carrying controls (from 1.1 to 1.8 vs 0.7 to 1.6). The efficiency of the d < 1.019 g/ml lipoprotein fraction of APOE*2(Lys146-Gln) FD patients to compete with low density lipoprotein (LDL) for binding to the LDL receptor is intermediate to that of controls and classical APOE*2(Arg158-Cys) homozygous FD patients. These findings suggest that in * Corresponding author. Tel.: +31 71 18 14 49; Fax: + 31 71 18 19 04. Abbreviations: FD, familial dysbetalipoproteinemia; Apo, apolipoprotein; DMEM, Dulbecco’s Modified Eagle’s Medi-

um; FCS, foetal calf serum; HSA, human serum albumin; LDL, low density lipoprotein; VLDL, very low density lipoprotein; IDL, intermediate density lipoprotein; HDL, high density lipoprotein; LPI+ lipoprotein lipase.

0021-9150/94/%07.00 0 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 002 1-9 150(94)05264-J

184

hf. Mulder et al. /Atherosclerosis IO8 (1994) 183-192

APOE*2(Lys146-Gln) allele carriers, the conversion of VLDL into IDL is impaired due to an inefficient lipolysis, possibly in combination with a retarded CETP activity. As a consequence, this might result in a less efficient cellular processing of these lipoproteins via the LDL receptor and in this way explain the dominant behaviour of the APOE*2(Lys146-Gln) allele in the expression of FD. Keywords: Familial dysbetalipoproteinemia;

Dominant

1. Introduction

In normal subjects, chylomicrons and very low density lipoproteins (VLDL) are partly lipolysed by the enzyme lipoprotein lipase (LPL). The resulting chylomicron remnants and the major portion of the VLDL remnants (or intermediate density lipoproteins, IDL) are rapidly removed from the circulation by receptor-mediated endocytosis in the liver. The remaining VLDL remnants are converted into low density lipoproteins (LDL) [l]. Apo E is the ligand for the binding of these remnants to hepatic lipoprotein receptors, and thus plays a crucial role in the remnant metabolism [2]. Apo E is a polymorphic protein of which, by use of isoelectric focusing, three major isoforms, E2, E3 and E4, can be separated [3,4]. Apo E2 differs from the most common apo E3 variant by exhibiting a cysteine residue at position 158 instead and is designated apo of an arginine, E2(Argl58-Cys). Apo E4 exhibits an arginine at residue 112 instead of a cysteine and is designated apo E4(Cysl12-Arg) [5]. The common apo E isoforms are encoded by three codominant alleles at a single APOE gene locus on chromosome 19 [4]. Hence, six common phenotypes can be distinguished: E2E2, E3E3, E4E4, E3E2, E4E2, and E4E3. Familial dysbetalipoproteinemia (FD) is characterized by high serum cholesterol and triglyceride concentrations, due to the accumulation in the plasma of chylomicron- and VLDL-remnants. Patients with FD have been shown to develop premature atherosclerosis involving both coronary and peripheral arteries [6]. Most FD patients for the homozygous (> 80%) are APOE*2(Argl58-Cys) allele [7,8]. The underlying metabolic defect in these patients is a disturbed interaction of apo E2(Arg158-Cys) with hepatic

mode of inheritance; Lipolysis; Lipoprotein

remnants

lipoprotein receptors [9,10]. However, of all APOE*2(Arg158- Cys) homozygotes, representing 1% of the total population, only 4% eventually develop hyperlipidemic FD [7]. FD is only rarely associated with the E3E2 or E4E2 phenotypes. Genotyping and DNA sequencing of our E3E2 heterozygous FD patients revealed that they exhibit the rare APOE*2(Lys 146- Gln) allele [ 11,121. This apo E variant was first described by Rall et al. [13]. Family studies have revealed that, in contrast to the most frequently occurring APOE*2(Arg158-Cys) allele, heterozygosity for the allele commonly APOE*2(Lys146-Gln) cosegregates with FD, indicating that this variant behaves like a dominant trait in the expression of the disease [12]. Hence, subjects heterozygous for the APOE*2(Lys146-Gin) allele frequently develop hyperlipidemic FD, despite the presence of a normal apo E3 allele. The present paper deals with the mechanism underlying the dominant behaviour of the apo E2(Lys146-Gln) variant. We found that in APOE*2(Lys146-Gln) allele carriers, the addition of LPL to complete serum results in a less dramatic increase in the cholesterol/triglyceride ratio of the d < 1.019 g/ml lipoprotein fraction as compared to that in non-carriers. We present evidence that this defect in conversion of triglyceriderich particles into more cholesterol-rich particles leads to a less efficient cellular processing of these lipoproteins and thus may explain the dominant behaviour of the APOE*2(Lys146_Gln) allele in the expression of FD.

2. Materials and methods 2.1. Subjects The probands

previously

described

as familial

M. Mulder et al, /Atherosclerosis 108 (1994), 183-192

dysbetalipoproteinemic (FD) patients with heterozygosity for the APOE*2(Lys146-Gln) allele [12], their APOE*2(Lys146-Gln) allele carrying and non-carrying relatives and the apo E2(Argl58 - Cys) homozygous FD patients were admitted to the lipid clinic in Leiden, or were visited at their homes. Blood was obtained by venepuncture after an overnight fast, and was allowed to clot for 1 h at 37°C. Serum was then separated from blood cells by centrifugation at 500 x g for 10 min at room temperature. Patients with FD were diagnosed on the basis of the presence of hyperlipidemia (cholesterol > 7.5 mmol/l; triglycerides > 2.0 mmol/l), concomitant with floating beta lipoproteins and an elevated VLDL cholesterol/plasma triglyceride ratio (> 0.69 on a molar basis). Serum cholesterol and triglyceride were determined using enzymatic methods (Boehringer Mannheim, Germany). 2.2. Apo E phenotyping and genotyping Apo E phenotyping was performed exactly as described before [14]. APOE genotyping of the common polymorphisms at codons 112 and 158 was performed by polymerase chain reaction (PCR) of the region encompassing both polymorphic sites, digestion of the PCR products with the restriction enzyme HhaI and electrophoresis on a polyacrylamide gel as described earlier [ 15,161. Identification of APOE*2(Lys146-Gin) allele carriers was performed by PCR using a mutagenic amplification primer assay. Primer 3012 5’GGCATCGCGGAGGAGAGCAGCT-3 ’ (nucleotides 3848-3867, non-coding strand) was designed with a nucleotide mismatch (underlined) as compared to the wild type APOE sequence [ 171. In the case of the APOE*2(Lys146-Gln) allele, a PvuII restriction site is introduced due to a base pair substitution specific for this allele and the nucleotide mismatch in the primer. PCR was performed using this primer and primer 398 5 ‘-GCGGGCACGGCTGTCCAAGG-3 ’ (nucleotides 3678-3697, coding strand). The reaction mixture included 50 pmol of each primer, 0.5 pg genomic DNA, 0.2 mM deoxyribonucleotide triphosphates (dNTPs), 0.5 mM MgCl,, 50 mM KCl, 10 mM Tris-HCl pH 8.3, 200 &ml BSA, 0.1 U of Taq polymerase (superTaq, Boehringer Mannheim) and 10% di-

185

methylsulphoxide (v/v) in a total volume of 50 ~1. Amplification was performed for 32 cycles of 1 min at 95°C 30 s at 55°C and 1 min 30 s at 72°C with an initial denaturation period of 4 min. Some 15 ~1 of PCR product was digested with restriction enzyme PvuII according to the recommendations of the supplier (Pharmacia). Thereafter, fragments were separated on a neutral 7.5% polyacrylamide bromide, and gel, stained with ethidium photographed on a UV source. Digestion of the PCR product of a heterozygous APOE*2(Lys146-Gln) carrier with P&I, will result in patterns consisting of three fragments: an undigested fragment of 189 bp, originating from the normal allele, and 171 bp and 18 bp fragments, originating from the mutant allele. 2.3. Lipid and lipoprotein analysis For the isolation of VLDL + IDL (d < 1.019 g/ml), 2 ml of plasma was brought to a density of 1.019 g/ml with potassium bromide and overlayered with a 3.5 ml solution of sodium chloride (d = 1.019 g/ml) in a 10.4 ml centrifuge tube fitting the 50 Ti fixed-angle rotor (Beckman Instruments, Geneva, Switzerland). VLDL + IDL was aspirated from the top (I ml fraction) after centrifugation at 106 000 x g for 16 h at 4°C. For the isolation of VLDL (d < 1.006 g/ml), 2 ml of plasma was overlayered with a 3.5 ml solution of NaCl (d = 1.006 g/ml) in a 10.4 ml tube fitting the 50 Ti fixed-angle rotor (Beckman Instruments). VLDL was aspirated from the top (1 ml fraction) after centrifugation at 106 000 x g for 16 h at 4°C. High density lipoprotein (HDL) was determined in the infranatant after precipitation of IDL and LDL [18]. Plasma and lipoprotein cholesterol concentrations in VLDL (d < 1.006), VLDL + IDL (d < 1.019) and HDL fractions were measured using the CHOD-PAP kit (No. 236691, Boehringer Mannheim, Germany). Plasma and lipoprotein triglycerides were measured using the GPO-PAP kit (No. 701904, Boehringer Mannheim, Germany). IDL-cholesterol was calculated as the difference between VLDL + IDL-cholesterol (d < 1.019 g/ml) minus VLDL-cholesterol (d < 1.006 g/ml). LDL-cholesterol (1.019 < d < 1.063 g/ml) was calculated using the formula: LDL-cholesterol =

M. Mulakr et al. /Atherosclerosis 108 (1994) 183-192

186

The mixtures were incubated for 2.5 h at 37°C. To stop the reaction, the mixtures were put on ice and solid KBr was added to adjust the density of the solutions to 1.21 g/ml. Lipoprotein fractions with density C 1.019 g/ml were isolated from both solutions (with and without lipolysis). Therefore, the 1.21 g/ml solutions of above were placed in 12.5 ml tubes fitting the 40 Ti swing-out rotor under a discontinuous gradient of salt solutions, of densities 1.063 and 1.019 g/ml and with a volume ratio of 0.95:1:1 from bottom to top. After centrifugation at 40 000 rev./min for 16 h at 4°C the d < 1.019 g/ml lipoprotein fraction was aspirated from the top (1 ml fraction).

plasma-cholesterol - (VLDL-cholesterol + IDLcholesterol + HDL-cholesterol). Free fatty acids were determined using a Nefa C kit from Wako Chemicals GmbH (Neuss, Germany). Protein contents of the lipoprotein fractions were determined according to Lowry et al. [19]. The apo E and apo B levels in lipoprotein fractions were estimated using enzyme-linkedimmunosorbent-assays [20,2 11. 2.4. Increase of cholesterol to triglyceride ratio of the d < 1.019 g/ml lipoprotein fraction by addition of LPL to complete serum

LPL used was purified from bovine milk [22]. To an aliquot of serum, corresponding to 7 mg of triglyceride, 400 mg of free fatty acid-free human serum albumin (HSA) was added. Thereafter, the volume was adjusted to a final volume of 4 ml with O.lM Tris-HCl buffer, pH 8.5. To this mixture a fixed amount of LPL, dissolved in the same buffer, was added. Adding the same volume of buffer instead of the LPL solution was used as control incubation.

2.5. Lubelling of the LDL with “‘I Immediately after isolation of the LDL according to Redgrave et al. [23], the lipoprotein preparations were used for iodination by the 1251C1method according to Bilheimer et al. [24]. After iodination, the LDLs were dialysed against phosphate-buffered saline for 4 h (4 x 500 ml). They were then stabilized by adding HSA (I%,

A

_

.,

controls

E2(lys146-gln) =

h0

f 0

C

proband&

3 -FD probands

8 _B

0

2.6 0

2-

z

---I-

L

F F

1.5

0.8

’

*

--$-

1.1

@

8

0

0

8 0’

6

0

befire

af4ter

6

0

befire

afier

0’

I

0

6

befire

afkr

Fig. 1. Ratio of cholesterol to triglyceride in the d c 1.019 lipoprotein fraction, before and after incubation of serum with bovineLPL. Sera from APOE*2(Lys146-Gin) heterozygous FD probands, APOE*2(Argl58-Cys) homozygous FD probands and from controls were incubated for 2.5 h either in the presence or in the absence of LPL, as described in Materials and methods. Lipoproteins with a density of d < 1.019 were then isolated from both solutions (designated before and after lipolysis) as described in Materials and methods. Thereafter, in these lipoprotein fractions, the cholesterol to triglyceride ratios (CTg: mmol/mmol) were determined. (A), three FD probands heterozygous for APOE*2(Lys146-Gln). For two probands the assay was repeated after 2 years, giving comparable results. (B), twelve APOE*2(Argl58-Cys) homozygous FD probands. (C), fourteen controls.

M. Mulder

et al. /Atherosclerosis

w/v) and further dialysed overnight against culture medium supplemented with 20 mM HEPEs buffer (pH 7.4), penicillin and streptomycin. The 1251labelled LDLs were stored at 4°C. Their specific radioactivity was 200-500 counts/min/ng of lipoprotein protein. The labelled LDLs were used within 2 weeks. When not labelled with 1251,lipoproteins were stabilized immediately with the addition of 1% HSA and, subsequently, extensively dialysed against culture medium as mentioned above.

108 (1994)

187

183-192

2.00 ”

controls

T-1 L !,.-. T z ._ I’ 5 E ‘.O” T-l/’ x q--y Yip” T/ 0.50 E ,i’i E2(Lys14&Gln) m :

=

3 0.00 0

25

50

incubation

2.6. Measurement of competition of lipoproteins with 12jI-LDL for association to HepG2 cells

HepG2 cells were cultured as previously described [25]. Competition experiments were performed by incubating HepG2 cells for a period of 3 h at 37°C with “‘1-LDL (10 pg/ml of protein) in the presence or in the absence of increasing amounts of unlabelled lipoproteins, as indicated. Cell association was measured as previously described 1251.

75

100

Table I Lipid and apolipoprotein composition of the carrying family members

d <

Fig. 2. Free fatty acid (ffa) release upon incubation of serum from APOE*2(Lysl46-Gln) heterozygous FD probands and from control subjects with LPL. From each subject an aliquot of serum containing 7 mg of triglyceride was adjusted to a fixed volume of 4 ml by the addition of phosphate buffered saline to a fixed volume of 4 ml. LPL-treatment was further performed as described in Materials and methods. At indicated time intervals, samples of 20 pl were taken in quadruplicate and immediately stored at -20°C until measurement of free fatty acid (ffa) content. The values represent the mean ?? S.D. of three APOE*2(Lysl46-Gln) heterozygous FD probands 0, and of four control subjects (two E2E2, one E3E3 and one E3E2) ?? .

before and after the addition of LPL. Before addition, in APOE*2(Lys146-Gin) allele carrying FD probands, this ratio was much less elevated than

1.0919 lipoprotein fraction of APOE*2(Lysl46-Gln)

allele carrying and non-

C/Tg (mmol/mmol; mean f SD.)

Tg/apo B (mmohmg; mean f SD.)

apo E/ape B (mgmg; mean f S.D.)

Lipolysis

Lipolysis

Lipolysis

Before

Carriers 1.1 f 0.3 Non0.7 f 0.3 carriers

125

time (min)

3. Results

Fig. 1 shows the ratios of cholesterol to triglyceride of the d < 1.019 g/ml lipoprotein fraction

”

1.50

After

Rel. increase %

P*

Before

After

Rel. increase %

P*

1.8 f 0.6 1.6 f 0.9

160 250

o,o,

8.7 f 5.0 9.0 f 5.1

5.5 f 3.3 3.5 f I.5

37 61

o oo5 2.8 f 2 ’ 0.8 f 0.6

Before

After

2.0 f 1.6 0.55 f 0.4

Sera from family members carrying the APOE*2(Lysl46-Gln) allele (n = 35) and from family members not carrying this allele (n = 15) were incubated for 2 h in the presence or absence of LPL, as described in Materials and methods. Lipoproteins with density d < 1.019 were then isolated, as described in Materials and methods, and cholesterol to triglyceride ratios (CTg; mmol/mmol), and triglyceride to apo B ratios (Tg/apo B, mmol/ug), and apo E to apo B ratios (apo E/ape B, mg/mg) were measured in this d < 1.019 fraction. Values are presented as the mean f S.D. *P: differences in relative change between the carriers and the non-carriers, as calculated with the Wilcoxon signed Ranks test.

188

M. Mulder et al. /Atherosclerosis

that in the classical FD patients with homozygosity for the common APOE*2(Arg158-Cys) allele (0.8 * 0.2 vs. 1.4 * 0.5, mmol/mmol; in control subjects: 0.7 ??0.3 mmol/mmol). We wondered whether this relatively low ratio in these patients could be due to an inefficient conversion of triglyceride-rich lipoproteins into cholesterol-rich lipoproteins, as a consequence of either an inefficient lipolysis, an inefficient CETP activity or a combination of both. To test this hypothesis, the LPL-mediated lipolysis of d < 1.019 g/ml lipoproteins of APOE*2(Lys146-Gln) carrying FD probands was compared with that of classical FD patients and non-FD control subjects. In order to keep the experimental conditions as physiological as possible, LPL was added directly to complete serum. Complete serum contains amongst others CETP that will interfere with LPL activity in increasing the cholesterol to triglyceride ratio. An excess of free fatty acid-free albumin was added to prevent inhibition of lipolysis as a result of the release of free fatty acids. We considered the lipid composition of the whole d < 1.019 g/ml lipoprotein density fraction as, upon lipolysis, the density of the lipoproteins might shift from d < 1.006 g/ml to the density between 1.006 and 1.019 g/ml. As shown in Fig. 1, in the APOE*2(Lys146-Gln) FD probands the cholesterol to triglyceride ratio did not change substantially upon treatment with LPL (from 0.8 to l.l), while in the homozygous APOE*2(Arg158-Cys) FD patients and the non-FD subjects (controls), these ratios increased remarkably (from 1.4 to 2.6 and from 0.7 to 1.5, respectively). This suggests that the d C 1.019 lipoproteins of APOE*2(Lys146-Gln) allele carrying FD patients are indeed relatively resistant to conversion into cholesterol-rich lipoproteins. In line with this, the results presented in Fig. 2 indicate that, upon incubation with LPL, the generation of free fatty acids in serum of the APOE*2(Lys146_Gln) allele carrying FD probands is impaired, when compared with that of the control group. In Fig. 3, the distribution of cholesterol among the different lipoprotein fractions is presented for APOE*2(Lys146-Gln) allele carrying FD probands, classical FD patients with homozygosity

IO8 (1994)

183-192

E2(Lys146-Gln) .”

0

“t

I B

C

VLDL

Control

IDL

LDL

HDL

Fig. 3. The distribution of plasma cholesterol over the various lipoprotein fractions. The amount of cholesterol in the VLDL (d < 1.006 g/ml), IDL (1.006 < d < 1.019 g/ml), LDL (1.019 < d < 1.063 g/ml) and HDL fractions are shown for APOE*2(Lysl46-Gln) heterozygous FD probands (A), APOE*2(ArglSS-Cys) homozygous FD patients (B), and controls representing the relatives of the APOE*2(Lysl46-Gln) FD patients not carrying the mutant allele (C). The values are mean f S.D.

M. Mulder et al. /Atherosclerosis

for the APOE*2(Arg158-Cys) allele [26] and control subjects. It is obvious that the APOE*2(Lys146-Gln) FD patients display, like classical FD patients, relatively elevated cholesterol levels in the VLDL fraction. However, the APOE*2(Lys146-Gln) carriers showed only mildly elevated amounts of cholesterol in the IDL with classical fraction, as compared APOE*2(Argl58-Cys) homozygous FD subjects. Hence, these results also suggest a relative resistance of the triglyceride rich lipoproteins in APOE*2(Lys-Gln) allele carriers to conversion into cholesterol-rich lipoproteins. In order to sustain this hypothesis, we collected blood samples from APOE*2(Lys146_Gln) allele carrying and non-carrying family members. In Table 1 the lipid and apolipoprotein compositions of the d c 1.019 g/ml lipoprotein fractions are It is obvious that in the presented. APOE*2(Lys146-Gln) allele carrying family members, the ratio of cholesterol to triglycerides in the d c 1.019 lipoprotein fraction is less dramatically increased upon addition of LPL than that in the non-carriers (160% vs. 250%). This observation is in agreement with the results found for the APOE*2(Lys146-Gln) carrying FD probands (Fig. 1). Concomitantly, a less efficient conversion into cholesterol-rich d < 1.019 lipoproteins in APOE*2(Lys146-Gln) carriers is also reflected

by a less dramatic decrease of the triglycerideiapo B ratio in these lipoproteins (63% vs. 39%) and by the absence of any effect on the cholesterol to apo B ratio (not shown). Strikingly, by measuring apo E and apo B levels in the respective d < 1.019 g/ml lipoprotein fractions before lipolysis, we observed that in the APOE*2(Lys146-Gln) allele carrying subjects, the relative amount of apo E in this lipoprotein fraction was much higher than in the corresponding fraction of non-carriers (apo E/ape B ratios of 2.8 vs. 0.8, by weight). We wondered whether a relative defect in conversion of triglyceride-rich lipoproteins into cholesterol-rich lipoproteins results in a relative defect of these lipoproteins to bind to the LDL receptor. The results presented in Table 2 indeed show that d < 1.019 g/ml lipoproteins from APOE*2(Lys146-Gln) FD patients are less efficient in competing with ‘251-labeled LDL for binding to the LDL receptor of HepG2 cells than the corresponding lipoprotein fraction of control subjects. However, the APOE*2(Lys146-Gln) lipoprotein fraction is more efficient in binding than the corresponding fraction of the classical E2E2 homozygous FD patients. 4. Discussion Familial dysbetalipoproteinemia

Table 2 Competition of unlabeled lipoproteins with a density of d < 1.019 (g/ml) from individuals with lZSI-LDL for association to HepG2 cells Unlabelled lipoprotein

Apo E phenotype

LDL d < I.019 lipoproteins: E3E3 E2(lysl46-gln)E3 E2E2

rg unlabelled

lipoprotein

0

IO

4

100

54*

3 5 6

100 100 100

66 * I 19 ?? 7 90 f IO

n

189

108 (1994) 183-192

with different

added

(FD) is com-

apo E phenotypes,

and LDL

per ml 50

II

32 f 4 55 ?? 4 60 zt I 74 * 20

After preincubation for 20 h in medium supplemented with 1% (w/v) HSA, the cells were incubated for 3 h at 37”C, with IO &ml of ‘2SI-LDL in the presence of unlabelled lipoproteins with a density of less than I.019 g/ml, as indicated. Thereafter, receptormediated cell association was measured as previously described [25]. Values represent cell association expressed as percentage of the control cell association. The control association is the association in the absence of unlabelled lipoprotein (100%). Binding of each lipoprotein sample was carried out in triplicate. n, number of different lipoprotein samples (subjects) tested. Each value represents the mean f S.D.

190

ki. Mulder et al. /Atherosclerosis 108 (1994) 183-192

monly associated with homozygosity for the APOE*2(Argl58 -Cys) allele. The underlying metabolic defect is a disturbed function of the apo E2(Argl58-Cys) isoform as ligand for lipoprotein receptors. Since homozygosity for this allele is required for the development of FD, this variant has come to be associated with a recessive mode of inheritance of FD. Heterozygosity for the APOE*2(Lys146-Gln) allele is also associated with FD. In this case, FD exhibits a dominant mode of inheritance as only one defective APOE*2(Lys146-Gln) allele is required. Thus, in this case FD is expressed despite the presence of normal apo E [ 121, suggesting that the presence of the abnormal apo E2(Lys146-Gln) variant itself is involved in the expression of FD. It has been shown that the basic amino acid residues in the region 131 to 150 of apo E are necessary for binding to the LDL receptor [27]. In line with this, purified apo E2(Lys146-Gln) protein, when complexed with artificial phospholipid vesicles, displays only about 40% of the binding activity of normal apo E3/phospholipid complexes [13]. We found that the native d < 1.019 lipoproteins of APOE*2(Lys146-Gln) heterozygous FD probands contain about equal amounts of normal and mutant apo E and that these lipoproteins bound to the LDL receptor rather efficiently when compared with the corresponding lipoprotein fraction of normal APOE* homozygotes (Table 2). In contrast, under the same experimental conditions, the d < 1.019 lipoproteins from APOE*2(Arg158-Cys) homozygotes were much more defective in binding to the receptor, comparable to the results described by Schneider et al. 191. Earlier, Chappell et al. [28] also found that VLDL isolated from a subject heterozygous for the APOE*2(Lys146-Gln) allele was not severely defective in its ability to compete with LDL for binding to the LDL receptor. Our present results show that the d < 1.019 lipoprotein fraction of the APOE*2(Lys146-Gln) heterozygous FD probands, and also of their relatives carrying this mutant allele, are less efliciently converted into cholesterol-rich lipoproteins upon addition of LPL to complete serum than the corresponding lipoproteins of their relatives, not carrying this apo E variant. This observation

could be due to a lower suitability of these lipoproteins as substrate for LPL. A less suitable LPL substrate could be due to a higher basal cholesterol to triglyceride ratio in these lipoproteins as compared to that in non-carriers. However, the observation that the d < 1.019 lipoproteins of APOE*2(Argl58- Cys) homozygotes are still relatively good substrates for LPL, despite its strongly elevated basal cholesterol to triglyceride ratio (Fig. l), argues against this possibility. This is in agreement with the statement of Demant et al. [29] saying that in APOE*2(Argl58 -Cys) homozygotes the conversion of IDL into LDL is disturbed, rather than the conversion of VLDL into VLDL-remnants or IDL. Notwithstanding these arguments, some studies do claim that the triglyceride-rich lipoproteins in homozygous APOE*2(Arg158-Cys) FD patients are relatively resistant to lipolysis [30,3 1,321, although this is not directly related to the apo E protein itself [31]. Apo C2 is known to be an activator of LPL (for review, see Ref. 33). Isoelectric focusing followed by protein-staining showed that the d < 1.019 lipoprotein fractions of the APOE*2(Lys146-Gln) heterozygous FD probands, do contain apo C2 (see Fig. 2 in Ref. 11). Thus the lack of susceptibility of these lipoproteins to LPL-mediated lipolysis is not due to a lack of apo C2. Under our experimental conditions the LPL was added to complete serum rather than to isolated lipoprotein fractions, in order to keep the condition as physiological as possible. Thus, CETP activity is also present during the LPL treatment and could therefore also be involved in the conversion of d < 1.019 g/ml lipoprotein fraction into more cholesterol-rich lipoproteins. Consequently, a defect in this conversion might be due to a defect in lipolysis, in CETP activity or in a combination of both. In the present study, we found that the d < 1.019 lipoproteins of the APOE*2(Lys146-Gln) allele carriers contain a relatively high amount of apo E per particle or per apo B molecule (Table l), which is about equally contributed by the apo EZ(Lys146-Gln) variant and the normal apo E protein [ 111. This relatively high amount of total apo E protein per lipoprotein particle might be the direct cause of poor conversion of the d < 1.019

hf. Mulder et al. /Atherosclerosis 108 (1994) 183-192

g/ml lipoproteins into more cholesterol-rich lipoproteins. Next to an enhanced apo E content, the d < 1.019 lipoprotein fraction of APOE*2(Lys146-Gln) allele carriers also contains relatively high amounts of unesterified cholesterol (results not shown). It has been suggested that increased cholesterol will lead to a competition of cholesterol with triglyceride for a place at the surface of the lipoprotein particle [34]. Thus, unesteritied cholesterol could also inhibit the LPL(and CETP-) mediated conversion. Previously, we have reported that the increase of the cholesterol to triglyceride ratio of the d < 1.019 lipoprotein fraction of normolipidemic subjects upon treatment with LPL, leads to an enhanced cellular processing of these lipoproteins via the LDL receptor [25]. From the present results we obtained evidence that a retarded conversion of the d < 1.019 lipoprotein fraction in APOE*2(Lys 146- Gln) allele carrying FD patients indeed predisposes to a less efficient hepatic catabolism of these lipoproteins. Such a retarded hepatic processing of remnant lipoproteins might help to explain the dominant mode of inheritance of FD in subjects with heterozygosity for the APOE*2(Lys146-Gln) allele. 5. Acknowledgements We thank Dr. Anton Stalenhoef and Dr. Guus Smelt (Lipid Clinics of the University Hospitals of Nijmegen and Leiden, respectively) and Leny van Mourik for providing us with blood samples. This supported by the study was financially Netherlands Heart Foundation (projects no.s 87.025 and 88.086).

4

5

6

7

8

9

10

11

12

13

6. References 14

I

2

3

Brown, MS., Kovanen, P.T. and Goldstein, J.L., Regulation of plasma cholesterol by lipoprotein receptors, Science, 212 (1981) 628. Sherrill, B.C., Innerarity, T.L. and Mahley, R.W., Rapid hepatic clearance of the canine lipoproteins containing only the E apolipoprotein by a high affinity receptor, J. Biol. Chem., 255 (1980) 1804. Utennann, G., Hees, M. and Steinmetz, A., PolymorE and occurrence of phism of apolipoprotein dysbetalipoproteinemia in man, Nature (Lond.), 269 (1977) 604.

15

16

191

Zannis, V.I., and Breslow. J.L., Human VLDL apolipoprotein E isoprotein polymorphism is explained by genetic variation and posttranslational modification, Biochemistry, 20 (1981) 1033. Rall, S.C., Weisgraber, K.H. and Mahley, R.W., Human apolipoprotein E. The complete amino acid sequence, J. Biol. Chem., 257 (1982) 4171. Mahley, R.W. and Rall, S.C., Jr., Type III hyperlipoproteinemia (dysbetalipoproteinemia): The role of apolipoprotein E in normal and abnormal metabolism. In: Striver, C.R., Beaudet, A.L., Sly, W.S. and Valle, D. (Eds.), The Metabolic Basis of Inherited Disease, McGraw-Hill, New York, 1989, p. 1195. Utermann, G., Vogelberg, K.H., Schoenborn, W. et al., Polymorphism of apolipoprotein E: II. genetics of hyperlipoproteinemia type III, Clin. Genet., 15 (1979) 37. Uterrnann, G., Genetic polymorphism in apolipoprotein E - Impact on plasma lipoprotein metabolism. In: Crepaldi, G., Tiengo, A. and Baggio, G. (Eds.), Diabetes, Obesity and Hyperlipidemias III, Amsterdam 1985: Excerpta Medica, I. Schneider, W.J., Kovanen, P.T., Brown, M.S. et al., Familial dysbetalipoproteinemia. Abnormal binding of mutant apolipoprotein E to low density receptors of human libroblasts and membranes from liver and adrenals of rats, rabbits and cows, J. Clin. Invest., 68 (1981) 1075. Rall, S.C., Weisgraber, K.H., Innerarity, T.L. and Mahley, R.W., Structural basis of receptor binding heterogeneity of apolipoprotein E from type III hyperlipoproteinemic subjects, Proc. Natl. Acad. Sci. USA, 79 (1982) 4695. Smit, M., de Knijff, P., Frants, R.R., Klasen, E.C. and Havekes, L.M., Familial dysbetalipoproteinemic subjects with the E3/E2 phenotype exhibit an E2 isoform with only one cysteine residue, Clin. Genet., 32 (1987) 335. Smit, M., de Knijff, P., van der Kooy-Meys, E. et al., Genetic heterogeneity in familial dysbetalipoproteinemia. The E2(Lysl46-Gin) variant results in a dominant mode of inheritance, J. Lipid Res., 31 (1990) 45. Rall, S.C., Weisgraber, K.H., Innerarity, T.L., Bersot, T.P. and Mahley, R.W., Identification of a new structural variant of human apolipoprotein E, E2(Lysl46-Gin) in a type III hyperlipoproteinemic subject with the E3/E2 phenotype, J. Clin. Invest., 72 (1983) 1288. Havekes, L.M., de Knijff, P., Beisiegel, U., Havinga, J., Smit, M. and Klasen, T.E., A rapid micromethod for apolipoprotein E phenotyping directly in serum, J. Lipid Res., 28 (1987) 455. De Knijff, P., van den Maagdenberg, A.M.J.M., Stalenhoef, A.F.H. et al., Familial dysbetalipoproteinemia associated with apolipoprotein E3-Leiden in an extended multigeneration pedigree, J. Clin. Invest., 88 (1991) 643. Hixson, J.E. and Vernier, D.T., Restriction isotyping of human apolipoprotein E by gene amplilication and cleavage with HhaI, J. Lipid Res., 31 (1990) 545.

192

17

18

19

20

21

22

23

24

25

26

IU. Mulder et al. /Atherosclerosis

Paik, Y.-K., Chang, D.J., Reardon, C.A., Davies, G.E., Mahley, R.W. and Taylor, J.M., Nucleotide sequence and structure of the human apolipoprotein E gene, Proc. Natl. Acad. Sci. USA, 88 (1985) 3445. Lopez-Virella, M.F., Stone, P., Ellis, S. and Colwell, J.A., Cholesterol determination in high-density lipoproteins separated by three different methods, Clin. Chem., 23 (1977) 882. Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J., Protein measurement with the folin phenol reagent, J. Biol. Chem., 193 (1951) 265. Van Vlijmen, B.J.M., van den Maagdenberg, A.M.J.M., Gijbels, M.J.J. et al., Diet-induced hyperlipoproteinemia and atherosclerosis in apolipoprotein E3-Leiden transgenic mice, J. Clin. Invest. (1994) in press. Kaptein, A., Roodenburg, L. and Princen, H.M.G., Butyrate stimulates the secretion of apolipoprotein A-I and B-100 by the human hepatoma cell line HepG2, Biochem. J., 278 (1991) 557. Pagan0 Mirani-Oostdijk, C., Havekes, L., Den Boer, P., De Wit, E. and Terpstra, J., Diurnal course of the suitability of serum samples obtained from normal subjects and from patients with type IV hyperlipidaemia to serve as substrate for lipoprotein lipase, Stand. J. Clin. Lab. Invest., 45 (1985) 207. Redgrave, T.G., Roberts, D.C.D. and West, C.E., Separation of plasma lipoproteins by density gradient ultracentrifugation, Anal. Biochem., 65 (1975) 42. Bilheimer, D.W., Eisenberg, S. and Levy, RI., The metabolism of very low density lipoproteins. 1. Preliminary in vitro and in vivo observations, Biochim. Biophys. Acta, 260 (1972) 212. Mulder, M., De Wit, E. and Havekes, L.M., The binding of human lipoprotein hpase treated VLDL by the human hepatoma cell line HepG2, B&him. Biophys. Acta, 1081 (1991) 308. Zhao, S.-P., Smelt, A.H.M., Gevers Leuven, J.A., van den Maagdenberg, A.M.J.M., van der Laarse, A. and van’t

108 (1994) 183-192

27

28

29

30

31

32

33

34

Hooft, F.M., Lipoproteins in familial dysbetalipoproteinemia. Variations of serum cholesterol level associated with VLDL concentration, Arteriosclerosis Thromb., 13 (1993) 316. Wilson, C., Wardell, M.R., Weisgraber, K.H., Mahley, R.W., Agard, D.A., Three-dimensional structure of the LDL receptor binding domain of human apohpoprotein E, Science, 252 (1991) 1817. Chappell, D.A., High receptor binding affinity of hpoproteins in atypical dysbetalipoproteinemia (type III hyperlipoproteinemia), J. Clin. Invest., 84 (1989) 1906. Demant, T., Bedford, D., Packard, C.J. and Shepherd, J., Influence of apolipoprotein E polymorphism on apolipoprotein B-100 metabolism in normolipidemic subjects, J. Clin. Invest., 88 (1991) 1490. Byung, H.C. and Segrest, J.P., Resistance of a very low density lipoprotein subpopulation from familial dysbetalipoproteinemia to in vitro lipolytic conversion to the low density lipoprotein density fraction, J. Lipid Res., 24 (1983) 1148. Ehnholm, C., Mahley, R.W., Chappell, D.A., Weisgraber, K.H., Ludwig, E. and Witztum, J.L., Role of apolipoprotein E in the lipolytic conversion of P-very low density hpoproteins to low density lipoproteins in type III hyperlipoproteinemia, Proc. Natl. Acad. Sci., USA, 81 (1984) 5566. Saheki, S., Takahashi, I., Murase, M., Takeuchi, N. and Uchida, K., Composition of very low density lipoproteins and in vitro effect of lipoprotein lipase, Clin. Chim. Acta, 204 (1991) 155. Wang, C.-S., Hartsuck, J. and McConathy, W., Structure and functional properties of lipoprotein lipase, Biochim. Biophys. Acta, 1123 (1992) 1. Miller, K.W. and Small, D.M., Structure of triglyceriderich lipoproteins: an analysis of core and surface phases. In: Gotto, A.M. (Ed.), Plasma Lipoproteins, Elsevier Science Publishers B.V., Amsterdam, 1987, p. 1.