A

awivTw

vphe

h©

K

1

tt

s

0d

Atherosclerosis 194 (2007) e116–e122

Effect of ezetimibe coadministered with statins ingenotype-confirmed heterozygous FH patients

Livia Pisciotta a, Tommaso Fasano b, Antonella Bellocchio a, Letizia Bocchi b,Raffaella Sallo a, Raffaele Fresa a, Isabella Colangeli c, Alfredo Cantafora d,

Sebastiano Calandra b,∗∗, Stefano Bertolini a,∗a Department of Internal Medicine, University of Genoa, Viale Benedetto XV 6, I-16132 Genoa, Italy

b Department of Biomedical Sciences, University of Modena and Reggio Emilia, Via Campi 287, I-41100 Modena, Italyc Merck Sharp & Dohme, Rome, Italy

d National Institute of Health, Rome, Italy

Received 7 September 2006; received in revised form 31 October 2006; accepted 31 October 2006Available online 30 November 2006

bstract

We investigated the effect of statins and statins plus ezetimibe in 65 FH heterozygotes carrying LDLR-defective or LDLR-negative mutationss well as the effect of ezetimibe monotherapy in 50 hypercholesterolemic (HCH) patients intolerant to statins. PCSK9 and NPC1L1 genesere analysed to assess the role of genetic variants in response to therapy. In FH patients combined therapy reduced LDL-C by 57%,

rrespective of the type of LDLR mutation. The additional decrease of plasma LDL-C induced by ezetimibe showed wide inter-individualariability (from −39% to −4.7%) and was negatively correlated with percent LDL-C decrease due to statin alone (r = −0.713, P < 0.001).he variable response to statins was not due to PCSK9 gene variants associated with statin hyper-sensitivity. The highest response to ezetimibeas observed in a carrier of R174H substitution in NPC1L1, which had been found to be associated with high cholesterol absorption.In HCH patients, ezetimibe monotherapy induced a variable decrease of plasma LDL-C (from −47.7% to −13.4%). To investigate this

ariability, we sequenced NPC1L1 gene in patients with the highest and the lowest response to ezetimibe. This analysis showed a higherrevalence of the G allele of the c.816 C>G polymorphism (L272L) in hyper-responders, an observation confirmed also in FH patientsyper-responders to ezetimibe. In both FH and HCH patients, the G allele carriers tended to have a higher LDL-C reduction in response tozetimibe.

These observations suggest that in FH heterozygotes LDL-C reduction following combined therapy reflects a complex interplay betweenepatic synthesis and intestinal absorption of cholesterol.

2006 Elsevier Ireland Ltd. All rights reserved.

eywords: Familial hypercholesterolemia; LDLR mutations; Statins; Ezetimibe; NPC1L1 gene; PCSK9 gene

hd

. Introduction

Familial hypercholesterolemia (FH), caused by muta-ions in the gene encoding the LDL receptor (LDLR), ishe most frequent form of autosomal co-dominant hyperc-

∗ Corresponding author. Tel.: +39 010 3537992; fax: +39 010 3537797.∗∗ Corresponding author. Tel.: +39 059 2055423; fax: +39 059 2055426.

E-mail addresses: sebcal@unimo.it (S. Calandra),tefbert@unige.it (S. Bertolini).

ld[iimm(

021-9150/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.atherosclerosis.2006.10.036

olesterolemia. The lack of functioning LDLRs results inecreased catabolism of LDL and the accumulation of thisipoprotein in plasma, a condition, which predisposes to theevelopment of premature coronary artery disease (CAD)1]. In homozygous and heterozygous FH individuals thencrease of plasma LDL as well as the severity of the clin-

cal phenotype are more pronounced in the presence ofutations abolishing the receptor function (receptor-negativeutations) than in mutations resulting in defective receptors

receptor-defective mutations) [1–3].

sclerosis 194 (2007) e116–e122 e117

eagiabicottgawe

td1mtIt[cdssing

tgoriwpe

2

2h

w(rapsi

Table 1Clinical and biochemical characteristics of genotype confirmed heterozy-gous FH patients before treatment

Gender P

Males Females

Number of patients 30 35Age (years) 51.4 ± 13.2 57.1 ± 12.8 NSBMI (kg/m2) 24.4 ± 2.5 24.4 ± 4.2 NST-C (mmol/L) 10.56 ± 1.84 10.94 ± 1.71 NSLDL-C (mmol/L) 8.53 ± 1.73 8.80 ± 1.65 NSHDL-C (mmol/L) 1.26 ± 0.39 1.45 ± 0.36 <0.05TG (mmol/L) 1.71 ± 0.74 1.50 ± 0.71 NSTx (%) 56.7 74.3 NSCAD (%) 46.6 34.3 NSCA-ATS (%) 56.7 57.1 NS

The values are mean ± S.D. Tx, tendon xanthomas; CAD, coronary arterydisease; CA-ATS, carotid atherosclerosis.

Table 2Clinical and biochemical features of receptor-defective and receptor-negative heterozygous FH patients before treatment

Receptor Unclassified

Defective Negative

Males/females 5/8 19/24 6/3Age (years) 53.6 ± 15.0 55.8 ± 12.7 49.4 ± 12.7BMI (kg/m2) 24.1 ± 3.4 24.3 ± 3.7 25.2 ± 2.7T-C (mmol/L) 9.15 ± 1.31* 11.17 ± 1.66 11.12 ± 1.62LDL-C (mmol/L) 7.16 ± 1.17* 9.08 ± 1.58 8.95 ± 1.64HDL-C (mmol/L) 1.35 ± 0.37 1.39 ± 0.39 1.25 ± 0.36TG (mmol/L) 1.44 ± 0.70 1.56 ± 0.69 2.02 ± 0.84Tx (%) 15.4* 76.7 88.8CAD (%) 15.4† 41.8 66.6CA-ATS (%) 38.4 60.4 66.6Simvastatin/atorvastatin 7/6 12/31 4/5

Tdo

w4tpatewwmt

2i

L. Pisciotta et al. / Athero

HMG-CoA reductase inhibitors (statins) have been veryffective in lowering plasma LDL cholesterol (LDL-C) [4]nd reducing the risk for coronary mortality [5] in heterozy-ous FH patients. However, FH patients show wide variationsn inter-individual plasma LDL-C responses to statins, whichre independent of the type of statin and the dose used. It haseen suggested that the type of LDL receptor mutation couldnfluence the response to statins. This issue, however, is stillontroversial as some studies provided evidence in favourf [6,7] and others against [8,9] this hypothesis. Irrespec-ive of the large inter-individual variability in the responseo statins, many FH patients are not achieving their LDL-Coal for an effective reduction of CAD risk. To achieve desir-ble LDL-C levels in FH patients statins have been combinedith inhibitors of cholesterol absorption, such as plant stanol

sters, bile salt sequestrant resins or ezetimibe [10,11].Ezetimibe is a new hypolipidaemic drug, which selec-

ively inhibits cholesterol absorption in enterocyte. Recentata suggest that this drug binds to the Niemann-Pick C1 Like(NPC1L1) protein, which is located in the brush-borderembrane of the enterocyte where it plays a major role in

he intestinal uptake of cholesterol and plant sterols [12–14].n patients with the clinical diagnosis of primary hypercholes-erolemia ezetimibe has been combined with simvastatin15], pravastatin [16], lovastatin [17] or atorvastatin [18]. Thisombined therapy, regardless of the type of statins and theiroses, substantially enhanced the LDL-C lowering effect oftatins by 12–19% [15–18]. Ezetimibe co-administered withimvastatin or atorvastatin has also been used successfullyn homozygous FH [19,20]. However, at present there areo data concerning the effect of the combined therapy inenotype-confirmed heterozygous FH patients.

The present study was designed to assess the effect ofhe addition of ezetimibe to simvastatin or to atorvastatin inenotyped heterozygous FH patients and to investigate somef the factors underlying the inter-individual variability in theesponse to this combined therapy. To investigate the inter-ndividual variability in response to ezetimibe monotherapye conducted a parallel study in a group of non-genotypedatients with primary hypercholesterolemia who were intol-rant to statins.

. Subjects and methods

.1. Molecularly characterized patients with familialypercholesterolemia (FH)

Sixty-five heterozygous FH patients who had been treatedith statins for at least 2 years were included in the study

Table 1). Fifty-six of them were classified as carriers ofeceptor-negative or of receptor-defective LDLR mutations,

ccording to the previously described criteria [2]. Nineatients were carriers of unclassified LDLR gene mutationsince cultured fibroblasts were not available for LDLR activ-ty assay (Table 2 and Supplementary Table 1).

hysm

he values are mean ± S.D. Tx, tendon xanthomas; CAD, coronary arteryisease; CA-ATS, carotid atherosclerosis. *Significantly different from thether two groups; †significantly different from receptor-unclassified group.

The distribution of patients in relation to statin treatmentas the following: simvastatin 20 mg/day no. 5, simvastatin0 mg/day no. 18, atorvastatin 20 mg/day no. 14, atorvas-atin 40 mg/day no. 28. Four plasma lipid determinations wereerformed at 2-month interval; their mean value was takens the reference parameter to evaluate the response to statinreatment. All patients maintained their statin regimen whenzetimibe (10 mg/day) was added. The combined treatmentas carried out for 3 months; plasma lipid concentrationsere determined monthly. The mean value of the three deter-inations was used to evaluate the effects of the combined

reatment.

.2. Patients with primary hypercholesterolemia (HCH)ntolerant to statin therapy

Fifty patients with the clinical diagnosis of primary

ypercholesterolemia (19 males and 31 females, 57.3 ± 13.7ears of age, BMI 22.9 ± 4.3 kg/m2) and intolerant totatins (i.e. with a history of adverse events during treat-ent with statins) were treated with ezetimibe (10 mg/day)

e scleros

fdsgctwocw

ght

2

mBMmcapw

2

fadaph8rp

itFaRcjstSspeari

wUw2

sLt(e

a(iDnn

2

(bevowtpcbla

3

3

itFrrofrtat−

118 L. Pisciotta et al. / Athero

or 3 months. The analysis of LDLR gene was not con-ucted in this group of patients, which most likely includedubjects with heterozygous FH as well as subjects with poly-enic hypercholesterolemia. Plasma lipid and apolipoproteinoncentrations were determined monthly during ezetimibereatment and the mean values of the three determinationsere compared to the pre-treatment values. In the HCH groupvert coronary artery disease was present in eight patients;arotid atherosclerosis with plaques causing >25% stenosisas documented in 48% of the cases.Informed consent was obtained from all subjects investi-

ated. The study protocol was approved by the institutionaluman investigation committee of each participating institu-ion.

.3. Biochemical analysis

Plasma cholesterol (T-C) and triglycerides (TG) wereeasured enzymatically and apolipoproteins AI and

by immunoturbidimetry (Roche Diagnostics GmbH,annheim, Germany) using an automated analyzer (Hitachiodel 912, Hitachi Ltd., Tokyo). High-density lipoprotein

holesterol (HDL-C) was measured in plasma supernatantfter precipitation of apo B-containing lipoproteins byhosphotungstate-MgCl2. LDL-C was calculated by Friede-ald’s formula.

.4. Patient genotyping

The sequence of LDLR and PCSK9 genes was per-ormed as previously described [2,3,21]. NPC1L1 gene wasmplified using appropriately designed primers and PCR con-itions specified in Supplementary Table 2. Patients werelso genotyped for APOE polymorphism [22]. The com-lete sequence of PCSK9 gene was performed in FH patientsyper-responders (with percent LDL-C decrease above the5th percentile of the distribution) and in FH patients hypo-esponders (with percent LDL-C decrease below the 15thercentile) to statin.

Since it has been reported that some sequence variantsn the coding region of NPC1L1 gene influence choles-erol absorption and plasma LDL-C [23], we screened allH and HCH patients for four rare non conservative aminocid variants reported to be associated with low (c.916 C>T,306C; c.1249 C>T, R417W) or high (c.521 G>A, R174H;.845 C>T, P282L) cholesterol absorption in Caucasian sub-ects [23]. The R174H, P282L and R306C mutations werecreened by amplification and direct sequencing of the cen-ral portion of exon 2 using Ex 2b primers reported inupplementary Table 2. This procedure also allowed us tocreen all FH and HCH patients for the presence of a silentolymorphism g.1679 C>G (c.816 C>G, L272L) located in

xon 2 [24]. The screening for R417W was performed bymplification of the 3′-half of exon 2 using Ex 2c primerseported in Supplementary Table 2. Since the mutation elim-nates an Age I restriction site, the PCR product was digested

i−rs

is 194 (2007) e116–e122

ith 10 U of Age I (New England Biolabs, Beverly, MA,SA) and the digestion products (381 and 246 bp for theild type and 627 bp for the mutant allele) were separated on% agarose gel.

The complete sequence of NPC1L1 gene was performed inix FH and five HCH patients hyper-responders (with percentDL-C decrease above the 90th percentile of the distribu-

ion) and in six FH and five HCH patients hypo-responderswith percent LDL-C decrease below the 10th percentile) tozetimibe.

The gene sequence variants were designatedccording to the Human Genome Variation Societywww.hgvs.org/mutnomen); the numerical series of codonsncludes the sequence of the signal peptide. For genomicNA and cDNA, the A of the ATG of the initiator methio-ine codon is denoted nucleotide +1. Nucleotide 5′ to +1 isumbered −1.

.5. Statistical analysis

Statistical analysis was performed by using the SPSS 13.0SPSS Inc., Chicago, IL) program. Differences in the distri-ution of categorical variables were assessed by χ2 or Fisher’sxact tests. Statistically significant differences between lipidalues before and after treatment were evaluated by ANOVAr by Student’s t-test for paired data. Triglyceride valueshich were not distributed normally were logarithmically

ransformed before analysis. Multiple comparisons amongairs of means were performed by t-tests with Bonferroni’sorrection. Least squares mean ± S.E. percent changes fromaseline are reported. Median percentage change from base-ine was calculated for triglycerides, since this parameter issymmetrically distributed.

. Results

.1. Response of FH patients to combined treatment

Table 3 shows the changes of lipid parameters dur-ng statin monotherapy and after 3 months of combinedherapy with statin plus ezetimibe in the whole group ofH patients. There were no significant differences in theesponse of LDL-C to statin therapy among patients witheceptor-defective (−32.5%), receptor-negative (−38.2%)r unclassified mutations (−35.6%). The same was trueor the combined therapy (−54.5%, −56.9% and −58.7%,espectively). Statins (simvastatin 20/40 mg day−1; atorvas-atin 20/40 mg day−1) induced a substantial decrease of T-Cnd LDL-C (mean decrease −29.7% and −36.7%, respec-ively) and a moderate decrease of TG (median decrease

14.4%). The addition of ezetimibe (10 mg/day) resulted

n a further reduction of T-C and LDL-C (mean decrease

46.4% and −56.7%, respectively) and a more significanteduction of plasma TG (median decrease −25.9%). Duringtatin treatment none of the patients achieved the recom-

L. Pisciotta et al. / Atherosclerosis 194 (2007) e116–e122 e119

Table 3Plasma lipid concentrations in heterozygous FH patients after treatment with statin and statin plus ezetimibe

Baseline Statin Statin + ezetimibe

T-C (mmol/L) 10.76 ± 1.76 7.50 ± 1.24 (−29.7 ± 1.1%)a 5.70 ± 0.85 (−46.4 ± 0.9%)a,b

LDL-C (mmol/L) 8.68 ± 1.68 5.42 ± 1.13 (−36.7 ± 1.3%)a 3.69 ± 0.72 (−56.7 ± 0.9%)a,b

HDL-C (mmol/L) 1.36 ± 0.38 1.47 ± 0.39 1.47 ± 0.39TG (mmol/L) 1.60 ± 0.73 1.34 ± 0.48 (−14.4%)c 1.18 ± 0.47 (−25.9%)d

T and Bonferroni’s test for multiple comparisons; aP < 0.0001 vs. baseline values;b eline values. Least squares mean ± S.E. percent changes from baseline are reportedi

mpappa

a

3r

bdb−watLibwt

Fg2porac

Fb

tP

3

he values are mean ± S.D. Statistical analysis was performed by ANOVAP < 0.0001 vs. statin values; cP < 0.05 vs. baseline value; dP < 0.001 vs. basn brackets. Median percent change is reported for TG.

ended plasma level of LDL-C (2.59 mmol/L); in only sevenatients (10.7%) plasma LDL-C level was between 2.59nd 4.14 mmol/L. After 3 months of combined treatment 5atients (7.7%) had LDL-C levels below 2.59 mmol/L, 15atients (23%) had LDL-C between 2.59 and 3.36 mmol/Lnd 23 (35.4%) had LDL-C between 3.37 and 4.14 mmol/L.

No clinical or biochemical side effects of combined ther-py were observed.

.2. Relation between the response to statin and theesponse to ezetimibe in FH patients

The additional percent decrease of plasma LDL-C inducedy ezetimibe (calculated as the difference between percentecrease induced by combined therapy minus that inducedy statin) showed a wide inter-individual variability (from39.2% to −4.7%), as reported in Fig. 1. In order to definehether the magnitude of the response to ezetimibe was

ffected by the pre-existing treatment with statin we plot-ed the percentage response to statin against the additionalDL-C reduction (given as percentage) induced by ezetim-

be. We observed a highly significant inverse correlationetween these parameters (Fig. 2). This inverse correlationas confirmed when we restricted the analysis to patients

reated with simvastatin 40 mg/day (r = −0.567; P < 0.02) or

ig. 1. Individual LDL-C response to ezetimibe (% decrease) in 65enotype confirmed heterozygous FH patients treated with simvastatin0/40 mg day−1 or atorvastatin 20/40 mg day−1. Grey columns indicateatients considered high- or low-responders to ezetimibe (percent reductionf LDL-C due to ezetimibe above the 90th and below the 10th percentile,espectively). R174H (boxed) indicates the subject carrying this rare aminocid substitution of NPC1L1 protein reported to be associated with highholesterol absorption (Ref. [23]).

s

iitp

aoba

TE(

TLHTAA

TLb

ig. 2. Pearson’s correlation between percent decrease of LDL-C inducedy statins and by ezetimibe in genotype confirmed heterozygous FH patients.

o patients treated with atorvastatin 40 mg/day (r = −0.792;< 0.0001).

.3. Response to ezetimibe of HCH patients intolerant totatin

Table 4 shows the response to the treatment with ezetimiben 50 patients with primary hypercholesterolemia who werentolerant to statins. The baseline plasma level of LDL-C inhis group of patients was lower than that of genotyped FHatients (P < 0.001).

Ezetimibe induced an almost 30% decrease in LDL-C and

less pronounced decrease of Apo B and TG. In this groupf patients the LDL-C decrease was higher than that inducedy ezetimibe in FH patients with pre-existing statin ther-py (−30.2 ± 1.0% versus −19.9 ± 1.0%, P < 0.001). Also in

able 4ffect of ezetimibe treatment in patients with primary hypercholesterolemia

HCH) intolerant to statin therapy

Baseline Ezetimibe

-C (mmol/L) 7.93 ± 1.05 6.19 ± 0.88 (−22.0 ± 0.6%)a

DL-C (mmol/L) 5.54 ± 0.99 3.88 ± 0.84 (−30.2 ± 1.0%)a

DL-C (mmol/L) 1.61 ± 0.40 1.62 ± 0.41G (mmol/L) 1.73 ± 0.57 1.43 ± 0.49 (−19.5%)a

po AI (mg/dl) 167.5 ± 28.6 170.7 ± 29.8po B (mg/dl) 142.7 ± 25.0 112.9 ± 20.8 (−20.8 ± 1.1%)a

he values are mean ± S.D. aP < 0.0001 vs. baseline values (paired t-test).east squares mean ± S.E. percent changes from baseline are reported inrackets. Median percent change is reported for TG.

e120 L. Pisciotta et al. / Atheroscleros

Fig. 3. Individual LDL-C response (% decrease) to ezetimibe monotherapyittp

Hie

3t

ts−�nhi

3t

iGarvCi(t

3r

awcgot

tmaaC

hatthfscovhhtte(frewiv

4

cctdvr(e

oppmdartwT

n patients with primary hypercholesterolemia (HCH) who were intoleranto statin. Grey columns indicate patients considered high- or low-responderso ezetimibe (percent reduction of LDL-C above the 90th and below the 10thercentile, respectively).

CH patients the response to ezetimibe showed a wide inter-ndividual variability (Fig. 3). No clinical or biochemical sideffects of ezetimibe therapy were observed.

.4. Relation between APOE genotype and the responseo treatment

When we stratified FH patients according to APOE geno-ype we found the following plasma LDL-C decrease aftertatin or combined treatment respectively: �2�3 (no. 4)44.4% and −57.4%; �3�3 (no. 47) −35.6% and −57.0%;

3�4 (no. 14) −36.7% and −55.6%. These differences wereot statistically significant. Similarly, the APOE genotypead no effect on the response to ezetimibe in HCH patientsntolerant to statins.

.5. Sequence variants in PCSK9 gene and the responseo statins

It has been recently reported that some missense mutationsn the coding region of PCSK9 gene (c.137 G>T, R46L; c.316>A, G106R; c.471 C>A, N157K; c.709 C>T, R237W) are

ssociated with hypocholesterolemia and possibly increasedesponse to statin therapy in FH patients [25]. The amino acidariants found in both groups were the following: c.42 43insTG (ins L at position 24, previously reported as 15 16

nsL [26]), c.158 C>T (A53V) and c.1420 A>G (I474V)Supplementary Table 3). No carriers of the missense muta-ions associated with hyper-sensitivity to statins were found.

.6. Sequence variants in NPC1L1 gene and theesponse to ezetimibe

The screening of FH and HCH patients for the rare aminocid variants of NPC1L1 protein, reported to be associatedith low (R306C and R417W) or high (R174H and P282L)

holesterol absorption [23], revealed that only one patient (aenotype confirmed heterozygous FH) was a carrier of onef these variants (R174H); this patient was found to havehe highest response (−39.2%) to ezetimibe (Fig. 1). During

(pct

is 194 (2007) e116–e122

his screening we also genotyped all patients for the silentutation c.816 C>G (L272L). The genotype distribution was

s follows: in the FH group 50 patients were CC, 14 were CGnd 1 GG; in the HCH group 39 patients were CC, 9 wereG and 2 GG.

The sequence of NPC1L1 gene in FH and HCH subjectsyper- or hypo-responders to ezetimibe showed several vari-nts, which are reported in Supplementary Table 4. Most ofhese variants had been previously reported [24,27]; two ofhem (Ex 15 g.22658, c.3201 C>T, L1067L, found in an HCHyper-responder, and Ex 17 g.24362, c.3405 CA, I1135I,ound in an HCH hypo-responder) are novel. This analy-is also showed a higher prevalence of the G allele of the.816 C>G polymorphism (see above) in hyper-respondersf both groups with respect to hypo-responders (FH: 0.25ersus 0.00; HCH 0.50 versus 0.10) and the presence of twoomozygotes for the rare allele (GG) among the five HCHyper-responders. In view of this finding we asked whetherhis polymorphism had any effect on the LDL-C responseo ezetimibe. Comparing the LDL-C reduction induced byzetimibe in subjects homozygous for the common alleleCC) with that observed in heterozygous and homozygousor the rare allele in both groups, we observed that car-iers of the G allele tended to have a higher response tozetimibe. The LDL-C percent decrease (mean ± S.E.M.)as as follows: −18.9 ± 1.0% in CC versus −23.3 ± 2.3%

n CG+GG (P < 0.07) in FH patients; −29.2 ± 1.0% in CCersus −33.7 ± 2.6% in CG+GG (P < 0.06) in HCH patients.

. Discussion

In the present study we documented that in genotype-onfirmed heterozygous FH patients treated with statins: (1)o-administration of ezetimibe induced a substantial addi-ional decrease of plasma LDL-C; (2) ezetimibe-dependentecrease of LDL-C showed a high degree of inter-individualariability; (3) there was a negative correlation between theesponse to statins and the subsequent response to ezetimibepatients hyper-responders to statins were poor-responders tozetimide and vice versa).

Although many studies have been conducted on the effectf combined therapy (statin plus ezetimibe) on the lipidrofile in patients with primary hypercholesterolemia, theresent work is the first, which reports the effects of this treat-ent in a large group of patients with molecularly defined

iagnosis of heterozygous FH. We thought that a well char-cterized group of FH patients with receptor-negative oreceptor-defective mutations would provide a unique oppor-unity to ascertain whether the response to this new treatmentas affected by the type of defect in the LDL receptor.he mean reduction of LDL-C obtained in our FH patients

−56.7%) was similar to that observed in patients withrimary hypercholesterolemia, who had been treated withomparable doses of simvastatin or atorvastatin together withhe standard dose of ezetimibe [15–18]. Our study showed

scleros

ttabwt[os

roOmpncsctcatohrweppiaihiuiaqhtrwsRapesa

aweps

shLeCscaaAw(iiGpt(tooa

opiIatdgiiwr

A

f

A

f2

R

L. Pisciotta et al. / Athero

hat the response to statins as well as the response to combinedherapy was highly variable among FH subjects. This vari-bility was not affected by the type of LDL receptor defect ory the APOE genotype. We also ruled out that this variabilityas due to the presence of mutations in PCSK9 gene, reported

o be associated with a good response to statin treatment25]. These negative results suggest that other environmentalr genetic factors are involved, which influence cholesterolynthesis or cholesterol absorption.

Previous studies showed that FH patients who were goodesponders to statins had a high basal level of plasma meval-nic acid, an indicator of a higher cholesterol synthesis [10].ther studies showed that statins reduced serum cholesterolore efficiently in those coronary patients with the lowest

lasma level of cholestanol, indicating a less efficient intesti-al cholesterol absorption [28]. Indeed the strong negativeorrelation between the LDL-C response to statins and theubsequent response to ezetimibe found in our FH patientsan be explained in terms of variability in cholesterol syn-hesis and cholesterol absorption. It is likely that the rate ofholesterol absorption and the rate of cholesterol synthesisre biologically linked, in that subjects with low choles-erol absorption capacity presumably have a reduced deliveryf intestinal cholesterol to the liver and a higher rate ofepatic cholesterol synthesis (due to a less pronounced down-egulation of HMG-CoA reductase) and vice versa. Patientsith low cholesterol absorption are expected to respond more

fficiently to statins and less efficiently to ezetimibe, whereasatients with high cholesterol absorption are expected to beoor responders to statins but good responders to ezetim-be. In this context the genetic factors controlling cholesterolbsorption might play an important role. NPC1L1 proteins one of the key factors in this process as recent studiesave showed that amino acid sequence variants of this proteinnfluence the cholesterol absorption and plasma LDL levelsnder physiological conditions [23], as well as during ezetim-be treatment [24,27]. For example, some non-conservativemino acid substitutions in NPC1L1 protein, found more fre-uently in African-Americans from the Dallas-Heart Study,ave been shown to be associated with reduced choles-erol absorption and lower plasma LDL-C; conversely, otherare amino acid variants have been found to be associatedith high cholesterol absorption [23]. Our FH patients were

creened for four rare amino acid variants (R174H, P282L,306C, R417W) associated with variations in cholesterolbsorption in Caucasian subjects from the Dallas-Heart Studyopulation [23]. We found that the FH patient with the high-st response to ezetimibe was, in fact, a carrier of the R174Hubstitution, which has been associated with high cholesterolbsorption [23].

To ascertain whether NPC1L1 sequence variants couldffect the LDL-C response to ezetimibe we sequenced the

hole NPC1L1 gene in FH and HCH patients with the high-

st (above the 90th percentile) and the lowest (below the 10thercentile) LDL-C response to ezetimibe. We found severalequence variants (Supplementary Table 4), the rare alleles of

is 194 (2007) e116–e122 e121

ome of which (g.-18 A, g.1679 G, g.25286 C and g.27621 C)ad been previously found to be associated with a higherDL-C response to ezetimibe [24,27]. In our patients, how-ver, three of these rare alleles (g.1679 G, g.25286 C, g.27621) were present in both hyper- and hypo-responders, thus

uggesting that their role in response to ezetimibe does notlearly emerge in a small sample of subjects. Nevertheless,mong HCH hyper-responders one subject was found to becarrier of all the above mentioned four rare alleles (g.-18, g.1679 G, g.25286 C and g.27621 C) and two subjectsere homozygous for the rare G allele of the g.1679 C>G

c.816 C>G) polymorphism. The latter polymorphism, whennvestigated in all FH and HCH patients, seems to have somenfluence on response to ezetimibe, as carriers of the rare

allele showed the tendency to have a higher reduction oflasma LDL-C following ezetimibe therapy. It is unlikelyhat this polymorphism, which results in a silent mutationL272L), directly affects the function of the NPC1L1 pro-ein. It is possible that it is linked to another mutationutside the coding region, which might affect the expressionf NPC1L1 gene and consequently the rate of cholesterolbsorption.

The mean reduction of LDL-C induced by the additionf ezetimide to statins therapy (i.e. the difference betweenercent decrease induced by combined therapy minus thatnduced by statin) was about −20% in FH patients (Fig. 1).n the group of HCH patients intolerant to statins we observedhigher percent reduction of LDL-C (about 30%) in response

o ezetimibe administration (Fig. 3). This difference might beue to the heterogeneity of the metabolic defect in the HCHroup, as compared to the group of genotyped FH patients. Its possible that the HCH group includes a high percentage ofndividuals with a high intestinal absorption of cholesterol,ho, for the reasons illustrated above, are expected to be good

esponders to ezetimibe treatment.

cknowledgement

This work was partially supported by grant unrestrictionalrom Merck Sharp & Dohme, Italy to S.B.

ppendix A. Supplementary data

Supplementary data associated with this article can beound, in the online version, at doi:10.1016/j.atherosclerosis.006.10.036.

eferences

[1] Rader DJ, Cohen J, Hobbs HH. Monogenic hypercholesterolemia:new insights in pathogenesis and treatment. J Clin Invest2003;111:1795–803.

[2] Bertolini S, Cantafora A, Averna M, et al. Clinical expression of familialhypercholesterolemia in clusters of mutations of LDL-receptor gene

e scleros

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

122 L. Pisciotta et al. / Athero

causing receptor-defective or receptor-negative phenotype. ArteriosclerThromb Vasc Biol 2000;20:e41–52.

[3] Bertolini S, Cassanelli S, Garuti R, et al. Analysis of LDL receptorgene mutations in Italian patients with homozygous familial hyperc-holesterolemia. Arterioscler Thromb Vasc Biol 1999;19:408–18.

[4] Vuorio AF, Kovanen PT, Gylling H. Hypolipidemic treatment of het-erozygous familial hypercholesterolemia: a lifelong challenge. ExpertRev Cardiovasc Ther 2004;2:405–15.

[5] Scientific Steering Committee on behalf of the Simon BroomeRegister Group. Mortality in treated heterozygous familial hyperc-holesterolemia: implication for clinical management. Atherosclerosis1999;142:105–12.

[6] Jeenah M, September W, van Roggen FG, et al. Influence of specificmutations at the LDL-receptor gene locus on the response to simvastatintherapy in Afrikaner patients with heterozygous familial hypercholes-terolemia. Atherosclerosis 1993;98:51–8.

[7] Kajinami K, Yagi K, Higashikata T, et al. Low-density lipoproteinreceptor genotype-dependent response to cholesterol lowering by com-bined pravastatin and cholestyramine in familial hypercholesterolemia.Am J Cardiol 1998;82:113–7.

[8] Sun XM, et al. Influence of genotype at the low density lipoprotein(LDL) receptor gene locus on the clinical phenotype and responseto lipid-lowering drug therapy in heterozygous familial hypercholes-terolemia. Atherosclerosis 1998;136:175–85.

[9] Sijbrands EJ, Lombardi MP, Westendorp RG, et al. Similar response tosimvastatin in patients heterozygous for familial hypercholesterolemiawith mRNA negative and mRNA positive mutations. Atherosclerosis1998;136:247–54.

10] O’Neill FH, Patel DD, Knight BL, et al. Determinants of variableresponse to statin treatment in patients with refractory familial hyper-cholesterolemia. Arterioscler Thromb Vasc Biol 2001;21:832–7.

11] Thompson GR, O’Neill F, Seed M. Why some patients respond poorlyto statins and how this might be remedied. Eur Heart J 2002;23:200–6.

12] Garcia-Calvo M, Lisnock JM, Bull HG, et al. The target of ezetimibeis Niemann-Pick C1-Like 1 (NPC1L1). PNAS 2005;102:8132–7.

13] Altmann SW, Davis HR, Zhu L-J, et al. Niemann-Pick C1 Like1 protein is critical for intestinal cholesterol absorption. Science2004;303:1201–4.

14] Sane AT, Sinnett D, Delvin E, et al. Localization and role of

NPC1L1 in cholesterol absorption in human intestine. J Lipid Res2006;47:2112–20.

15] Davidson MH, McGarry T, Bettis R, et al. Ezetimibe coadministeredwith simvastatin in patients with primary hypercholesterolemia. J AmColl Cardiol 2002;40:2125–34.

[

is 194 (2007) e116–e122

16] Melani L, Mills R, Hassman D, et al. Efficacy and safety of ezetimibecoadministered with pravastatin in patients with primary hypercholes-terolemia: a prospective, randomised, double-blind trial. Eur Heart J2003;24:717–28.

17] Kerzner B, Corbelli J, Sharp S, et al. Efficacy and safety of ezetimibecoadministered with lovastatin in primary hypercholesterolemia. Am JCardiol 2003;91:418–24.

18] Ballantyne CM, Houri J, Notarbartolo A, et al. Effect of ezetim-ibe coadministered with atorvastatin in 628 patients with primaryhypercholesterolemia. A prospective, randomized, double-blind trial.Circulation 2003;107:2409–15.

19] Gagne C, Gaudet D, Bruckert E. Efficacy and safety of ezetimibe coad-ministered with atorvastatin or simvastatin in patients with homozygousfamilial hypercholesterolemia. Circulation 2002;105:2469–75.

20] Yamamoto A, Harada-Shiba M, Endo M, et al. The effect of ezetimibeon serum lipids and lipoproteins in patients homozygous familial hyper-cholesterolemia undergoing LDL-apheresis therapy. Atherosclerosis2006;186:126–31.

21] Pisciotta L, Oliva CP, Cefalu AB, et al. Additive effect of mutationsin LDLR and PCSK9 genes on the phenotype of familial hypercholes-terolemia. Atherosclerosis 2006;186:433–40.

22] Hixson JE, Vernier DT. Restriction isotyping of human apolipopro-tein E by gene amplification and cleavage with HhaI. J Lipid Res1990;31:545–8.

23] Cohen JC, Pertsemlidis A, Fahmi S, et al. Multiple rare variants inNPC1L1 associated with reduced sterol absorption and plasma low-density lipoprotein levels. PNAS 2006;103:1810–5.

24] Hegele RA, Guy J, Ban MR, Wang J. NPC1L1 haplotype is associ-ated with inter-individual variation in plasma low-density lipoproteinresponse to ezetimibe. Lipids Health Dis 2005;4:16.

25] Berge KE, Ose L, Leren TP. Missense mutations in the PCSK9 gene areassociated with hypocholesterolemia and possibly increased responseto statin therapy. Arterioscler Thromb Vasc Biol 2006;26:1094–100.

26] Leren TP. Mutations in the PCSK9 gene in Norwegian sub-jects with autosomal dominant hypercholesterolemia. Clin Genet2004;65:419–22.

27] Simon JS, Karnoub MC, Devlin DJ, et al. Sequence variation inNPC1L1 and association with improved LDL-cholesterol lowering in

response to ezetimibe treatment. Genomics 2005;86:648–56.

28] Miettinen TA, Strandberg TE, Gylling H. Non-cholesterol sterols andcholesterol lowering by long-term simvastatin treatment in coronarypatients. Relation to basal serum cholestanol. Arterioscler Thromb VascBiol 2000;20:1340–6.