Clinical Science (2006) 110, 443–458 (Printed in Great Britain) doi:10.

1042/CS20050323 443

R E V I E W

Searching for genetic clues to the

causes of pre-eclampsia

Sally CHAPPELL and Linda MORGAN

Department of Clinical Chemistry, Institute of Genetics, University of Nottingham, Nottingham NG7 2UH, U.K.

A B S T R A C T

Pre-eclampsia and its related syndromes are significant causes of maternal and fetal death,
but much remains unclear about the underlying disease mechanisms. Epidemiological research
has consistently demonstrated a familial predisposition to pre-eclampsia, which has encouraged
genetic research in this area. The goal is the discovery of susceptibility genes which will inform
understanding of the pathophysiology of pre-eclampsia, and may prove to be targets for therapeutic
or preventative strategies. This review examines the application of molecular technologies to the
search for genetic clues in pre-eclampsia and emphasizes the importance of integrative approaches.
The results of recent genome-wide linkage studies have been particularly encouraging, identifying
a number of loci which merit closer examination. Candidate gene studies have proved less fruit-
ful, generating conflicting and inconclusive results. Possible explanations and remedies for this
deficiency are discussed with a view to stimulating closer collaboration between researchers in
this field.

INTRODUCTION teinuria. Because pre-eclampsia resolves postpartum,

premature delivery of the baby may be essential to
Pre-eclampsia is a potentially life-threatening multi- safeguard the mother’s life. Consequently, many of
system disorder which affects approx. 3 % of pregnant the infants born of pre-eclamptic pregnancies require
women in the Western world [1]. Recognized clinically costly support on special care baby units. The burden
by the onset of hypertension and loss of protein in of pre-eclampsia on health care resources is therefore
the urine, pre-eclampsia is nevertheless phenotypically substantial. Progress in the prevention and treatment of
diverse and much remains to be understood about its this condition will require advances in our understanding
causes. The imperative for research into pre-eclampsia of the pathophysiology of the disorder at the molecular
is its unacceptable burden on maternal and fetal health, level.
much of which is borne by the developing world. It has been recognized for many years that pre-
Globally, the hypertensive disorders of pregnancy are eclampsia has a familial component, and the identification
responsible for up to 50 000 maternal deaths and as many of susceptibility genes is one of a number of strategies
as 900 000 perinatal deaths annually [2]. Prompt diagnosis designed to elucidate the underlying pathogenetic mech-
and intervention are of vital importance in reducing anisms. Increasingly, researchers of complex disorders are
maternal mortality and have guided the development turning to integrated approaches, which must include an
of antenatal care facilities for regular monitoring of appreciation of underlying biological processes, but can
maternal blood pressure and early detection of pro- also utilize the power of genome-wide screening, RNA

Key words: fetal genotype, genome-wide screen, hypertension, maternal genotype, pre-eclampsia, pregnancy, susceptibility gene.
Abbreviations: Ang II, angiotensin II; AT1, Ang II type 1; CI, confidence interval; eNOS, endothelial nitric oxide synthase;
GOPEC, Genetics Of Pre-EClampsia; IGF, insulin-like growth factor; IGFBP-1, IGF-binding protein-1; KIR, killer Ig-like receptor;
NK, natural killer; PlGF, placental growth factor; sFlt-1, soluble Flt-1; SNP, single nucleotide polymorphism; TDT, transmission
disequilibrium testing; TNFα, tumour necrosis factor α; uNK, uterine NK; VEGF, vascular endothelial growth factor.
Correspondence: Dr Linda Morgan (email Linda.Morgan@nottingham.ac.uk).

C 2006 The Biochemical Society
444 S. Chappell and L. Morgan

microarray technology and proteomics to generate novel Treatment is directed initially towards stabilization of
hypotheses. blood pressure and prevention of convulsions. Methyl-
Comprehensive reviews of genetic studies in pre- dopa and labetolol are widely used to control blood pres-
eclampsia are available [3,4], and no attempt will be made sure in pre-eclampsia, supplemented by nifedipine.
to reproduce these excellent articles in detail here. The Magnesium sulphate is the drug of choice for the treat-
aim of this review is to present some of the recent genetic ment and prevention of eclamptic convulsions. There is
studies in the context of earlier research, and to explore now good evidence that it is superior to the anticon-
the impact of developing technologies and bioinformatic vulsants phenytoin and diazepam [2]. If the mother’s
resources on evolving research strategies. condition can be stabilized, it may be possible to pro-
long pregnancy to allow further maturation of the
fetus. However, pre-eclampsia does not begin to resolve
PRE-ECLAMPSIA PHENOTYPES until after delivery, and rapidly progressive disease may
demand premature delivery to protect the life and health
Pre-eclampsia is a syndrome observed only at the time of both mother and baby.
of pregnancy. The condition resolves following delivery
and there is no recognized counterpart in males or non-
parous women. Pre-eclampsia has been defined as the PATHOPHYSIOLOGY OF PRE-ECLAMPSIA
onset of persistent hypertension exceeding 140/90 mmHg
associated with proteinuria exceeding 300 mg/day after Established pre-eclampsia is characterized by reduced
the 20th week of gestation [5]. Although this definition vascular perfusion of major organs resulting from wide-
is useful for research purposes, it forces diagnostic thres- spread maternal vascular endothelial dysfunction, which
holds on the continuous distributions of blood pressure antedates the onset of clinical disease [13]. Although pre-
and proteinuria, and masks the phenotypic diversity eclampsia presents clinically after the 20th week of gesta-
which characterizes clinical practice. Additional features tion, in many cases it has its origins much earlier in
affecting sub-groups of women include thrombocyto- pregnancy, at the time of trophoblastic invasion into
penia, deranged liver function, renal impairment and maternal tissues. In normal pregnancy, specialized extra-
pulmonary oedema. A cluster of three features, Haem- villous trophoblast cells infiltrate the maternal spiral
olysis, Elevated Liver enzymes and Low Platelets, usually arteries supplying the placental bed, converting them
but not invariably associated with hypertension and pro- from muscular vessels responsive to vasopressor stimuli
teinuria, is known as HELLP syndrome [6]. Eclampsia, into flaccid high-capacity sinuses, ideal for the exchange
which affects a minority of women, includes some or of nutrients and gases between maternal and fetal blood
all of the features of pre-eclampsia, in association with [14]. The process of spiral artery remodelling is incom-
convulsions. Pre-eclampsia may run a fulminating course, plete in many pre-eclamptic pregnancies, pointing to
necessitating swift, frequently premature, delivery of the importance of the interaction between fetal and
the child. In other cases, the disorder progresses more maternal factors in the origin of the disease. The resulting
slowly, allowing time to control maternal blood pressure impairment of placental perfusion triggers maternal
and deliver the child at a later gestational age. Up to endothelial activation in those women who subsequently
one-third of infants born of pre-eclamptic pregnancies develop pre-eclampsia. It is worth noting that impaired
are affected by intrauterine growth restriction; in other spiral artery remodelling does not invariably lead to
pregnancies, fetal growth does not appear to be seriously pre-eclampsia; similar features are observed in growth-
impaired [7]. It is clear that pre-eclampsia, eclampsia restricted pregnancies without maternal disease. The
and HELLP are related but phenotypically varied syn- implication is that maternal factors, including genetic
dromes. predisposers, determine whether impaired placentation
In clinical practice, a substantial proportion of the results in pre-eclampsia (Figure 1).
women affected by pre-eclampsia have predisposing Research into the link between placental ischaemia
medical conditions, which include chronic hypertension and maternal endothelial activation is on-going, but it
[8], diabetes [9], renal disease [10] and dyslipidaemia [11]. is clear that oxidative stress, both in the placenta and
Furthermore, long-term follow-up studies have shown within the maternal systemic circulation, is an important
that women who were unaffected by these conditions component of the pathophysiology [15]. There is evi-
prior to conception are nevertheless more likely to dence for increased trafficking of syncytiotrophoblast
develop them later in life following an episode of pre- microfragments into maternal plasma in pre-eclamptic
eclampsia [12]. It is worth noting that inherited liability pregnancies, with the release of inflammatory cytokines
to essential hypertension, diabetes and dyslipidaemia [16]. The ability of the mother to mount an adequate
is well-recognized, raising the possibility that pre- response to released pro-oxidant molecules may be of
eclampsia may share susceptibility genes with these key importance. This view is endorsed by the increased
conditions. risk of pre-eclampsia in women with pre-existing

C 2006 The Biochemical Society
 Searching for genetic clues to the causes of pre-eclampsia 445

Figure 1 Suggested pathophysiological mechanisms in pre-eclampsia

conditions associated with oxidative stress, including trophoblastic invasion, and genetic approaches may be
obesity [17], chronic hypertension [8] and diabetes [9], appropriate for testing this hypothesis.
where antioxidant reserves may be inadequate. These
observations have led to trials of the prophylactic effects
of antioxidants, including vitamins C and E and selenium. EVIDENCE FOR A GENETIC BASIS TO
A recent Cochrane review of seven randomized trials [18] PRE-ECLAMPSIA
reported a 39 % reduction in the risk of pre-eclampsia as-
sociated with oral antioxidant supplements. The authors A familial predisposition to pre-eclampsia has been con-
stress that these findings should be interpreted with firmed in numerous studies, from the U.S.A., Scotland,
caution as some data were from poor quality studies. The Iceland, Scandinavia and Australia, which record a 2–5-
results of large randomized controlled trials are awaited. fold increase in risk to first-degree relatives of women
A promising proposal for a placental factor which with pre-eclampsia [1,20–23]. Segregation analysis as-
might trigger endothelial dysfunction is the soluble form suming a Mendelian mode of inheritance has yielded a
of the VEGF (vascular endothelial growth factor) recep- number of possible models, including maternal recessive
tor sFlt-1. Flt-1 is encoded by VEGF1; the soluble form or maternal dominant with partial penetrance [22]. It
is a splice variant lacking the cytoplasmic and transmem- was noted that factors which modify the penetrance of
brane domains. VEGF and PlGF (placental growth maternal genes could include fetal genetic effects.
factor), which play a role in maintaining vascular endo- Alternative suggested genetic mechanisms include a
thelial integrity, are both bound by circulating sFlt-1 mitochondrial mode of inheritance [24], although epi-
(soluble Flt-1), which thus acts as an antagonist for demiological data indicate that mitochondrial genes do
these factors. In a study of women with established pre- not contribute to the inheritance of pre-eclampsia in the
eclampsia, placental sFlt-1 production was up-regulated, majority of families [1]. Graves [25] proposed a suscept-
and sFlt-1 was detectable at increased concentrations ibility gene which is paternally imprinted; only the
in maternal plasma, with a corresponding reduction in maternally inherited allele is active. Inheritance of a
circulating free VEGF and PlGF [19]. Furthermore, mutated copy of the gene from a heterozygous mother
injection of sFlt-1 into pregnant and non-pregnant rats would effectively result in loss of gene activity in the
was followed by the development of hypertension and fetus (Figure 2). This proposal combines both maternal
proteinuria. sFlt-1 production is induced by hypoxia, and fetal gene effects, and is prescient in the light of
providing a plausible link between placental ischaemia the results of linkage studies recently reported by Dutch
and maternal disease. It is also possible that primary researchers [26], discussed below. In view of the diversity
dysfunction of the VEGF system contributes to reduced of inheritance models which have been suggested, it is

C 2006 The Biochemical Society
446 S. Chappell and L. Morgan

Medical Birth Registry from 1967 to 1992, which ident-

ified almost 400 000 women who had had at least two
pregnancies [1]. It was possible from the records to
distinguish between pregnancies which had the same
parents, and those where either the mother or the father
differed in the second pregnancy. Only 1.3 % of women
who had not had pre-eclampsia in their first pregnancy
developed pre-eclampsia in the second pregnancy. This
risk was significantly increased (2.9 %; P = 0.005) if they
had a new partner who had previously fathered a pre-
eclamptic pregnancy, consistent with an effect of paternal
genes acting via the fetus. Further support for a role for
paternal genes comes from a study which used records
on the Utah Population Database to identify men and
women born of pre-eclamptic pregnancies and to match
them with controls born of healthy pregnancies [29].
Men who were born of a pre-eclamptic pregnancy were
at increased risk of fathering a pre-eclamptic pregnancy
[odds ratio 2.1 (95 % CI, 1.0–4.3)], suggesting that sus-
ceptibility can be transmitted via paternal genes, pre-
Figure 2 Genetic imprinting
sumably acting in the fetus. A recent analysis of 438 597
A model of paternal imprinting of gene A in pre-eclampsia. The superscript ‘m’
mother–baby pairs and 286 945 father–baby pairs from
or ‘p’ represents alleles of maternal or paternal origin respectively. The red line
the Norwegian birth registry confirmed a 1.5-fold in-
indicates imprinting (silencing) of that allele. In this example, fetal expression of
crease in risk of fathering an affected pregnancy to men
allele A is required for a healthy pregnancy. The mother is heterozygous for a
born of a pre-eclamptic pregnancy, and a 2.2-fold increase
mutant allele ‘a’. If the mutant maternal allele is transmitted to the fetus, loss of
in risk to women born of an affected pregnancy [30]. The
gene function results, as the wild-type paternally inherited allele is not expressed.
authors reasoned that the risk to women is influenced by
genes active both in the mother and in the fetus, whilst
paternal genes can act only via the fetus. In support of
common to regard pre-eclampsia as a complex disorder, this hypothesis, they showed that sisters who were not
which allows for Mendelian inheritance in a subgroup of born of pre-eclamptic pregnancies nevertheless were at
pedigrees and interactions of multiple susceptibility genes increased risk in their own pregnancies, due to maternally
in other families. active genes. In contrast, brothers born of unaffected
The classic approach to dissecting the contribution pregnancies, who presumably had not inherited fetal
of environmental and inherited components to disease susceptibility genes, had a similar risk to men with no
susceptibility is to compare the concordance rates in family history of fathering pre-eclamptic pregnancies.
monozygotic and dizygotic twins, who share 100 % and The data also demonstrated an increased risk of pre-
50 % respectively of their genetic makeup. Early studies eclampsia presenting before the 34th week of gestation
cast doubt on the existence of maternal genetic factors, if the father or mother had been born of an affected pre-
as concordance rates for pre-eclampsia among parous gnancy, suggesting that genetic factors are predictive of
monozygotic twin sisters were low [27]. Estimates of severe disease.
heritability – the proportion of phenotypic variability A large study using data from the population-based
attributable to genetic causes – had wide CIs (confidence Swedish Birth Register, which prospectively records
intervals; 95 % CI, 0–0.49) due to the relatively small antenatal, obstetric and neonatal data, estimated heritabil-
number of affected women ascertained. The largest pub- ity due to maternal genes as 0.35 (95 % CI, 0.33–0.36) and
lished twin study cross-linked data from the Swedish that due to fetal genes as 0.20 (95 % CI, 0.11–0.24) [31].
Twin Register and the Swedish Medical Birth Register Paternally and maternally inherited genes contributed
[28]: 917 pairs of parous monozygotic and 1199 pairs of equally to fetal genetic effects. Overall, these data suggest
dizygotic twin sisters were identified. The estimate of the that genetic factors are responsible for over half of the
heritability of pre-eclampsia was 0.54, but even with this liability to pre-eclampsia.
large sample the CIs were wide (95 % CI, 0–0.71). When
non-proteinuric gestational hypertension was considered
as a mild form of pre-eclampsia, the heritability was 0.47 CANDIDATE GENE STUDIES
(95 % CI, 0.13–0.61).
Evidence of a role for paternally inherited fetal genes A PubMed search conducted in September 2005 using
comes from an analysis of data from the Norwegian the search terms ‘pre-eclampsia, pregnancy hypertension,

C 2006 The Biochemical Society
 Searching for genetic clues to the causes of pre-eclampsia 447

Table 1 Candidate gene studies in pre-eclampsia

AT2, Ang II type 2.

Gene Selected Gene Selected

Gene name symbol references Gene name symbol references

Thrombophilia Interleukin 1β IL1B [53]

Factor V Leiden F5 [32–35] Interleukin 1 receptor antagonist IL1RN [54]
Prothrombin 20210 F2 [34,36] Interleukin 10 IL10 [55]
Methylene tetrahydrofolate reductase MTHFR [32–35,37,38] T-lymphocyte-associated protein 4 CTLA4 [56]
Cystathione β-synthase CBS [33] TNF-receptor superfamily member 6 FAS [57]
Plasminogen activator inhibitor 1 SERPINE1 [34,38] Oxidative stress
β-Fibrinogen FGB [39] Microsomal epoxide hydrolase EPHX1 [58]
Platelet glycoprotein IIIa ITGB3 [36] Glutathione S-transferase pi GSTP1 [38,59]
Thrombomodulin THBD [40] Glutathione S-transferase mu1 GSTM1 [59,60]
Factor VII F7 [39] Glutathione S-transferase theta1 GSTT1 [59,60]
Platelet collagen receptor α2β1 ITGA2 [34] Myeloperoxidase MPO [60]
Factor XIII A-subunit F13A1 [41] Manganese superoxide dismutase SOD2 [60]
Haemodynamics Cytochrome 1A1 CYP1A1 [59,60]
Angiotensinogen AGT [35,38,42] Haptoglobin HP [61]
Renin REN [43] p22phox CYBA [62]
Angiotensin-converting enzyme ACE [38,42] Lipid metabolism
AT1 receptor AGTR1 [35,38,42] Lipoprotein lipase LPL [63]
AT2 receptor AGTR2 [35,44] Apolipoprotein E APOE [64]
Epithelial sodium channel SCNN1B [45] Peroxisome-proliferator-activated receptor γ PPARG [65]
Endothelial function Cholesteryl ester transfer protein CETP [66]
eNOS NOS3 [35,46] β 3 -Adrenergic receptor ADRB3 [67]
Endothelin 1 EDN1 [47] Endocrine
Dimethylarginine dimethylaminohydrolase 1 DDAH1 [48] Oestrogen receptor α ESR1 [68]
Dimethylarginine dimethylaminohydrolase 2 DDAH2 [48] Oestrogen receptor β ESR2 [69]
G-protein β3 GNB3 [49] Inhibin α INHA [70]
Cytokines Angiogenesis
TNFα TNF [35,38,50] VEGF VEGF [71]
IGF II IGF2 [51] Matrix metallopeptidase 1 MMP1 [72]
Interleukin 1α IL1A [52]

gene, genetic, polymorphism and molecular’ identified angiotensin-converting enzyme and angiotensin re-
over 130 English language reports of candidate gene ceptors); inherited thrombophilias (coagulation factor V
studies in pre-eclampsia (Table 1, and Supplementary Leiden variant, prothrombin and methylene tetrahydro-
Table 1, at http://www.clinsci.org/cs/110/cs1100443add. folate reductase); the NOS3 gene regulating the syn-
htm). The majority of candidate gene studies adopted the thesis of the vasorelaxant eNOS (endothelial nitric oxide
simplest approach to the identification of genetic asso- synthase); and the gene encoding the cytokine TNF
ciation, comparing the frequency of genetic variants in (tumour necrosis factor) α.
cases and controls. Many investigated a single poly- It must be conceded that over a decade of molecular
morphism in a single candidate gene; a minority tested genetic research has failed to identify a single universally
several genes, or multiple polymorphisms in one or more accepted susceptibility gene for pre-eclampsia. There has
genes. been a lack of consistent reproducibility of results, as
Although studies of over 50 candidate genes have been has been the experience with many complex disorders.
reported, eight genes account for 70 % of published re- Possible explanations have been discussed widely [73],
search into candidate genes for pre-eclampsia. These have and studies of pre-eclampsia share some of these generic
been suggested by the features of established disease: problems. Progress in this field will require attention to
genes encoding elements of the renin–angiotensin sys- both study design and choice of candidate genes, and
tem, which regulates blood pressure (angiotensinogen, these aspects are considered in this review.

C 2006 The Biochemical Society
448 S. Chappell and L. Morgan

Phenotype ments for high-throughput genotyping at multiple SNPs.

Considerable effort is required to minimize phenotypic This approach has only recently been applied to genetic
heterogeneity in both case and control groups in genetic studies in pre-eclampsia and is mandatory if genes are
studies, in the belief that this will ensure greater gen- to be confidently excluded from the list of possible
etic homogeneity and thus maximize the power of the candidates.
study. Internationally recognized definitions of pre-
eclampsia are available [5], but strict application of the Statistical significance and power
definition can be difficult as pre-eclampsia has a tran- There is ongoing debate about the level of statistical
sient phenotype. Retrieving data from incomplete medi- significance which is required to declare significant asso-
cal records in order to recruit retrospectively may be ciation in genetic studies. Over 4 million verified SNPs
unreliable, and recruitment at the time of diagnosis is have been identified in the human genome, but there is
likely to lead to more reliable phenotyping. no consensus on an appropriate statistical correction for
multiple testing. A Bayesian approach has been suggested
Ethnicity [73], which recognizes that a gene selected at random has
Allele frequencies at many polymorphisms differ a lower prior probability of association than a candidate
between ethnic groups. If disease prevalence differs be- gene supported by biochemical, pathophysiological or
tween subgroups of differing ancestry, population admix- linkage data. Using this approach, the level of statistical
ture may result in differences in allele frequencies between significance required in a study of a gene with a moderate
cases and controls, which may be wrongly interpreted prior probability of association would be of the order
as indicating genetic susceptibility. Typing of multiple of 5 × 10−4 , whereas the equivalent value for a gene sel-
unlinked genetic markers has been advocated as a means ected at random would be 5 × 10−5 or lower. It is
of detecting and correcting for population stratification apparent that only large studies including hundreds of
[74], but these have not been applied to studies of pre- cases and controls are adequately powered to meet this
eclampsia. requirement. For example, testing of a biallelic poly-
The overwhelming majority of candidate gene studies morphism with a minor allele frequency of 0.3 would
in pre-eclampsia have investigated women of white require over 300 cases and 300 controls to achieve 80 %
Western European descent. Very few studies have in- power to detect a doubling of risk of disease associated
cluded black women, which is both surprising and re- with the minor allele at a significance level of 5 × 10−4 .
grettable in view of the high risk of pre-eclampsia in black To detect a genetic risk ratio as low as 1.3, which is not
women and the high incidence of eclampsia in Africa. unrealistic in complex disorders, the required number
of cases and controls is over 2000. From over 130 case-
Maternal or fetal genes? control studies identified by PubMed searching, only five
One challenge which is a specific feature of disorders of included more than 300 cases. Three yielded negative
pregnancy is that both maternal and fetal genes may play results [34,42,45]. One large study reported an association
a role, but it is not possible to distinguish between these with the platelet glycoprotein IIIa gene ITGB3 (P < 0.01)
using a case-control design; genes which are active in the [36], which has not yet been re-tested in an adequately
mother will also be over-represented in the fetus, and vice powered study. Another large study from Colombia
versa. In fact, remarkably few of the published molecular found an association with the eNOS gene NOS3
genetic studies of pre-eclampsia have investigated the fetal (P = 0.001) [46]. The mixed ethnic background of subjects
genotype. in this study is a limitation acknowledged by the authors,
who advocated the use of unlinked polymorphic markers
Genotype or haplotype? to assess confounding by ethnicity in future studies. Their
In the early years of molecular genetics, researchers positive findings were not replicated in a large study from
focused on previously reported polymorphisms, often the U.K. using TDT (transmission disequilibrium testing)
one or two per candidate gene, for which assays [35].
were readily available. More recently, initiatives such as
the HapMap project (http://www.hapmap.org/) and the Meta-analysis
establishment of the dbSNP database (http://www.ncbi. It has been argued that such stringent criteria may be
nlm.nih.gov/projects/SNP/) have identified multiple impractical and that combining the results of several
SNPs (single nucleotide polymorphisms) in each gene. smaller studies in a meta-analysis is an alternative power-
The combination of alleles at each polymorphic site deter- ful approach. One recent meta-analysis failed to find
mines the gene haplotype. any evidence for an increased risk of pre-eclampsia
Comprehensive genetic analysis demands a thorough associated with the MTHFR 677C > T variant {pooled
evaluation of haplotypic associations with disease odds ratio, 1.01 (95 % CI, 0.79–1.29)] [76]; an earlier
through testing of numerous polymorphisms in each gene meta-analysis suggested that 677C > T may be associ-
[75]. Technologies are now available to meet the require- ated with severe pre-eclampsia only [diastolic blood

C 2006 The Biochemical Society
 Searching for genetic clues to the causes of pre-eclampsia 449

pressure
110 mm Hg; odds ratio, 1.41 (95 % CI, 1.03– eclampsia for any of the genes tested. The CIs were
1.73)] [77]. Meta-analysis of prothrombin 20210G > A narrow and suggested that these genes contribute little
did not support a role for this polymorphism in pre- or nothing to the risk of pre-eclampsia.
eclampsia [odds ratio, 1.37 (95 % CI, 0.72–2.57)] [76].
Factor V Leiden was associated with an approx. 2-fold
increase in risk of pre-eclampsia in all of three Immunogenetics
meta-analyses incorporating many of the same studies The fetus expresses both paternal and maternal genes,
[76,78,79]. Some important general points emerged from and mechanisms have evolved to protect the fetal semi-
these meta-analyses. There was considerable variation in allograft from immune rejection by the mother. There has
the recruitment protocols for case-control studies and been speculation that an inappropriate maternal immune
in the phenotypic profile of affected women. Further- response might lead to impaired trophoblast invasion and
more, there was statistical evidence of heterogeneity pre-eclampsia. Studies of maternal–fetal immunological
between the results of studies; in particular, large studies interactions must take into account the unusual repertoire
and studies published within the last few years tended to of HLA antigens expressed by trophoblastic cells.
yield negative results, whereas the majority of positive Trophoblasts at the maternal fetal junction do not express
results came from earlier smaller studies. classical HLA antigens (HLA-A, -B and -DR), rather
they express a distinctive repertoire of HLA-C, HLA-E
and HLA-G antigens. HLA-G has attracted particular
The GOPEC (Genetics Of Pre-EClampsia) attention in the investigation of pre-eclampsia and other
study pregnancy-related disorders, as HLA-G expression may
The U.K. GOPEC study was designed to address some protect trophoblast cells from NK (natural killer)-
of the problems inherent in genetic studies of pre-eclamp- mediated cell lysis [81].
sia [35]. Involving a consortium of researchers from ten Studies of HLA antigens in pre-eclampsia, which
U.K. universities, GOPEC recruited over 1000 women have been comprehensively reviewed recently [82], have
affected by pre-eclampsia and their families. Importantly, yielded conflicting results. Such studies are inherently
all were recruited at the time of their illness using rigo- difficult to perform; the extensive polymorphism within
rously defined criteria. A TDT approach to genetic ana- HLA-A, -B and -DR classes results in many small sub-
lysis was planned [80]. TDT evaluates the transmission groups and loss of statistical power. In a review of
of allelic variants from heterozygous parents; under the this topic, Saftlas and co-workers [82] highlight the
null hypothesis of no association, the expected frequency urgent need for large studies examining interactions
of transmission of either allele to an affected offspring is between maternal, paternal and fetal genotypes in order
50 %. Significant deviations from this expected frequency to adequately assess the role of HLA in pre-eclampsia
across affected families indicate the presence of a genetic [82].
variant associated with disease. By genotyping affected A novel hypothesis has been advanced in a recent study
women and their parents, it is possible to conduct TDT of focusing on fetal HLA-C genes and their interaction
the maternal genotype. Genotyping the affected woman, with maternal KIRs (killer Ig-like receptors) expressed
her partner and baby enables independent testing of the by specialized maternal uNK (uterine NK) cells at
fetal genotype. Furthermore, TDT is not affected by the maternal–fetal interface. uNK cells are believed to
unsuspected population admixture, in contrast with case- facilitate and regulate trophoblast invasion through the
control studies. As is increasingly the case in studies of secretion of cytokines. HLA-C are ligands for KIRs, and
complex disorders, the GOPEC study is designed to test binding results in activation or inhibition of cytokine
multiple polymorphisms in each candidate gene in order production depending on the repertoire of KIR genes
to capture the majority of haplotypic diversity within the possessed by the NK cell. KIR haplotypes are grouped
population. into haplotype A, which mediates predominantly inhibi-
The first phase of the GOPEC study sought to provide tory signals, and haplotype B, which contains additional
definitive results for candidate genes which have yielded KIR genes with an activating profile. HLA-C antigens
conflicting data from earlier studies [35]. These included are expressed by invasive trophoblasts and the different
AGT, AGTR1, AGTR2, NOS3, MTHFR, TNF and the alleles can conveniently be classified as C1 or C2 based
Factor V Leiden variant. A haplotyping approach was on their interaction with KIRs. A case-control study
adopted for all genes with the exceptions of the Factor V of mother–child pairs detected a significant increase
gene and MTHFR, where analysis was restricted to in maternal homozygosity for the KIR A (inhibitory)
variants of known functionality. Twenty-eight SNPs were haplotype in pre-eclamptic pregnancies where the fetus
analysed in 657 families, providing 85 % power to detect possessed one or two copies of an HLA-C2 allotype
a genotype relative risk of 1.6 at a significance level of [83]. This combination of maternal KIR and fetal HLA-
5 × 10−4 . There was no evidence in either maternal or C genotypes might possibly result in inhibition of uNK
fetal genotypes or haplotypes for association with pre- cytokine production, a mechanism that could explain the

C 2006 The Biochemical Society
450 S. Chappell and L. Morgan

Figure 3 Integrated approaches to pre-eclampsia research

impaired trophoblast invasion which can lead to pre- in genetic studies to date. A handful of genes involved in
eclampsia. antioxidant response mechanisms have been considered
(Table 1), but these have mostly been relatively small
studies of one or two polymorphisms per gene. There
CLUES TO NOVEL CANDIDATE GENES has, as yet, been no attempt to undertake comprehensive
haplotype screening of antioxidant pathways in a large
Where should researchers be looking for novel suscept-
candidate gene study.
ibility genes? Some important pathways suggested by the
disease pathophysiology remain relatively unexplored.
However, the identification of candidate genes from Linkage studies
biological hypotheses is limited by our current under- Until recently, genetic linkage analysis has been the only
standing of the pathways involved. Molecular biology is approach available to geneticists for screening the entire
now providing global screening strategies which generate genome for susceptibility loci. Linkage analysis has
biological hypotheses (Figure 3). Genome-wide scans proved its worth in identifying the molecular basis of
are vital weapons in the molecular genetics armoury, many Mendelian disorders, but has been less successful
and screening of pre-eclamptic pedigrees has pointed to in guiding the search for susceptibility genes for com-
several loci which merit fine mapping with a view to the plex disorders. This may be due to underlying genetic
identification of positional candidates. Rapidly develop- heterogeneity (susceptibility loci may differ between
ing genotyping technologies and the availability of affected pedigrees) or to the relatively low genetic risk
haplotype maps have made genome-wide screening using of disease conferred by individual genes in a complex
dense SNP markers a reality which could be applied mode of inheritance. Ascertainment of multicase families
to large carefully phenotyped case-control collections. presents difficulties in studies of pre-eclampsia, as there
Other molecular strategies which do not require a is no known male phenotype, and susceptibility in
priori aetiological hypotheses include the study of gene females is apparent only during pregnancy. Nevertheless,
expression profiles (transcriptomics) and proteomics. research groups from Iceland, Australia, Finland and
These techniques hold great promise, but have been the Netherlands have undertaken genome-wide linkage
applied only relatively recently to the study of pre- screens which have yielded encouraging results (Figure 4).
eclampsia and other disorders of pregnancy.
Chromosome 2
Clues from pathophysiology A study of 124 Icelandic pedigrees was the first to report
Current pathogenetic models of pre-eclampsia highlight linkage to chromosome 2p [84]. Within these families,
two mechanisms which merit detailed genetic analysis: affected women were classified as having mild disease
impaired placentation and oxidative stress [13]. Placenta- (gestational hypertension with no proteinuria) or severe
tion requires the orchestrated expression of a range disease (hypertension with proteinuria and/or seizures,
of proteinases, cytokines, angiogenic factors, adhesion or severe gestational hypertension). The Icelandic group
molecules and their receptors which have barely figured analysed the linkage data both with and without inclusion

C 2006 The Biochemical Society
 Searching for genetic clues to the causes of pre-eclampsia 451

Figure 4 Chromosomal loci linked to pre-eclampsia

Chromosomal position is given according to NCBI (National Center for Biotechnology Information) Build 35.1. LOD scores (red) or non-parametric linkage scores (blue)
are shown; the highest scores provide the strongest evidence of linkage. The results of linkage analysis using both strict diagnostic criteria (S, pre-eclampsia/eclampsia
only) and general diagnostic criteria (G, includes pregnancy-induced hypertension) are shown.

of women with mild pre-eclampsia, which they termed tiple SNPs within two positional candidate genes:
general and strict criteria respectively. This dual approach TACR1, which encodes the neurokinin B receptor, and
to pedigree analysis has been adopted in a number of TCF7L1, which encodes a transcription factor with a
published linkage studies of pre-eclampsia. Linkage to possible role in embryo implantation. There was no
the locus on chromosome 2p was apparent under both evidence that any of the SNPs tested was associated with
general and strict criteria, and was chiefly attributable pre-eclampsia.
to two large multicase pedigrees. When these families A genome-wide screen of 15 Finnish pedigrees [88]
were removed from the analysis, a different locus on has implicated a susceptibility locus on chromosome 2p
chromosome 2q demonstrated suggestive linkage under distinct from the loci identified by the Australian and
the general criteria. A subsequent study of 34 pedigrees Icelandic groups. As a possible explanation for these
from Australasia and New Zealand also demonstrated findings, the authors raise the interesting possibility
two linkage peaks on chromosome 2p and 2q close to that members of the same gene family with overlapping
the regions identified in the Icelandic study [85]. It is functions, arising from historical gene duplication events,
possible that these two studies have identified identical may co-exist on chromosome 2.
loci in the pericentromeric region of chromosome 2,
strengthening the case for a susceptibility gene in this Chromosome 4
region. Interestingly, suggestive linkage to the 2q region A study of 15 Australian pedigrees identified a locus with
was reported in a genome-wide screen of British families suggestive linkage on chromosome 4q [89]. A subsequent
affected by essential hypertension [86], a reminder that genome-wide scan of 34 families [85], including those
gestational hypertension and non-pregnancy hyperten- screened in the earlier report [89], used a denser
sion may share some inherited features. panel of markers. This implicated chromosome 2, as
Bioinformatics databases record over 400 genes, in- discussed above, but linkage to chromosome 4q was
cluding many of unknown function, within the 2p peak less significant, and there was no evidence of linkage
region and over 100 at the 2q peak, indicating the to this region in the Icelandic pedigrees. More recently,
magnitude of the challenge to identify individual suscept- the genome-wide screen of Finnish pedigrees [88] has
ibility genes [87]. The Australian group tested mul- demonstrated suggestive linkage to chromosome 4q, very

C 2006 The Biochemical Society
452 S. Chappell and L. Morgan

close to the linked markers identified in the Australian non-invasive polyploid syncytiotrophoblast. A common
study. tyrosine > histidine polymorphism at codon 153 in
the DNA-binding domain of STOX1 was consistently
Chromosome 9 shared by affected sisters in pre-eclamptic pedigrees, and
The Finnish study demonstrated linkage to chromosome importantly was almost invariably transmitted to the
9p under both the general and strict diagnostic criteria child born of the affected pregnancy. As the tropho-
[88]. In view of the increased risk of pre-eclampsia as- blast is fetal tissue, the possession of a mutant mater-
sociated with diabetes, it is of interest that linkage to this nally derived STOX1 allele, together with a silenced
region of chromosome 9p has also been demonstrated paternally imprinted allele, would lead to loss of function.
in Finnish families affected by Type II diabetes [90]. Two The number of informative pregnancies in this fascin-
other markers on chromosome 9q displayed weak linkage ating study was limited, but if the findings can be
under the general diagnostic criteria only. independently confirmed they have far-reaching impli-
cations for research into the pathogenesis of pre-
Chromosome 10 eclampsia. Interestingly, a paralogue of the gene, STOX2,
A genome-wide screen undertaken on Dutch women is located on chromosome 4q close to the suggestive
yielded evidence for linkage with pre-eclampsia on region identified in genome-wide screens in Australia and
chromosome 10q [91]. Interestingly, in families affected Finland, and merits further investigation.
by pre-eclampsia associated with the HELLP syndrome,
there was suggestive evidence for a separate susceptibility Transcriptomics
locus on chromosome 12q. In order to reduce the 132 Microarray technology is widely used to study the
positional candidate genes within the linked locus on expression patterns of thousands of genes simultaneously.
chromosome 10q to a manageable number, this group Investigating the expression of genes in both diseased
adopted an approach which illustrates the combined and healthy tissue may provide clues to the cause of the
power of bioinformatics and integrative biology. They disease and allow further research to focus on specific
hypothesized that pre-eclampsia may be associated with genes or pathways. Although this may be reasonably
genetic imprinting. Re-analysis of their data under this straightforward for diseases such as cancer, where samples
model confirmed excess sharing of maternally inherited of affected and healthy tissue can be taken from the
alleles at the 10q locus between sisters affected by pre- same patient, conditions such as pre-eclampsia are more
eclampsia, suggesting a paternally imprinted maternally difficult, since variation between individuals has also to
active susceptibility gene [92]. They subjected the be taken into account. A PubMed search identified nine
132 genes at the 10q locus to homology (BLAST) studies published between 2002 and 2005 which have
searches to detect DNA sequence features characteristic used microarrays to compare placental gene expression
of imprinted genes. As a further strategy to identify in pre-eclamptic and normotensive pregnancy [93–101].
the most likely positional candidate genes, homology However, the lists of genes produced by each study show
searches were conducted between the 10q locus and the a discouraging lack of overlap.
genomic regions on chromosomes 2p and 9p, which The first problem with using this technology for the
have demonstrated significant linkage in genome-wide study of pre-eclampsia is tissue availability. It is clear that
screens, on the assumption that similar genes on different the placenta plays an integral role in the disease, but the
chromosomes may cause similar disease phenotypes in disease is established early in pregnancy with a failure
different populations. Fifty-five genes met one or more of trophoblast invasion into the uterine tissue. Since
of the specified criteria for further investigation. To microarray studies use placental tissue which is obtained
test whether these genes are imprinted in trophoblast, following delivery, usually early in the third trimester,
mRNA expression was compared between first trimester it is difficult to establish whether changes in gene
placental tissue from terminated pregnancies and tissue expression are a cause or a consequence of the disease.
obtained from first trimester hydatidiform molar preg- A related problem is obtaining control tissue matched
nancies, which contain solely paternally derived nuclear for gestational age, as patients with pre-eclampsia often
DNA. Two clusters of genes were shown to be tran- require premature delivery. Although it is possible to
scribed in normal trophoblast, but not in andro- obtain samples from normotensive pregnancies which
genetic placentas, implying preferential expression of deliver before term, it is possible that the factors which
maternally derived genes [92]. DNA sequencing of 17 resulted in premature delivery might themselves alter
genes identified 55 SNPs and enabled the group to gene expression. For example, in one microarray study
identify a minimum critical region of allele sharing be- three out of six controls delivered early due to premature
tween affected sisters, which included the paternally rupture of membranes, which is often associated with
imprinted gene STOX1 [26]. This encodes a putative infection. This may lead to an increased expression of
DNA-binding protein, which appears to function in the genes involved in inflammation or the immune response.
transition of invasive diploid extravillous trophoblast into As it seems likely that onset of labour will affect gene

C 2006 The Biochemical Society
 Searching for genetic clues to the causes of pre-eclampsia 453

expression, cases and controls should be matched for pression patterns of genes shown in Table 2 were similar
mode of delivery. The majority of studies included in in at least two independent studies. Some encode pro-
the present review used samples obtained following teins which have been associated previously with pre-
Caesarean section, with one exception, which used both eclampsia; however, two of the genes in Table 2 have
pre-eclamptic and control samples obtained after vaginal not been implicated previously in pre-eclampsia. TNF
delivery [97]. receptor (ligand) superfamily member 10 [TRAIL (TNF-
There has been variation in the microarray design used related apoptosis-inducing ligand)] is expressed at a
for expression studies of pre-eclampsia. Two types of significant level in most normal tissues. It has been
array have been used: cDNA arrays [93–98,100,101] and shown to preferentially induce apoptosis in transformed
oligonucleotide arrays [99]. This terminology refers to and tumour cells, but does not appear to kill normal
the type of DNA probe that is immobilized on to the cells (reviewed in [104]). AXL receptor tyrosine kinase
solid support. cDNA arrays have spots of cDNA ranging is a member of the receptor tyrosine kinase subfamily.
from 500–5000 bp, whereas the oligonucleotide arrays use Members of this subfamily play an important role
shorter molecules of around 15–70 bp. Furthermore, the in cellular proliferation and differentiation, and may
number of genes represented on an array has ranged from mediate the effects of Ang II (angiotensin II) via the AT1
a selected group, e.g. cytokines [93], to much larger chips, (Ang II type 1) receptor, although the exact mechanisms
such as the Affymetrix U133A array which contains are unclear [105].
14 500 well-characterized genes [99]. Lack of replication
of results between studies may therefore be due to the Proteomics
absence of a gene from the array used in some studies. Evolving MS-based technologies now facilitate screening
A third variable is the method of labelling and detection. and identification of multiple proteins in tissues and
Many studies have used radioactive labelling to detect biofluids. The application of these techniques to disorders
hybridization [97,98,101] which mandates hybridization of pregnancy is in its early stages [115] and a detailed
of samples from pre-eclamptic and normal patients to discussion of these methodologies is beyond the scope
separate copies of the array. Labelling of samples from of this review, but the integration of proteomic and
cases and controls with different fluorescent dyes enables genetic approaches to the study of pre-eclampsia holds
hybridization of both to a single array, reducing the much promise. The spatial and temporal expression
experimental variability [93–96,99,100]. patterns of molecules at the maternal–fetal interface in
Pooling of RNA from a number of subjects is early pregnancy will be of particular interest. Screening
commonly used in hybridization experiments to reduce of maternal plasma collected prospectively from early
costs, to compensate for low amounts of starting material pregnancy may detect biomarkers which predate the
or to reduce variation between individual samples. onset of clinical pre-eclampsia. Genetic testing of novel
Pooling of all samples from one group has been used in candidates suggested by the results of proteomic screen-
a number of studies [93–96,100,101], an approach which ing will be a powerful complementary strategy in the
cannot provide an estimate of variability among arrays. identification of causative mechanisms of disease.
Sub-pooling, the pooling and hybridization of multiple
subsets of samples selected at random, is preferable, as Transgenic animal models
the variation between arrays can be considered, although There is no naturally occurring animal model of pre-
obviously a larger number of patient samples will be eclampsia; however, recent advances have enabled the use
required [103]. Pooling RNA from different subjects of transgenic methods to study the impact of specific
results in a permanent loss of information and has also genes.
been shown to result in a loss of sensitivity and an One of the first transgenic models of pre-eclampsia
increase in false-positive results. In spite of the financial arose from the generation of mice expressing human
implications, hybridization of individual RNA samples angiotensinogen or human renin [116]. When female mice
and pooling of results for statistical analysis was the carrying the human angiotensinogen gene were mated
approach used in two studies [97,98]. with males carrying the human renin gene, the females
Although it is now widely accepted that microarray developed hypertension and proteinuria which resolved
experiments should be replicated in independent experi- after delivery. Pregnant females derived from all other
ments to confirm changes in gene expression, it appears mating combinations did not develop these symptoms.
that the majority of experiments using placental tissue In affected females, renin was localized to the trophoblast
have been performed only once or repeated using the cells in the placenta and was secreted into the maternal
same samples and the same membrane-based array [101]. circulation. It was suggested that the secretion of placental
After considering all the differences between the nine (paternal) human renin contributes to the regulation
array experiments included in the present review, it is of maternal blood pressure, providing the female was
disappointing, but perhaps not surprising, that there expressing human angiotensinogen. This experiment has
is so little agreement between them. However, the ex- also been carried out in rats with similar results [117].

C 2006 The Biochemical Society
454 S. Chappell and L. Morgan

Table 2 Genes showing altered placental expression in pre-eclampsia in at least two independent studies
IFN, interferon; LFA-1: the integrin lymphocyte function-associated antigen 1; hCG, human chorionic gonadothropin; FSH, follicle-stimulating hormone; LH, luteinizing
hormone; TSH, thyroid-stimulating hormone; TRAIL, TNF-related apoptosis-inducing ligand.

Fold increase in
Genes with altered expression in expression in
pre-eclampsia Gene symbol pre-eclampsia Previous association with pre-eclampsia

Intercellular adhesion molecule 1 ICAM1 21.2 [94] Serum level increased prior to onset of clinical symptoms [106].
2.8 [98]
Vascular cell-adhesion molecule 1 VCAM1 39.6 [94] Serum level increased prior to onset of clinical symptoms [106].
2.7 [98]
Integrin α 1 ITGA1 43.7 [100] Expected up-regulation of α 1 β 1 integrin is not seen in
16.2 [94] pre-eclampsia [107]. This contrasts with the array results, but the
α 1 integrin monomer could be forming other dimers.
Integrin α L ITGAL 1.5 [94] LFA-1 has α L and β 1 subunits. Higher levels of LFA-1 observed
2.7 [98] within the uteroplacental circulation compared with peripheral
venous blood in pre-eclamptic patients [108].
Glycoprotein hormones α polypeptide CGA 2.1 [101] Subunit of hCG, FSH, LH, TSH. Serum hCG concentration is
22.7 [98] increased in pre-eclampsia [109].
Inhibin β A INHBA 4.9 [100] Subunit of inhibin A and activin A. Serum concentrations of
2.4 [93] inhibin A (αβ A ) and activin A(β A β A ) are increased in
pre-eclampsia [111].
IFNγ receptor 2 IFNGR2 6.9 [93] Normal up-regulation of the receptor does not occur in
2.2 [98] pre-eclampsia [112] in contrast with the microarray results.
Secreted phosphoprotein 1 (osteopontin) SPP1 1.6 [101] Increased production of osteopontin in cytotrophoblasts from
9.4 [95] pre-eclamptic placentas [113].
Prostaglandin I2 (prostacyclin) synthase PTGIS 1.9 [96] Reduced placental prostacyclin secretion in pre-eclampsia
2.4 [98] antedates clinical disease [114].
TNF-receptor (ligand) superfamily member 10 TNFSF10 3.2 [93] Novel candidate.
(TRAIL) 3.6 [98]
AXL receptor tyrosine kinase AXL 4.1 [100] Novel candidate.
8.3 [95]

In addition to developing hypertension and proteinuria, be minimal. Although there is a modest reduction of
the female rats also produce auto-antibodies which bind fetal weight early in pregnancy in transgenic fetuses, by
to and activate the AT1 receptor. Women with pre- late pregnancy the wild-type and transgenic fetuses had
eclampsia produce anti-(AT1 receptor) antibodies [118] identical growth characteristics. In addition, placental
and it has been suggested that these may play a role in the growth was similar for both throughout gestation.
pathogenesis of pre-eclampsia, making this a potentially However, maternal IGFBP-1 genotype appears to have
valuable animal model. some effect on placental growth, with a greater placental
Transgenic mice have also been used to assess the weight in transgenic females at all time points irrespective
contribution of the IGFBP-1 [IGF (insulin-like growth of fetal genotype. In addition, trophoblast invasion is
factor)-binding protein-1] to placental development and less pronounced in transgenic females compared with
intrauterine growth restriction [119]. These mice can wild-type. Fetal growth in early pregnancy is impaired
be used to gain a better understanding of the factors in transgenic females, but growth is similar in both wild-
important for correct placentation, which is thought to type and transgenic females by later time points. There
be dysfunctional in pre-eclampsia. The contribution of also appears to be an increase in fetal death from matings
fetal genotype to placental and fetal development was involving a transgenic female.
assessed by studying the characteristics of transgenic The suggestion that pre-eclampsia may be caused by
and wild-type fetuses derived from a female wild-type a paternally imprinted gene [25] is supported by the
mouse. Similarly, the effect of maternal genotype was results from mice deficient in p57Kip2 , which is a paternally
studied using the characteristics of wild-type fetuses imprinted gene in both humans and mice, i.e. only
derived from both wild-type and transgenic females. expressed from the maternally derived chromosome. A
These experiments revealed that the fetal IGFBP-1 decrease in expression has been associated with abnormal
contribution to placental and fetal growth appears to proliferation of trophoblast cells [120], which may be

C 2006 The Biochemical Society
 Searching for genetic clues to the causes of pre-eclampsia 455

relevant to pre-eclampsia. Heterozygous female mice to play in finding genetic clues to the aetiology of pre-
that were mated with heterozygous males showed signs eclampsia; truly a task for dedicated teamwork. The goal
of pre-eclampsia, including hypertension, proteinuria is the identification of pathophysiological mechanisms
and increased endothelin levels in late pregnancy. These which will inform future strategies for the prevention
features were also noted in heterozygous females which and treatment of this devastating condition.
were mated with wild-type males, but not in wild-type
females mated with either heterozygous or wild-
type males, i.e. symptoms were only seen in females ACKNOWLEDGMENTS
carrying p57Kip2 -deficient pups regardless of their own
L. M. is the clinical co-ordinator of the GOPEC study,
p57Kip2 -expression status. This indicates that the fetal
funded by the British Heart Foundation.
genotype may play a role in the development of pre-
eclampsia.
The results of animal models reveal that symptoms REFERENCES
similar to those exhibited by women with pre-eclampsia
can be initiated either by maternal factors (IGFBP-1 1 Lie, R. T., Rasmussen, S., Brunborg, H., Gjessing, H. K.,
Lie-Nielsen, E. and Irgens, L. M. (1998) Fetal and
mice), by factors originating in the fetus (p57Kip2 mice) maternal contributions to risk of pre-eclampsia:
or by an interaction of both maternal and fetal factors population based study. Br. Med. J. 316, 1343–1347
(renin/angiotensinogen rodents). If the same holds true 2 Altman, D., Carroli, G., Duley, L. et al. (2002) Do women
with pre-eclampsia, and their babies, benefit from
for humans, with patterns of genetic influences differing magnesium sulphate? The Magpie Trial: a randomised
between pedigrees, this will complicate the analysis of the placebo-controlled trial. Lancet 359, 1877–1890
3 Lachmeijer, A. M. A., Dekker, G. A., Pals, G., Aarnoudse,
underlying genetics of the disorder further. J. G., ten Kate, L. P. and Arngrimsson, R. (2002)
Searching for preeclampsia genes: the current position.
Eur. J. Obstet. Gyn. 105, 94–113
4 Wilson, M. L., Goodwin, T. M., Pan, V. L. and Ingles,
CONCLUSIONS S. A. (2003) Molecular epidemiology of preeclampsia.
Obstet. Gynecol. Surv. 58, 39–66
5 Brown, M. A., Lindheimer, M. D., de Swiet, M.,
A major challenge in the study of pre-eclampsia is Van Assche, A. and Moutquin, J. M. (2001) The
to disentangle the causes of the disease from its con- classification and diagnosis of the hypertensive disorders
of pregnancy: statement from the International Society
sequences. In this respect, genetic studies are particularly for the Study of Hypertension in Pregnancy (ISSHP).
valuable, as genetic variation may be the cause of a Hypertens. Pregnancy 20, IX–XIV
6 Weinstein, L. (1982) Syndrome of hemolysis, elevated
pathophysiology, but not its consequence. Genetics has liver enzymes, and low platelet count: a severe
a dual role in the search for disease mechanisms: genome- consequence of hypertension in pregnancy. Am. J.
Obstet. Gynecol. 142, 159–167
wide screens can suggest novel hypotheses, and candidate 7 Walker, J. J. (2000) Pre-eclampsia. Lancet 356, 1260–1265
gene studies can be used to test hypotheses suggested by 8 Rey, E. and Couturier, A. (1994) The prognosis of
pathophysiology or global screening strategies. Recent pregnancy in women with chronic hypertension. Am. J.
Obstet. Gynecol. 171, 410–416
genome-wide screens of families affected by pre- 9 Garner, P. R., D’Alton, M. E., Dudley, D. K., Huard, P.
eclampsia have yielded encouraging results. By contrast, and Hardie, M. (1990) Preeclampsia in diabetic
pregnancies. Am. J. Obstet. Gynecol. 163, 505–508
the credibility of candidate gene studies has been under- 10 Jones, D. C. and Hayslett, J. P. (1996) Outcome of
mined by conflicting and inconclusive results. It is ap- pregnancy in women with moderate or severe renal
insufficiency. N. Engl. J. Med. 335, 226–232
parent that larger studies of adequate statistical power to 11 Barden, A. E., Beilin, L. J., Ritchie, J., Walters, B. N. and
detect small genetic effects are needed to reliably identify Michael, C. (1999) Does a predisposition to the metabolic
syndrome sensitize women to develop pre-eclampsia?
or exclude susceptibility genes. This invites a multicentre J. Hypertens. 17, 1307–1315
collaborative approach between clinicians and geneticists 12 Seely, E. W. (1999) Hypertension in pregnancy: a
to develop common recruitment protocols which will potential window into long-term cardiovascular risk in
women. J. Clin. Endocrinol. Metab. 84, 1858–1861
lead to the establishment of large DNA resources, 13 Roberts, J. M. and Cooper, D. W. (2001) Pathogenesis and
or at least enable statisticians to conduct meaningful genetics of pre-eclampsia. Lancet 357, 53–56
14 Pijnenborg, R., Dixon, G., Robertson, W. B. and
meta-analyses. Continuing epidemiological research in- Brosens, I. (1980) Trophoblastic invasion of human
dicates that both the maternal and fetal genotypes are decidua from 8 to 18 weeks of pregnancy. Placenta 1, 3–19
15 Redman, C. W. and Sargent, I. L. (2000) Placental debris,
valid targets for study. The adoption of current global oxidative stress and pre-eclampsia. Placenta 21, 597–602
screening technologies will challenge bioinformaticians 16 Knight, M., Redman, C. W., Linton, E. A. and Sargent,
to produce software to interpret and integrate massive I. L. (1998) Shedding of syncytiotrophoblast microvilli
into the maternal circulation in pre-eclamptic
amounts of data. The role of the statistical geneticist is pregnancies. Br. J. Obstet. Gynaecol. 105, 632–640
vital in developing sophisticated analytical tools which 17 Wolf, M., Kettyle, E., Sandler, L., Ecker, J. L., Roberts, J.
and Thadhani, R. (2001) Obesity and preeclampsia: the
will tackle the complexities of haplotype estimation, potential role of inflammation. Obstet. Gynecol. 98,
gene–gene interactions and the interplay of fetal and ma- 757–762
18 Rumbold, A., Duley, L., Crowther, C. and Haslam, R.
ternal genotypes. Clinicians, geneticists, epidemiologists, (2005) Antioxidants for preventing pre-eclampsia.
bioinformaticians and statisticians therefore all have a part Cochrane Database Syst. Rev. CD004227

C 2006 The Biochemical Society
456 S. Chappell and L. Morgan

19 Maynard, S. E., Min, J. Y., Merchan, J. et al. (2003) Excess 37 Pegoraro, R. J., Chikosi, A., Rom, L., Roberts, C. and
placental soluble fms-like tyrosine kinase 1 (sFlt1) may Moodley, J. (2004) Methylenetetrahydrofolate reductase
contribute to endothelial dysfunction, hypertension, gene polymorphisms in black South Africans and the
and proteinuria in preeclampsia. J. Clin. Invest. 111, association with preeclampsia. Acta Obstet.
649–658 Gynecol. Scand. 83, 449–454
20 Chesley, L. C., Annitto, J. E. and Cosgrove, R. A. (1968) 38 Levesque, S., Moutquin, J. M., Lindsay, C., Roy, M. C.
The familial factor in toxemia of pregnancy. and Rousseau, F. (2004) Implication of an AGT haplotype
Obstet. Gynecol. 32, 303–311 in a multigene association study with pregnancy
21 Sutherland, A., Cooper, D. W., Howie, P. W., Liston, hypertension. Hypertension 43, 71–78
W. A. and MacGillivray, I. (1981) The incidence of severe 39 Laasanen, J., Hiltunen, M., Punnonen, K., Mannermaa, A.
pre-eclampsia amongst mothers and mothers-in-law of and Heinonen, S. (2002) Fibrinogen and factor VII
pre-eclamptics and controls. Br. J. Obstet. Gynaecol. 88, promoter polymorphisms in women with preeclampsia.
785–791 Obstet. Gynecol. 100, 317–320
22 Arngrimsson, R., Bjornsson, S., Geirsson, R. T., 40 Borg, A. J., Higgins, J. R., Brennecke, S. P. and Moses,
Bjornsson, H., Walker, J. J. and Snaedal, G. (1990) E. K. (2002) Thrombomodulin Ala455Val dimorphism is
Genetic and familial predisposition to eclampsia and not associated with pre-eclampsia in Australian and New
pre-eclampsia in a defined population. Br. J. Zealand women. Gynecol. Obstet. Invest. 54, 43–45
Obstet. Gynaecol. 97, 762–769 41 Clark, P., Freeman, D. J., Streja, E., Sattar, N., Walker,
23 Cincotta, R. B. and Brennecke, S. P. (1998) Family history I. D. and Greer, I. A. (2003) The G-to-T point mutation
of pre-eclampsia as a predictor for pre-eclampsia in in codon 34 of the factor XIII gene and the risk of
primigravidas. Int. J. Gynaecol. Obstet. 60, 23–27 pre-eclampsia. Blood Coagulation Fibrinolysis 14,
24 Folgero, T., Storbakk, N., Torbergsen, T. and Oian, P. 155–157
(1996) Mutations in mitochondrial transfer ribonucleic 42 Roberts, C. B., Rom, L., Moodley, J. and Pegoraro, R. J.
acid genes in preeclampsia. Am. J. Obstet. Gynecol. 174, (2004) Hypertension-related gene polymorphisms in
1626–1630 pre-eclampsia, eclampsia and gestational hypertension in
25 Graves, J. A. M. (1998) Genomic imprinting, development Black South African women. J. Hypertens. 22, 945–948
and disease – is pre-eclampsia caused by a maternally 43 Maruyama, A., Nakayama, T., Furuya, K., Mizutani, Y.
imprinted gene? Reprod. Fertility Develop. 10, 23–29 and Yamamoto, T. (2005) Association study between
26 van Dijk, M., Mulders, J., Poutsma, A. et al. (2005) the human Renin gene and preeclampsia.
Maternal segregation of the Dutch preeclampsia locus at Hypertens. Pregnancy 24, 39–48
10q22 with a new member of the winged helix gene 44 Plummer, S., Tower, C., Alonso, P. et al. (2004)
family. Nat. Genet. 37, 514–519 Haplotypes of the angiotensin II receptor genes AGTR1
27 Thornton, J. G. and Macdonald, A. M. (1999) Twin and AGTR2 in women with normotensive pregnancy and
mothers, pregnancy hypertension and pre-eclampsia. women with preeclampsia. Hum. Mutat. 24, 14–20
Br. J. Obstet. Gynaecol. 106, 570–575 45 Pegoraro, R. J., Roberts, C. B., Rom, L. and Moodley, J.
28 Salonen Ros, H., Lichtenstein, P., Lipworth, L. and (2004) T594M mutation of the epithelial sodium channel
Cnattingius, S. (2000) Genetic effects on the liability of beta-subunit gene in pre-eclampsia and eclampsia in Black
developing pre-eclampsia and gestational hypertension. South African women. Br. J. Obstet. Gynaecol. 111,
Am. J. Med. Genet. 91, 256–260 1012–1013
29 Esplin, M. S., Fausett, M. B., Fraser, A. et al. (2001) 46 Serrano, N. C., Casas, J. P., Diaz, L. A. et al. (2004)
Paternal and maternal components of the predisposition Endothelial NO synthase genotype and risk of
to preeclampsia. N. Engl. J. Med. 344, 867–872 preeclampsia: a multicenter case-control study.
30 Skjaerven, R., Vatten, L. J., Wilcox, A. J., Ronning, T., Hypertension 44, 702–707
Irgens, L. M. and Lie, R. T. (2005) Recurrence of 47 Barden, A. E., Herbison, C. E., Beilin, L. J., Michael,
pre-eclampsia across generations: exploring fetal and C. A., Walters, B. N. and Van Bockxmeer, F. M. (2001)
maternal genetic components in a population based Association between the endothelin-1 gene Lys198Asn
cohort. Br. Med. J. 331, 877 polymorphism blood pressure and plasma endothelin-1
31 Cnattingius, S., Reilly, M., Pawitan, Y. and levels in normal and pre-eclamptic pregnancy.
Lichtenstein, P. (2004) Maternal and fetal genetic factors J. Hypertens. 19, 1775–1782
account for most of familial aggregation of preeclampsia: 48 Akbar, F., Heinonen, S., Pirskanen, M., Uimari, P.,
a population-based Swedish cohort study. Am. J. Med. Tuomainen, T. P. and Salonen, J. T. (2005) Haplotypic
Genet. A 130, 365–371 association of DDAH1 with susceptibility to
32 O’Shaughnessy, K. M., Fu, B., Ferraro, F., Lewis, I., pre-eclampsia. Mol. Hum. Reprod. 11, 73–77
Downing, S. and Morris, N. H. (1999) Factor V Leiden 49 Jansen, M. W., Bertina, R. M., Vos, H. L. et al. (2004)
and thermolabile methylenetetrahydrofolate reductase C825T polymorphism in the human G protein β3 subunit
gene variants in an East Anglian preeclampsia cohort. gene and preeclampsia; a case control study.
Hypertension 33, 1338–1341 Hypertens. Pregnancy 23, 211–218
33 Kim, Y. J., Williamson, R. A., Murray, J. C. et al. (2001) 50 Lachmeijer, A. M. A., Crusius, J. B. A., Pals, G., Dekker,
Genetic susceptibility to preeclampsia: roles of G. A., Arngrimsson, R. and ten Kate, L. P. (2001)
cytosine-to-thymine substitution at nucleotide 677 of the Polymorphisms in the tumor necrosis factor and
gene for methylenetetrahydrofolate reductase, 68-base lymphotoxin-α gene region and preeclampsia.
pair insertion at nucleotide 844 of the gene for Obstet. Gynecol. 98, 612–619
cystathionine β-synthase and factor V Leiden mutation. 51 Bermingham, J., Jenkins, D., McCarthy, T. and
Am. J. Obstet. Gynecol. 184, 1211–1217 O’Brien, M. (2000) Genetic analysis of insulin-like
34 Morrison, E. R., Miedzybrodzka, Z. H., Campbell, D. M. growth factor II and HLA-G in pre-eclampsia.
et al. (2002) Prothrombotic genotypes are not associated Biochem. Soc. Trans. 28, 215–219
with pre-eclampsia and gestational hypertension: results 52 Haggerty, C. L., Ferrell, R. E., Hubel, C. A.,
from a large population-based study and systematic Markovic, N., Harger, G. and Ness, R. B. (2005)
review. Thromb. Haemostasis 87, 779–785 Association between allelic variants in cytokine genes and
35 GOPEC Consortium. (2005) Disentangling fetal and preeclampsia. Am. J. Obstet. Gynecol. 193, 209–215
maternal susceptibility for pre-eclampsia: a British 53 Lachmeijer, A. M., Nosti-Escanilla, M. P., Bastiaans, E. B.
multicenter candidate-gene study. Am. J. Hum. Genet. 77, et al. (2002) Linkage and association studies of IL1B and
127–131 IL1RN gene polymorphisms in preeclampsia.
36 O’Shaughnessy, K. M., Fu, B., Downing, S. and Morris, Hypertens. Pregnancy 21, 23–38
N. H. (2001) Thrombophilic polymorphisms in 54 Faisel, F., Romppanen, E. L., Hiltunen, M. et al. (2003)
pre-eclampsia: altered frequency of the functional Polymorphism in the interleukin 1 receptor antagonist
98C > T polymorphism of glycoprotein IIIa. gene in women with preeclampsia. J. Reprod. Immunol.
J. Med. Genet. 38, 775–777 60, 61–70

C 2006 The Biochemical Society
 Searching for genetic clues to the causes of pre-eclampsia 457

55 de Groot, C. J., Jansen, M. W., Bertina, R. M., 74 Pritchard, J. K., Stephens, M., Rosenberg, N. A. and
Schonkeren, J. J., Helmerhorst, F. M. and Huizinga, T. W. Donnelly, P. (2000) Association mapping in structured
(2004) Interleukin 10 −2849AA genotype protects against populations. Am. J. Hum. Genet. 67, 170–181
pre-eclampsia. Genes Immun. 5, 313–314 75 Morris, R. W. and Kaplan, N. L. (2002) On the advantage
56 Samsami Dehaghani, A., Doroudchi, M., Kalantari, T., of haplotype analysis in the presence of multiple disease
Pezeshki, A. M. and Ghaderi, A. (2005) Heterozygosity susceptibility alleles. Genet. Epidemiol. 23, 221–233
in CTLA-4 gene and severe preeclampsia. Int. J. 76 Lin, J. and August, P. (2005) Genetic thrombophilias and
Gynaecol. Obstet. 88, 19–24 preeclampsia: a meta-analysis. Obstet. Gynecol. 105,
57 Sziller, I., Nguyen, D., Halmos, A., Hupuczi, P., Papp, Z. 182–192
and Witkin, S. S. (2005) An A > G polymorphism at 77 Kosmas, I. P., Tatsioni, A. and Ioannidis, J. P. (2004)
position-670 in the Fas (TNFRSF6) gene in pregnant Association of C677T polymorphism in the
women with pre-eclampsia and intrauterine growth methylenetetrahydrofolate reductase gene with
restriction. Mol. Hum. Reprod. 11, 207–210 hypertension in pregnancy and pre-eclampsia:
58 Zusterzeel, P. L. M., Peters, W. H. M., Visser, W., a meta-analysis. J. Hypertens. 22, 1655–1662
Hermsen, K. J. M., Roelofs, H. M. J. and Steegers, E. A. P. 78 Kosmas, I. P., Tatsioni, A. and Ioannidis, J. P. A. (2003)
(2001) A polymorphism in the gene for microsomal Association of Leiden mutation in Factor V gene with
epoxide hydrolase is associated with pre-eclampsia. hypertension in pregnancy and pre-eclampsia: a meta-
J. Med. Genet. 38, 234–237 analysis. J. Hypertens. 21, 1221–1228
59 Zusterzeel, P. L., Visser, W., Peters, W. H., Merkus, H. W., 79 Dudding, T. E. and Attia, J. (2004) The association
Nelen, W. L. and Steegers, E. A. (2000) Polymorphism in between adverse pregnancy outcomes and maternal
the glutathione S-transferase P1 gene and risk for factor V Leiden genotype: a meta-analysis.
preeclampsia. Obstet. Gynecol. 96, 50–54 Thromb. Haemostasis 91, 700–711
60 Kim, Y. J., Park, H. S., Park, M. H., Suh, S. H. and Pang, 80 Spielman, R. S., McGinnis, R. E. and Ewens, W. J. (1993)
M. G. (2005) Oxidative stress-related gene polymorphism Transmission test for linkage disequilibrium: the insulin
and the risk of preeclampsia. Eur. J. Obstet. Gynecol. gene region and insulin-dependent diabetes mellitus
Reprod. Biol. 119, 42–46 (IDDM). Am. J. Hum. Genet. 52, 506–516
61 Raijmakers, M. T., Roes, E. M., te Morsche, R. H., 81 Pazmany, L., Mandelboim, O., Vales-Gomez, M., Davis,
Steegers, E. A. and Peters, W. H. (2003) Haptoglobin and D. M., Reyburn, H. T. and Strominger, J. L. (1996)
its association with the HELLP syndrome. J. Med. Genet. Protection from natural killer cell-mediated lysis by
40, 214–216 HLA-G expression on target cells. Science 274, 792–795
62 Raijmakers, M. T. M., Roes, E. M., Steegers, E. A. P. and 82 Saftlas, A. F., Beydoun, H. and Triche, E. (2005)
Peters, W. H. M. (2002) The C242T-polymorphism of the Immunogenetic determinants of preeclampsia and
NADPH/NADH oxidase gene p22phox subunit is not related pregnancy disorders: a systematic review.
associated with pre-eclampsia. J. Hum. Hypertens. 16, Obstet. Gynecol. 106, 162–172
423–425 83 Hiby, S. E., Walker, J. J., O’Shaughnessy K, M. et al.
63 Kim, Y. J., Williamson, R. A., Chen, K., Smith, J. L., (2004) Combinations of maternal KIR and fetal HLA-C
Murray, J. C. and Merrill, D. C. (2001) Lipoprotein lipase genes influence the risk of preeclampsia and reproductive
gene mutations and the genetic susceptibility of
success. J. Exp. Med. 200, 957–965
preeclampsia. Hypertension 38, 992–996
84 Arngrimsson, R., Sigurardttir, S., Frigge, M. L. et al.
64 Chikosi, A. B., Moodley, J., Pegoraro, R. J., Lanning, P. A.
(1999) A genome-wide scan reveals a maternal
and Rom, L. (2000) Apolipoprotein E polymorphism in
susceptibility locus for pre-eclampsia on chromosome
South African Zulu women with preeclampsia.
Hypertens. Pregnancy 19, 309–314 2p13. Hum. Mol. Genet. 8, 1799–1805
65 Laasanen, J., Heinonen, S., Hiltunen, M., Mannermaa, A. 85 Moses, E. K., Lade, J. A., Guo, G. et al. (2000) A genome
and Laakso, M. (2002) Polymorphism in the peroxisome scan in families from Australia and New Zealand
proliferator-activated receptor-γ gene in women with confirms the presence of a maternal susceptibility locus
preeclampsia. Early Hum. Dev. 69, 77–82 for pre-eclampsia, on chromosome 2. Am. J. Hum. Genet.
66 Belo, L., Gaffney, D., Caslake, M. et al. (2004) 67, 1581–1585
Apolipoprotein E and cholesteryl ester transfer protein 86 Caulfield, M., Munroe, P., Pembroke, J. et al. (2003)
polymorphisms in normal and preeclamptic pregnancies. Genome-wide mapping of human loci for essential
Eur. J. Obstet. Gynecol. Reprod. Biol. 112, 9–15 hypertension. Lancet 361, 2118–2123
67 Malina, A. N., Laivuori, H. M., Agatisa, P. K. et al. (2004) 87 Fitzpatrick, E., Goring, H. H. H., Liu, H. et al. (2004)
The Trp64Arg polymorphism of the β3-adrenergic Fine mapping and SNP analysis of positional candidates
receptor is not increased in women with preeclampsia. at the preeclampsia susceptibility locus (PREG1) on
Am. J. Obstet. Gynecol. 190, 779–783 chromosome 2. Hum. Biol. 76, 849–862
68 Malamitsi-Puchner, A., Tziotis, J., Evangelopoulos, D. 88 Laivuori, H., Lahermo, P., Ollikainen, V. et al. (2003)
et al. (2001) Gene analysis of the N-terminal region of the Susceptibility loci for preeclampsia on chromosomes 2p25
estrogen receptor alpha in preeclampsia. Steroids 66, and 9p13 in Finnish families. Am. J. Hum. Genet. 72,
695–700 168–177
69 Maruyama, A., Nakayama, T., Sato, N., Mizutani, Y., 89 Harrison, G. A., Humphrey, K. E., Jones, N. et al. (1997)
Furuya, K. and Yamamoto, T. (2004) Association study A genomewide linkage study of preeclampsia/eclampsia
using single nucleotide polymorphisms in the estrogen reveals evidence for a candidate region on 4q. Am. J.
receptor beta (ESR2) gene for preeclampsia. Hum. Genet. 60, 1158–1167
Hypertens. Res. 27, 903–909 90 Lindgren, C. M., Mahtani, M. M., Widen, E. et al. (2002)
70 Ciarmela, P., Florio, P., Battistini, S. et al. (2005) Genomewide search for type 2 diabetes mellitus
Mutational analysis of the inhibin α gene in preeclamptic susceptibility loci in Finnish families: the Botnia study.
women. J. Endocrinol. Invest. 28, 30–33 Am. J. Hum. Genet. 70, 509–516
71 Papazoglou, D., Galazios, G., Koukourakis, M. I. et al. 91 Lachmeijer, A. M. A., Arngrimsson, R., Bastiaans, E. J.
(2004) Vascular endothelial growth factor gene et al. (2001) A genome-wide scan for preeclampsia in the
polymorphisms and pre-eclampsia. Mol. Hum. Reprod. Netherlands. Eur. J. Hum. Genet. 9, 758–764
10, 321–324 92 Oudejans, C. B., Mulders, J., Lachmeijer, A. M. et al.
72 Jurajda, M., Kankova, K., Muzik, J. et al. (2001) Lack of (2004) The parent-of-origin effect of 10q22 in
an association of a single nucleotide polymorphism in the pre-eclamptic females coincides with two regions
promoter of the matrix metalloproteinase-1 gene in clustered for genes with down-regulated expression in
Czech women with pregnancy-induced hypertension. androgenetic placentas. Mol. Hum. Reprod. 10, 589–598
Gynecol. Obstet. Invest. 52, 124–127 93 Pang, Z. J. and Xing, F. Q. (2003) Comparative study on
73 Colhoun, H. M., McKeigue, P. M. and Davey Smith, G. the expression of cytokine-receptor genes in normal and
(2003) Problems of reporting genetic associations with preeclamptic human placentas using DNA microarrays.
complex outcomes. Lancet 361, 865–872 J. Perinat. Med. 31, 153–162

C 2006 The Biochemical Society
458 S. Chappell and L. Morgan

94 Pang, Z. J. and Xing, F. Q. (2003) Expression profile of 108 Mellembakken, J. R., Aukrust, P., Olafsen, M. K.,
trophoblast invasion-associated genes in the Ueland, T., Hestdal, K. and Videm, V. (2002) Activation
pre-eclamptic placenta. Br. J. Biomed. Sci. 60, 97–101 of leukocytes during the uteroplacental passage in
95 Pang, Z. J. and Xing, F. Q. (2004) DNA microarrays preeclampsia. Hypertension 39, 155–160
detect the expression of apoptosis-related genes in 109 Davidson, E. J., Riley, S. C., Roberts, S. A., Shearing,
preeclamptic placentas. J. Perinat. Med. 32, 25–30 C. H., Groome, N. P. and Martin, C. W. (2003) Maternal
96 Pang, Z. J. and Xing, F. Q. (2004) Comparative profiling serum activin, inhibin, human chorionic gonadotrophin
of metabolism-related gene expression in pre-eclamptic and α-fetoprotein as second trimester predictors of
and normal pregnancies. Arch. Gynecol. Obstet. 269, pre-eclampsia. Br. J. Obstst. Gynaecol. 110, 46–52
91–95 110 Reference deleted
97 Anteby, E. Y., Ayesh, S., Shochina, M. et al. (2005) 111 Muttukrishna, S., Knight, P. G., Groome, N. P., Redman,
Growth factor receptor-protein bound 2 (GRB2) C. W. and Ledger, W. L. (1997) Activin A and inhibin A as
upregulation in the placenta in preeclampsia implies a possible endocrine markers for pre-eclampsia. Lancet 349,
possible role for ras signalling. Eur. J. Obstet. Gynecol. 1285–1288
Reprod. Biol. 118, 174–181 112 Banerjee, S., Smallwood, A., Moorhead, J. et al. (2005)
98 Heikkila, A., Tuomisto, T., Hakkinen, S. K., Placental expression of interferon-γ (IFN-γ ) and its
Keski-Nisula, L., Heinonen, S. and Yla-Herttuala, S. receptor IFN-γ R2 fail to switch from early hypoxic to
(2005) Tumor suppressor and growth regulatory genes are late normotensive development in preeclampsia. J. Clin.
overexpressed in severe early-onset preeclampsia: an array Endocrinol. Metab. 90, 944–952
study on case-specific human preeclamptic placental 113 Gabinskaya, T., Salafia, C. M., Gulle, V. E., Holzman,
tissue. Acta Obstet. Gynecol. Scand. 84, 679–689 I. R. and Weintraub, A. S. (1998) Gestational
99 Hviid, T. V., Larsen, L. G., Hoegh, A. M. and Bzorek, M. age-dependent extravillous cytotrophoblast osteopontin
(2004) HLA-G expression in placenta in relation to immunolocalization differentiates between normal and
HLA-G genotype and polymorphisms. Am. J. preeclamptic pregnancies. Am. J. Reprod. Immunol. 40,
Reprod. Immunol. 52, 212–217 339–346
100 Reimer, T., Koczan, D., Gerber, B., Richter, D., Thiesen, 114 Fitzgerald, D. J., Entman, S. S., Mulloy, K. and
H. J. and Friese, K. (2002) Microarray analysis of FitzGerald, G. A. (1987) Decreased prostacyclin
differentially expressed genes in placental tissue of biosynthesis preceding the clinical manifestation of
pre-eclampsia: up-regulation of obesity-related genes. pregnancy-induced hypertension. Circulation 75,
Mol. Hum. Reprod. 8, 674–680 956–963
101 Tsoi, S. C., Cale, J. M., Bird, I. M. and Kay, H. H. (2003) 115 Shankar, R., Gude, N., Cullinane, F., Brennecke, S.,
cDNA microarray analysis of gene expression profiles in Purcell, A. W. and Moses, E. K. (2005) An emerging
human placenta: up-regulation of the transcript encoding role for comprehensive proteome analysis in
muscle subunit of glycogen phosphorylase in human pregnancy research. Reproduction 129,
preeclampsia. J. Soc. Gynecol. Invest. 10, 496–502 685–696
102 Reference deleted 116 Takimoto, E., Ishida, J., Sugiyama, F., Horiguchi, H.,
103 Peng, X., Wood, C. L., Blalock, E. M., Chen, K. C., Murakami, K. and Fukamizu, A. (1996) Hypertension
Landfield, P. W. and Stromberg, A. J. (2003) Statistical induced in pregnant mice by placental renin and
implications of pooling RNA samples for microarray maternal angiotensinogen. Science 274,
experiments. BMC Bioinformatics 4, 26 995–998
104 Ashkenazi, A. and Dixit, V. M. (1999) Apoptosis control 117 Dechend, R., Gratze, P., Wallukat, G. et al. (2005)
by death and decoy receptors. Curr. Opin. Cell Biol. 11, Agonistic autoantibodies to the AT1 receptor in a
255–260 transgenic rat model of preeclampsia. Hypertension 45,
105 Berk, B. C. (1999) Angiotensin II signal transduction in 742–746
vascular smooth muscle: pathways activated by specific 118 Wallukat, G., Homuth, V., Fischer, T. et al. (1999) Patients
tyrosine kinases. J. Am. Soc. Nephrol. 10 (Suppl. 11), with preeclampsia develop agonistic autoantibodies
S62–S68 against the angiotensin AT1 receptor. J. Clin. Invest. 103,
106 Austgulen, R., Lien, E., Vince, G. and Redman, C. W. 945–952
(1997) Increased maternal plasma levels of soluble 119 Crossey, P. A., Pillai, C. C. and Miell, J. P. (2002) Altered
adhesion molecules (ICAM-1, VCAM-1, E-selectin) in placental development and intrauterine growth restriction
preeclampsia. Eur. J. Obstet. Gynecol. Reprod. Biol. 71, in IGF binding protein-1 transgenic mice. J. Clin. Invest.
53–58 110, 411–418
107 Zhou, Y., Damsky, C. H., Chiu, K., Roberts, J. M. and 120 Takahashi, K., Kobayashi, T. and Kanayama, N. (2000)
Fisher, S. J. (1993) Preeclampsia is associated with p57Kip2 regulates the proper development of labyrinthine
abnormal expression of adhesion molecules by invasive and spongiotrophoblasts. Mol. Hum. Reprod. 6,
cytotrophoblasts. J. Clin. Invest. 91, 950–960 1019–1025

Received 21 October 2005/11 January 2006; accepted 16 January 2006

Published on the Internet 15 March 2006, doi:10.1042/CS20050323

C 2006 The Biochemical Society