Keywords
preeclampsia, genetics, polymorphism, linkage analysis, candidate gene, genomics, polygenic, DNA sequencing
Editors’ comment: Recognition that preeclampsia might have a genetic predeliction dates back to late 1800s reports of familial clustering of eclampsia. But in his 1941 treatise on “The Toxemias of Pregnancy” Dieckmann does not mention genetic or familial propensity. Chesley’s first-edition text summarized early investigations in mother–daughter pairs, and he himself reported that among daughters of women with eclampsia, 26% had preeclampsia in their first pregnancies. By contrast, the daughters-in-law control group had only an 8% rate of first-pregnancy preeclampsia. This comprehensively updated chapter will “fast-forward” the reader 35 years into the post-genomic era. Rendering and analysis of increasingly precise molecular platforms and bioinformatic tools to manage megadata have galvanized the field of preeclampsia genetics, which promises to revolutionize how we think about the complexity of the etiologies, management strategies and prevention of this syndrome.
Preeclampsia is a complex familial disorder and likely involves multiple genes in multiple biological pathways. Various types of genetic studies have been employed to solve the preeclampsia puzzle, including family reports, twin studies, segregation analyses, linkage analyses, association studies, and next-generation sequencing studies. The association studies comprise most of the literature on the genetics of preeclampsia, but because they typically study only a few pre-selected genes at a time and utilize fewer than 500 human subjects, they are insufficient for advancing the research. Genome-wide studies are the future of preeclampsia genetic research. Multicenter efforts are needed to define clinical and pathologic subsets, and high-dimensional systems biology needs to be employed so that these studies can account for preeclampsia’s heterogeneity. New directions in research could lead to a greater understanding of preeclampsia’s primary pathophysiology and the development of genetic screening and diagnostic tests and appropriate treatments and therapies.
Dedication
Current interest in the genetics of preeclampsia can be traced back to the signal study by Leon Chesley, who single-handedly followed the remote course of 267 women who survived eclampsia during the years 1931–1951. He postulated that in the absence of a renal biopsy, a convulsion – especially in a primiparous hypertensive patient – was the most convincing clinical evidence that diagnosis of preeclampsia was correct. These patients were interviewed and reexamined periodically, the data from some participants spanning more than 40 years after their eclamptic convulsion. In the course of his evaluations, Dr. Chesley noted the increased occurrence of preeclampsia–eclampsia within families. His observations and his reviews of other data that suggested familial factors are involved in the etiology of preeclampsia caught the fancy of genetic investigators. With the advent of molecular genetics and the still more recent mapping of the human genome, this field is advancing rapidly. This chapter, dedicated to the pioneering studies of Leon Chesley, summarizes research into the genetics of preeclampsia through 2013. A glossary of genetics terminology is provided for readers who are less familiar with genetic concepts.
Introduction
Preeclampsia and eclampsia *
* Throughout the rest of this chapter, preeclampsia will be the word used when the discussion could refer to either preeclampsia or eclampsia. Distinctions will be made between the two conditions as necessary.
are familial, as genetic research on these conditions over the last century has shown. Due to monumental improvements in neonatal care and decreased mortality rates of newborns over the last 50 years, generational trends can now be observed: eclamptic or preeclamptic mothers, aunts, and grandmothers have had female descendants who show an increased risk of preeclampsia over the general population. Preeclampsia tends to cluster in families; a heritability study using a Utah genealogy database determined the coefficient of kinship for preeclampsia cases to be more than 30 standard deviations higher than for controls (and unpublished data). The recurrence risk for preeclampsia in the daughters of either eclamptic or preeclamptic mothers is in the 20–40% range. For sisters it is in the 11–37% range. Much lower rates are seen in relatives by marriage, such as daughters in-law and mothers in-law. African-American mothers at all socioeconomic levels experience a higher rate of preeclampsia than the general population in the United States, suggesting that ethnicity, rather than socioeconomic status, has a greater impact on incidence of preeclampsia. And finally, twin studies estimate that approximately 22% to 47% of preeclampsia risk is heritable, as opposed to environmentally influenced.While the ultimate causes of preeclampsia remain unknown, it is perhaps obvious that genes should play a role. As discussed in other chapters of this book, preeclampsia occurs when placental ischemia or inflammation causes various mediators to be released directly into the maternal circulation. The maternal endothelium and arterioles respond initially in an adaptive manner, but ultimately cause profound dysfunction of various major organs. Any number of disorders with a genetic component might interfere with maternal vascular responses, affect trophoblast function, or increase the placental mass, causing fetal demands to outpace the supply. Every ligand, every receptor, every amplification cascade, every aspect of the programmed responses that orchestrate the pathophysiological response is under the control of either the mother’s or the fetus’s genes.
Geneticists finally have the technology and genomic data to find practical clinical solutions for the genetic testing, prevention, and treatment of preeclampsia. The completion of the Human Genome Project has enabled researchers to investigate the genetic causes of diseases both rare and common, and new technologies and statistical models are able to detect increasingly subtle variations within the human genome, gene effects on physiology, and their interactions with environmental factors.
Studies conducted over the last decade suggest that the genetic contributions to preeclampsia are likely to be complex, involving non-Mendelian transmission and numerous variants, gene–gene interactions, and environmental variables. The genes involved would not directly cause preeclampsia but rather would lower a woman’s biological threshold at which she would develop the condition.
Indeed, as noted throughout this chapter, preeclampsia is best understood as a multifactorial, polygenic condition. In a Mendelian disease (e.g., cystic fibrosis), an allelic variant or mutation is directly involved in causing the disease, and penetrance is complete or nearly complete for the relatively few persons in the population with that disease genotype. By contrast, complex diseases are the result of numerous common variants, usually at multiple loci, which contribute to varying degrees to a person’s susceptibility to that disease. Because of environmental and other variables, genotype alone may not result in phenotypic manifestation of the disease, but it will increase disease susceptibility and risk. These polygenic multifactorial conditions are familial, usually affecting multiple generations, but their inheritance does not follow Mendelian ratios. Polygenic disease occurs under diverse environmental conditions, and the genes underlying the disease can show highly variable expression usually resulting in a continuum of phenotypes with patients above some threshold classified as having the disease.
Genetic background plays a critical role in polygenic inheritance since several factors must collaborate to cause a bodily function to go awry, and only after these factors reach some critical point is the phenotypic effect seen. Thus recurrence risk of a polygenic multifactorial disease is higher within populations with a high incidence of the disorder. In disorders with a relatively high heritability, the recurrence risk of the disorder approximates the square root of the population incidence. Marked variation in the incidence and expression may occur in different ethnic groups. Within families, the greater the number of family members who have already been affected with a multifactorial condition, the more likely it is that the genetic background is favorable for expression of this condition. Consanguinity also increases the risk because of the greater likelihood of deleterious genes being shared. The severity of the disorder often correlates with the recurrence risk.
In the case of preeclampsia, the relative importance of each environmental risk factor would vary among women, depending on their genotype. One woman may have a genotype that requires an accumulation of several environmental variables, such as substance abuse and multiple gestation, to exceed this threshold. Another woman may have no identifiable environmental risk factors but may have prepregnancy diabetes, which may result in an extremely low, easily exceeded preeclampsia threshold.
Since the diagnostic criteria of preeclampsia force arbitrary thresholds on the continuous distributions of blood pressure and proteinuria, it is likely that no single etiology or genetic marker will account for all cases of preeclampsia. Given the clinical heterogeneity of the condition, it is all the more important to first correctly classify preeclampsia cases into subtypes, based on apparent etiology. This concept is stressed in Chapter 2 , Chapter 8 in the current edition of this text.
Table 4.1 summarizes many of the genetic terms you will find in this chapter.
| Term | Definition |
|---|---|
| Allele | The DNA sequence on a single chromatid at a particular genetic locus; at each locus, a single allele is inherited separately from each parent. |
| Base pair (bp) | Two nucleotides on opposite complementary DNA or RNA strands that are connected via hydrogen bonds. The human genome is composed of approximately 3 billion base pairs. Another unit frequently used is the kilobase (kb) , which denotes 1000 base pairs. |
| Candidate gene | A gene researched and suspected of being involved in a particular trait or disease. Frequently identified in gene expression or genetic association studies. |
| Coefficient of kinship | The probability that a gene at a given locus, picked at random from each of two individuals, will be identical due to familial relationship. |
| Epigenesis | Environment-induced variations in the expression or operation of a functional gene even though the underlying genomic code is stable and unchanging. |
| Epistasis | Interaction between or among non-allelic genes in which one combination of such genes has a dominant effect over other combinations (for instance when one gene suppresses the expression of another) |
| Exon | Only about 1% of our genome is the “genes” which are “translated” into the proteins. The roughly 22,000 human genes are divided into 180,000 functional segments referred to as exons. |
| Exome | The portion of the genome (the 180,000 exons) that codes for proteins. Currently, over two-thirds of the DNA errors known to cause genetic disorders occur in the exome. |
| Family clustering | The repeated occurrence of a phenotype in a given family. Often detected in genetic family studies conducted for this purpose. |
| Fine-mapping | The determination of the sequence of nucleotides and their relative distances from one another in a specific region or locus of the genome. |
| Founder effect | The loss of genetic variation that occurs when a relatively small number of individuals establish a new colony distinct from their original, larger population. As a result of the loss of genetic variation, the new population may be genetically and phenotypically different from the parent population but have relatively low genetic variation within itself. It may show increased sensitivity to genetic drift and an increase in inbreeding. |
| Gene expression | The process by which gene-coded information is converted into structural proteins and enzymes. Expressed genes are transcribed into mRNA, which in turn is translated into protein, or they are transcribed into RNA that does not translate into protein (e.g., transfer, noncoding and ribosomal RNAs). |
| Genetic association | The hypothesis that a given candidate gene or SNP causes or is otherwise related to a particular genetic condition, based on the differences in single-locus alleles or genotype frequency between a case group (with the genetic condition) and a control group (without the genetic condition). |
| Genetic drift | The evolutionary process of change in allele frequencies that occurs entirely from chance, from one generation to the next. |
| Genomic | Of or relating to the entire genome, which in humans comprises 23 chromosome pairs, or 3.2 billion DNA base pairs. |
| Genotype | The specific allele makeup of an individual, usually referring to one or more particular alleles being studied. |
| GWAS | Genome-wide association study – a discovery approach, widely used in recent genetics research, in which genetic variations (genotypes) across the genomes of many cases and controls are assayed in order to find genetic variants associated with a disease or trait (phenotypes). |
| Haplotype | 1. Two or more alleles or SNPs at distinct loci on one chromosome that are transmitted together. 2. A set of SNPs on a single chromatid that is statistically associated. |
| Heritability | The proportion of phenotypic variation in a population that is attributable to genetic variation among individuals. |
| Heterogeneity | The variability of phenotypes despite identical genetic information. |
| Heterozygous | When, of the two possible alleles at a given locus on homologous chromosomes, one of each allele is present. |
| Human Genome Project | A 13-year research project completed in 2003 and coordinated by the US Department of Energy and the National Institutes of Health with contributions from numerous industrial nations. It identified all genes in the human genome (approximately 23,000) and determined the sequence of the 3 billion nucleotide base pairs that make up human DNA. It also developed infrastructure for the storage of these data and their future analysis, which is expected to continue in the private sector for many years. |
| Immunogenetic | 1. The genetic basis of susceptibility to immune response and disease. 2. The relationship between immunity to disease and genetic makeup. |
| Linkage | Occurs when particular genetic loci on a chromosome, or alleles, are inherited jointly due to their close proximity to one another. It is measured by the percentage recombination between loci (unlinked genes showing 50% recombination). A linkage analysis studies this phenomenon in the context of candidate genes and markers. |
| Linkage disequilibrium (LD) | The occurrence of some alleles at two or more loci (like linkage, though not necessarily on the same chromosome) more or less often than would be expected, presumably because the combination confers some selective (evolutionary) advantage or disadvantage, respectively. Non-random associations between polymorphisms at different loci are measured by the degree of LD and can be demonstrated with haplotype analysis. |
| Locus (plural is loci) | The position on a chromosome of a gene or other marker; also, the DNA at that position. The meaning of locus is sometimes restricted to regions of DNA that are expressed (see Gene expression ). |
| Logarithm of the odds (LOD) score | A measure of the likelihood of two loci being within a measurable distance of each other. |
| Marker | An identifiable physical location on the genome, the inheritance of which can be monitored. Markers can be expressed regions of DNA (genes), a restriction enzyme cutting site, or some segment of DNA with no known coding function but with a distinguishable inheritance pattern. Markers must be linked with a clear-cut phenotype; they are used as a point of reference when mapping new variants. |
| Multifactorial | A disease or condition influenced in its expression by many factors, both genetic and environmental. |
| Mutation | Any heritable, permanent change in DNA sequence; also, the process by which genes undergo a structural change. |
| Next-generation sequencing | DNA sequencing is the process of determining the order of the four nucleotide bases – adenine, guanine, cytosine, and thymine – in a strand of DNA. Next generation sequencing uses high-throughput, automated methods to sequence hundreds of thousands of sequences in parallel so that an entire human genome can be sequenced in a day. |
| Nucleotide | One of the monomeric units from which DNA or RNA polymers are constructed; it consists of a purine or pyrimidine base, a pentose sugar, and a phosphoric acid group. |
| Pedigree | A diagram of the relationships within a given family with symbols to represent people and lines to represent genetic relationships. Often used to determine the mode of inheritance (dominant, recessive, etc.) of genetic diseases. |
| Penetrance | The extent to which individuals who carry the gene for a particular genetic condition express that gene as an expected phenotype. Penetrance may be described as complete, incomplete, low, or high, or may be quantitatively expressed as a percentage. |
| Polygenic disorder | A genetic disorder resulting from the combined action of alleles of more than one gene (e.g., heart disease, diabetes, and some cancers). Although such disorders are inherited, they depend on the simultaneous presence of several alleles; thus the hereditary patterns are usually more complex than those of single-gene disorders. Also referred to as complex disorders . |
| Polymorphism | Genetic differences in the DNA sequence that naturally occur among individuals. A genetic variation that occurs in more than 1% of a population would be considered a useful polymorphism for genetic linkage analysis. |
| Population admixture | The unwitting inclusion of members of a genetic population different from the genetic population being studied, having been selected to increase genetic homogeneity. It has the potential to give false-positive results in studies of genes underlying complex traits or to mask, change, or even reverse true genetic effects. |
| Predisposition | An increased likelihood of, or an advanced tendency toward, a specific medical condition. |
| Promoter | A site on DNA to which RNA polymerase will bind and initiate transcription; classically 5′ to its coding region. |
| Quantitative trait loci (QTL) | Stretches of DNA that are closely linked to the gene(s) that underlie the trait in question, though they are not necessarily genes themselves. Can be molecularly identified to help map regions of the genome that contain genes involved in specifying a quantitative trait. |
| Race/Ethnicity | Indicators of evolutionary ancestral geographic origin (by continent, in the broadest terms) that can be detected via certain markers on a person’s genome. In population genetics, these distinctions are useful in increasing the homogeneity of a population for detection of differences between diseased and control subjects. |
| Recessive | A gene that is phenotypically manifest in the homozygous state but is masked in the presence of a dominant allele. |
| Recombination | The process by which offspring derive a specific combination of genes different from that of either parent. |
| Recurrence risk | The chance that a genetic disease present in a family will recur in that family and affect another person (or persons). |
| Segregation analysis | A statistical test to determine the pattern of genetic inheritance for a trait or genetic condition (e.g., Mendelian, dominant autosomal, epistatic, polygenic, age-dependent). |
| Single nucleotide polymorphism (SNP) | A common (occurring in more than 1% of the population), single base substitution in the DNA sequence that may or may not cause a difference in gene expression (called functional or nonfunctional, respectively). Functional SNPs include: (1) SNPs in coding regions of genes resulting in amino acid substitutions (non-synonymous SNPs), which may in turn alter protein sequence, structure, function, or interaction; enzyme stability; catalytic activity; and/or substrate specificity; (2) SNPs in the non-coding, regulatory regions of genes that affect the genes’ transcription, translation, regulation, or mRNA stability (synonymous SNPs); (3) SNPs in genes that are duplicated, resulting in higher product levels; (4) SNPs in genes that are completely or partially deleted, resulting in no product; (5) SNPs that cause splice site variants that result in truncated or alternatively spliced protein products. Also look up promoter , enhancer , repressor , regulator , intron , exon , codon , deletion , duplication , insertion , inversion , translocation , and copy number variant (CNV) . |
| Single-gene disorder | Hereditary disorder caused by a mutant allele of a single gene (e.g., cystic fibrosis, myotonic dystrophy, sickle cell disease). |
| Subtype | One of two or more genetic pathways to the same disease phenotype. A disease may be defined by clinical characteristics, but once its genetic contributions are discovered, that same disease may be separated into subtypes based on which genes contribute to which type. This classification may or may not lead to refinements in the definition of the phenotype and/or clinical characteristics. |
| Wellcome Trust Case Control Consortium | A collaboration of 24 human geneticists who analyzed over 19,000 DNA samples from patients suffering from different complex diseases to identify common genetic variations for each condition. Conditions included tuberculosis, coronary heart disease, type 1 diabetes, type 2 diabetes, rheumatoid arthritis, Crohn’s disease, bipolar disorder, and hypertension. Two thousand patients were recruited for each disease, and 3000 were recruited as controls. The research was conducted at a number of institutes throughout the UK. |
Biological Pathways of Preeclampsia
Although preeclampsia may be a fairly homogeneous entity when it is defined by glomerular endotheliosis – the characteristic histopathologic feature of the condition – we are forced to depend on other criteria for probing the genetic aspects of the condition, as the diagnostic signs are nonspecific. The diagnostic blood pressure and proteinuria criteria in common use are, in fact, arbitrary cutoffs along a continuous distribution of values. Even if perfect diagnostic criteria existed, there is a great deal of debate about what the proper phenotype is for study, whether it is proteinuric hypertension, gestational hypertension, a placental phenotype such as reduced trophoblast invasion, a renal phenotype such as glomerular endotheliosis, or other phenotypes. Up to one-third of infants born of preeclamptic pregnancies are affected by intrauterine growth restriction, and this subset may have different variants involved. Some have argued that early-onset disease needs to be considered a different subtype. Co-morbidities related to diabetes, thrombophilia, renal disease, auto-immunity, et cetera might allow more optimal classification. The non-genetic research of preeclampsia suggests that its essential pathophysiology comprises oxidative stress and endothelial decompensation originating from and/or contributing to poor placental perfusion, which cumulatively lead to the observable symptoms of the disease.
Preeclampsia is a difficult disorder to study using genetic methodologies. A major lesson of modern genetics is that syndromes defined on the basis of clustering of clinical symptoms often reveal marked heterogeneity until they are better understood at a molecular level. In this respect, the boundaries around preeclampsia, gestational hypertension, and HELLP syndrome are likely to be redrawn once genetic determinants can be examined directly.
Research in the last decade has suggested that preeclampsia, despite its appearance only in pregnancy, has a systemic pathophysiology involving distinct yet diverse biological pathways in both the fetus and the mother. Any of these pathways – including immunologic, inflammatory, hypoxic, and thrombophilic pathways – can push inadequate placentation or pregnancy-induced hypertension over the threshold into preeclampsia, and all of these pathways are influenced by the biological products of genetic expression and genotype. As in all genetic studies of pregnancy disorders, not just one but two genotypes must be considered: the genotype of the fetus as well as the mother.
Any genetic hypothesis of preeclampsia must explain the first pregnancy effect. It is widely known that most women will not have preeclampsia with future pregnancies unless another condition exists (e.g., underlying renal disease, twins, diabetes). This has suggested an immunogenetic mechanism to many investigators, in the form of desensitization or tolerance to paternal antigens in subsequent gestations. An increased risk for first pregnancies with new partners (i.e., primipaternity with new paternal antigens presented in the placenta) has also been noted. Limited evidence exists that couples who use condoms for contraception, who accept an ovum donation, or who have a shorter length of cohabitation prior to conception have an increased risk of preeclampsia; couples that practice oral sex and women who have had multiple blood transfusions have a lower risk. Other explanations for the first pregnancy phenomenon must also be considered, however: certain enzymes in pregnancy are permanently induced and never go back to baseline levels after delivery. Similarly, permanent changes occur in maternal blood volume and in the vascular architecture of the uterus after a term gestation, so that a greater volume expansion is generally achieved with subsequent pregnancies at later parities.
While increased compared to normotensive deliveries, remote essential hypertension in first pregnancy preeclamptic women is surprisingly low and may hold a clue as to why preeclamptic patients do not have a particularly high rate of developing essential hypertension. If, for example, the AGT T235 polymorphism increases the risk of both preeclampsia and essential hypertension through a blood volume mechanism, sustained changes in baseline blood volumes that occur after delivery might explain both the first-pregnancy effect and possible reduction in an affected woman’s lifelong risk of essential hypertension.
Fetal/Placental Components of Preeclampsia
There have been numerous reports of circumstantial evidence of a paternal/fetal genetic effect. Astin et al. reported a man who lost two consecutive wives to eclampsia, and had a severely preeclamptic third wife. The observation that preeclampsia is extremely common in molar pregnancies (in which all the fetal chromosomes are derived from the father) is considered further evidence for the role of paternal genes. Triploidy, although unusual in advanced gestations, frequently presents with preeclampsia. The increase in paternal genetic material associated with the triploid diandric placenta may support the role of paternal genes in the development of preeclampsia. Hydatidiform moles, which have two sets of paternal chromosomes, commonly cause a preeclampsia-like illness.
In the segregation analyses described above, Cooper found an increased rate of preeclampsia if the proband’s own mother was eclamptic with their pregnancy. Others have investigated a small increase in the incidence of preeclampsia in the daughters in-law of women who had pregnancy-induced hypertension.
Males presumably do not manifest a phenotype when they carry susceptibility genes, which may be much greater in number than the existing literature indicates, considering how few studies have been done on paternal carriers. Considering that for most of human history, eclampsia was a common but frequently fatal disease for mother or child or both, a large percentage of modern cases must be due to new mutations without a direct familial pattern.
Immunogenetic Factors (see also Chapter 8 )
The suppression of the maternal immune response to the fetus/placenta in normal and abnormal pregnancies has been frequently and thoroughly studied. The immunological aspects of preeclampsia get to the heart of this query because symptoms and other epidemiological evidence of the mother’s abnormal immune response to paternal antigens – and the fetoplacental allograft derived from those antigens – are curiously prominent in many cases of preeclampsia:
- ●
Preeclampsia is more common during a first conception.
- ●
Preeclampsia may be more common after a mother has switched partners.
- ●
Oral or long-term semen exposure decreases the risk of preeclampsia.
- ●
Preeclampsia is more common in women who use barrier contraception.
- ●
The frequency of preeclampsia in the cases of donated gametes is very high.
- ●
HIV-related T cell immune deficiency is associated with a low rate of preeclampsia.
On the molecular level, immune mediators are closely involved in many aspects of pregnancy from implantation and placentation to labor. A normal pregnancy is accompanied by a pregnancy-specific, immunomodulated inflammatory response to the antigenic stimulus presented by the fetal-placental semiallograft. The largest surface area of contact between maternal immunocompetent T cells and the fetus is at the level of the villous trophoblasts. These cells originate in the embryo and lack expression of major histocompatibility complex (MHC) class I and class II antigens. The extravillous trophoblasts (EVT) only express human leukocyte antigens (HLA) C (weakly), Ib, G, F, and E, rather than the strong transplantation antigens HLA-A, -B, -D, -Ia and -II. Of these, only HLA-C is signaling paternal (foreign) alloantigens. There is new evidence that maternal immune cells cross the placenta, colonize fetal lymph nodes and remain to tolerize fetal T regulatory cells until early adulthood. Other inflammatory factors in preeclampsia are an abnormal immunological maternal response, comprising a change in the role of monocytes and natural killer (NK) cells for the release of circulating cytokines and an activation of proinflammatory angiotensin II subtype 1 (AT1) receptors. The significance of these receptors is discussed further in Chapter 15 by Dechend, LaMarca and Taylor. Activated neutrophils, monocytes, and NK cells initiate inflammation, which in turn induces endothelial dysfunction, if activated T cells support inadequate tolerance during pregnancy.
In preeclampsia, genetic or non-genetic factors such as hypoxia or oxidative stress can induce necrosis or aponecrosis of trophoblasts. Macrophages or dendritic cells that phagocytose these trophoblasts produce type 1 cytokines such as tumor necrosis factor alpha (TNFα), interleukin (IL) 12, and IFN-γ that augment inflammation, such that the cytokine profile in preeclampsia, in contrast to that of normal pregnancy, is one of type 1 proinflammatory cytokine dominance, while the production of type 2 immunomodulatory cytokines is suppressed. This is supported by studies reporting elevated plasma levels of TNFα and IL-1β in preeclampsia. Furthermore, if type 2 regulatory cytokines are reduced system-wide regulatory T cell function can be inhibited.
Redman and Sargent argue that the immunological interaction between the mother and fetus might be mediated predominantly by NK cells instead of T cells (see their chapter – Chapter 8 ). For example, the invading trophoblasts predominantly encounter maternal decidual lymphocytes that are NK cells with an unusual phenotype. Notably, the NK cells express receptors (such as killer immunoglobulin-like receptors, or KIRs) that recognize the exact combination of HLAs associated with invasive cytotrophoblasts, particularly polymorphic HLA-C. These NK cells express a unique array of KIRs for binding the combinations of HLAs expressed by intermingling cytotrophoblast and NK cells that mediate immune recognition.
Numerous haplotypes, differing in gene content and allele combinations, are associated with the multigene KIRs. Some haplotypes inhibit NK-cell function (cytokine production in these cells), whereas others are stimulatory, depending on both the KIR phenotype of the NK cells and the HLA-C phenotype of the stimulating cells. KIR gene haplotypes can be divided into two functional groups: the simpler A group codes mainly for inhibitory KIR, and the more complex B group codes for receptors that stimulate NK cells. Preeclampsia is much more prevalent in women homozygous for the inhibitory KIR haplotypes (AA) than in women homozygous for the stimulator KIR BB. The effect is strongest if the fetus is homozygous for the HLA-C2 haplotype.
Essentially, normal placentation is more likely and preeclampsia is less likely when trophoblasts strongly stimulate uterine (maternal) NK cells. This interaction between trophoblasts and NK cells serves as an essential component of immunity suppression/stimulation issues in normal and abnormal pregnancies, specifically in preeclampsia.
There is evidence for increased release of syncytiotrophoblast microvesicles and other cellular “debris” into the maternal plasma that also influence immune stimulation, cytokine production and vascular remodeling. The ability of the mother to mount an adequate response to these released pro-oxidant molecules may be of key importance for oxygen supply throughout pregnancy. This view is endorsed by the increased risk of preeclampsia in women with preexisting medical conditions that frequently lead to oxidative stress, including chronic hypertension, diabetes, and renal disease. Furthermore, long-term follow-up studies have shown that women who were unaffected by these conditions prior to conception are nevertheless more likely to develop them later in life following an episode of preeclampsia, as reviewed in Chapter 3 . It is now well known that chronic hypertension and diabetes have a genetic component, which raises the possibility that preeclampsia shares susceptibility genes with these conditions, particularly related to oxidative stress.
Types of Genetic Studies Conducted
Family Reports
The first hints that preeclampsia is a genetic disease came from reports of familial clustering. Numerous, early case series were summarized by Chesley. Elliott was the first to report familial incidence of eclampsia in 1873. He reported a woman who died of eclampsia during her fifth pregnancy. Three of her four daughters subsequently died of eclampsia as well. One of the first systematic studies of the genetics of preeclampsia was presented in 1960 by Humphries, who studied mother–daughter pairs delivering at the Johns Hopkins Hospital. Preeclampsia occurred in 28% of the daughters of women who had toxemia in pregnancy, compared with 13% of the comparison group. As noted above, Chesley’s own remarkable study was well under way in the 1960s, when he reported data on the pregnancies of the daughters, daughters in-law, and sisters of the eclamptic probands, in whom he observed a relative risk of eight-fold for eclampsia in the daughters. Remarkably persistent in this effort, Dr. Chesley was able to find information on 96% of all the daughters, greatly reducing the possibility of ascertainment bias. At the time of his death, he was preparing to publish additional data on the granddaughters and granddaughters-in-law of the original cohort.
Another large body of information comes from the Aberdeen Maternity Hospital in Scotland. Research teams headed by Adams, Cooper, and Sutherland studied this population in 1961, 1979, and 1981, respectively. These studies were unique because of consistent diagnostic criteria and classification and careful recording of births through several decades. Arngrimsson studied the Icelandic population. Because of the population’s small size and interest in genealogy, as well as the concentration of maternity records at only one hospital, relatively complete information was available on the relatives of the index cases with preeclampsia. Daughters had a prevalence of preeclampsia or eclampsia of 23%, whereas those syndromes occurred in just 10% of the daughters-in-law.
Alexander notes that both male and female children of preeclamptic mothers are at an increased risk of having or fathering a preeclamptic pregnancy in turn, providing evidence of fetal contribution to preeclampsia susceptibility. However, according to pedigrees and heritability studies, this susceptibility remains greater in female than in male offspring, suggesting that transmission from mother to daughter is a critical component.
Table 4.2 summarizes the family reports discussed in this section.
| Author(s) ref | Year | Focus of Study |
|---|---|---|
| Humphries | 1960 | Mother–daughter pairs. |
| Adams and Finlayson | 1961 | Disease of preeclampsia, sisters of preeclamptic women. |
| Chesley et al. | 1968 | Pregnancies of daughters and granddaughters of eclamptics compared to daughters and granddaughters “in-law.” |
| Cooper and Liston | 1979 | “Severe” preeclampsia. |
| Sutherland et al. | 1981 | Increased preeclampsia in mothers and daughters of preeclamptic women. |
| Armgrimsson et al. | 1990 | Increased rate of preeclampsia in mothers and daughters of preeclamptic women. |
| Esplin et al. | 2001 | Utah database: both men and women who were born to a mother with preeclampsia were significantly more likely to have a child who was the product of a pregnancy complicated by preeclampsia. |
| Cnattingius et al. | 2004 | Swedish registries. Heritability estimated at 0.55, maternal genes contribute more than fetal genes. |
| Alexander | 2007 | Mother–fetus pairs, including male transmission. |
Twin Studies
Once there are indications that a disorder is genetic, twin studies can be used to measure the heritability of the condition (the proportion of preeclampsia risk that is attributable to genetics, as opposed to environmental factors). Several studies wholly or partially dedicated to this question regarding preeclampsia have been conducted. Using twin concordance to estimate heritability is never a straightforward method because of the mechanisms and associations underlying the monozygotic twinning process. Indeed, Hall postulated that the development of a discordant cell line (by any mechanism including chromosomal, single gene mutation, mitochondria1 mutation, uniparental disomy, somatic crossing over, X-inactivation, imprinting, etc.) early in embryonic development may in fact be an underlying cause of human monozygotic (MZ) twinning; thus artificially inflating the number of discordant MZ twin pairs, leading to biased estimates of heritability.
Thornton and Macdonald studied a cohort of female twins in the UK to estimate the maternal genetic contribution to preeclampsia and non-proteinuric hypertension. However, their study ran into complications when many of the twin pairs’ self-reported preeclampsia diagnoses could not be confirmed with medical records. Based on self-reported diagnoses only, preeclampsia had a heritability of 0.221, whereas non-proteinuric hypertension heritability was 0.198. Using hospital records and various models to predict the heritability of preeclampsia based on the heritability of non-proteinuric hypertension (0.375), the study concluded that neither preeclampsia nor non-proteinuric hypertension was as heritable as previously believed, and certainly non-Mendelian in transmission.
O’Shaughnessy et al. conducted a study in the UK on monozygotic twin concordance for preeclampsia, using four pairs of monozygotic twin mothers and one monozygotic triplet mother. The study indicated that concordant monozygotic siblings were no more likely to develop preeclampsia than discordant ones. Finally, an Australian study examined a large cohort of twin pairs to determine the maternal versus fetal genetic causes of preeclampsia by evaluating concordance among several degrees of relatives: between monozygotic and dizygotic female co-twins, between female partners of male monozygotic and dizygotic twin pairs, and between female twins and partners of their male co-twins in dizygotic, opposite-sex pairs. The study determined preeclampsia to have a very low genetic recurrence risk; the maternal genetic contribution was also much lower than expected.
The accumulated evidence on twin studies of preeclampsia, including those conducted by Thornton and Onwude, Lachmeijer et al., and Salonen Ros et al., suggests that penetrance in preeclampsia is generally less than 50%, and that the accumulated confidence interval is quite wide (95% CI, 0–0.71). This may suggest a greater diversity of inheritance models and modes across the spectrum of women who exhibit the preeclampsia phenotype, encompassing Mendelian, single-gene dominance through polygenic, multifactorial inheritance.
Segregation Analyses
Segregation analyses attempt to fit the recurrence risk data from family studies into a genetic model and are useful, if not for proving inheritance via a particular model, for eliminating alternative models. Several segregation analyses of preeclampsia have been published, with varying conclusions.
In the aggregate, pre-2000 segregation analyses consistently suggested a relatively common allele acting as a “major gene” conferring susceptibility to preeclampsia. The marked increase in the incidence of preeclampsia in blood relatives but not in relatives by marriage implies that maternal genes are more important than fetal genes. Alternatively, this inheritance pattern supports the hypothesis of transmission of preeclampsia from mother to fetus by a recessive gene. More recent analyses suggest a multifactorial, polygenic inheritance with strong epigenetic contributions. Some examples of methylated CpG islands in introns or promoters of differentially expressed genes have been proposed, e.g., the STOX1 transcription factor alleles (see below), but these remain controversial.
One alternative not adequately addressed in these models is whether a very high new mutation rate (as would be expected for a common but deadly condition) exists. Chesley has reviewed the recorded history of eclampsia, and it is clear that mortality from eclampsia was high until the last few generations. Given its high lethality, it is unclear how a preeclampsia gene would become so common in the population. With the exception of one report describing a gorilla pedigree with preeclampsia, we could find no evidence that preeclampsia occurs spontaneously in our recent primate ancestors. (However, surgical induction of uteroplacental ischemia in baboons and rhesus can result in a preeclampsia-like syndrome.) Usually the only way a lethal gene will stay common in the population is if there are frequent new mutations or if the gene is positively selected on some other basis. Both of these are possibilities have been suggested in preeclampsia.
New studies are determining which preeclampsia-causing alleles are common in the population, and alleles at which mutations are frequent by tracking their occurrence through populations of women with and without preeclampsia. In one of the most significant new sequence analyses, van Dijk et al. narrowed a minimal critical region at 10q22 linked with preeclampsia in women of Dutch ancestry to 444 kb. The STOX1 gene in this region contains five different missense mutations, identical between affected sisters, cosegregating with the preeclamptic phenotype and following matrilineal inheritance. The predominant Y153H variation of this gene is highly mutagenic by conservation criteria but subject to incomplete penetrance. Given the maternal effect, a causal relation of this gene with preeclampsia can exist only if the mutations are maternally derived and transmitted to the children born from the affected pregnancies. Substitution of the conserved amino acid (either tyrosine or phenylalanine) from the Y153H mutation may lead to disease.
Oudejans et al. provide a concise review of segregation studies conducted as well as models for the transcription mechanisms at the various sites throughout the genome. Regarding the van Dijk analysis, Oudejans’s study group proposes that a second variation, also on 10q22, is needed to explain the full phenotype in Dutch females. Oudejans’s study group also noted that all of the susceptibility loci with significant linkage to preeclampsia detected previously – 2p12 (Iceland), 2p25 (Finland), and 9p13 (Finland), in addition to the 10q22 region – display evidence of epigenetic effects, which could contribute to the persistence of preeclampsia despite its historically high mortality. To date, this work has not been confirmed in other populations.
Table 4.3 summarizes the segregation analyses discussed in this section.
| Author(s) ref | Year | Study Conclusions |
|---|---|---|
| Rana et al. | 2003 | Familial forms of focal segmental glomerulosclerosis (FSGS) determined to be primarily autosomal recessive; preeclampsia developed in successive pregnancies. |
| Van Dijk et al. | 2005 | Narrowed a minimal critical region at 10q22 linked with preeclampsia in women of Dutch ancestry. |
| Laivuori | 2007 | Polygenic; also notes maternally inherited missense mutations in the STOX1 gene of the fetus. |
| Oudejans et al. | 2007 | Polygenic, with segregation of preeclampsia into early-onset, placental and late-onset. Notes contribution of epigenetics. |
| Berends et al. | 2008 | Cosegregation of preeclampsia with intrauterine growth restriction. High rate of consanguinity. |
Linkage Analyses
Linkage studies use a “positional” approach to gene identification using regions of the genome that segregate with the disease of interest, in contrast to physiologic hypotheses. Linkage studies require an accurate diagnosis of the disease under study and precise histories of family relationships among the study participants. Furthermore, linkage analysis of pedigrees requires that the appropriate model be used in the LOD (logarithm of the odds) score calculation. Markers are tested in family studies to find any violations of Mendel’s second law, which states that independent traits segregate independently. Whenever two independent traits are closely located on the same chromosome, Mendel’s second law is violated. These aberrations from independent segregation can be used to map the chromosomal location of a disease gene by comparing these markers with the other chromosomes.
With technological advances in the last decade, genome-wide linkage studies have become possible and preferable to analysis of pre-selected regions on pre-selected chromosomes for detecting susceptibility loci for complex diseases. After a genome-wide scan highlights certain regions and rules out others, the fine mapping and sequencing of the suspect regions can zero in on candidate single nucleotide polymorphisms (SNPs) and genes and detect parent-of-origin effects. Below is a summary of the genome-wide linkage studies and the genomic regions they have investigated. Other studies have been conducted based on biological hypotheses regarding functional genes in predetermined regions of the genome, or in an effort to fine-map areas first detected with genome-wide studies.
- ●
Harrison et al. published a genome-wide linkage study of 15 Australian family pedigrees in 1997. They found a 2.8-cM candidate region between D4S450 and D4S610 on 4q with a barely significant, maximum multipoint LOD score of 2.9. Because of uncertainties concerning inheritance and diagnosis, four different inheritance models were used to carry out LOD score analysis.
- ●
Arngrimsson et al. published their results of a genome-wide linkage study of 124 Icelandic family pedigrees in 1999. They found a maternal susceptibility locus on 2p13 with a significant LOD score of 4.70. Their data supported a primarily dominant inheritance model.
- ●
Moses et al. published their results of a medium-density genome-wide scan of 34 Australian/NZ families in 2000. They found loci on chromosome 2 at an LOD score of 2.58 (tentatively supporting Arngrimsson’s group ) and at 11q23–24 (LOD score 2.02). The model used was multipoint nonparametric.
- ●
Lachmeijer et al. published their results of a genome-wide scan of 67 Dutch families in 2001. The highest LOD score of 1.99 was determined on the long arm of chromosome 12, associated primarily with HELLP-afflicted families. When HELLP families were eliminated, the remaining 38 families were evaluated and returned LOD scores of 2.38 on chromosome 10q and of 2.41 on chromosome 22q. No chromosome 2 loci were detected.
- ●
Laivuori et al. published results of a genome-wide scan of 15 Finnish families in 2003. Two loci were detected: 2p25, with a nonparametric linkage (NPL) score of 3.77, and 9p13, with an NPL score of 3.74. A third locus of slightly weaker score was found at 4q32 (NPL 3.13).
- ●
In 2006, Kalmyrzaev et al. published results of a genome-wide scan on a single Kyrgyz family, so selected for the notable early onset of preeclampsia and the geographic isolation of the Kyrgyz population, which was expected to reduce heterogeneity of the condition. Nonparametric analysis detected a region at 2q23–q37, and with 2-point parametric analysis an LOD score of 2.67 was obtained at 2q24.3.
- ●
Moses et al. published results of a genome scan of 34 Australian and New Zealand families in 2006, in which they detected a significant locus at 2q with an LOD score of 3.43.
- ●
In 2007, Johnson et al. reanalyzed a previous genome-wide scan of 34 Australian/New Zealand families with a more refined and powerful variance components model represented by quantitative trait loci (QTL). Doing this analysis returned two novel QTLs at 5q and 13q, with LOD scores of 3.12 and 3.10, respectively.
Table 4.4 summarizes the linkage analyses discussed in this section.
