点击显示 收起
From the Department of Medicine (F.M., R.E.P., V.J.D., C.-C.L.), Brigham and Women’s Hospital and Harvard Medical School, Boston, Mass; the Department of Medicine (T.W.A.d.B., C.J.H.v.d.K.), Cardiovascular Research Institute Maastricht, Maastricht, the Netherlands; and the Medical Genetics Institute (J.I.R., G.A.H.) and the Department of Medicine and Obstetrics/Gynecology (Y.-D.I.C.), Cedars-Sinai Research Institute, Los Angeles, Calif.
ABSTRACT
Objectives— The genetic background of familial combined hyperlipidemia (FCHL) is currently unclear. We propose transcriptional profiling as a complementary tool for its understanding. Two hypotheses were tested: the existence of a disease-specific modification of gene expression in FCHL and the detectability of such a transcriptional profile in blood derived cell lines.
Methods and Results— We established lymphoblastic cell lines from FCHL patients and controls. The cells were cultured in fixed conditions and their basal expression profile was compared using microarrays; 166 genes were differentially expressed in FCHL-derived cell lines compared with controls, with enrichment in metabolism-related genes. Of note was the upregulation of EGR-1, previously found to be upregulated in the adipose tissue of FCHL patients, the upregulation of DCHR-7, the downregulation of LYPLA2, and the differential expression of several genes previously unrelated to FCHL. A cluster of potential EGR-1–regulated transcripts was also differentially expressed in FCHL cells.
Conclusion— Our data indicate that in FCHL, a disease-specific transcription profile is detectable in immortalized cell lines easily obtained from peripheral blood and provide complementary information to classical genetic approaches to FCHL and/or the metabolic syndrome.
We studied the transcriptional profile of FCHL-derived lymphoblast cell lines using microarrays. Of note was the differential expression of EGR-1, several lipid metabolism-related genes, and many novel genes.
Key Words: dyslipidemia ? blood cells ? microarray ? genomics ? genetics
Introduction
Familial combined hyperlipidemia (FCHL) is the most common form of genetically determined dyslipidemia. FCHL is characterized by familial aggregation of multiple dyslipidemic profiles in association with early-onset coronary artery disease.1–4 Consistent findings in FCHL patients are increased apolipoprotein B (apoB) and reduced high-density lipoprotein (HDL) cholesterol.5 Furthermore, FCHL patients frequently have features of the insulin resistance syndrome such as abdominal obesity, impaired glucose tolerance, and hyperinsulinemia.6
Although FCHL was first identified 30 years ago,7 its genetic background is still to be fully elucidated. Gene expression profiling represents a powerful tool for the identification of genes involved in complex diseases. This approach has already been used in several conditions, including diabetes mellitus,8 heart failure,9 and cancer,10 as well as in Tangier disease, a monogenic dyslipidemia.11 With our study, we tested 2 distinct hypotheses. First, the existence of a primary genetically driven modification of gene expression in FCHL. Second, the ability to detect such a transcriptional pattern in immortalized cell lines obtained from peripheral blood.
Methods
Subjects
The study protocol was approved by the Review Committees of the University Hospital of Maastricht and all subjects gave their informed consent. FCHL patients and controls were recruited through the Lipid Clinic of the Maastricht University Hospital. FCHL patients met the criteria described previously.12 Normal subjects were recruited in the spouse and normolipidemic relative group, representing an environment-, nutrition-, and age-matched control group. The controls were also matched for sex and body mass index with the FCHL patients. The anthropometric and metabolic characteristics of the individuals at the time of B-cell isolation are summarized in Table 1.
TABLE 1. Clinical Data of the 24 Individuals at the Time of Lymphocyte Isolation (Average±SD)
Cell Lines, RNA Isolation, Microarrays
Human lymphoblast cell lines were derived from peripheral blood lymphocytes and immortalized with Epstein–Barr virus as previously described.13 Cells were grown in suspension in T flasks in RPMI-1640 medium (GibcoBRL) containing 2 mmol/L L-glutamine, 100 μg/mL streptomycin, and 10% fetal calf serum. Total RNA was extracted from cultured cells using Trizol Reagent (Life Technologies) and further purified with affinity resin columns (Qiagen).
The microarray platform used was Genechip HG-U133A (Affymetrix), an oligonucleotide array comprising >22 000 oligonucleotide probe sets. Briefly, hybridization and scanning were performed according to the manufacturer’s protocol. A high-stringency multistep procedure was used for data analysis (GeneSpring Software ver. 6.1; SI Genetics and SAM, Stanford University Labs14). For additional information, please see the supplementary data (available at http://atvb.ahajournals.org). Immortalization, cell culture, and microarray experiments were performed at Cedars-Sinai Research Institute.
Quantitative Reverse-Transcription Polymerase Chain Reaction
Quantitative RT-PCR (qRT-PCR) was performed as previously described9 on RNA samples newly extracted from 2 separate culture sets of the 24 cell lines. Standard curves were prepared for both the target and the endogenous control, and the expression level of each gene was expressed in arbitrary units relative to the expression level of GAPDH.
Results
Multistep Data Analysis
A multistep filtering procedure was applied to the microarray data. First, we only considered for analysis the genes reliably detected in a minimum of 6 hybridizations in both groups. A total of 7647 genes passed this filter. Second, we retained for further analysis only the genes with a fold change higher than the 95th percentile or lower than the 5th percentile of the fold change distribution (Figure I, available online at http://atvb.ahajournals.org). Using this criterion, 834 genes were selected. Third, a 2-sided Student t test was used to compare the gene expression levels between the 2 groups of cell lines. This resulted in 166 genes (0.8% of array) with P<0.05, which were called as differentially expressed in FCHL versus control cells. The false-discovery rate had median and 90th percentile values of 5.8% and 7.1%, respectively.
Differentially Expressed Genes
The 166 differentially expressed transcripts (Table I, available online at http://atvb.ahajournals.org) included 152 known genes and 14 expressed sequence tags, which did not match to any currently known gene. Of these 166 genes, 92 (55%) were upregulated and 74 (45%) were downregulated. The distribution of gene ontology annotations15 for the differentially expressed genes is represented in Figure 1. The most represented category was metabolism (41%), followed by cell communication and cellular physiological process (both 13.9%) in the functional category of biological process. In the categories of molecular function, binding (39.2%), catalytic activity (27.1%), and transcription regulation (9%) were the most represented. The list of differentially expressed genes was examined for over-representation of functional classes using the EASE tool.16 Of note, the metabolism category scored a statistically significant EASE score (P=0.002), indicating over-representation of metabolism-related genes among the differentially expressed transcripts.
Figure 1. Distribution of gene ontology annotations for the differentially expressed genes in the categories of biological process (A) and molecular function (B) among the differentially expressed genes. Numbers in brackets indicate the number of genes in the corresponding categories.
Within the metabolism category, lipid and carbohydrate metabolism-related genes were differentially regulated in FCHL lymphocytes. Genes functionally involved in lipid metabolism were 7-dehydrocholesterol reductase (DHCR7), phosphatidylinositol glycan class C (PIGC), the ?-regulatory subunit A of protein phosphatase 2 (PPP2R1B), lysophospholipase 3 (LYPLA2), lipoyltransferase (LIPT1), GM2 ganglioside activator protein (GM2A), and CDP-diacylglycerol synthase 2 (CDS2). Genes related to carbohydrate metabolism were sorbitol dehydrogenase (SORD), mannosidase -class 2C-member 1 (MAN2C1), and ribose 5-phosphate isomerase A (RPIA).
A previous study17 identified 28 genes differentially expressed in the adipose tissue of FCHL individuals. Affymetrix Genechip HG-U133A contains probe sets for 15 of these genes. We thus examined the expression levels of these genes in our study. Notably, early growth response-1 (EGR-1), a transcription factor that is upregulated in the adipose tissue of FCHL subjects, was significantly upregulated also in FCHL lymphoblasts (fold change 1.49; P=0.037), a result confirmed by qRT-PCR (fold change 1.56; P=0.01). This independent finding provides evidence that the upregulation of EGR-1 may be a disease-specific feature of FCHL. Also, GADD45A (growth arrest and DNA-damage-inducible alpha) and APC (adenomatosis polyposis coli) showed a similar trend in our study, although neither reached statistical significance.
Correlation Analysis
A correlation analysis was performed between the expression levels of each differentially expressed gene and the main clinical variables. Plasma triglyceride levels were log-transformed to obtain a normal distribution. Table I reports the correlation coefficients of the transcripts showing a statistically significant correlation between mRNA levels and the lipid parameters. To increase the stringency and biological significance of our analysis, we only considered as true positives the transcripts that both gained a significant R value across the entire population of cell lines and presented concordant R values in each of the 2 groups of cell lines. Among the differentially expressed genes, 25 (15%) significantly correlated with log (triglycerides), 28 (17%) with total cholesterol, 28 (17%) with apoB, 14 (8%) with 2 of 3 of the lipid variables, and 5 (3%) with all 3 variables. None had a significant correlation with age, whereas 6 (4%) correlated with body mass index.
EGR-1–Related Genes
Given the previous report of EGR-1 upregulation in the adipose tissue of FCHL individuals, we next asked if any EGR-1 target gene was differentially regulated in FCHL-derived cell lines. Transforming growth factor (TGF)b1 is a known target of EGR-1 involved in the pathobiology of atherosclerosis.18 By qRT-PCR, TGFb-1 mRNA levels were slightly but significantly higher in FCHL cell lines than in controls (fold change 1.25; P=0.03). This suggested that an EGR-1–related cluster of genes may be differentially expressed in FCHL-derived cell lines.
We next asked whether any of the differentially regulated genes contained an EGR-1 element in their promoter region. For this purpose, the surrounding genomic sequences of the 166 genes were obtained. Subsequently, 10 kilobases of flanking sequence were queried for the presence of consensus EGR-1 elements (5'-CGCCCCCGC-3'), using the 5' end of the expressed sequence tags as an anchor. Of note, the results of this query (Table 2) indicated 16 genes that contain 1 (N=14) or more (N=2) consensus EGR-1 elements within 10 kilobases of their 5' regulatory sequences. Of the differentially regulated genes potentially regulated by EGR-1, 11 were upregulated (including EGR-1 and DHCR-7) and 5 were downregulated (including LYPLA2) in FCHL-derived cell lines. Some of these genes (EGR-1, ATP1B1) have been previously shown to be similarly regulated by EGR-1 in other cell types.19 Despite the absence of a consensus EGR-1 element in the queried 10 kb, PIGC, which was downregulated in our study, was previously reported to be similarly regulated by EGR-1.20 Another possible EGR-1 target genes is CDC2L5.20 Overall, these genes constitute a potential EGR-1–regulated cluster of transcripts differentially expressed in FCHL cells.
TABLE 2. Differentially Regulated Genes Containing EGR-1 Elements in Their Upstream Region
Cluster Analysis
A hierarchical clustering analysis of the arrays was performed using the list of 166 differentially expressed transcripts. To test the transcriptional profile also on a set of samples distinct from the training set, we introduced in the analysis other 12 arrays from control subjects. None of the second cohort of individuals was recruited from the relative or spouse group of the previous one. Unlike the control subjects from the first set, the latter were not matched to the FCHL patients for age and body mass index. There was no significant difference between the levels of total cholesterol, low-density lipoprotein cholesterol, and triglycerides between the 2 control groups.
The hierarchical clustering analysis resulted in the tree displayed in Figure 2. The transcriptional profile yielded a substantial separation of FCHL and control samples, both from the first and from the second set. This result provides evidence that, in our study, the 166 differentially expressed genes constitute a complex gene expression profile able to discriminate, with correlative algorithms, FCHL from normal cell lines. Two main clusters were obtained. A first cluster (referred to as NL1) included 19 of the 24 normal control cell lines. A second cluster (referred to as FCHL1) comprised 7 FCHL cell lines. The remaining individuals were grouped into 2 minor FCHL clusters (FCHL2 and FCHL3, including 2 and 3 cell lines, respectively) and a distinct normal control cluster (NL2, 5 cell lines).
Figure 2. Hierarchical clustering of the 36 cell lines (12 FCHL, 12 paired controls, 12 nonpaired controls) based on the differentially expressed gene list.
To detect a possible association between FCHL clusters and clinical phenotypes, we then examined the distribution of the clinical variables in the identified clusters. The average clinical values, measured at the time of blood collection for cell line generation, are shown in Table 3. FCHL1 included all the female cell lines but one, which clustered in FCHL3. However, the gender distribution in the 3 clusters did not reach statistical significance (P=0.16 with a 2 test). FCHL3 subjects shared the lowest apoB levels (P<0.01 with ANOVA) and the highest plasma triglycerides (although not gaining statistical significance, P=0.11). No significant difference in the clinical variables was present between NL1 and NL2 (data not shown).
TABLE 3. Clinical Variables of FCHL Subjects According to the Identified Gene Expression Clusters
Validation of Microarray Data
We analyzed the correlation between gene expression levels measured with qRT-PCR and microarrays. For every gene analyzed, a statistically significant correlation was found between the results achieved with qRT-PCR and with the microarray platform. We also compared the fold change values measured with qRT-PCR and the ones obtained with microarrays (Figure II, available online at http://atvb.ahajournals.org). The R value was 0.93 (P<0.0001). Overall, the results obtained with qRT-PCR paralleled those with microarrays, thus providing technical validation to the microarray-based data.
Discussion
Genetic modifications can affect the cellular pattern of gene expression in several ways, the most etiologically meaningful being the absence of a specific transcript as the direct consequence of a mutation. Differentially expressed genes may also include those functionally affected by the primitive defect, either as compensatory responses or as contributors to the disease pathophysiology. Thus, high-throughput gene expression profiling is an approach that can provide meaningful and complementary data to classical genetic analyses. Previously, the transcriptional profiling of cultured fibroblasts obtained from individuals affected by Tangier disease contributed to the identification of the genetic and biochemical background of this monogenic dyslipidemia,11 thus showing the potentials of a genomic approach in the understanding of a monogenic disease, as well as the possibility to detect complex disease-specific transcriptional abnormalities in cells that are not directly involved in the pathophysiology of the examined condition. We chose a similar strategy in our study of FCHL. Using a microarray platform, we assayed the transcriptional profile of FCHL-derived and control cells that had been obtained, grown, and harvested under the same experimental conditions. Exposing the cells to identical environment and stimuli, this experimental design allowed us to attribute the observed difference in gene expression to the underlying genetic background of the cell lines.
Our study was performed on lymphoblast cell lines obtained from peripheral blood lymphocytes. Being that FCHL is a disease that does not routinely require invasive procedures, tissues and cells from affected individuals are not easily available to investigators, with the only current exception of blood cells. EBV-immortalized lymphoblasts are cell lines easily obtained from blood samples with reproducible protocols.13 Although the immortalization process may influence gene expression, this factor is acting on every single cell line, including the controls. Thus, given the comparative nature of our analysis, a dependency of the observed gene expression profile on the transformation process should be ruled out. Despite that the particular role of blood cells in FCHL has not been investigated yet, we propose lymphoblasts as biologically meaningful cell lines for several reasons, including a gene expression profile comprising a wide variety of transcripts and the involvement of immune cells in the biology of atherosclerosis, a key clinical feature of FCHL.2–4 Clearly, no information on transcripts not expressed in lymphoblasts is provided by this study. Obvious limitations of the current approach are represented by the inability to evaluate cell-specific and wider tissue or organism-triggered responses.
Herein, we provide evidence that FCHL-derived cells present subtle but significant differences in basal gene expression compared with normal control-derived cells. Under the present experimental conditions, 166 genes were differentially expressed in FCHL-derived lymphoblasts, with a significant enrichment in metabolism related transcripts. The observed fold changes ranged under the 2-fold threshold, a result consistent with a basal and stimulus-deprived cellular state. However, even small changes in the expression of critical genes may result as biologically significant, especially if multiple gene patterns and cascade effects are involved. It is also possible that undetectable differences in gene expression in baseline conditions may burst to higher ones in the presence of specific stimuli.
A previous study examined the transcriptional profiling of subcutaneous adipose tissue of a small group of FCHL patients.17 Importantly, one of the genes upregulated in the adipose tissue of FCHL patients, the transcription factor EGR-1, is one of the most prominently upregulated in our study, as well. This finding provides external validation of our current data and suggests that lymphoblast cell lines may provide insight into disease-specific modifications of gene expression effective even in cells and tissues not directly involved in the pathophysiology of FCHL. EGR-1 is a widely expressed transcription factor21 directly implied in lipid metabolism.22 Interestingly, 2 EGR-1 binding sites are situated in the promoter of the apoAI gene.23,24 The role of the APOAI/CIII/AIV/AV gene cluster (11q23-q24) in FCHL has been recently suggested by 2 separate combined linkage and association studies, confirming original observations.25–27 EGR-1 targets include TGF?-1, plasminogen activator inhibitor-1, and fibronectin. Several studies support a role of these genes and especially of TGF? in atherosclerosis and vascular disease as mediators of inflammatory response and matrix deposition.18 In our study, TGF?-1 was in fact upregulated in FCHL-derived cells. Analyzing the flanking regions of the differentially regulated genes, we also showed that 16 genes contain EGR-1 elements in their regulatory region. Another gene (PIGC), despite lacking a consensus EGR-1 element in the queried sequence, has been previously shown to be similarly regulated by EGR-1.20 Overall, these data suggest that a cluster of EGR-1–regulated genes may be differentially regulated in FCHL cells, making EGR-1 an intriguing candidate for further studies on FCHL.
Within the metabolism category, several other lipid metabolism-related genes were differentially regulated in FCHL lymphocytes. Of note is the upregulation of DHCR7, whose gene product catalyzes the last step of cholesterol biosynthesis.28 Although the biosynthesis of cholesterol is complex and highly regulated, an increased activity of DHCR7 may contribute to excessive cholesterol production. LYPLA2 is a lysophospholipase structurally related to lecithin cholesterol acyltransferase (LCAT), the key enzyme in the reverse cholesterol transport on HDLs. LYPLA2 exists in the human plasma and has been associated with HDLs. LYPLA2 hydrolyzes lysophosphatidylcholine, a proatherogenic lipid.29,30 The observed downregulation of LYPLA2 in FCHL may be associated with a reduction in the reverse cholesterol transport and subsequent accumulation of atherogenic products in the vasculature. Overall, the transcriptional profile obtained in FCHL-derived lymphoblasts is consistent with a genetic background leading to increased cholesterol biosynthesis, reverse cholesterol transport, and accumulation of atherogenic products in the vasculature.
Of note, the majority of the differentially expressed genes are novel in the FCHL scientific literature, whereas none of the genes whose role has been previously hypothesized in the cause or pathophysiology of FCHL is included in this list. This result is not surprising, considering that most candidate genes have been proposed on the basis of well-known metabolic pathways and that no previous data regarding blood cells in FCHL are currently available. If not in the development of the dyslipidemic phenotype, these genes might be involved in the premature coronary artery disease associated with FCHL (eg, inflammation, foam cell formation, remodeling).
As shown by the cluster analysis, the set of differentially expressed genes could be used to distinguish samples obtained from FCHL patients and from normal controls, including individuals from a separate cohort. Recently, 2 distinct metabolic phenotypes of FCHL have been described in relation to small dense low-density lipoprotein and very low-density lipoprotein subclasses:31 one characterized by large, buoyant low-density lipoprotein with higher cholesterol and apoB plasma levels, and a second with predominantly small dense low-density lipoprotein showing hypertriglyceridemia and moderately elevated total cholesterol and apoB. Interestingly, in our study these patterns are observed in clusters FCHL1 and FCHL3, respectively. Albeit limited by the small number of individuals in each subgroup, our results suggest that distinct gene expression patterns may result in corresponding phenotypic clusters. However, given the limited power of our study for such a subgroup analysis, the current should be essentially regarded as preliminary data and further studies are needed to specifically address this issue.
Acknowledgments
F.M. was supported by the First School of Specialization in Internal Medicine of Turin University and by the Hypertension Unit of San Vito Hospital, Turin, Italy.
This work was supported by grants from the National Institute of Health (2P01-HL28481 and Cedars-Sinai Board of Governor’s Chair in Medical Genetics) (J.I.R.), the Microarray Core (Y.-D.I.C.), and GCRC (M01-RR00425) of the Cedars-Sinai Research Institute, by grants from the National Institutes of Health (5RO1-HL5851603, 5P5O-HL5931603, and 5RO1HL6166102) (V.J.D.), and grants from the Academic Hospital Maastricht and the Cardiovascular Research Institute Maastricht (T.B.D.). We thank all the fellows in the laboratory of V.J.D. for helpful suggestions, and the Cedars-Sinai Research Institute Microarray Core and Mr Daniel Fefer for their technical contributions.
References
Genest JJ Jr, Martin-Munley SS, McNamara JR, Ordovas JM, Jenner J, Myers RH, Silberman SR, Wilson PW, Salem DN, Schaefer EJ. Familial lipoprotein disorders in patients with premature coronary artery disease. Circulation. 1992; 85: 2025–2033.
Austin MA, McKnight B, Edwards KL, Bradley CM, McNeely MJ, Psaty BM, Brunzell JD, Motulsky AG. Cardiovascular disease mortality in familial forms of hypertriglyceridemia: a 20-year prospective study. Circulation. 2000; 101: 2777–2782.
Voors-Pette C, de Bruin TW. Excess coronary heart disease in familial combined hyperlipidemia, in relation to genetic factors and central obesity. Atherosclerosis. 2001; 157: 481–489.
Hopkins PN, Heiss G, Ellison RC, Province MA, Pankow JS, Eckfeldt JH, Hunt SC. Coronary artery disease risk in familial combined hyperlipidemia and familial hypertriglyceridemia: a case-control comparison from the National Heart, Lung, and Blood Institute Family Heart Study. Circulation. 2003; 108: 519–523.
de Graaf J, Stalenhoef AF. Defects of lipoprotein metabolism in familial combined hyperlipidaemia. Curr Opin Lipidol. 1998; 9: 189–196.
Vakkilainen J, Porkka KV, Nuotio I, Pajukanta P, Suurinkeroinen L, Ylitalo K, Viikari JS, Ehnholm C, Taskinen MR. Glucose intolerance in familial combined hyperlipidaemia. EUFAM study group. European J Clin Invest. 1998; 28: 24–32.
Goldstein JL, Schrott HG, Hazzard WR, Bierman EL, Motulsky AG. Hyperlipidemia in coronary heart disease. II. Genetic analysis of lipid levels in 176 families and delineation of a new inherited disorder, combined hyperlipidemia. J Clin Invest. 1973; 52: 1544–1568.
Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, Puigserver P, Carlsson E, Ridderstrale M, Laurila E, Houstis N, Daly MJ, Patterson N, Mesirov JP, Golub TR, Tamayo P, Spiegelman B, Lander ES, Hirschhorn JN, Altshuler D, Groop LC. PGC-1-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003; 34: 267–273.
Hwang JJ, Allen PD, Tseng GC, Lam CW, Fananapazir L, Dzau VJ, Liew CC. Microarray gene expression profiles in dilated and hypertrophic cardiomyopathic end-stage heart failure. Physiol Genom. 2002; 10: 31–44.
van de Vijver MJ, He YD, van’t Veer LJ, Dai H, Hart AA, Voskuil DW, Schreiber GJ, Peterse JL, Roberts C, Marton MJ, Parrish M, Atsma D, Witteveen A, Glas A, Delahaye L, van der Velde T, Bartelink H, Rodenhuis S, Rutgers ET, Friend SH, Bernards R. A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med. 2002; 347: 1999–2009.
Lawn RM, Wade DP, Garvin MR, Wang X, Schwartz K, Porter JG, Seilhamer JJ, Vaughan AM, Oram JF. The Tangier disease gene product ABC1 controls the cellular apolipoprotein-mediated lipid removal pathway. J Clin Invest. 1999; 104: R25–R31.
van der Kallen CJ, Cantor RM, van Greevenbroek MM, Geurts JM, Bouwman FG, Aouizerat BE, Allayee H, Buurman WA, Lusis AJ, Rotter JI, de Bruin TW. Genome scan for adiposity in Dutch dyslipidemic families reveals novel quantitative trait loci for leptin, body mass index and soluble tumor necrosis factor receptor superfamily 1A. Int J Obesity Related Metab Dis: J Int Assoc Study Obesity. 2000; 24: 1381–1391.
Pressman S, Rotter JI. Epstein-Barr Virus transformation of cryopreserved lymphocytes: prolonged experience with the technique. Am J Hum Genet. 1991; 49: 467.
Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A. 2001; 98: 5116–5121.
Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, Harris MA, Hill DP, Issel-Tarver L, Kasarskis A, Lewis S, Matese JC, Richardson JE, Ringwald M, Rubin GM, Sherlock G. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000; 25: 25–29.
Hosack DA, Dennis G Jr, Sherman BT, Lane HC, Lempicki RA. Identifying biological themes within lists of genes with EASE. Genome Biol. 2003; 4: R70.
Eurlings PM, Van Der Kallen CJ, Geurts JM, Kouwenberg P, Boeckx WD, De Bruin TW. Identification of differentially expressed genes in subcutaneous adipose tissue from subjects with familial combined hyperlipidemia. J Lipid Res. 2002; 43: 930–935.
McCaffrey TA. TGF-betas and TGF-beta receptors in atherosclerosis. Cytokine Growth Factor Rev. 2000; 11: 103–114.
Fu M, Zhu X, Zhang J, Liang J, Lin Y, Zhao L, Ehrengruber MU, Chen YE. Egr-1 target genes in human endothelial cells identified by microarray analysis. Gene. 2003; 315: 33–41.
Liao Y, Shikapwashya ON, Shteyer E, Dieckgraefe BK, Hruz PW, Rudnick DA. Delayed hepatocellular mitotic progression and impaired liver regeneration in early-growth-response-1 deficient mice. J Biol Chem. 2004.
Sukhatme VP, Cao XM, Chang LC, Tsai-Morris CH, Stamenkovich D, Ferreira PC, Cohen DR, Edwards SA, Shows TB, Curran T. A zinc finger-encoding gene coregulated with c-fos during growth and differentiation, and after cellular depolarization. Cell. 1988; 53: 37–43.
Zhang F, Ahlborn TE, Li C, Kraemer FB, Liu J. Identification of Egr1 as the oncostatin M-induced transcription activator that binds to sterol-independent regulatory element of human LDL receptor promoter. J Lipid Res. 2002; 43: 1477–1485.
Cui L, Schoene NW, Zhu L, Fanzo JC, Alshatwi A, Lei KY. Zinc depletion reduced Egr-1 and HNF-3? expression and apolipoprotein A-I promoter activity in Hep G2 cells. Am Journal of Physiology - Cell Physiology. 2002; 283: C623–C630.
Kilbourne EJ, Widom R, Harnish DC, Malik S, Karathanasis SK. Involvement of early growth response factor Egr-1 in apolipoprotein AI gene transcription. J Biol Chem. 1995; 270: 7004–7010.
Naoumova RP, Bonney SA, Eichenbaum-Voline S, Patel HN, Jones B, Jones EL, Amey J, Colilla S, Neuwirth CK, Allotey R, Seed M, Betteridge DJ, Galton DJ, Cox NJ, Bell GI, Scott J, Shoulders CC. Confirmed locus on chromosome 11p and candidate loci on 6q and 8p for the triglyceride and cholesterol traits of combined hyperlipidemia. Arterioscler Thromb Vasc Biol. 2003; 23: 2070–2077.
Dallinga-Thie GM, van Linde-Sibenius Trip M, Rotter JI, Cantor RM, Bu X, Lusis AJ, de Bruin TW. Complex genetic contribution of the Apo AI-CIII-AIV gene cluster to familial combined hyperlipidemia. Identification of different susceptibility haplotypes. J Clin Invest. 1997; 99: 953–961.
Gagnon F, Jarvik GP, Motulsky AG, Deeb SS, Brunzell JD, Wijsman EM. Evidence of linkage of HDL level variation to APOC3 in two samples with different ascertainment. Hum Genet. 2003; 113: 522–533.
Fitzky BU, Witsch-Baumgartner M, Erdel M, Lee JN, Paik YK, Glossmann H, Utermann G, Moebius FF. Mutations in the Delta7-sterol reductase gene in patients with the Smith-Lemli-Opitz syndrome. Proc Natl Acad Sci U S A. 1998; 95: 8181–8186.
Taniyama Y, Shibata S, Kita S, Horikoshi K, Fuse H, Shirafuji H, Sumino Y, Fujino M. Cloning and expression of a novel lysophospholipase which structurally resembles lecithin cholesterol acyltransferase. Biochem Biophysl Res Commun. 1999; 257: 50–56.
Hiraoka M, Abe A, Shayman JA. Cloning and characterization of a lysosomal phospholipase A2, 1-O-acylceramide synthase. J Biol Chem. 2002; 277: 10090–10099.
Georgieva AM, van Greevenbroek MM, Krauss RM, Brouwers MC, Vermeulen VM, Robertus-Teunissen MG, van der Kallen CJ, de Bruin TW. Subclasses of low-density lipoprotein and very low-density lipoprotein in familial combined hyperlipidemia: relationship to multiple lipoprotein phenotype. Arterioscler Thromb Vasc Biol. 2004; 24: 744–749.