差异基因 GO分析

Analysis of differentially expressed genes in placental tissues of preeclampsia patients using microarray combined with
the Connectivity Map database
Y.Song a,J.Liu a,*,S.Huang a,L.Zhang b
a Department of Obstetrics&Gynecology,Peking Union Medical College Hospital,Chinese Academy of Medical Science,Peking Union Medical College, Beijing,PR China
b BioChain(Beijing)Science&Technology Inc.,No.7A,Yongchang North Rd,Business Development Area,Beijing,PR China
a r t i c l e i n f o
Article history:
Accepted19September2013
Keywords:
Preeclampsia
Microarray
Expression profiles Connectivity Map database a b s t r a c t
Preeclampsia(PE),which affects2e7%of human pregnancies,causes significant maternal and neonatal morbidity and mortality.To better understand the pathophysiology of PE,the gene expression profiles of placental tissue from5controls and5PE patients were assessed using microarray.A total of224tran-scripts were significantly differentially expressed(>2-fold change and q value<0.05,SAM software). Gene Ontology(GO)enrichment analysis indicated that genes involved in hypoxia and oxidative and reductive processes were significantly changed.Three differentially expressed genes(DEGs)involved in these biological processes were further verified by quantitative real-time PCR.Finally,the potential therapeutic agents for PE were explored via the Connectivity Map database.In conclusion,the data obtained in this study might provide clues to better understand the pathophysiology of PE and to identify potential therapeutic agents for PE patients.
Ó2013Elsevier Ltd.All rights reserved.
1.Introduction
One of the leading causes of mortality and morbidity amongst pregnant women and their offspring is PE,which affects2e7%of all pregnancies,depending on the diagnostic criteria and the patients’ethnicity[1].PE is generally defined as new hyperten-sion and substantial proteinuria at or after20weeks gestation.It is generally accepted that PE involves two pathophysiological stages.Thefirst stage occurs when impaired trophoblasts invade the myometrial arteries,leading to reduced uteroplacental arte-rialflow and episodes of irregular placental perfusion,followed by ischemic hypoxia and oxidative stress.The second stage of PE is thought to be the maternal response to abnormal placentation. An intrinsic failure in trophoblast differentiation at different time points of ontogeny may lead to either a mild disorder with late-onset appearance or to intrauterine growth retardation,poten-tially complicated with maternal symptoms[2].Systemic endothelial dysfunction appears to be an important common factor[3,4].
Despite breakthroughs in the understanding of the patho-genesis of PE,the mechanisms that ultimately trigger the disease are still not clearly elucidated[5].It seems clear that the development of PE requires the placenta,given that PE can occur in molar pregnancies wherein a placenta,but no fetus,is present.Furthermore,the only effective treatment of PE is de-livery of the placenta[6,7].Previous studies also showed that PE has a strong familial predisposition,and it has been concluded that a hereditary element may exist.Therefore,it is clinically useful to compare the gene expression profiles of placentas with and without PE to better understand the pathophysiology of this disease and to aid in the development of new therapies [8,9].
Microarray technology is a powerful tool for simultaneously assessing the expression of a high number of genes and to pinpoint the patterns of gene expression associated with a disease[10].To date,a number of microarray studies have been conducted on the global gene expression profiles of the PE placenta,and several important genes have been identified using this strategy[11e18]. However,a comparison of different studies performed using DNA microarrays has shown poor overlap,and therefore,only a minority of DEGs are shared between these studies.The resulting differences may be due to the diverse genetic backgrounds of the study pop-ulation,the use of different platforms and protocols,and the impact of individual differences on expression signatures[16].Therefore, to better understand the events leading to PE,it is necessary to
*Corresponding author.
E-mail address:juntao_liu@(J.
Liu).Contents lists available at ScienceDirect
Placenta
journal homepage:www.elsev /locat e/placen
ta
0143-4004/$e see front matterÓ2013Elsevier Ltd.All rights reserved.
/10.1016/j.placenta.2013.09.013
Placenta34(2013)1190e1195
continue the search for genes involved in the pathophysiology of the disease.
The Connectivity Map(CMAP)is a collection of genome-wide transcriptional expression data collected from cultured human cells treated with bioactive small molecules.Its purpose is to identify associations between gene expression profiles due to the disease state and drug treatment state.The biological connection between disease and drug treatment states can be established by comparing the genome-wide expression profiling of the two states. Despite the fact that CAMP was established on cultured cell lines due to the ease of such an operation,wherein DEGs were identified by comparing the microarray hybridization signal between drug-treated cell lines and controls,CAMP was also found to be an effi-cacious method to query the DEGs originating from disease tissues vs.control tissues.It is believed that the strategy of using ratio values can diminish any sample differences.For example,if the gene-expression signature of a small molecular compound is found to be the opposite of that characterizing a disease state,then the compound might be used as a drug to treat that disease.CMAP has previously been used to discover mechanisms of drug action and disease pathogenesis,as well as to identify new potential thera-peutics[19e22].
To explore novel genes involved in the development of PE in the Chinese Han population,as well as to seek potential therapeutic agents for PE,we generated a global gene expression profile via DNA oligonucleotide microarrays for placental tissues from PE and control pregnancy tissues.Moreover,the resultant PE gene expression signature was used in silico by employing CMAP analysis to identify therapeutic agents that could potentially be effective against this disease.
2.Materials and methods
2.1.Patient clinical definitions
This case-control study was approved by the local ethics committee.Placental samples were collected from women whose pregnancies were complicated by se-vere preeclampsia(n¼5)and from women without PE(n¼5)at Peking Union Medical College Hospital.Given that gene expression in placental tissues may be influenced by gestational age and uterine contractions,we chose5preterm controls complicated with premature rupture of membranes(PROM).Allfive control cases were in good condition before PROM.After fetal lung maturating,C-sections were performed due to malpresentation(n¼4)and previous C-section(n¼1).None of the mothers were in plete blood count and C-reactive protein monitoring showed no signs of infection prior to C-section.Bacterial cultures of uterine cavity swabs during operation were negative.Severe preeclampsia was defined as blood pressure!160/100mmHg,measured at least6h apart,in combination with pro-teinuria!3g/24h or!2þ(dipstick),developing after20weeks of pregnancy. Women who had a history of cardiovascular,renal,and other hypertension-associated diseases were excluded;however,family history of PE was not recor-ded.C-sections were performed in all preeclampsia cases due to medical indications. All patients provided informed consent for the study.Patient data are summarized in Table1.2.2.Sample collection
Placental biopsies were obtained immediately after deliveries.Samples of approximately0.2e0.5cm3were excised from central cotyledons close to the um-bilical cord to reduce the possible bias due to regional differences in gene expres-sion.After extensive washing in saline solution to remove excess blood,the samples were stored in RNAlater reagent(Qiagen,Valencia,CA,USA).
2.3.RNA extraction from tissue samples
Total RNA was extracted from placental tissues using Trizol reagent(Life Tech-nologies,Gaithersburg,MD,USA).RNA quality was evaluated using formaldehyde agarose gel electrophoresis and was quantified via spectrophotometry(NanoDrop, Wilmington,DE,USA).
2.4.Microarray analysis
RNA was amplified and labeled according to a previous protocol.Briefly,total RNA was used to synthesize the double stranded cDNA.RNA was amplified by in vitro transcription using Ambion’s MessageAmpÔII aRNA Amplification Kits(Life Tech-nologies,Austin,TX,USA).Then,aRNA was reverse transcribed into cDNA and further labeled with Klenow enzyme.cDNA from PE placental tissue was labeled with Cy5-dCTP,and cDNA from control placental tissue was labeled with Cy3-dCTP. Fluorescent dye-labeled cDNA was hybridized to an Agilent SurePrint G3Human GE 8Â60K Microarray(Agilent Technologies,Santa Clara,CA,USA).Hybridization, scanning and washing were performed on Agilent’s Microarray Platform according to Agilent’s standard protocols.The array data were analyzed with Agilent Feature Extraction software.Although the dual-color microarray method was adopted here, which can save microarray slides,we extracted the single channel intensity instead of the ratio value from each dual-color slide for DEGs analysis.The strategy to extract hybridization signals from dual-color microarray for DEGs analysis has been re-ported in a few previous studies[23,24].
After global mean normalization,probes with an intensity<400werefiltered out for further analysis.SAM(Significance analysis of microarrays)software[25]was used to identify DEGs with a threshold set at a fold change>2and a q value<0.05. DEGs were further analyzed with function enrichment of the GO terms using R language package software.
2.5.Validation of differentially expressed genes
To validate the results of the microarray,ten DEGs involved in oxidative stress were selected for analysis by quantitative real-time PCR(qRT-PCR).Total RNA used in the microarray analysis was treated with DNase,and1m g of DNase-treated RNA was reverse transcribed to cDNA with oligo(dT)15using M-MLV reverse tran-scriptase(Life Technologies)in a total volume of20m L.After the reverse tran-scription reaction,1m L of the reaction mixture was used for a qPCR program of45 cycles consisting of melting(30s at94 C),annealing(30s at58 C)and extension (30s at72 C).The20m L reaction mixture contained1ÂPCR Buffer(Mg2þPlus), 0.5m M forward primer0.5m M reverse primer,200m M of each dNTP,and Eva Green Master Mix in a LightCyclerÒ480Real-Time PCR System(Roche Applied Science). The primers in this assay were designed with the primer software version5.0and
Table1
Maternal and fetal clinical characteristics.
Characteristic Controls(n¼5)Preeclampsia
patients(n¼5)
P value
Maternal age,years30.6Æ3.626.6Æ5.40.208 Gestational age,weeks32.0Æ1.931.7Æ3.10.854 Prepregnancy body mass index22.6Æ2.124.2Æ3.20.760 Body mass index at delivery27.8Æ2.829.9Æ3.30.344 Systolic blood pressure,mmHg121Æ8168Æ11<0.001 Diastolic blood pressure,mmHg68Æ9113Æ6<0.001 24h Proteinuria(g)Not tested 6.5Æ3.2e Proteinuria(dipstick)Negativeþþþ<0.001 Multipara22 1.000 Nullipara33 1.000 Fetal sex:male/female5/04/1 1.000 Infant birth weight(g)1886Æ5811482Æ7450.367 Small for gestational age030.167Table2
Primers used for quantitative real-time RT-PCR.
Gene symbol Sequences(50e30)Product sizes(bp)
FLT1F GTCAGTCCAAGAAGTGACACCG150
R GTTACACCACTGTCGGCCAA
LEP F CTGCTCTGGAAAATGTGACCC248
R CGCCATCTAGTGACCTTGTGAT
CXCL12F TCCAAATCCCCTAAGCAGACC177
R CCTTCTCCTGGACCATTTTCAC
SLC8A1F AAAGCTAGCCTAGAAGCACCAA163
R TGAAATCGACCAATGCAAGAA
TFRC F TGGTTCGGGTGTTACGCA181
R CAGGAATATTGAGCCTGTTAGCA
CYBB F CACTTAGGTCCAGCCTGTTCAC177
R TAGAAAGGGCCAATATTCTCAGA
SH3PXD2A F CCAGAATCAATGGTGGTGCTAA168
R CCAAAGAGTCCCGTCAAAGTG
CD36F AAACGGCTGCAGGTCAACC186
R CATTTCTATCAGGCCAAGGAGG
CYP11A1F TCAACCTCATTCTGATGCCTG161
R GGGACAGACGACTGAAGATGC
SEPP1F ACTGAAAGGTGATTGCAGCTTT151
R TCTTTGTTGTTCTTCCTCCATTC
b-actin F CCATCGTCCACCGCAAAT194
R GCTGTCACCTTCACCGTTC
F:forward primer;R:reverse primer.
Y.Song et al./Placenta34(2013)1190e11951191
were synthesized by Invitrogen (Invitrogen,Beijing,China),as shown in Table 2.Data were analyzed by the 2ÀDD Ct method.b -actin was used as a reference gene.All other results are shown as fold-change relative to the b -actin parison of gene expression pro files with the CMAP database
The DEGs of PE and control placental tissues were used to query the CMAP database (build 02),which contains more than 7000expression pro files representing 1309compounds [26].The similarity between the gene expression pro file of the query signature and that of a CMAP instance is measured by the connectivity score,which ranges from À1to 1.A high positive connectivity score indicates that the corre-sponding drug induces the expression of the query signature.A high negative con-nectivity score indicates that the corresponding drug reverses the expression of the query signature.Prior to the query in the CMAP database via the Internet (/cmap ),the probe ID of Agilent SurePrint G3Human GE 8Â60K Microarray was transformed into the probe ID de fined by the AffymetrixGeneChip Human Genome U133A array according to the probe ’s corresponding gene.
3.Results 3.1.DEGs in PE
We compared the global gene expression pro files of placental tissues from the PEs and controls using oligonucleotide micro-arrays.The original microarray data were submitted to GEO with an accession number of GSE47187.After normalization,DEGs were assessed using one class analysis in the SAM software,with the signi ficance threshold set to fold changes greater than 2.0and a q value less than 0.05.A total of 224transcripts were identi fied as
being signi ficantly differentially expressed (a full list of DEGs is presented in the Supplementary Material ).The heat map of DEGs is shown in Fig.1.Of these DEGs,91genes were up-regulated,and 133genes were down-regulated in PE placentas compared to control placentas.Speci fically,LEPTIN,FLT1,CRH,INHBA,INHA,BCL6,and pappalysin 2were signi ficantly up-regulated in the PE group,which is consistent with their previously described association with this disease.After function enrichment analysis,the top three signi fi-cantly functional group genes that explain the molecular mecha-nisms of PE were found to be involved in immune response,in flammatory response and chemotaxis.We also found that genes involved in oxidative stress were signi ficantly changed.These genes are listed in Table 3.
3.2.Validation of DEGs using qRT-PCR
To further validate the microarray results,especially the func-tional genes involved in oxidative stress as identi fied by microarray,ten of the DEGs related to oxidative stress were chosen for further validation using quantitative RT-PCR.The ratio of the average expression levels of each gene between the PE placentas and control placental tissues was measured using the same batch of samples used in the microarray.As shown in Fig.2,fold changes measured by microarray were closely correlated with those measured by quanti-tative RT-PCR,which con firmed that large differences in gene expression are present between the PE and control parison of gene expression pro files with the CMAP database
To better understand the pathogenesis of PE and to identify potential therapeutic agents,the DEGs were used to compare the expression signatures of 1309compounds in the CMAP
database.
Fig.1.Heat map of the 224genes that were signi ficantly increased and decreased in five preeclamptic placentas,as compared with five control placentas (fold change >2and q <0.05in the SAM software).The red color indicates up-regulation and green indicates down-regulation.
Table 3
List of altered genes related with oxidative stresses.Gene symbol Fold change GO molecular function FLT1
8.6
GO:0001666(response to hypoxia);GO:0008284
(positive regulation of cell proliferation);GO:0045766(positive regulation of angiogenesis)
LEP 124.5
GO:0001666(response to hypoxia);GO:0001819
(positive regulation of cytokine production);GO:0008284(positive regulation of cell proliferation)
CXCL120.27
GO:0001666(response to hypoxia);GO:0006955
(immune response);GO:0006874(cellular calcium ion homeostasis);GO:0008284(positive regulation of cell proliferation)
SLC8A10.49
GO:0001666(response to hypoxia);GO:0006874(cellular calcium ion homeostasis);GO:0042542(response to hydrogen peroxide)
TFRC 0.39GO:0001666(response to hypoxia);GO:0010035(response to inorganic substance)
CYBB
0.36
GO:0006954(in flammatory response);GO:0006801(superoxide metabolic process);GO:0022900(electron transport chain)
SH3PXD2A 2.5GO:0006801(superoxide metabolic process);GO:0007154(cell communication)
CD36
0.46
GO:2000121(regulation of removal of superoxide radicals);GO:0007263(nitric oxide mediated signal transduction);GO:2000379(positive regulation of reactive oxygen species metabolic process)CYP11A1 2.7GO:0042542(response to hydrogen peroxide);GO:0042493(response to drug)
SLC8A1
0.49
GO:0001666(response to hypoxia);GO:0006874(cellular calcium ion homeostasis);GO:0042542(response to hydrogen peroxide)
SEPP10.34
GO:0006979(response to oxidative stress)
GO:gene ontology.
Y.Song et al./Placenta 34(2013)1190e 1195
1192
The similarity of the gene expression pro file between the query signature and that of a CMAP instance was measured by the con-nectivity score,which ranges from À1to 1.A connection score of c ¼1signi fies that the gene signature has the maximum positive connection strength with the reference pro file,which indicates that the experimental condition that gives rise to this gene signature had the strongest possible correlation with the reference pro file.A connection score of c ¼À1indicates that the two experimental permutations were most inversely correlated.Because we intended to screen drugs that can reverse the disease status,we identi fied five compounds,cinchonine,hesperidin,probenecid,nifuroxazide and primidone,with an average connectivity scores less than À0.7(listed in Table 4),indicating their potential molecular signatures that may counteract that of PE.4.Discussion
The present study represents a comprehensive analysis of the expression pro files of placental tissues from five patients with PE and five healthy controls,and the findings provide further valida-tion of signi ficant DEGs using quantitative real-time RT-PCR.We identi fied 91genes that were up-regulated and 131genes that were down-regulated in PE placentas,compared to control placentas.GO enrichment analysis revealed that these genes were involved in regulating the immune response,in flammatory response,chemo-taxis,and hypoxia response,among other actions.These findings not only add to a growing body of literature demonstrating dys-regulated genes and biological pathways in PE pregnancies,but it also provides the basis for the identi fication of new biomarkers in PE that are worthy of future in-depth studies to elucidate their roles in PE.
Microarray technology has been widely used for identifying genes responsible for PE.These studies identi fied many genes associated with PE;however,the results of the different studies have shown varying degrees of overlap,which highlights the heterogeneity of PE.Potential explanations for the discrepancies include differences in maternal ethnicities,as well as variations in platforms and protocols [16].In addition,the placental sam-ples from the previous studies are mainly derived from Caucasian subjects [27].As the rates of PE are lower among Asian women compared to Caucasian women,it would be worth repeating the expression pro file analysis with placentas from Chinese women.As expected,our studies identi fied many of the same genes as highlighted by previous studies,such as LEPTIN,FLT1,CRH,INHBA,INHA,BCL6,and pappalysin 2.However,the CXCL12and CXCL9
genes that were reported previously to be up-regulated in PE [28,29]were identi fied as down-regulated genes in our present analysis.This discrepancy may be due to the complexity of the placenta.The placental samples consisted mainly of trophoblasts,as well as blood vessels,macrophages and fibroblasts.The dif-ferences in CXCL12expression between the present study and other experiments may re flect the different tissue components involved.In addition,the DEGs in PE placentas,which are known to be affected by oxidative stress,were also identi fied.These genes included SLC8A1,TFRC,CYBB,SH3PXD2A,CD36,CYP11A1,and SEPP1.Our results are consistent with previous studies and may highlight an important role for oxidative stress in the pathogenesis of PE [30].
The pathogenesis of PE,which represents a serious pregnancy complication,is not fully understood;however,it is obvious that oxidative stress and in flammation are important mechanisms in the development of this disease [6].Therefore,antioxidant sup-plementation or anti-in flammatory agents may be potential treatments for prevention or amelioration of PE.One small trial has evaluated high doses of vitamins C and E as anti-oxidant agents for prevention of PE [31].Unfortunately,the bene fits of these treatments were not con firmed in several subsequent studies [32,33].Therefore,new methods of preventing PE are still needed.
Increased uric acid level is also a key clinical feature of PE.Higher levels of uric acid are positively correlated with signi ficant maternal and fetal morbidity and mortality [34,35].Because high levels of uric acid can promote in flammation,oxidative stress and endothelial dysfunction,the hypothesis that hyperuricemia con-tributes to vascular damage in PE is generally accepted [34],despite the fact that its speci fic role in the pathogenesis of PE remains unclear.
In this study,we employed CMAP analysis to identify potential therapeutic agents for PE.These agents would reverse the gene expression signature of PE placental tissue to control levels.The top five compounds that demonstrated opposite correlations with DEGs in PE vs.control placentas include cinchonine,hesper-idin,probenecid,nifuroxazide and primidone.Hesperidin is one
of
parison of gene expression changes obtained with microarray and quan-titative real time RT-PCR.
Table 4
The top five agents showing highly negative correlations with PE in the CMAP database.Rank Agent P value a Score
Speci ficity b Function
1Cinchonine 0.00034À0.8890.0062It has been used for treatment of malaria 2
Hesperidin
0.00064À0.8630
It acts as an antioxidant and also can reduced cholesterol and blood pressure
3Probenecid 0.00131À0.8350
It increases uric acid excretion in the urine and was primarily used in treating gout and hyperuricemia.
4Nifuroxazide 0.00328À0.7980.0213It was used to treat colitis and diarrhea 5
Primidone
0.00559À0.7710
It is a mainstay anticonvulsant in
the treatment of partial and generalized seizures
a
The two tailed p value was calculated between a random gene signature and the speci fic gene signature in CMAP database,which indicated the proportion of random connection scores that are greater than the observed score [45].b
Speci ficity is an estimate of the uniqueness of the connectivity between the interrogated gene signature and a speci fic gene signature in CMAP.High speci ficity scores indicate that the extent of connectivity found between the interrogated in-stances and the interested signature in CMAP is unexceptional and/or the investi-gated agents have multiple biological effects.
Y.Song et al./Placenta 34(2013)1190e 11951193
the most abundant naturalflavonoids[36]and has anti-inflammatory and analgesic activities[37].Some reports indicate that hesperidin could also prevent oxidant injury by several mechanisms,such as scavenging for oxygen radicals,protecting against lipid peroxidation and chelating metal ions[38].Currently, it remains unclear whether vitamins C and E can prevent PE.One matter of debate is the optimal antioxidant dose.For daily sup-plements,40mg/day,45mg/day and60e95mg/day vitamin C are recommended by UK Food Standards Agency,World Health Orga-nization and US National Academy of Science,respectively.How-ever,some studies demonstrated that the lower doses of vitamin C were not confirmed as effective antioxidants,whereas the higher doses are effective.Levine and colleagues reported an improve-ment in endothelium-dependent vasodilation after administration of2000mg/day ascorbic acid[39],but kinetic studies show that plasma saturation occurs at1000mg/day[40].Polidori and col-leagues also reported that plasma and body saturation with 1000mg/day vitamin C in humans appears desirable to maximize antioxidant protection and lower risk of oxidative damage[41]. Thesefindings suggest that the sole intake of a higher dose of vitamin C above1000mg/day does not increase the ascorbic acid concentration in plasma.Interestingly,a previous study has shown that certainflavonoids increase the biological value of ascorbic acid [42].Another study also showed that a daily supplement consisting of a commercial bioflavonoid mixture,at a dose of200mg three times a day for a month,resulted in a significant increase in the ratio of reduced to oxidized ascorbic acid in pregnant women compared to controls,who received lactose placebo capsules[43]. One potential explanation for thisfinding is thatflavonoids can increase ascorbic acid storage.A second explanation is thatflavo-noids may reduce the absorption of heavy metal ions or increase the levels of reduced glutathione.Therefore,it is interesting to consider the possibility that a combination of vitamin C andfla-vonoids can increase the biological value of vitamin C in PE clinical management.
CMAP analysis also indicated that probenecid showed an opposite correlation with DEGs in PE vs.control tissues.Probenecid restrains the active reabsorption of urate in the renal tubular,in-creases the excretion of urate,and reduces the concentration of urate in the blood.Thus,probenecid can promote the excretion of uric acid and reduce the deposition of uric acid in serum[44].The CMAP analysis found that the two compounds that are most effective in reversing the PE phenotype were also related to inflammation,oxidation or hyperuricemia.These data indirectly imply that genes involved in oxidative stress and inflammation also play an important role in PE pathophysiology.However,given the design principles of CMAP,potential compounds were identified based on similarity comparisons of gene expression signatures. Therefore,the method is,so far,unable to identify genes that are key-points of pathogenesis or therapeutic targets.
In conclusion,we identified a DEGs signature in PE using Chinese Han placental tissues.These observations confirm and expand upon the current knowledge of the pathophysiology of PE and may provide the basis for developing novel diagnostic bio-markers of PE to improve reproductive outcomes.We also used the CMAP to identify compounds with molecular signatures that can counteract those of PE.The expression profiling,combined with the CMAP analysis,suggest possible therapeutic agents for clinical trials to develop novel prevention strategies and treat-ments for PE.
Acknowledgments
This study was supported by the National Nature Science Foundation of China(81100446).Appendix A.Supplementary data
Supplementary data related to this article can be found at http:// /10.1016/j.placenta.2013.09.013.
References
[1]Steegers EA,von Dadelszen P,Duvekot JJ,Pijnenborg ncet
2010;376(9741):631e44.
[2]Huppertz B.Placental origins of preeclampsia:challenging the current hy-
pothesis.Hypertension2008;51(4):970e5.
[3]Roberts JM,Cooper DW.Pathogenesis and genetics of ncet
2001;357(9249):53e6.
[4]Sun CJ,Zhang L,Zhang WY.Gene expression profiling of maternal blood in
early onset severe preeclampsia:identification of novel biomarkers.J Perinat Med2009;37(6):609e16.
[5]James JL,Whitley GS,Cartwright JE.Pre-eclampsia:fitting together the
placental,immune and cardiovascular pieces.J Pathol2010;221(4):363e78.
[6]Young BC,Levine RJ,Karumanchi SA.Pathogenesis of preeclampsia.Ann Rev
Pathol2010;5:173e92.
[7]Kajii T,Ohama K.Androgenetic origin of hydatidiform mole.Nature
1977;268(5621):633e4.
[8]Salonen Ros H,Lichtenstein P,Lipworth L,Cnattingius S.Genetic effects on the
liability of developing pre-eclampsia and gestational hypertension.Am J Med Genet2000;91(4):256e60.
[9]Moses EK,Fitzpatrick E,Freed KA,Dyer TD,Forrest S,Elliott K,et al.Objective
prioritization of positional candidate genes at a quantitative trait locus for pre-eclampsia on2q22.Mol Hum Reprod2006;12(8):505e12.
[10]Schadt EE,Zhang B,Zhu J.Advances in systems biology are enhancing our
understanding of disease and moving us closer to novel disease treatments.
Genetica2009;136(2):259e69.
[11]Centlow M,Wingren C,Borrebaeck C,Brownstein MJ,Hansson SR.Differential
gene expression analysis of placentas with increased vascular resistance and pre-eclampsia using whole-genome microarrays.J Pregnancy2011;2011: 472354.
[12]Han JY,Kim YS,Cho GJ,Roh GS,Kim HJ,Choi WJ,et al.Altered gene expression
of caspase-10,death receptor-3and IGFBP-3in preeclamptic placentas.Mol Cells2006;22(2):168e74.
[13]Gack S,Marme A,Marme F,Wrobel G,Vonderstrass B,Bastert G,et al.Pre-
eclampsia:increased expression of soluble ADAM12.J Mol Med(Berl) 2005;83(11):887e96.
[14]Heikkila A,Tuomisto T,Hakkinen SK,Keski-Nisula L,Heinonen S,Yla-
Herttuala S.Tumor suppressor and growth regulatory genes are overex-pressed in severe early-onset preeclampsia e an array study on case-specific human preeclamptic placental tissue.Acta Obstet Gynecol Scand 2005;84(7):679e89.
[15]Herse F,Dechend R,Harsem NK,Wallukat G,Janke J,Qadri F,et al.Dysre-
gulation of the circulating and tissue-based renin-angiotensin system in preeclampsia.Hypertension2007;49(3):604e11.
[16]Kang JH,Song H,Yoon JA,Park DY,Kim SH,Lee KJ,et al.Preeclampsia leads to
dysregulation of various signaling pathways in placenta.J Hypertens 2011;29(5):928e36.
[17]Nishizawa H,Ota S,Suzuki M,Kato T,Sekiya T,Kurahashi H,et parative
gene expression profiling of placentas from patients with severe pre-eclampsia and unexplained fetal growth restriction.Reprod Biol Endocrinol 2011;9:107.
[18]Meng T,Chen H,Sun M,Wang H,Zhao G,Wang X.Identification of differential
gene expression profiles in placentas from preeclamptic pregnancies versus normal pregnancies by DNA microarrays.OMICS2012;16(6):301e11. [19]Fayad W,Fryknas M,Brnjic S,Olofsson MH,Larsson R,Linder S.Identification
of a novel topoisomerase inhibitor effective in cells overexpressing drug efflux transporters.PLoS One2009;4(10):e7238.
[20]Gheeya J,Johansson P,Chen QR,Dexheimer T,Metaferia B,Song YK,et al.
Expression profiling identifies epoxy anthraquinone derivative as a DNA topoisomerase inhibitor.Cancer Lett2010;293(1):124e31.
[21]Hieronymus H,Lamb J,Ross KN,Peng XP,Clement C,Rodina A,et al.Gene
expression signature-based chemical genomic prediction identifies a novel class of HSP90pathway modulators.Cancer Cell2006;10(4):321e30.
[22]Slonim DK,Koide K,Johnson KL,Tantravahi U,Cowan JM,Jarrah Z,et al.
Functional genomic analysis of amnioticfluid cell-free mRNA suggests that oxidative stress is significant in Down syndrome fetuses.Proc Natl Acad Sci U S A2009;106(23):9425e9.
[23]t Hoen PA,Turk R,Boer JM,Sterrenburg E,de Menezes RX,van Ommen GJ,
et al.Intensity-based analysis of two-colour microarrays enables efficient and flexible hybridization designs.Nucleic Acids Res2004;32(4):e41.
[24]Xia Q,Cheng D,Duan J,Wang G,Cheng T,Zha X,et al.Microarray-based gene
expression profiles in multiple tissues of the domesticated silkworm,Bombyx mori.Genome Biol2007;8(8):R162.
[25]Tusher VG,Tibshirani R,Chu G.Significance analysis of microarrays applied to
the ionizing radiation response.Proc Natl Acad Sci U S A2001;98(9):5116e21.
[26]Lamb J,Crawford ED,Peck D,Modell JW,Blat IC,Wrobel MJ,et al.The con-
nectivity map:using gene-expression signatures to connect small molecules, genes,and disease.Science2006;313(5795):1929e35.
Y.Song et al./Placenta34(2013)1190e1195 1194。