Mol. Hum. Reprod. Advance Access originally published online on June 29, 2006
Molecular Human Reproduction 2006 12(8):505-512; doi:10.1093/molehr/gal056
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Objective prioritization of positional candidate genes at a quantitative trait locus for pre-eclampsia on 2q22
1Department of Genetics, Southwest Foundation for Biomedical Research, San Antonio, TX, USA, 2Department of Perinatal Medicine and the University of Melbourne Department of Obstetrics & Gynaecology, Royal Women’s Hospital, Carlton, 3Australian Genome Research Facility Ltd, Parkville and 4ChemGenex Pharmaceuticals, Geelong, Victoria, Australia
5 Present address: Wellcome Trust Centre for Human Genetics, Oxford OX3 7BN, UK
6 To whom correspondence should be addressed at: Department of Genetics, Southwest Foundation for Biomedical Research, San Antonio, TX 78227-5301, USA. E-mail: moses{at}darwin.sfbr.org
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Pre-eclampsia/eclampsia (PE/E) is a common, serious medical disorder of human pregnancy. Familial association of PE/E has been recognized for decades, but the genetics are complex and poorly understood. In an attempt to identify PE/E susceptibility genes, we embarked on a positional cloning strategy using 34 Australian and New Zealand PE/E pedigrees. An initial 10-cM resolution genome scan revealed a putative susceptibility locus spanning a broad region on chromosome 2 that overlaps an independently determined linkage signal seen in Icelandic PE pedigrees. Subsequent fine mapping using 25 additional short tandem repeat (STR) markers in this region and non-parametric multipoint linkage analysis did not change the overall position. Under a strict diagnosis of PE, we obtained significant evidence of linkage on 2q with a peak log-of-odds ratio score (LOD) of 3.43 near marker D2S151 at 155 cM. To prioritize positional candidate genes at the 2q locus for detailed analysis, we applied an objective prioritization strategy that integrates quantitative bioinformatics, assessment of differential gene expression and association analysis of single-nucleotide polymorphisms (SNPs). Highest priority was assigned to the activin receptor gene ACVR2. This gene also showed >10-fold differential gene expression in human decidual tissue from normotensive and PE individuals. We genotyped five known SNPs in this gene in our pedigrees and performed tests for association and linkage disequilibrium. One SNP (rs1424954) showed strong preliminary evidence of association with PE (P = 0.007), whereas two others (rs1364658 and rs1895694) exhibited nominal evidence (P < 0.05). Haplotype analysis revealed no additional association information. There was evidence of weak linkage disequilibrium among these SNPs. The highest observed LD occurred between the two strongest associated SNPs, suggesting that the observed signals may be the signature of an observed functional variant.
Key words: association/polymorphisms/pre-eclampsia/QTL/susceptibility gene
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Pre-eclampsia (PE) is a complex and serious disorder of human pregnancy characterized by the development in the latter half of pregnancy of new-onset hypertension which resolves post-partum (Davey and MacGillivray, 1988
; Roberts and Redman, 1993; Witlin and Sibai, 1997). In its severe form, there is also significant proteinuria, oedema and multi-organ dysfunction. There is no reliable predictive test, nor is there an effective medicinal therapy to cure this serious condition. The most worrisome complication of PE is its unpredictable progression to eclampsia (E), a convulsive condition that is life threatening for both mother and baby.
Although the full aetiology and pathogenesis of PE/E are unknown (Cooper et al., 1993; Higgins and Brennecke, 1998; Roberts and Cooper, 2001), the evidence to date suggests that abnormal interaction between maternal and fetal tissues during placentation may be the initiating factor in the complex chain of events that constitute the syndrome of PE/E. In a normotensive pregnancy, cytotrophoblasts invade the spiral arterioles in the decidua, replacing the endothelium and removing the muscular covering of the arterioles. The arterioles thus dilate and the decidual-trophoblast blood flow is increased. In PE/E pregnancies, the invasion of the spiral arterioles by the cytotrophoblasts is limited, and they remain small in diameter and retain their muscular covering, with a consequent restriction of maternal placental blood flow (Brosens et al., 1972; Redman, 1991; Zhou et al., 1997). This defect in PE/E pregnancies precedes the onset of clinical features and offers a cogent explanation of their rather generalized nature. The reduced maternal placental blood flow leads to relative placental hypoxia, with consequent release of toxic factors from the placenta into the maternal circulation, which causes a generalized maternal vascular endothelial cell dysfunction. In turn, this dysfunction perturbs the balance of endothelial cell-derived vasoactive autacoids such as prostacyclin, nitric oxide and endothelin, thereby leading to the generalized vasospasm, diminished tissue perfusion and hypoxia of organs such as the brain, the kidneys and the liver, which is characteristic of PE/E (Redman, 1991; Roberts and Redman, 1993; Arbogast et al., 1994; Higgins and Brennecke, 1998).
The familial association of PE has been recognized for decades (Cooper et al., 1993) with recent estimates of heritability greater than 0.5 being reported in several populations (Ros et al., 2000; the current study). However, while it is clear that genetic factors play a major role in susceptibility to PE/E, the mode of inheritance is the subject of continued speculation (Cooper et al., 1993; Pipkin, 1999; Ros et al., 2000; Roberts and Cooper, 2001). A large Norwegian nationwide study involving all 1.7 million births between 1967 and 1992 showed that the relative risk of developing PE was 2.2 for full sisters, 1.8 for paternal half sisters and 1.8 if pregnant to a man who had fathered a PE pregnancy in another woman, leading to the conclusion that mother and fetus contribute equally to genetic risk, the contribution of the fetus being effected by paternal genes (Lie et al., 1998). This is now the most widely held view, and PE/E, like many common human diseases, is therefore a ‘complex trait’ that does not involve simple Mendelian monogenic inheritance but rather is likely to involve the contribution of multiple genetic and environmental components and their interactions. In this context, a role for imprinting has recently been reported, whereby maternally inherited missense mutations in the paternally inactive STOX1 gene, affecting the putative DNA-binding protein STOX1, are thought to predispose to PE, at least in some Dutch families, possibly by interfering with differentiation of the trophoblast cells (Van Dijk et al., 2005). This Dutch study has provided the strongest evidence to date for specific allelic variation having a role in susceptibility to PE and lends support to endeavours, such as described in this article, to identify genetic risk factors having a role in other populations.
The widespread relevance of the STOX1 gene residing on chromosome 10q is not yet known, with other population-based linkage studies in Australia/New Zealand (Moses et al., 2000; Fitzpatrick et al., 2004), Iceland (Arngrimsson et al., 1999) and Finland (Laivuori et al., 2003) all providing significant evidence of linkage to chromosome 2. Perhaps, the greatest challenge now faced by most investigators having identified and resolved a linkage region using the positional cloning approach is which of the possibly several hundred resident positional candidate genes should be studied exhaustively to identify causal genetic variation for the disease under study. This is the situation with which we were faced after conventional fine mapping of a putative PE susceptibility locus on chromosome 2 that we initially identified after undertaking an short tandem repeat (STR)-marker-based 10-cM resolution genome scan in a cohort of human PE/E pedigrees from Australia and New Zealand (Moses et al., 2000; Fitzpatrick et al., 2004). In this article, we present our further refinement of this putative susceptibility locus on chromosome 2 using a variance components-based linkage approach and our objective prioritization of the positional candidate genes using a combination of quantitative bioinformatics and assessment of differential gene expression in decidual tissue from PE and normotensive women.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
PE/E pedigrees
The 34 PE/E pedigrees (26 Australian and 8 New Zealand) were the subject of our previous 10-cM resolution genome scan that provided strong evidence for a susceptibility locus on chromosome 2 (Moses et al., 2000
). These families were recruited over a 15-year period (1984–1999) through the Monash Medical Centre and the Royal Women’s Hospital, Melbourne, via media advertisements in Sydney, Australia, and through the National Women’s Hospital, Auckland, New Zealand. Subsets of the 26 Australian families have been described in earlier studies (Wilton et al., 1990; Harrison et al., 1997; Guo et al., 1999; Lade et al., 1999). The 34 families included 13 women with E, 74 women with severe PE, 34 women who had hypertensive first pregnancies without proteinuria (mild PE) and 71 women who had normotensive first pregnancies. Diagnosis was based on clinical assessment using the criteria of the Australasian Society for the Study of Hypertension in Pregnancy (Brown et al., 1993), which were very similar to those used in the deCODE Genetics genome scan from Iceland, reported just prior to publication of the Australian study (Arngrimsson et al., 1999). Pregnant women were considered to have severe PE if they had (i) a rise from baseline systolic blood pressure of at least 25 mmHg and/or a rise from baseline diastolic pressure of at least 15 mmHg or (ii) the presence of a systolic pressure of at least 140 mmHg and/or a diastolic pressure of at least 90 mmHg. These levels had to occur on at least two occasions 6 h or more apart.
Variance components-based linkage analysis
Variance components-based linkage analysis was performed using the SOLAR program (Almasy and Blangero, 1998). The variance component linkage method is based upon the classical quantitative genetic decomposition of a phenotype. While most often used for continuous traits, it also can be used for dichotomous traits such as affection status (as used for our primary trait of PE) by assuming an underlying continuous distribution of liability or risk of disease (Williams et al., 1999). Using this approach, it is possible to use data from pedigree structures of arbitrary complexity to make inference regarding the localization and effect sizes of quantitative trait loci (QTL). All parameter estimation and hypothesis testing is performed using a likelihood framework. Using the variance component model, we test the null hypothesis that the additive genetic variance due to the i-th QTL equals zero (no linkage) by comparing the likelihood of this restricted model with that of a model in which the variance due to the i-th QTL is estimated. The difference between the two log10 likelihoods provides the log-of-odds ratio score (LOD) score which is a measure of the support for the hypothesis of linkage over that of ‘no linkage’ at a particular chromosomal location. For this test, P-values are obtained by considering twice the difference in loge likelihoods of these two models which yields a test statistic that is asymptotically distributed as a : mixture of a chi21 variable and a point mass at zero (Self and Liang, 1987). Markedly non-normally distributed traits (and especially leptokurtic traits) may lead to excessive Type I errors when normality is assumed (Allison et al., 1999). Several methods have been developed to handle non-normality of trait data including use of the multivariate t-distribution (Blangero et al., 2001) and the robust LOD score method that can be used for any trait distribution (Blangero et al., 2000, 2001). For a discrete trait such as the primary definition of PE, the robust LOD score method is preferred and is employed in this study (and will provide results that are effectively equivalent to empirical P-values).
Estimation of IBD probability matrices
For the PE pedigrees, we calculated an exact estimate of the location-specific Identity by Descent (IBD) probability matrix using the computer program, Loki (Heath, 1997). These IBD matrices were utilized in SOLAR using specialized import procedures that we have developed.
Error checking of genotyping
Genotypes were checked for Mendelian consistency using the PEDSYS (Dyke, 1989) programs INFER and GENTEST. Unlike most genotype checking programs, GENTEST utilizes all possible pedigree information to search for errors. The program SimWalk2 (Sobel and Lange, 1996), which uses Markov Chain Monte Carlo and simulated annealing algorithms to assign probabilities of mistyping to each genotype, was used to make decisions about the appropriate genotypes to blank.
Patient samples
Human decidua basalis (20–400 mg) was collected at the time of Caesarean section from normotensive (n = 7) and PE (n = 7) pregnancies. At the time of the Caesarean section, the decidua (uterine wall lining) was sampled subjacent to the site of placental attachment to the uterus. This sampling was undertaken through the incision made in the uterus for the routine, clinical purpose of the Caesarean section. Immediately following the delivery of the baby, 10 IU oxytocin was given intravenously to the mother and the placenta was then located by manual palpation. After spontaneous separation of the placenta from the uterine wall, the placenta was gently removed, and one piece of decidual tissue (approximately 1 cm3) from the identified placental bed was then collected and immediately frozen.
RNA isolation
Total RNA was extracted from these samples using QIAGEN RNeasy Midi kits (Qiagen). To remove genomic DNA contamination, the RNA was digested with proteinase K followed by DNase treatment. The concentration of RNA was determined spectrophotometrically, whereas the integrity of the resuspended RNA was visually determined by agarose gel electrophoresis. Samples within each of the two groups were combined for the microarray analysis.
Microarray analysis
Following RNA isolation and pooling, biotin-labelled cRNAs were synthesized and hybridized to CodeLink Human Whole Genome Bioarrays (Amersham Biosciences), which comprise 
55 000 30-mer oligonucleotide probes designed to conserved exons across the transcripts of targeted genes, according to manufacturer’s instructions (CodeLink Expression Assay Kit; Amersham Biosciences). Two bioarrays were processed in parallel, one hybridized with RNA derived from normotensive decidua and the other from PE decidua. The Cy5TM-streptavidin-stained bioarray slides (Cy5TM-streptavidin obtained from Amersham Biosciences) were scanned on a GenePix Array Scanner using CodeLink Expression Scanning Software (Amersham Biosciences). The images were analysed using CodeLink Expression Analysis Software (Amersham Biosciences). Expression values were globally normalized to the median expression value of the whole array spots.
GeneSniffer analysis
GeneSniffer is a computer program that was specifically developed to assist with the prioritization of candidate disease susceptibility genes within defined genomic intervals (www.genesniffer.org). For each gene within a given genomic interval, GeneSniffer downloads appropriate webpages from the NCBI’s Gene, OMIM and PubMed databases and interrogates the text using a list of disease-specific keywords provided by the investigator (assigned a score between 1 and 10 depending on significance). Homologues of each gene are identified by BLAST, and these are scored and weighted according to the degree of homology. A cumulative hitscore is calculated for each gene from the database hits and the weighted homologue database hits, and the output is presented as a webpage (in HTML format) to provide the investigator with information as to the source of the hits and links to relevant external webpages. The method also can employ the observed LOD score function in the region of the QTL to use the localization data as a weighting function, in which genes would be considered more relevant the closer they are to the observed QTL peak.
SNP genotyping
Five single-nucleotide polymorphisms (SNPs) from ACVR2 were genotyped in all members of the 34 PE families. The SNPs were chosen either from the realSNPTM database (Sequenom, realSNP.com) or from the Applied Biosystems Assay-on-demand web site that uses SNPs found in the Celera database. The SNPs were selected to be spaced across the gene, and only SNPs validated in more than one chromosome were used. SNP PCR assays from the RealSNP databases were performed on a Sequenom MassArrayTM system housed at the Australian Genome Research Facility. Genotyping of SNPs chosen from the Celera database was performed using commercially available TaqMan Assays (Assays-on-Demand, Applied Biosystems) and utilizing the ABI 7700 Real Time PCR machine.
Association analysis using the QTDT
Association analyses were performed using the quantitative transmission disequilibrium tests (QTDT) originally described by Abecasis et al. (2000). This procedure has now been implemented in SOLAR for pedigrees of arbitrary size and has been modified to work with discrete traits using a threshold model (similar to the variance component approach described above). The QTDT approach was also employed using haplotypes.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Variance components-based linkage analysis
Variance components linkage analysis implemented in the SOLAR program was used to perform multipoint linkage analysis on the STR fine mapping dataset we previously described (Fitzpatrick et al., 2004
). The benefit of using this approach is that (i) it employs a biologically reasonable threshold model that can directly consider severity and is general enough to capture the critical informational features of more highly parametric models, (ii) it can allow for prevalence differences that vary by age and other covariates and (iii) it utilizes all of the available pedigree information (whereas our previous GENEHUNTER analyses had to drop individuals from the analysis to run under computer memory constraints). Using the variance component approach to analyse the strict diagnosis of PE, we obtained significant evidence of linkage on 2q with a peak LOD score of 3.43 being displayed near marker D2S151 at 155 cM (with LODs greater than 3.3 observed from 154 to 158 cM). Under the general diagnostic model, nominal evidence of linkage was seen on 2p with an LOD score of 1.36 at 108 cM and on 2q at 150 cM, reaching an LOD score of 1.74 (Figure 1).
|
Prioritization of positional candidates using linkage-directed gene expression analysis
As discussed earlier, a primary defect in PE/E pregnancies is insufficient modification of the maternal spiral arteries in the decidua. We believe that it is highly likely that PE/E susceptibility gene(s) are involved in this process and that perturbed mRNA expression in the decidua may be a recognizable signature. There have been major improvements in the methodologies available for quantifying gene expression, and we used CodeLink Human Whole Genome Bioarrays (Amersham Biosciences) that comprise 55 000 30-mer oligonucleo-tide probes designed to conserved exons across the transcripts of targeted genes. Two bioarrays were processed in parallel, one hybridized with RNA derived from normotensive decidua and the other from PE decidua. In this study, our focus has been on genes residing within our region of significant linkage on 2q. After comparing the normalized intensity of each transcript in this region, there were 17 transcripts that showed at least a 2.5-fold change in expression in association with PE, with 15 of the 17 being underexpressed in the PE samples compared to normotensive tissue (Table I). Two of these genes (ACVR2 and ACVR1C) were assigned the highest ‘hitscore’ after interrogation of the region using the bioinformatics tool GeneSniffer (described below).
|
2.5-fold differential expression
Using GeneSniffer, a bioinformatics tool to help prioritize positional candidate genes for SNP typing
Traditionally, a 1-LOD-drop support interval flanking the best linkage result is used to define the critical region of an observed linkage peak. Using the 1-LOD-drop support interval as a guide would define the critical region on 2q from D2S112 (at 144.6 cM) to D2S2330 (at 173.3 cM) under the severe diagnostic criteria. Our preliminary GeneSniffer analysis identified and interrogated 120 genes within the 2q critical region. Positional candidates with the highest hitscore were the activin receptor genes ACVR2 and ACVR1/C (Figure 2).
|
Association analysis of SNPs within positional candidate genes ACVR2 and PE
We performed preliminary SNP analysis on the positional candidate gene ACVR2, found to be under-expressed in PE versus normotensive human decidua and to have obtained the highest hitscore in our GeneSniffer analysis. We genotyped five known validated SNPs in this gene in our cohort of 34 PE/E pedigrees and performed tests for association and linkage disequilibrium using the QTDT test, as implemented in SOLAR using the threshold model. Table II summarizes the SNPs employed, their physical locations, their minor allele frequencies with standard errors and the results of the association analyses. One SNP (rs1424954) showed strong preliminary evidence of association with PE (P = 0.007), whereas two others (rs1364658 and rs1895694) exhibited nominal evidence (P = 0.04 and P = 0.05, respectively). Haplotype analysis revealed no additional association information.
|
There was evidence of weak linkage disequilibrium among these SNPs as summarized in Table III. Linkage disequilibrium was measured by the correlation among alleles (or equivalently among SNP genotypes). The highest observed LD occurs amongst the three associated SNPs, suggesting that the observed signals may be the signature of an observed functional variant.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
In this study, we are pursuing a putative susceptibility gene for PE/E at a locus we previously identified on chromosome 2 following a genome scan in Australian and New Zealand multigenerational PE/E pedigrees (Moses et al., 2000
). Our initial localization evidence was broad (85 cM) and spanned the centromere leading us to subsequently attempt fine mapping after genotyping an additional 25 microsatellite markers at this locus. However, the overall position and shape of the localization evidence obtained using model-free multipoint linkage analysis, as implemented in GENEHUNTER, did not change from that seen previously in our 10-cM resolution genome scan, with two peaks being displayed, one on chromosome 2p at marker D2S388 (107.46 cM) and the other on chromosome 2q at marker D2S2313 (151.5 cM)(Fitzpatrick et al., 2004).
In this study, we hypothesized that although the primary PE phenotype is dichotomous, we can assume that the underlying liability is inherently quantitative. Thus, the chromosome 2 susceptibility gene represents a QTL. This led us to consider the application of variance component procedures using a threshold model for PE, a radical departure from the statistical methodologies used in our previous linkage studies (Moses et al., 2000; Fitzpatrick et al., 2004) and in PE linkage studies in other populations (Arngrimsson et al., 1999; Lachmeijer et al., 2001; Laivouri et al., 2003). Using this strategy has allowed us to both resolve and strengthen our linkage evidence on chromosome 2q.
There are multiple benefits to this statistical approach that could be exploited in future studies. The variance component approach can easily handle covariate effects such as age, baseline blood pressure and smoking status. Such effects are generally ignored in classical model-based and model-free linkage approaches but are critical because they lead to differences in expected prevalences for different subsets of individuals. Additionally, the variance component approach could be used to jointly consider additional quantitative variables such as change in blood pressure during the pregnancy event along with the dichotomous disease status. Such bivariate analyses are more powerful than univariate analysis. Like model-free analyses, variance component procedures can be made completely robust to distributional violations. Unlike model-free linkage analyses, variance component linkage analysis provides a true LOD score with all of the attendant characteristics. Model-free approaches approximate the likelihood. In addition, the variance component approach also allows a unification of the association and linkage approaches. This has the benefit of permitting likelihood-based conditional linkage analyses in which tests for residual linkage are conditioned upon observed associations.
Having resolved our linkage evidence to 2q, we, like most investigators having mapped a QTL for a common human disorder, were faced with the challenge of selecting positional candidates (from the most likely hundreds to choose from) for further genetic and ultimately functional analyses. Rather than relying solely on our own intuition, we decided to develop an objective prioritization strategy to assist us with this task. Combining mapping information with transcriptional profiling is an approach that has been widely speculated in the review literature as having the potential to reduce the overall effort needed in identifying genes causally associated with complex traits (Bogardus et al., 2002; Williams et al., 2002). Successful adoption of this strategy has been reported for a variety of traits, including high-density lipoprotein cholesterol in baboons (Cox et al., 2002), ovariole number in Drosophila melanogaster (Wayne and McIntyre, 2002), human type 2 diabetes (Yang et al., 2002), Parkinson’s disease (Hauser et al., 2003), hypertension (Hubner et al., 2005; Yagil et al., 2005) and chronic obstructive pulmonary disease (Demeo et al., 2006). Our combined use of transcriptional profiling with the bioinformatics tool GeneSniffer is a novel approach that has provided us with consistent gene prioritization data at the PE/E QTL on chromosome 2q. A plausible role for the top-ranked activin receptor genes ACVR2 and ACVR1C in the pathophysiology of PE/E can be readily assigned. ACVR2 and ACVR1C are receptors for the transforming growth factor-ß (TGF-ß) superfamily of signalling proteins, which are known to regulate a variety of cellular functions and play critical roles in many developmental and physiological processes (Massagué and Chen, 2000, Zimmerman and Padgett, 2000; Chang et al., 2002), including placental development (Graham and Lala, 1991; Graham et al., 1992; Peng, 2003). Although ACVR2 and ACVR1C have different ligands, the signalling pathways are linked through the common activation of Smad family proteins, which translocate the signal into the nucleus and act as a transcription factor (Heldin et al., 1997; Attisano and Wrana, 1998; Kretzschmar and Massague, 1998). ACVR1C (activin receptor-like kinase-7) is a type I receptor and has recently been shown to be the physiological ligand of Nodal (Reissman et al., 2001). The role of ACVR1C is largely unknown, but its expression in the placenta from early to late gestation suggests a role for ACVR1C during pregnancy (Roberts et al., 2003). A recent study demonstrated that Nodal acting through ACVR1C has an inhibitory effect on trophoblast cell proliferation and induces apoptosis of trophoblast cells (Munir et al., 2004).
ACVR2 (activin A receptor, type IIA) is a type II receptor for the cell signalling protein activin A that is known to be an important regulator of reproductive function. In early pregnancy, activin A facilitates implantation by promoting trophoblast differentiation towards an invasive phenotype and stimulating the production of paracrine agents involved in invasion (Caniggia et al., 1997). Activin A also specifically promotes decidualization in a dose-dependent manner in endometrial stromal cells (Jones et al., 2002) and, in other cell types, has been seen to stimulate the production of many factors associated with decidualization, such as prostaglandin E2, matrix metalloproteinase-2 and fibronectin (Petraglia et al., 1993; Caniggia et al., 1997). There have been several studies looking at maternal serum of PE patients which have shown increased circulating activin A concentrations compared with gestation-matched controls (Petraglia et al., 1995a; Petraglia et al., 1995b; Muttukrishna et al., 1997; Fraser et al., 1998; Laivuori et al., 1999; Silver et al., 1999; D’Antona et al., 2000). The main source of activin A in the maternal blood during pregnancy is thought to originate from the placenta (Petraglia et al., 1987). Alteration of the expression of ACVR2 and the subsequent effect that it would have on the action of activin A could disturb the process of decidualization effecting the regulation of trophoblast invasion resulting in the insufficient remodelling of the spiral arteries, the pathological hallmark of PE.
Although there are no other published studies on the genetics of PE/E that have combined transcriptional profiling with genetic linkage as a prioritization strategy, and as far as we are aware no other comprehensive microarray gene expression analyses of human decidual tissues from normotensive and pre-eclamptic pregnancies, there have been several such expression studies on human placental tissues. A variety of differentially expressed genes have been reported, including genes associated with obesity (Reimer et al., 2002), glycogen metabolism (Tsoi et al., 2003), hormone and redox metabolism (Pang and Xing, 2004) and hypoxia (Soleymanlou et al., 2005). The ACVR1 and ACVR2 genes were not reported to be differentially expressed in these studies, nor in a candidate gene study that specifically measured ACVR1 and ACVR2 gene expression in placental tissue from pre-eclamptics and controls (Casagrandi et al., 2003). However, there are several obvious differences between these previous studies and our study that may be relevant to these observed differences, including the tissues studied, the population sampled and the complexity of the microarrays that were employed, notwithstanding the fact that PE/E is a complex disease with multiple genetic and environmental factors likely to be at play in conferring susceptibility. It is also important to emphasize that our application of microarray transcriptional profiling, using only two pooled RNA samples, was specifically for the purpose of providing data to be seen as supportive data in the context of our other prioritization strategies.
In conclusion, we have used for the first time in a genetic study of PE/E a variance components-based linkage approach to resolve a susceptibility QTL on chromosome 2q22. Our novel objective prioritization strategy combining linkage-directed transcriptional profiling and bioinformatics analysis with the GeneSniffer program provided consistent top ranking of the activin receptor genes ACVR2 and ACVR1C at this QTL. Our preliminary association/linkage disequilibrium data obtained using known SNPs within ACVR2 are strong encouragement for further genetic investigation. We suggest that extensive re-sequencing of ACVR2 in the 34 PE/E families used in our previous linkage studies to map the 2q QTL is now an appropriate strategy to maximize the chances of identifying the putative functional variant(s) at play in this gene.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
We gratefully acknowledge the support of the clinicians, research midwives and patients who contributed to this study. This work was supported through fundraising from the University of Melbourne, the Royal Women’s Hospital and a start grant from the Southwest Foundation for Biomedical Research. E.F. is the recipient of the Royal Women’s Hospital Postgraduate Research Degree Scholarship.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Abecasis GR, Cookson WO and Cardon LR (2000) Pedigree tests of transmission disequilibrium. Eur J Hum Genet 8,545–551.[CrossRef][Web of Science][Medline]
Allison DB, Heo M, Kaplan N and Martin ER (1999) Sibling-based tests of linkage and association for quantitative traits. Am J Hum Genet 64, 1754–1763.[CrossRef][Web of Science][Medline]
Almasy L and Blangero J (1998) Multipoint quantitative trait linkage analysis in general pedigrees. Am J Hum Genet 62,1198–1211.[CrossRef][Web of Science][Medline]
Arbogast BW, Leeper SC, Merrick RD, Olive KE and Taylor RN (1994) Which plasma factors bring about disturbance of endothelial function in pre-eclampsia? Lancet 343,340–341.[CrossRef][Web of Science][Medline]
Arngrimsson R, Siguroardottir S, Frigge ML, Bjarnadttir RI, Jonsson T, Stefansson H, Baldursdottir A, Einardsdottir AS, Palsson B, Snorradottir S et al. (1999) A genome-wide scan reveals a maternal susceptibility locus for pre-eclampsia on chromosome 2p13. Hum Mol Genet 8,1799–1805.
Attisano L and Wrana JL (1998) Mads and Smads in TGF beta signalling. Curr Opin Cell Biol 10,188–194.[CrossRef][Web of Science][Medline]
Blangero J, Williams JT and Almasy L (2000) Robust LOD scores for variance component-based linkage analysis. Genet Epidemiol 19,S8–S14.
Blangero J, Williams JT and Almasy L (2001) Variance components methods for detecting complex trait loci. Adv Genet 42,151–181.[Medline]
Bogardus C, Baier L, Permana P, Prochazka M, Wolford J and Hanson R (2002) Identification of susceptibility genes for complex metabolic diseases. Ann N Y Acad Sci 967,1–6.[CrossRef][Web of Science][Medline]
Brosens IA, Robertson WB and Dixon HG (1972) The role of the spiral arteries in the pathogenesis of pre-eclampsia. Obstet Gynecol Annu 1,177–191.[Medline]
Brown MA, Gallery EDM, Gatt SP, Leslie G and Robinson J (1993) Management of hypertension in pregnancy: executive summary. Med J Aust 158,700–702.[Web of Science][Medline]
Caniggia I, Lye SJ and Cross JC (1997) Activin is a local regulator of human cytotrophoblast cell differentiation. Endocrinology 138,3976–3986.
Casagrandi D, Bearfield C, Geary J, Redman CW and Muttukrishna S (2003) Inhibin, activin, follistatin, activin receptors and beta-glycan gene expression in the placental tissue of patients with pre-eclampsia. Mol Hum Reprod 9,199–203.
Chang H, Brown CW and Matzuk MM (2002) Genetic analysis of the mammalian transforming growth factor-beta superfamily. Endocr Rev 23,787–823.
Cooper DW, Brennecke SP and Wilton AN (1993) Genetics of pre-eclampsia. Hyperten Preg 12,1–23.
Cox LA, Birnbaum S and VandeBerg JL (2002) Identification of candidate genes regulating HDL cholesterol using a chromosomal region expression array. Genome Res 12,1693–1702.
D’Antona D, Reis FM, Benedetto C, Evans LW, Groome NP, de Kretser DM, Wallace EM and Petraglia F (2000) Increased maternal serum activin A but not follistatin levels in pregnant women with hypertensive disorders. J Endocrinol 165,157–162.[Abstract]
Davey DA and MacGillivray I (1988) The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 158,892–898.[Web of Science][Medline]
Demeo DL, Mariani TJ, Lange C, Srisuma S, Litonjua AA, Celedon JC, Lake SL, Reilly JJ, Chapman HA, Mecham BH et al. (2006) The SERPINE2 gene is associated with chronic obstructive pulmonary disease. Am J Hum Genet 78,253–264.[CrossRef][Web of Science][Medline]
Dyke B (1989) PEDSYS. A Pedigree Data Management System. Users Manual. PGL Technical Report no. 2. San Antonio, Texas: Southwest Foundation for Biomedical Research.
Fitzpatrick E, Göring HH, Liu H, Borg A, Forrest S, Cooper DW, Brennecke SP and Moses EK (2004) Fine mapping and SNP analysis of positional candidates at the pre-eclampsia susceptibility locus (PREG1) on chromosome 2. Hum Biol 76,849–862.[CrossRef][Web of Science][Medline]
Fraser RF, McAsey ME and Coney P (1998) Inhibin-A and pro-alpha C are elevated in preeclamptic pregnancy and correlate with human chorionic gonadotropin. Am J Reprod Immunol 40,37–42.
Graham CH and Lala PK (1991) Mechanism of control of trophoblast invasion in situ. J Cell Physiol 148,228–234.[CrossRef][Web of Science][Medline]
Graham CH, Lysiak JJ, McCrae KR and Lala PK (1992) Localization of transforming growth factor-beta at the human fetal–maternal interface: role in trophoblast growth and differentiation. Biol Reprod 46,561–572.[Abstract]
Guo G, Lade JA, Wilton AN, Moses EK, Grehan M, Fu Y, Qiu H, Cooper DW and Brennecke SP (1999) Genetic susceptibility to pre-eclampsia and chromosome 7q36. Hum Genet 105,641–647.[CrossRef][Web of Science][Medline]
Harrison GA, Humphrey KE, Jones N, Badenhop R, Guo G, Elakis G, Kaye JA, Turner RJ, Grehan M, Wilton AN et al. (1997) A genomewide linkage study of pre-eclampsia/eclampsia reveals evidence for a candidate region on 4q. Am J Hum Genet 60,1158–1167.[Web of Science][Medline]
Hauser MA, Li YJ, Takeuchi S, Walters R, Noureddine M, Maready M, Darden T, Hulette C, Martin E, Hauser E et al. (2003) Genomic convergence: identifying candidate genes for Parkinson’s disease by combining serial analysis of gene expression and genetic linkage. Hum Mol Genet 12,671–677.
Heath SC (1997) Markov chain Monte Carlo segregation and linkage analysis for oligogenic models. Am J Hum Genet 61,748–760.[Web of Science][Medline]
Heldin CH, Miyazono K and ten Dijke P (1997) TGF-beta signalling from cell membrane to nucleus through SMAD proteins. Nature 390,465–471.[CrossRef][Medline]
Higgins JR and Brennecke SP (1998) Pre-eclampsia – still a disease of theories? Curr Opin Obstet Gynecol 10,129–133.[CrossRef][Web of Science][Medline]
Hubner N, Wallace CA, Zimdahl H, Petretto E, Schulz H, Maciver F, Mueller M, Hummel O, Monti J, Zidek V et al. (2005) Integrated transcriptional profiling and linkage analysis for identification of genes underlying disease. Nat Genet 37,209–210.[CrossRef][Web of Science][Medline]
Jones RL, Salamonsen LA and Findlay JK (2002) Activin A promotes human endometrial stromal cell decidualization in vitro. J Clin Endocrinol Metab 87,4001–4004.
Kretzschmar M and Massague J (1998) SMADs: mediators and regulators of TGF-beta signaling. Curr Opin Genet Dev 8,103–111.[CrossRef][Web of Science][Medline]
Lachmeijer AM, Arngrimsson R, Bastiaans EJ, Frigge ML, Pals G, Sigurdardottir S, Stefansson H, Palsson B, Nicolae D, Kong A et al. (2001) A genome-wide scan for pre-eclampsia in the Netherlands. Eur J Hum Genet 9,758–764.[CrossRef][Web of Science][Medline]
Lade JA, Moses EK, Guo GL, Wilton AN, Grehan M, Cooper DW and Brennecke SP (1999) The eNOS gene: a candidate for the pre-eclampsia susceptibility locus? Hyperten Preg 18,81–93.[Web of Science][Medline]
Laivuori H, Kaaja R, Turpeinen U, Stenman UH and Ylikorkala O (1999) Serum activin A and inhibin A elevated in pre-eclampsia: no relation to insulin sensitivity. Br J Obstet Gynaecol 106,1298–1303.[Web of Science][Medline]
Laivuori H, Lahermo P, Ollikainen V, Widen E, Haiva-Mallinen L, SundstromH, Laitinen T, Kaaja R, Ylikorkala O and Kere J (2003) Susceptibility loci for pre-eclampsia on chromosomes 2p25 and 9p13 in Finnish families. Am J Hum Genet 72,168–177.[CrossRef][Web of Science][Medline]
Lie RT, Rasmussen S, Brunborg H, Gjessing HK, Lie-Nielsen E and Irgens LM (1998) Fetal and maternal contributions to risk of pre-eclampsia: population based study. BMJ 316,1343–1347.
Massagué J and Chen YG (2000) Controlling TGF-beta signaling. Genes Dev 14,627–644.
Moses EK, Lade JA, Guo G, Wilton AN, Grehan M, Freed K, Borg A, Terwilliger JD, North R, Cooper DW et al. (2000) A genome scan in families from Australia & New Zealand confirms the presence of a maternal susceptibility locus for pre-eclampsia, on chromosome 2. Am J Hum Genet 67,1581–1585.[CrossRef][Web of Science][Medline]
Munir S, Xu G, Wu Y, Yang B, Lala PK and Peng C (2004) Nodal and ALK7 inhibit proliferation and induce apoptosis in human trophoblast cells. J Biol Chem 279,31277–31286.
Muttukrishna S, Child TJ, Groome NP and Ledger WL (1997) Source of circulating levels of inhibin A, pro alpha C-containing inhibins and activin A in early pregnancy. Hum Reprod 12,1089–1093.
Pang ZJ and Xing FQ (2004) Comparative profiling of metabolism-related gene expression in pre-eclamptic and normal pregnancies. Arch Gynecol Obstet 269,91–95.[CrossRef][Medline]
Peng C (2003) The TGF-beta superfamily and its roles in the human ovary and placenta. J Obstet Gynaecol Can 25,834–844.[Medline]
Petraglia P, Sawchenko P, Lim AT, Rivier J and Vale W (1987) Localization, secretion, and action of inhibin in human placenta. Science 237,187–189.
Petraglia F, Anceschi MM, Calza L, Garuti GC, Fusaro P, Giardino L, Genazzani AR and Vale W (1993) Inhibin and activin in human fetal membranes: evidence for a local effect on prostaglandin release. J Clin Endocrinol Metab 77,542–548.[Abstract]
Petraglia F, Aguzzoli L, Gallinelli A, Florio P, Zonca M, Benedetto C and Woodruff K (1995a) Hypertension in pregnancy: changes in activin A maternal serum concentration. Placenta 16,447–454.[CrossRef][Web of Science][Medline]
Petraglia F, De Vita D, Gallinelli A, Aguzzoli L, Genazzani AR, Romero R and Woodruff K (1995b) Abnormal concentration of maternal serum activin A in gestational diseases. J Clin Endocrinol Metab 80,558–561.[Abstract]
Pipkin FB (1999) What is the place of genetics in the pathogenesis of pre-eclampsia? Biol Neonat 76,325–330.
Redman CWG (1991) Pre-eclampsia and the placenta. Placenta 12,301–308.[Web of Science][Medline]
Reimer T, Koczan D, Gerber B, Richter D, Thiesen HJ and Friese K (2002) Microarray analysis of differentially expressed genes in placental tissue of pre-eclampsia: up-regulation of obesity-related genes. Mol Hum Reprod 8,674–680.
Reissmann E, Jornvall H, Blokzijl A, Andersson O, Chang C, Minchiotti G, Persico MG, Ibanez CF and Brivanlou AH (2001) The orphan receptor ALK7 and the Activin receptor ALK4 mediate signaling by Nodal proteins during vertebrate development. Genes Dev 15,2010–2022.
Roberts JM and Cooper DW (2001) Pathogenesis and genetics of pre-eclampsia. Lancet 357,53–56.[CrossRef][Web of Science][Medline]
Roberts JM and Redman CW (1993) Pre-eclampsia: more than pregnancy-induced hypertension. Lancet 341,1447–1451.[CrossRef][Web of Science][Medline]
Roberts HJ, Hu S, Qiu Q, Leung PCK, Cannigia I, Gruslin A, Tsang B and Peng C (2003) Biol Reprod 68,1719–1726.
Ros HS, Lichtenstein P, Lipworth L and Cnattingius S (2000) Genetic effects on the liability of developing pre-eclampsia and gestational hypertension. Am J Med Genet 91,256–260.[CrossRef][Web of Science][Medline]
Self SG and Liang K-Y (1987) Asymptotic properties of maximum likelihood estimators and likelihood ratio tests using nonstandard conditions. J Am Stat Assoc 82,605–610.[CrossRef][Web of Science]
Silver HM, Lambert-Messerlian GM, Star JA, Hogan J and Canick JA (1999) Comparison of maternal serum total activin A and inhibin A in normal, preeclamptic, and nonproteinuric gestationally hypertensive pregnancies. Am J Obstet Gynecol 180,1131–1137.[CrossRef][Web of Science][Medline]
Sobel E and Lange K (1996) Descent graphs in pedigree analysis: applications to haplotyping, location scores, and marker sharing statistics. Am J Hum Genet 58,1323–1337.[Web of Science][Medline]
Soleymanlou N, Jurisica I, Nevo O, Ietta F, Zhang X, Zamudio S, Post M and Caniggia I (2005) Molecular evidence of placental hypoxia in preeclampsia. J Clin Endocrinol Metab 90,4299–4308.
Tsoi SC, Cale JM, Bird IM and Kay HH (2003) cDNA microarray analysis of gene expression profiles in human placenta: up-regulation of the transcript encoding muscle subunit of glycogen phosphorylase in preeclampsia. J Soc Gynecol Investig 10,496–502.[CrossRef][Web of Science][Medline]
Van Dijk M, Mulders J, Potsma A, Konst AAM, Lachmeijer AMA, Dekker GA, Blankenstein MA and Oudejans CBM (2005) Maternal segregation of the Dutch preeclampsia locus at 10q22 with a new member of the winged helix gene family. Nat Genet 37,514–519.[CrossRef][Web of Science][Medline]
Wayne ML and McIntyre LM (2002) Combining mapping and arraying: an approach to candidate gene identification. Proc Natl Acad Sci USA 99,14903–14906.
Williams JT, Van Eerdewegh P, Almasy L, Blangero J (1999) Joint multipoint linkage analysis of multivariate qualitative and quantitative traits. I. Likelihood formulation and simulation results. Am J Hum Genet 65,1134–1147.[CrossRef][Web of Science][Medline]
Williams NM, O’Donovan MC and Owen MJ (2002) Genome scans and microarrays: converging on genes for schizophrenia? Genome Biol 3,1011.1–1011.5.
Wilton AN, Cooper DW, Brennecke SP, Bishop SM and Marshall P (1990) Absence of close linkage between maternal genes for susceptibility to pre-eclampsia/eclampsia and HLA DR? Lancet 336,653–657.[CrossRef][Web of Science][Medline]
Witlin AG and Sibai BM (1997) Hypertension in pregnancy: current concepts of pre-eclampsia. Annu Rev Med 48,115–127.[CrossRef][Web of Science][Medline]
Yagil C, Hubner N, Monti J, Schulz H, Sapojnikov M, Luft FC, Ganten D and Yagil Y (2005) Identification of hypertension-related genes through an integrated genomic-transcriptomic approach. Circ Res 96,605–606.
Yang X, Pratley RE, Tokraks S, Bogardus C and Permana PA (2002) Microarray profiling of skeletal muscle tissues from equally obese, non-diabetic insulin-sensitive and insulin-resistant Pima Indians. Diabetologia 45, 1584–1593.[CrossRef][Web of Science][Medline]
Zhou Y, Damsky CH and Fisher SJ (1997) Pre-eclampsia is associated with failure of human cytotrophoblasts to mimic a vascular adhesion phenotype. One cause of defective endovascular invasion in this syndrome? J Clin Invest 99,2152–2164.[Web of Science][Medline]
Zimmerman CM and Padgett RW (2000) Transforming growth factor beta signaling mediators and modulators. Gene 249,17–30.[CrossRef][Web of Science][Medline]
Submitted on April 14, 2006; resubmitted on May 17, 2006; accepted on May 23, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Fitzpatrick, M.P. Johnson, T.D. Dyer, S. Forrest, K. Elliott, J. Blangero, S.P. Brennecke, and E.K. Moses Genetic association of the activin A receptor gene (ACVR2A) and pre-eclampsia Mol. Hum. Reprod., March 1, 2009; 15(3): 195 - 204. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.P. Johnson, E. Fitzpatrick, T.D. Dyer, J.B.M. Jowett, S.P. Brennecke, J. Blangero, and E.K. Moses Identification of two novel quantitative trait loci for pre-eclampsia susceptibility on chromosomes 5q and 13q using a variance components-based linkage approach Mol. Hum. Reprod., January 1, 2007; 13(1): 61 - 67. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||