Dysfunctional Co-Expression Network Analysis of Familial Hypercholesterolemia

Journal of Cardiology 62 (2013) 58–62

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

Contents lists available at SciVerse ScienceDirect provided by Elsevier - Publisher Connector

Journal of Cardiology

jo urnal homepage: www.elsevier.com/locate/jjcc

Original article

Dysfunctional co-expression network analysis of familial hypercholesterolemia

a,1 b,1 a d

Yizhou Ye (MD, PhD) , Kefei Li (MD) , Jian Liu (MD, PhD) , Mingqiu Li (MD, PhD) ,

c d d d

Wei Wang (MD, PhD) , Ruxing Wang (MD, PhD) , Jian Zou (MD, PhD) , Ping Xie (MD, PhD) ,

d d,∗ a,∗

Liuyan Wei (MD, PhD) , Guoqing Jiao (MD, PhD) , Zhongxiang Yuan (MD, PhD)

Department of Cardiovascular Surgery, Afﬁliated Shanghai 1st People’s Hospital, Shanghai Jiaotong University, Shanghai 210008, People’s Republic of China

Department of Cardiac Surgery, Shanghai East Hospital. No.150 Jimo Road, 200120 Shanghai People’s Republic of China

Department of Cardiovascular Surgery, National Center for Cardiovascular Disease, Beijing 200000, People’s Republic of China

Department of Cardiovascular Surgery, Afﬁliated Wuxi People’s Hospital, Nanjing Medical University, Qingyang Road 299, Wuxi City, Jiangsu Province 214023, People’s Republic of

China

a r t i c l e i n f o a b s t r a c t

Article history: Familial hypercholesterolemia (FH) is an inherited disorder of blood lipid metabolism characterized by

Received 22 June 2012

high serum low-density lipoprotein cholesterol levels and premature coronary artery disease. In this

Received in revised form 4 February 2013

study, we used a system biology approach to identify co-expressed gene pairs that were potentially

Accepted 11 February 2013

involved in the progression of FH and constructed a conserved co-expression network using these genes.

Available online 16 May 2013

A total of 4232 co-expressed relationships were identified and we verified the significance by random

permutation. FH patients showed differences in lipoprotein and cholesterol metabolism in circulating

Keywords:

monocytes and lymphocytes compared to healthy controls. We hope our study could aid in understanding

Co-expression network

of FH and could provide the basis for FH biomarker identiﬁcation.

Differentially expressed genes

Familial hypercholesterolemia

Introduction of FH at primary health care level according to the World Health

Organization.

Familial hypercholesterolemia (FH) is an inherited disorder of Various attempts have been made to understand the patho-

blood lipid metabolism characterized by high serum low-density genesis of FH in the past decades. FH is caused by mutations of

lipoprotein (LDL) cholesterol levels. It is the most common and genes in sterol and lipoprotein pathways and displays varying gene-

most severe of the single gene disorders of LDL metabolism. dose effects [8]. To date, several genes have been identiﬁed as

The high level of serum LDL cholesterol frequently results in diagnostic markers for FH. The most common genetic mutations

excess deposition of cholesterol in tissues, leading to accelerated are LDL receptor (LDLR) mutations [9,10], apolipoprotein (apo) B

atherosclerosis and increased risk of premature coronary artery mutations [11], proprotein convertase subtilisin/kexin 9 (PCSK9)

disease [1]. It is estimated that 85% of males and 50% of females mutations [12,13], and the autosomal recessive hypercholesterol-

with FH will suffer coronary artery disease before 65 years of age emia (ARH) adaptor protein mutations [14]. Besides, the presence

[2–4]. FH may be loosely classiﬁed into “heterozygous” (one mutant of mutations in other candidate genes has been postulated in recent

LDL receptor allele) and “homozygous” (two mutated LDL recep- years, such as CYP7A1 [15], SREP-2 [16], or SCAP [17]. But the evi-

tor alleles) clinical phenotypes. Heterozygous FH occurs in 1:500 dence that these mutations cause the phenotype is not strong. High

people in the general population of most countries; homozygous throughput molecular techniques have been widely used to bet-

FH is much rarer, occurring in 1 in a million births [5,6]. Although ter understand the pathogenesis and progression of FH in recent

statins have been shown to reduce cholesterol levels in FH, they years. Most previous studies have focused on differential expres-

have not yet proven to be successful to cure this disease [7]. This sion and have identiﬁed hundreds of differentially expressed genes

high prevalence is sufﬁcient to warrant the diagnosis and treatment (DEGs) [18]. However, these investigations mostly focus on individ-

ual genes without consideration of potential relationships among

these genes.

It has been shown that co-expressed genes are correlated with

∗

Corresponding authors at: Department of Cardiovascular Surgery, Nanjing Med- functional relationships, such as physical interaction between the

ical University Afﬁliated Wuxi People’s Hospital, No. 299, Qingyang Road, Wuxi,

encoded proteins [19,20]. From the perspective of system biology,

Jiangsu 214023, People’s Republic of China. Tel.: +86 0510 82700775;

gene co-expression analysis is useful for investigating gene inter-

fax: +86 0510 82732445/85350555.

connection at the expression level. In this study, we implemented

E-mail address: [email protected] (G. Jiao).

Co-ﬁrst authors. a system biology approach and analyzed gene expression proﬁles

Y. Ye et al. / Journal of Cardiology 62 (2013) 58–62 59

in 23 monocytes samples and 23 T cell samples from patients with To determine genes that represent biological processes and

FH compared with those from healthy controls. We hope our study pathways that were most affected in FH patients, the genes most

could aid in understanding of FH and could provide the basis for FH signiﬁcantly changed in expression level were mapped to the GO

biomarker identiﬁcation. and KEGG using the GOstats package [29] in R. The GOstats package

identiﬁed the over-represented GO categories in biological process

Materials and methods and KEGG pathway based on the hypergeometric distribution [30].

It was supposed that there were a total of N genes in all GO bio-

Affymetrix microarray data logical processes (or KEGG pathways), and n of them were DEGs,

m represented the number of DEGs in a given GO biological pro-

Two published microarray datasets (GSE6054 and GSE6088) of cess (or KEGG pathway), and then the p-value which k genes were

different leukocyte subsets from FH patients and healthy controls overlapping at least was calculated as follow:

were used in our study. These datasets were downloaded from NCBI

k−1 CmCN−m

GEO database (http://www.ncbi.nlm.nih.gov/geo/) which is based

p k n−k

= −

1 N

on Affymetrix Human Genome U133 Plus 2.0 Array. Both of the two Cn

proﬁles were from Institute of Vascular Medicine, University Hospi-

tal Juna. Gene expression proﬁles of homozygous and heterozygous

The over-represented GO term with FDR less than 0.001 were

FH from patients with documented genetic defects in the LDLR

selected as the signiﬁcant category. The pathways that fulﬁlled the

gene were compared with healthy volunteers. The transcription

criteria of FDR less than 0.01 were identiﬁed.

proﬁle of GSE6054 contains 23 monocyte samples: 4 homozygous

FH, 6 heterozygous FH, and 13 control participants. The transcrip-

Results

tion proﬁle of GSE6088 contains 23 T cell samples: 3 homozygous

FH, 7 heterozygous FH, and 13 control participants.

Differentially expressed genes analysis

For the GSE6054 and GSE6088 datasets, the t-test was used to

identify the DEGs, respectively. The FDR less than 0.05 was chosen

For the GSE6054 and GSE6088 datasets, the t-test was used to

as the cut-off criterion. Finally, a total of 1161 genes were identiﬁed

identify DEGs, respectively. We preprocessed the CEL source ﬁles by

as DEGs of GSE6054 and a total of 6131 genes were identiﬁed as

RMA (Robust Multi-Array Average) algorithm [21] with defaulted

DEGs of GSE6088. Of these, a total of 1024 overlapping genes were

parameters in R [22] Affy package. Probe sets were mapped to NCBI

used in the following analysis.

entrez genes using DAVID [23]. If there were multiple probe sets

that correspond to the same gene, the expression values of those

Dysfunctional co-expression network construction

probe sets were averaged. If one probe corresponded to multiple

genes, the probe was deleted. All of the expression values were

By repeating the permutation procedure 100 times, 4232

then converted to fold changes (FC) with log2 base. To circumvent

co-expressed gene pairs between 871 DEGs were identiﬁed as

the multi-test problem which might induce too many false positive

signiﬁcant dysfunctional pairs in FH (p < 0.001). Based on the

results, the BH method [24] was used to adjust the raw p-values into

above relationships, a co-expression network was constructed by

false discovery rate (FDR). The DEGs only with FDR less than 0.05

Cytoscape (Fig. 1).

were selected.

Functional annotation of the dysfunctional co-expression network

Co-expression dysfunctional network construction

In order to investigate the co-expression network on a more

For demonstrating the potential regulatory relationship

functional level, we used the GOstats and observed signiﬁcant

between DEGs, the Pearson correlation coefﬁcient (PCC) [25] was

enrichment of these genes in multiple GO categories (Table 1).

calculated for all pair-wise comparisons of gene-expression values

Eighty-one GO biological processes were enriched with a FDR less

in FH samples and healthy controls. In order to verify the signiﬁ-

than 0.001 and the top 20 signiﬁcant GO terms are shown in Table 1.

cance of these co-expressed relationships, we randomly permuted

The most signiﬁcant enrichment was the GO category of primary

the gene expression proﬁles 100 times and then calculated the

metabolic process with FDR = 4.62E−07. The other signiﬁcant GO

PCC of each gene pair after permutation. If the real PCC was equal

categories included cellular metabolic process (FDR = 6.20E−07)

or larger than the permuted PCC with a p-value less than 0.001,

and cellular metabolic process (FDR = 3.03E−06). In fact, all sig-

then we considered this relationship was signiﬁcant. And then, the

niﬁcant GO category clusters were related to metabolic process,

signiﬁcant co-expressed gene pairs were selected to construct a

immune response, localization, and cell communication.

co-expression dysfunctional network using Cytoscape [26].

To gain further insights into the dysregulated pathways in

FH patients, we performed pathway enrichment analysis using

Gene ontology and pathway enrichment analysis

the KEGG pathways. Fourteen KEGG pathways were enriched

with a FDR less than 0.01 (Table 2), including chronic myeloid

The gene ontology (GO) project is a collaborative effort to

leukemia (FDR = 0.0006808), neurotrophin signaling pathway

address the need for consistent descriptions of gene products in

(FDR = 0.0011128), and SNARE interactions in vesicular transport

different databases and utilizes a controlled GO vocabulary in a

(FDR = 0.0014316).

curated database [27]. GO provides three structured networks of

deﬁned terms to describe gene product attributes: biological pro-

cess, molecular function, and cellular compartment. KEGG (Kyoto Discussion

Encyclopedia of Genes and Genomes) is a collection of online

databases dealing with genomes, enzymatic pathways, and bio- The advent of DNA microarray technology, which facilitates

logical chemicals [28]. The PATHWAY database in KEGG records investigation to identify genes and biological pathways that are

networks of molecular interactions in the cells and variants of them associated with clinically deﬁned conditions [31], has provided

speciﬁc to particular organisms. a powerful tool to understand the complex pathogenesis of

60 Y. Ye et al. / Journal of Cardiology 62 (2013) 58–62

Fig. 1. The dysfunctional co-expression network in FH.

Table 1

heterogeneous diseases. However, most previous studies have

Classiﬁcation of differentially co-expressed genes between healthy controls and

focused on differential expression and have produced large sets of

familial hypercholesterolemia patients according to gene ontology (GO) terms with

a false discovery rate (FDR) < 0.001 (Top 20). genes without consideration of potential relationships among these

genes. In this study, we constructed a conserved co-expression net-

GO-ID Category FDR

work of DEGs in FH by combining DNA expression data sets from

GO:0044238 Primary metabolic process 4.62E−07

human T cells and monocytes, and eventually characterized the

GO:0044260 Cellular macromolecule metabolic process 6.20E−07

dysregulated biological processes and pathways.

GO:0044237 Cellular metabolic process 3.03E−06

GO:0006139 Nucleobase-containing compound metabolic 3.64E−06 Co-expressed genes are biologically related, grouping these

process highly connected genes by network analysis may shed light on

GO:0043170 Macromolecule metabolic process 3.75E−06

underlying functional processes in a manner complementary to

GO:0002376 Immune system process 6.32E−06

standard differential expression analyses [32]. Thus, identiﬁcation

GO:0016070 RNA metabolic process 1.39E−05

of the co-expressed genes is important to analyze FH mechanisms.

GO:0045087 Innate immune response 1.68E−05

GO:0050794 Regulation of cellular process 1.79E−05 In this study, we identiﬁed a total of 4232 co-expression relation-

GO:0090304 Nucleic acid metabolic process 2.40E−05

ships between 871 genes and veriﬁed the signiﬁcance by random

GO:0034641 Cellular nitrogen compound metabolic process 2.53E−05

permutation. These co-expression relationships were dysregulated

GO:0006955 Immune response 2.74E−05

in the progression of FH, and functional annotation of these rela-

−

GO:0006950 Response to stress 2.75E 05

GO:0006807 Nitrogen compound metabolic process 3.88E−05 tionships will shed light on the mechanisms underlying FH.

−

GO:0046907 Intracellular transport 4.11E 05 Functional annotation of the co-expression network enriched a

GO:0080134 Regulation of response to stress 4.42E−05

total of 81 biological processes and the top 20 signiﬁcant terms are

GO:0009987 Cellular process 4.46E−05

shown in Table 1. We can ﬁnd that metabolic process accounts for a

GO:0071310 Cellular response to organic substance 4.98E−05

high proportion of the enriched GO categories which suggests that

GO:0008152 Metabolic process 5.14E−05

GO:0008104 Protein localization 6.57E−05 metabolism is disordered in the progression of FH. Our result was in

Y. Ye et al. / Journal of Cardiology 62 (2013) 58–62 61

line with a previous study which suggested that FH is an autosomal

codominant inherited disorder of lipoprotein metabolism charac-

terized by high LDL cholesterol levels [33]. Many patients with FH

have mutations in the LDLR gene that encodes the LDL receptor pro-

tein, which normally removes LDL from the circulation, or ApoB,

which is the part of LDL that binds with the receptor or mutations

in other genes [34]. All of these mutations have in common resulted

in defective clearance of LDL within a complex system of lipid and

lipoprotein metabolism and regulation [8,35]. Our results suggest

that circulating monocytes and lymphocytes show differences in

lipoprotein and cholesterol metabolism.

Pathway analysis of the differentially co-expressed genes

showed that chronic myeloid leukemia, neurotrophin signaling HLADOA;

pathway, and SNARE interactions in vesicular transport belong to

the most prominently regulated signaling pathways in FH. So far,

no studies have reported these pathways to be dysregulated in

DYNC1H1; FH. However, we observed that many members of MAPK family

were differentially expressed in FH, such as MAPK1, MAP3K1, and

TUBB; MAPK9. Previous publications have shown that the MAPK pathways

are crucial for the development of atherosclerosis and are related CTSS;

by oxdized LDL in a scavenger receptor-dependent manner [36].

The pathway of SNARE interactions in vesicular transport was

also enriched in our result with a p-value of 0.0014316. SNARE

DYNC1LI2; (soluble N-ethylmaleimide-sensitive factor attachment protein

receptor) proteins are a family of membrane proteins that allow

fusion between membranes [37]. SNARE proteins play important

PIKFYVE;

roles in mediating vesicle fusion through full fusion exocytosis or

kiss-and-run fusion exocytosis. That is, the exocytosis of cellular

STX7;

transport vesicles with the cell membrane at the porosome or with

a target compartment [38,39]. In healthy individuals, cells ingest

cholesterol by endocytosis of LDL bound to the LDLR. After endo- PIK3C3;

cytosis, the LDLR uncouples from its ligand and returns to the cell

surface, while the LDL is catabolized [40]. In our study, we observed ACTB;

many genes related to endocytosis were differentially expressed

in FH, such as members of syntaxin family (STX7, STX8, SNAP23),

golgi SNAP receptor complex (GOSR2), and vesicle-associated pro- GENE MADH4;GAB2;CRKL;MAPK1;PIK3CG;PIK3CA;SMAD4;CTBP1;SMAD3;RELA;STAT5B;PTPN11 STX7;SNAP23;GOSR2;VAMP8;STX16;VTI1A;STX8 CHP;MAPK9;DLG1;MAPK1;NFAT5;PIK3CG;MAP3K8;MAPK14;PIK3CA;TEC;RELA;PAK2;CHP1 MAPK9;CASP8;NLRC4;MAPK1;CCL2;MAPK14;HSP90B1;RELA HLA-DRA; HLA-DRA;JAK2;MAPK1;MAPK14;JAK1;RELA;HLADOA;HLADMA;HLADRA;FCGR2C CD44;MAPK9;ATG5;ACTB;CRKL;MAPK1;MAPK14;RELA MADH4;MAPK9;MAPK1;PIK3CG;PIK3CA;APC;SMAD4;SMAD3;SMAD2 MAPK9;CASP8;MAPK1;TLR8;PIK3CG;MAP3K8;MAPK14;PIK3CA;LBP;RIPK1;RELA

HLADMA;TAP2;HLADRA;ATP6V1G2;FCGR2C;M6PR

tein (VTI1A, VAMP8). From our results, we hypothesized that the

transport of LDLR via the pathway of SNARE interactions in vesic-

ular transport is dysfunctional in FH. However, as the number of

subjects was limited in our study, the evidence remains incon-

clusive. Further experimental analysis is needed to conﬁrm our

hypothesis. 0.0006808 0.0014316 0.0056533 0.0063234 0.0068287 0.0074453 0.008575 0.0094447 0.0098529

Conclusion 0.01.

This study used a system biology approach to identify co-

expressed genes that are potentially involved in the development than

of FH and constructed a conserved co-expression network of DEGs.

less

reticulum 0.0058739 MAPK9;CUL1;EIF2AK2;ATXN3;DNAJA2;SEL1L;NFE2L2;UBE4B;PDIA6;PDIA4;UBE2J2;RPN1;HSP90B1;MAN1A2;RNF5;ATF6B

This approach moves beyond single gene investigation to provide

transport

a system level perspective on the potential relationships between (FDR)

pathway

pathway members of a network. Our results suggest that FH patients showed

rate

pathway

pathway 0.0011128 MAP3K1;MAPK9;PSEN1;CRKL;MAPK1;PIK3CG;GAB1;MAPK14;PIK3CA;SORT1;MAPKAPK2;YWHAZ;FRS2;RELA;PTPN11 differences in lipoprotein and cholesterol metabolism in circulating

vesicular

endoplasmic

monocytes and lymphocytes compared to healthy controls. A fur- proteolysis 0.0054165 MAP3K1;CDC27;CUL1;ITCH;ANAPC5;FBXW7;UBE3A;UBE2B;UBE4B;HERC3;DET1;CUL4B;UBE2Z;UBE2J2;UBE2H in

in signaling

signaling

leukemia ther investigation of our co-expression network may contribute to

discovery signaling

signaling

understanding the complex interacting mechanisms of transcrip-

cancer 0.0018022 MADH4;MAPK9;MAPK1;PIK3CG;PIK3CA;SMAD4;RALB;JAK1;SMAD3;RELA;SMAD2 cancer

false mediated tion factors and their regulated genes in FH. receptor

name FDR receptor

myeloid

processing

interactions

receptor

with cycle 0.0066225 ORC2L;MADH4;CDC27;CUL1;MCM2;ANAPC5;TFDP2;ORC2;CCND2;SMAD4;YWHAZ;SMAD3;SMAD2;TFDP1;PRKDC

cell

Pathway Chronic Neurotrophin SNARE Pancreatic Ubiquitin T Protein NOD-like Cell Phagosome Leishmaniasis Shigellosis Colorectal Toll-like Acknowledgements pathways

This research project was supported by the research grant

of Wuxi Science and Technology Commission (CSZ00N1203), the

research grant of the Administrative Center of Wuxi’s Hospitals and 2

signiﬁcant

the research grant of Shanghai Science and Technology Commission 05220 04120 04130 04660 04110 05210 04620 05140 04621 04145 04722 05212 04141 05131 KEGG (114119b1600). Table The

62 Y. Ye et al. / Journal of Cardiology 62 (2013) 58–62

References [20] Jansen R, Greenbaum D, Gerstein M. Relating whole-genome expression data

with protein–protein interactions. Genome Res 2002;12:37–46.

[1] Ueda M. Familial hypercholesterolemia. Mol Genet Metab 2005;86:423–6. [21] Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed

[2] Civeira F. Guidelines for the diagnosis and management of heterozygous famil- TP. Exploration, normalization, and summaries of high density oligonucleotide

ial hypercholesterolemia. Atherosclerosis 2004;173:55–68. array probe level data. Biostatistics 2003;4:249–64.

[3] Itakura H, Nakaya N, Kusunoki T, Shimizu N, Hirai S, Mochizuki S, Ishikawa [22] Team RDC. R: A language and environment for statistical computing. Vienna:

T. Long-term event monitoring study of ﬂuvastatin in Japanese patients with R Foundation for Statistical Computing; 2011.

hypercholesterolemia: efﬁcacy and incidence of cardiac and other events in [23] Huang da W, Sherman BT, Lempicki RA. Systematic and integrative anal-

elderly patients (>/=65 years old). J Cardiol 2011;57:77–88. ysis of large gene lists using DAVID bioinformatics resources. Nat Protoc

[4] Takeda N, Maemura K. Circadian clock and cardiovascular disease. J Cardiol 2009;4:44–57.

2011;57:249–56. [24] Benjamini Y, Drai D, Elmer G, Kafkaﬁ N, Golani I. Controlling the false discovery

[5] Rader DJ, Cohen J, Hobbs HH. Monogenic hypercholesterolemia: new insights rate in behavior genetics research. Behav Brain Res 2001;125:279–84.

in pathogenesis and treatment. J Clin Invest 2003;111:1795–803. [25] Derrick TR, Bates BT, Dufek JS. Evaluation of time-series data sets using

[6] Burnett JR, Ravine D, van Bockxmeer FM, Watts GF. Familial hypercholestero- the Pearson product-moment correlation coefﬁcient. Med Sci Sports Exerc

laemia: a look back, a look ahead. Med J Aust 2005;182:552–3. 1994;26:919–28.

[7] Waters DD. Exploring new indications for statins beyond atherosclerosis: suc- [26] Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N,

cesses and setbacks. J Cardiol 2010;55:155–62. Schwikowski B, Ideker T. Cytoscape: a software environment for integrated

[8] Marais AD. Familial hypercholesterolaemia. Clin Biochem Rev 2004;25:49–68. models of biomolecular interaction networks. Genome Res 2003;13:2498–504.

[9] Tolleshaug H, Goldstein JL, Schneider WJ, Brown MS. Posttranslational [27] Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolin-

processing of the LDL receptor and its genetic disruption in familial hyper- ski K, Dwight SS, Eppig JT, Harris MA, Hill DP, Issel-Tarver L, Kasarskis A, Lewis

cholesterolemia. Cell 1982;30:715–24. S, et al. Gene ontology: tool for the uniﬁcation of biology. The Gene Ontology

[10] Hobbs HH, Russell DW, Brown MS, Goldstein JL. The LDL receptor locus in famil- Consortium. Nat Genet 2000;25:25–9.

ial hypercholesterolemia: mutational analysis of a membrane protein. Annu [28] Kanehisa M. The KEGG database. Novartis Found Symp 2002;247:91–101

Rev Genet 1990;24:133–70. [discussion-3, 19–28, 244–52].

[11] Tybjaerg-Hansen A, Steffensen R, Meinertz H, Schnohr P, Nordestgaard BG. [29] Falcon S, Gentleman R. Using GOstats to test gene lists for GO term association.

Association of mutations in the apolipoprotein B gene with hypercholes- Bioinformatics 2007;23:257–8.

terolemia and the risk of ischemic heart disease. N Engl J Med 1998;338: [30] Storey JD, Tibshirani R. Statistical signiﬁcance for genomewide studies. Proc

1577–84. Natl Acad Sci U S A 2003;100:9440–5.

[12] Pisciotta L, Priore Oliva C, Cefalu AB, Noto D, Bellocchio A, Fresa R, Cantafora A, [31] van Baarsen LG, Bos CL, van der Pouw Kraan TC, Verweij CL. Transcription

Patel D, Averna M, Tarugi P, Calandra S, Bertolini S. Additive effect of mutations proﬁling of rheumatic diseases. Arthritis Res Ther 2009;11:207.

in LDLR and PCSK9 genes on the phenotype of familial hypercholesterolemia. [32] Wang L, Tang H, Thayanithy V, Subramanian S, Oberg AL, Cunningham JM, Cer-

Atherosclerosis 2006;186:433–40. han JR, Steer CJ, Thibodeau SN. Gene networks and microRNAs implicated in

[13] Abifadel M, Varret M, Rabes JP, Allard D, Ouguerram K, Devillers M, Cruaud aggressive prostate cancer. Cancer Res 2009;69:9490–7.

C, Benjannet S, Wickham L, Erlich D, Derré A, Villéger L, Farnier M, Beucler I, [33] Smilde TJ, van Wissen S, Wollersheim H, Kastelein JJ, Stalenhoef AF. Genetic and

Bruckert E, et al. Mutations in PCSK9 cause autosomal dominant hypercholes- metabolic factors predicting risk of cardiovascular disease in familial hypercho-

terolemia. Nat Genet 2003;34:154–6. lesterolemia. Neth J Med 2001;59:184–95.

[14] Norman D, Sun XM, Bourbon M, Knight BL, Naoumova RP, Soutar AK. Charac- [34] Soutar AK, Naoumova RP. Mechanisms of disease: genetic causes of familial

terization of a novel cellular defect in patients with phenotypic homozygous hypercholesterolemia. Nat Clin Pract Cardiovasc Med 2007;4:214–25.

familial hypercholesterolemia. J Clin Invest 1999;104:619–28. [35] Thompson GR, Spinks T, Ranicar A, Myant NB. Non-stedy-state studies of low-

[15] Pullinger CR, Eng C, Salen G, Shefer S, Batta AK, Erickson SK, Verhagen A, Rivera density-liproprotein turnover in familial hypercholesterolaemia. Clin Sci Mol

CR, Mulvihill SJ, Malloy MJ, Kane JP. Human cholesterol 7alpha-hydroxylase Med 1977;52:361–9.

(CYP7A1) deﬁciency has a hypercholesterolemic phenotype. J Clin Invest [36] Lei ZB, Zhang Z, Jing Q, Qin YW, Pei G, Cao BZ, Li XY. OxLDL upregulates

2002;110:109–17. CXCR2 expression in monocytes via scavenger receptors and activa-

[16] Muller PY, Miserez AR. Identiﬁcation of mutations in the gene encoding sterol tion of p38 mitogen-activated protein kinase. Cardiovasc Res 2002;53:

regulatory element binding protein (SREBP)-2 in hypercholesterolaemic sub- 524–32.

jects. J Med Genet 2002;39:271–5. [37] Cosson P, Ravazzola M, Varlamov O, Sollner TH, Di Liberto M, Volchuk A, Roth-

[17] Durst R, Jansen A, Erez G, Bravdo R, Butbul E, Ben Avi L, Shpitzen S, Lotan man JE, Orci L. Dynamic transport of SNARE proteins in the Golgi apparatus.

C, Leitersdorf E, Defesche J, Friedlander Y, Meiner V, Miserez AR. The dis- Proc Natl Acad Sci U S A 2005;102:14647–52.

crete and combined effect of SREBP-2 and SCAP isoforms in the control of [38] Lynch KL, Gerona RR, Kielar DM, Martens S, McMahon HT, Martin TF.

plasma lipids among familial hypercholesterolaemia patients. Atherosclerosis Synaptotagmin-1 utilizes membrane bending and SNARE binding to drive

2006;189:443–50. fusion pore expansion. Mol Biol Cell 2008;19:5093–103.

[18] Alonso R, Defesche JC, Tejedor D, Castillo S, Stef M, Mata N, Gomez-Enterria P, [39] Chen YA, Scheller RH. SNARE-mediated membrane fusion. Nat Rev Mol Cell Biol

Martinez-Faedo C, Forga L, Mata P. Genetic diagnosis of familial hypercholes- 2001;2:98–106.

terolemia using a DNA-array based platform. Clin Biochem 2009;42:899–903. [40] Mosig S, Rennert K, Buttner P, Krause S, Lutjohann D, Souﬁ M, Heller R, Funke

[19] Lee HK, Hsu AK, Sajdak J, Qin J, Pavlidis P. Coexpression analysis of human genes H. Monocytes of patients with familial hypercholesterolemia show alterations

across many microarray data sets. Genome Res 2004;14:1085–94. in cholesterol metabolism. BMC Med Genomics 2008;1:60.