In other instances, the mechanism of the genetic susceptibility is less clear-cut. This can involve inheritance of a single gene polymorphism (p. 67) that leads to differences in susceptibility to a disease (e.g., acetaldehyde dehydrogenase activity and alcoholism). In addition, inherited single-gene polymorphisms appear to determine the response to as yet undefined environmental factors—for example, the antigens of the major histocompatibility (HLA) complex and specific disease associations (p. 200) such as type 1 diabetes, rheumatoid arthritis, and coeliac disease. Lastly, genetic susceptibility can determine differences in responses to medical treatment; isoniazid inactivation status in the treatment of tuberculosis (p. 186) is a good example.

Approaches to Demonstrating Genetic Susceptibility to Common Diseases

In attempting to understand the genetics of a particular condition, the investigator can approach the problem in a number of ways (Box 15.1). These can include comparing the prevalence and incidence in various different population groups, the effects of migration, studying the incidence of the disease among relatives in family studies, comparing the incidence in identical and nonidentical twins, determining the effect of environmental changes by adoption studies, and studying the association of the disease with DNA polymorphisms. In addition, study can be made of the pathological components or biochemical factors of the disease in relatives (e.g., serum lipids among the relatives of patients with coronary artery disease). Study of diseases in animals that are homologous to diseases that occur in humans can also be helpful (p. 73). Before considering the use of these different approaches in a number of the common diseases in humans, specific aspects of some of these approaches will be discussed in more detail.

Box 15.1

Types of Genetic Approach to the Common Diseases

Population/migration studies

Family studies

Twin studies

Adoption studies

Polymorphism associations

Biochemical studies

Animal models

Population/Migration Studies

Differences in the incidence of a particular disease in different population groups suggest the possibility of genetic factors being important. They could, however, also be explained by differences in environmental factors. Studies of migrant groups moving from a population group with a low incidence of a disease to one with a high incidence, in which the incidence of the disease in the migrant group rises to that of its new population group, would suggest that environmental factors are more important. Conversely, maintenance of a low incidence of the disease in the migrant group would suggest that genetic factors are more important.

Family Studies

Genetic susceptibility to a disease can be suggested by the finding of a higher frequency of the disease in relatives than in the general population. The proportion of affected relatives of a specific relationship—first degree, second degree, and so forth—can provide information for empirical recurrence risks in genetic counseling (p. 346), as well as evidence supporting a genetic contribution (see Chapter 9). Familial aggregation does not, however, prove a genetic susceptibility, since families share a common environment. The frequency of the disease in spouses who share the same environment but who will usually have a different genetic background can be used as a control, particularly for possible environmental factors in adult life.

Twin Studies

If both members of a pair of identical twins have the same trait, this could be thought to prove that the trait is hereditary. This is not necessarily so. Since twins tend to share the same environment, it is possible they will be exposed to the same environmental factors. For example, if one of a twin pair contracts a contagious disease such as impetigo, it is likely the other twin will also become affected. This problem can be partly resolved by comparing differences in the frequency of a disease or disorder between nonidentical or dizygotic (DZ) and identical or monozygotic (MZ) twin pairs.

Both members of a pair of twins are said to be concordant when either both are affected or neither is affected. The term discordant is used when only one member of a pair of twins is affected. Both types of twins will have a tendency to share the same environment but, whereas identical twins basically have identical genotypes (p. 106), nonidentical twins are no more similar genetically than brothers and sisters. If a disease is entirely genetically determined, then apart from rare events such as chromosome nondisjunction or a new mutation occurring in one of a twin pair, both members of a pair of identical twins will be similarly affected, but nonidentical twins are more likely to differ. If a disease is entirely caused by environmental factors, then identical and nonidentical twins will have similar concordance rates.

Although all twins tend to share the same environment, it is probable this is more likely in identical twins than in nonidentical twins. Similarities between identical twins can therefore reflect their shared environment as much as their identical genotypes. One way of getting round this difficulty is to study differences between identical twin pairs who, through unusual family circumstances, have been reared apart from an early age. If a particular disease is entirely genetically determined, then if one identical twin is affected, the other will also be affected, even if they have been brought up in different environments. It is rare, however, for identical twins to be separated from early childhood, so only a limited number of studies for any one disorder exist. In one study of identical twins reared separately, the data clearly showed that each pair of twins differed little in height but differed considerably in body weight. These observations suggest that heredity could play a bigger part in determining stature than it does in determining body weight.

Adoption Studies

Another approach that helps differentiate between genetic and environmental factors is to compare the frequency of a disease in individuals who remain with their biological parents with those who are adopted out of their biological family. Adopted individuals take their genes with them to a new environment. If the frequency of a disease in the individuals adopted out of a family is similar to that seen in those who remain with their biological parents, then genetic factors are likely to be more important. If, conversely, the frequency of the disease in the adopted individuals is similar to that of their adoptive parents, then environmental factors are likely to be more important.

Polymorphism Association Studies

The widespread existence of inherited biochemical, protein, enzyme, and DNA variants (p. 147) allows the possibility of determining whether particular variants occur more commonly in individuals affected with a particular disease than in the population in general, or what is known as association. Although demonstration of a polymorphic association can suggest that the inherited variation is involved in the aetiology of the disorder, such as the demonstration of HLA associations in the immune response in the causation of the autoimmune disorders (p. 200), it may only reflect that a gene nearby in linkage disequilibrium (p. 138) is involved in causation of the disorder.

The human genome contains approximately 10 million single nucleotide polymorphisms (SNPs). Developments in high-throughput microarray SNP genotyping, together with information about SNP haplotypes (from the HapMap project [p. 148]) and the availability of large collections of DNA samples from patients with common diseases, have collectively enabled genome-wide association (GWA) studies (p. 149) to reliably identify numerous loci harbouring such variants.

Biochemical Studies

Analysis of metabolite or enzyme activity levels in biochemical or metabolic pathways likely to be involved in the causation of a particular disorder can provide evidence of genetic contribution to some of the common diseases—the hormones involved in the control of blood pressure in hypertension or the regulation of lipid levels in atherosclerosis, for example. However, in disorders where we have a limited understanding of the biological basis (e.g., schizophrenia), this approach has been of very little use.

Animal Models

Recognition of the same disease/disorder that occurs in humans and in another species such as the mouse allows the possibility of experimental studies that are often not possible in humans. In many instances, however, the disorder in the animal model will have a single-gene basis that has been identified or bred for. Nevertheless, spontaneously occurring or transgenic animal models (p. 73), experimentally induced for mutations in single genes involved in the metabolic processes or disease pathways of common diseases, will provide vital insights into the genetic contribution to these disorders.

Disease Models for Multifactorial Inheritance

The search for susceptibility loci and polygenes, sometimes also referred to as quantitative trait loci, in human multifactorial disorders has met with increasing success in recent years. This is largely due to the success of GWA studies (p. 149). Examples of recent research in some common conditions will be considered to illustrate the progress to date and the extent of the challenges that lie ahead.

Diabetes Mellitus

There are two main forms of diabetes mellitus (DM) that are clinically distinct. Type 1 (T1DM) is the rarer juvenile-onset, insulin-dependent form (previous abbreviation IDDM) which affects 0.4% of the population and shows a high incidence of potentially serious renal, retinal, and vascular complications. T1DM has a peak age of onset in adolescence and can only be controlled by regular injection of insulin. Type 2 is the more common later-onset, non–insulin-dependent form that affects up to 10% of the population. It usually affects older persons and may respond to simple dietary restriction of carbohydrate intake, although many persons with T2DM require oral hypoglycemic medication, and some require insulin. An additional 1% to 2% of persons with diabetes have monogenic (single gene) forms of diabetes (Table 15.1).

Table 15.1 Subtypes of Diabetes

Around 1% to 3% of women develop glucose intolerance during pregnancy. This is known as gestational diabetes. Their abnormal glucose tolerance usually reverts to normal after the pregnancy, although approximately half to three-quarters of these women go on to develop T2DM later in life.

Diabetes can also occur secondary to a variety of other rare genetic syndromes and nongenetic disorders. Examples include Prader-Willi syndrome (p. 122), Bardet-Biedl syndrome, Wolfram syndrome, and Friedreich ataxia (Table 2.5, p. 24). Diabetes mellitus is therefore aetiologically heterogeneous.

Monogenic Forms of Diabetes

Rare forms of diabetes that show high penetrance within families are usually due to mutations in single genes. More than 20 monogenic forms of diabetes have been identified (see Table 15.1).

Maturity-Onset Diabetes of the Young

Maturity-onset diabetes of the young (MODY) is an autosomal dominant form of diabetes characterized by pancreatic β-cell dysfunction. It shows clinical heterogeneity that can now be explained by genetic heterogeneity. Mutations in the glucokinase gene cause mild hyperglycemia (blood glucose levels are usually between 5.5 and 8 mmol/L), which is stable throughout life and often treated by diet alone. Glucokinase is described as the pancreatic glucose sensor because it catalyses the rate-limiting step of glucose metabolism in the pancreatic β cell. It was therefore an obvious candidate gene. Many patients with glucokinase mutations are asymptomatic, and their hyperglycemia is detected during routine screening—for example, during pregnancy or employment medicals. The mild phenotype means that finding a glucokinase mutation is ‘good news.’

Mutations in five additional genes which encode transcription factors required for development of the β cell have been reported. The hepatocyte nuclear factor 1α (HNF1A) and hepatocyte nuclear factor 4α (HNF4A) genes were identified through positional cloning efforts and are associated with a more severe, progressive form of diabetes usually diagnosed during adolescence or early adulthood. These patients are sensitive to treatment with sulphonylurea tablets; this is an example of pharmacogenetics (see Chapter 12). Good glycemic control is important, as patients have a long duration of diabetes and may suffer from diabetic complications. Mutations in the HNF1A gene are the most common cause of MODY in most populations (65% of UK MODY), and HNF4A mutations are less frequent.

Hepatocyte nuclear factor 1β (HNF-1β) plays a key role in development of the kidney. Mutations cause renal cysts and diabetes (RCAD), and some female patients also have genital tract malformations. Insulin promoter factor 1 (IPF-1), NEUROD1, INS and CEL mutations are rare causes of MODY but highlight the possibility that further genes encoding β-cell transcription factors may be mutated in MODY (so-called MODYX genes).

Neonatal Diabetes

Analysis of HLA genotypes in children diagnosed with diabetes before the age of 6 months has shown that these patients have a similar frequency of high-risk alleles for type 1 diabetes as that found in the general population. This suggests that type 1 diabetes is rare before 6 months of age and implies a genetic cause.

Although definitions of the neonatal period vary, we know that diabetes is rare before 6 months of age, and the incidence is estimated at around 1 in 100,000 live births. There has been great progress over the past 10 years in defining the genetics of this rare condition.

Neonatal diabetes can be transient or permanent. More than 70% of cases of transient neonatal diabetes result from the overexpression of paternally expressed genes on chromosome 6q24. The inheritance and extent of this abnormality of imprinting (p. 121) are variable. However, in a small percentage of cases, patients have imprinting abnormalities besides 6q24 at multiple loci in the genome; these are associated with mutations in the zinc-finger transcription factor gene, ZFP57. Patients with a 6q24 abnormality are usually diagnosed in the first week of life and treated with insulin. Apparent remission occurs by 3 months, but there is a tendency for children to develop diabetes in later life.

Permanent neonatal diabetes does not remit, and until recently, patients were treated with insulin for life. The most common causes (>50%) are mutations in the KCNJ11 or ABCC8 genes which encode the Kir6.2 and SUR1 subunits of the adenosine triphosphate (ATP)-sensitive potassium (K-ATP) channel in the pancreatic β cell. Closure of these channels in response to ATP generated from glucose metabolism is the key signal for insulin release. The effect of activating mutations in these genes is to prevent channel closure by reducing the response to ATP, and hence insulin secretion. The most exciting aspect of this recent discovery is that most patients with this genetic aetiology can be treated with sulphonylurea drugs that bind to the channel and cause closure independently of ATP. Not only have patients been able to stop insulin injections and take sulphonylurea tablets instead, but they achieve better glycaemic control, which improves both their quality of life and later risk of diabetic complications.

The second most common cause of permanent neonatal diabetes is a mutation in the insulin gene (INS). Heterozygous INS mutations result in misfolding of proinsulin, which leads to death of the β cell secondary to endoplasmic reticulum stress and apoptosis. Homozygous INS mutations cause decreased biosynthesis of insulin through a variety of mechanisms, including gene deletion, mRNA instability, and abnormal transcription. They result in a more severe phenotype (e.g., earlier age of diagnosis) than the heterozygous mutations.

Neonatal diabetes is genetically heterogeneous, and mutations in the following genes have also been reported in a small number of patients: homozygous or compound heterozygous GCK or PDX1 (also known as IPF1) mutations (heterozygous mutations cause MODY); heterozygous HNF1B mutations (also cause renal cysts) or SLC2A2 mutations (cause Fanconi-Bickel syndrome); homozygous PTF1A or GLIS3 mutations result in diabetes with cerebellar aplasia or hypothyroidism respectively. Neonatal diabetes is also a feature of Wolcott-Rallison Syndrome (homozygous or compound heterozygous EIF2AK3 mutations) and the X-linked IPEX syndrome (FOXP3 mutations). There are still many patients without a genetic diagnosis, which suggests that there are likely to be other monogenic aetiologies.

Type 1 Diabetes

Initial research into the genetics of diabetes tended to focus on type 1 diabetes, where there is greater evidence for familial clustering (λ_s is 15 for T1DM versus 3.5 for T2DM [p. 146]). The concordance rates in monozygotic and dizygotic twins are around 50% and 12%, respectively. These observations point to a multifactorial aetiology with both environmental and genetic contributions. Known environmental factors include diet, viral exposure in early childhood, and certain drugs. The disease process involves irreversible destruction of insulin-producing islet β cells in the pancreas by the body’s own immune system, perhaps as a result of an interaction between infection and an abnormal genetically programmed immune response.

The first major breakthrough came with the recognition of strong associations with the HLA region on chromosome 6p21. The original associations were with the HLA B8 and B15 antigens that are in linkage disequilibrium with the DR3 and DR4 alleles (pp. 147, 200). It is with these that the T1DM association is strongest, with 95% of affected individuals having DR3 and/or DR4 compared with 50% of the general population. Following the development of PCR analysis for the HLA region, it was shown that the HLA contribution to T1DM susceptibility is determined by the 57th amino acid residue at the DQ locus, where aspartic acid conveys protection, in contrast to other alleles that increase susceptibility.

The next locus to be identified was the insulin gene on chromosome 11p15, where it was shown that variation in the number of tandem repeats of a 14-bp sequence upstream to the gene (known as the INS VNTR) influences disease susceptibility. It is hypothesized that long repeats convey protection by increasing expression of the insulin gene in the fetal thymus gland, thereby reducing the likelihood that insulin-producing β cells will be viewed as foreign by the mature immune system.

These two loci contribute λ_svalues of approximately 3 and 1.3, respectively. However, the total risk ratio for T1DM is around 15. Confirmation that other loci are involved came first of all from linkage analysis using breeding experiments with the nonobese diabetic (NOD) strain of mice. These mice show a very high incidence of T1DM, with immunopathological features similar to those seen in humans. These linkage data pointed to the existence of 9 or 10 different susceptibility loci in mice. Following this, the results of numerous genome-wide linkage scans in humans provided evidence for the existence of between 15 and 20 susceptibility loci. However, apart from HLA, many of the other regions were not consistently replicated in independent studies. The largest and most recent linkage study in 2009 (2496 families and 2658 affected sibling pairs, p. 147) provided evidence of linkage to only one previously identified region of linkage besides HLA and INS: a region near to CTLA4. This locus was one of the only three (others encompassing the PTPN22 and IL2RA [CD25] genes) identified and confirmed in candidate gene association studies conducted between 1996 and 2007.

Since 2006, GWA studies p. 149) of increasing size have led to an explosion in the number of T1DM susceptibility loci supported by robust statistical evidence, bringing the total to over 40 distinct genomic locations. It is likely that many more remain to be identified through future, even larger, efforts. Most of the identified loci confer a modest increase in the risk of T1DM, with odds ratios (p. 148) ranging from 1.1 to 1.3 for each inherited allele, in contrast with the much larger role of the HLA locus. In most cases, the causal genes and variants underlying the associations have yet to be identified. However, the regions of association often encompass strong biological candidates—for example, the interleukin genes, IL10, IL19, IL20, and IL27. In two notable cases, follow-up studies have already enabled the causal gene to be confirmed, deepening our understanding of the biological pathways behind the associations.

The first example was a study of the IL2RA (CD25) locus by Dendrou et al. (2009). It used the UK-based Cambridge BioResource, a collection of approximately 5000 volunteers who can be recalled to participate in research on the basis of their genotype. Using fewer than 200 of these individuals, and by means of flow cytometry to assay the levels of CD25 protein expressed on the surface of T-regulatory cells, the study showed that people with the T1DM-protective haplotype expressed higher CD25 levels. This confirmed that IL2RA is indeed the causal gene and that the genotype-phenotype association is mediated via differences in expression of the gene product.

In the second study by Nejentsev et al. (2009), the exons and splice sites of 10 candidate genes situated in regions of genome-wide association were resequenced in 480 T1DM patients and 480 controls. Variants identified were then tested for association with the disease in 30,000 further subjects. Four rare variants (minor allele frequency ≈ 1% to 2%) in the IFIH1 gene were identified, each of which independently reduced the odds of T1DM by about 50%. This finding demonstrated that the IFIH1 gene is important in the aetiology of T1DM. Since its function is to mediate the induction of an interferon response to viral RNA, it adds to the evidence implicating viral infection in the development of the disease. These results also demonstrate that there may be both high- and low-frequency susceptibility variants at the same locus, with varying effect sizes. Future follow-up by resequencing of other loci, both in T1DM and in other diseases, should lead to the identification of even more of these variants and a better understanding of the loci.

Type 2 Diabetes

The prevalence of T2DM is increasing and is predicted to reach 300 million affected worldwide by 2025. Although commonly believed to be more benign than the earlier-onset, insulin-dependent type 1 diabetes, patients with T2DM are also prone to both macrovascular and microvascular diabetic complications, with corresponding excess morbidity and mortality.

Table 15.2 lists the known susceptibility loci for T2DM. There is no overlap with the T1DM loci, illustrating that these two diseases have very different aetiologies. Unlike the HLA and INS VNTR loci in T1DM, there are no major predisposing loci associated with T2DM. Most odds ratios are modest (between 1.05 and 1.3 per allele). As a result, large candidate gene and GWA studies (p. 146) have enabled progress in identifying the loci, whereas linkage studies (p. 146) have been underpowered and therefore unsuccessful.

Table 15.2 Known Susceptibility Loci for Type 2 Diabetes

Human models have proven useful for identifying candidate genes in T2DM. Mutations in all five of the genes identified in this way (PPARG, KCNJ11, WFS1, HNF1B, HNF1A) cause rare monogenic forms of diabetes. One of the variants is population specific: the G319S variant in the HNF1A gene has only been found in the Oji-Cree population in Ontario, Canada.

The TCF7L2 locus has the largest odds ratio of all T2DM loci found in multiple populations. Individuals who inherit two risk alleles (approximately 9% of Europeans) are at nearly twice the risk of T2DM as those who inherit none. The locus was discovered in large-scale association studies of a region on chromosome 10, which was originally identified in linkage studies. However, the TCF7L2 variant does not account for the linkage in this region, suggesting that other rarer but more penetrant variants may be close by. As with many of the other loci, the TCF7L2 risk allele is associated with impaired β-cell function, highlighting the importance of the β cell in T2DM aetiology.

By far the greatest progress in identifying T2DM susceptibility loci has come from GWA studies. To date, 16 loci have been confirmed in T2DM case-control GWA studies, and 6 loci that were originally associated genome-wide with fasting glucose levels were later shown to predispose to T2DM. Many of the genes situated in the identified loci were never thought to be biological candidates, so their discovery has opened up new avenues for research. For example, the FTO gene, which harbours variants associated with body fat mass, was of previously unknown function. Subsequent research has shown, using bioinformatics and animal models, that it has a potential role in nucleic acid demethylation and is expressed in the hypothalamic nuclei of the brain, which govern energy balance and appetite.

In 2008, analysis of the combined effects of 18 of these loci by Lango, Weedon, and colleagues suggested that they increase the likelihood of disease in an additive way. The 1.2% of Europeans in the study who inherited more than 24 risk alleles were over 4 times more likely to develop T2DM than the 2% who inherited only 10 to 12 risk alleles. It is likely that many more loci will be identified through future meta-analyses of GWA studies and that detailed follow-up of the associated regions will lead to identification of the causal variants. The large number of predisposing loci highlights multiple targets for intervention, but there is much work to be done to translate these findings into useful clinical applications.

Crohn Disease

Inflammatory bowel disease (IBD) includes two clinical subtypes: Crohn disease and ulcerative colitis. Its prevalence in Western countries is 1% to 2%, and the estimated λ_s is 25. Positional cloning for IBD identified a striking linkage peak at chromosome 16p12, which was linked to Crohn disease but not ulcerative colitis in the majority of studies. Crohn disease is characterized by perturbed control of inflammation in the gut and with its interaction with bacteria.

In 2001, two groups working independently and using different approaches identified disease-predisposing variants in the CARD15 gene (previously known as NOD2). One of the groups, Ogura et al., had previously identified a Toll-like receptor (p. 193), NOD2, which activates nuclear factor Kappa-B (NFκB) (p. 201), making it responsive to bacterial lipopolysaccharides. The CARD15 gene is located within the 16p12 region and was therefore a good positional and functional candidate. Sequence analysis revealed three variants (R702W, G908R and 3020insC) that were shown by case-control and transmission disequilibrium tests to be associated with Crohn disease. The second group, Hugot et al., fine-mapped the 16p12 region by genotyping SNPs within the 20Mb interval and also arrived at the same variants within the CARD15 gene. These variants are found in up to 15% of patients with Crohn disease but only 5% of controls. The relative risk conferred by heterozygous and homozygous genotypes was approximately 2.5 and 40, respectively. For therapy, drugs which target the NFκB complex (p. 201) are already the most effective drugs currently available.

Since 2006, GWA studies have identified over 30 susceptibility loci for Crohn disease, all of which confer more modest risks of disease than the CARD15 variants (odds ratios per allele between 1.1 and 2.5). Discoveries of loci containing the IRGM and ATG16L1 genes were particularly exciting findings, as these genes are essential for autophagy, a biological pathway whose relevance to the disease was previously unsuspected. Further studies of the IRGM locus by McCarroll and colleagues (2008) identified that the causal variant is a 20-kb deletion immediately upstream of IRGM, which is in linkage disequilibrium (p. 138) with the associated SNPs. The deletion results in altered patterns of gene expression, which in turn were shown to modulate the autophagy of bacteria inside cells. Efforts are already underway to translate this finding into therapeutic applications.

Hypertension

Hypertension (chronically elevated blood pressure) leads to increased morbidity and mortality through a greater risk of stroke and coronary artery and renal disease. Various studies have shown that between 10% and 25% of the population is hypertensive, but the prevalence is age dependent, with up to 40% of 75- to 79-year-olds being hypertensive. Elevated blood pressure may contribute up to 50% of the global cardiovascular disease epidemic. There is substantial evidence that treatment of hypertension prevents development of these complications.

Persons with hypertension fall into two groups. In one, the onset is a consequence of another disorder such as kidney disease. In the other more common group, hypertension usually begins in middle age and has no recognized cause. This is known as essential hypertension. The following discussion is concerned with only essential hypertension.

Environmental Factors in Hypertension

Environmental factors such as high dietary sodium levels, obesity, alcohol intake, and reduced exercise are recognized as being associated with an increased risk of hypertension. Hypertension is also more prevalent in persons from poorer socioeconomic groups. Studies of adopted children have shown lower correlation of their blood pressure with their biological parents than with children remaining with their biological parents. In addition, migration studies involving persons moving from a population with a low prevalence to one with a high prevalence of hypertension have shown that the immigrant group acquires the frequency of hypertension of their new population group during the course of one to two generations. This suggests that environmental factors are of major importance in the aetiology of hypertension.

Genetic Factors in Hypertension

Family and twin studies have shown that hypertension is familial (Table 15.3) and that blood pressure correlates with the degree of relationship (Table 15.4). These findings suggest the importance of genetic factors in the aetiology of hypertension. In addition, there are differences in the prevalence of hypertension between populations, hypertension being more common in persons of Afro-Caribbean origin and less common in Eskimos, Australian Aborigines, and Central and South American Indians.

Table 15.3 Recurrence Risks for Hypertension

Group	%
Population	5
2 Normotensive parents	4
1 Hypertensive parent	8–28
2 Hypertensive parents	25–45

From Burke W, Motulsky AG 1992 Hypertension. Chapter 10 in King RA, Rotter JI, Motulsky AG eds The genetic basis of common diseases. New York: Oxford University Press

Table 15.4 Coefficient of Correlation for Blood Pressure in Various Relatives

Group	Correlation Coefficient
Siblings	0.12–0.34
Parent/child	0.12–0.37
Dizygotic twins	0.25–0.27
Monozygotic twins	0.55–0.72

From Burke W, Motulsky AG 1992 Hypertension. Chapter 10 in King RA, Rotter JI, Motulsky AG eds The genetic basis of common diseases. New York: Oxford University Press

Susceptibility Genes

Rare homozygous mutations in the renal salt-handling genes, SLC12A3, SLC12A1, and KCNJ1, cause recessive diseases characterized by severe reductions in blood pressure. However, resequencing of these genes by Ji et al. (2008) showed that heterozygous rare variants (minor allele frequency < 0.1%) are present in healthy individuals and contribute to blood pressure variation in the general population. These variants cause clinically relevant reductions in blood pressure and protect against hypertension.

Common genetic variants also influence normal blood pressure variation. In 2009, very large meta-analyses and replication of GWA studies (p. 149) were published by the Global Blood Pressure Genetics (Global BPgen) consortium (N > 100,000 Europeans and > 12,000 Indian Asians) and the Cohorts for Heart and Aging Research in Genome Epidemiology (CHARGE) consortium (N > 29,136 Europeans). Fourteen genetic loci were robustly associated with either systolic or diastolic blood pressure, and all showed evidence of association with hypertension risk. Although the causal genes have not yet been confirmed, the loci highlighted likely candidates, including CYP17A1, rare mutations of which cause a form of adrenal hyperplasia (p. 174) characterized by hypertension.

A particularly interesting association was found on chromosome 12q24. The region of association is large, with linkage disequilibrium extending over 1.6 Mb and encompassing 15 genes. The locus is intriguing because other GWA studies have reported association between the same haplotype that raises blood pressure and type 1 diabetes, coeliac disease, myocardial infarction, eosinophil count, and platelet count. These pleiotropic effects are thought to be the result of selection which rapidly increased the frequency of the haplotype in Europeans approximately 3400 years ago at a time when human settlements were enlarging. Further studies will clarify whether the haplotype contains multiple functional variants that give rise to the pleiotropy.

Coronary Artery Disease

Coronary artery disease is the most common cause of death in industrialized countries and is rapidly increasing in prevalence in developing countries. It results from atherosclerosis, a process taking place over many years which involves the deposition of fibrous plaques in the subendothelial space (intima) of arteries, with a consequent narrowing of their lumina. Narrowing of the coronary arteries compromises the metabolic needs of the heart muscle, leading to myocardial ischemia, which if severe, results in myocardial infarction.

For the majority of persons, their risk of coronary artery disease is multifactorial or polygenic in origin. A variety of different genetic and environmental risk factors have been identified that predispose to early onset of the atherosclerotic process, including lack of exercise, dietary saturated fat, and smoking.

Lipid Metabolism

The metabolic pathways by which the body absorbs, synthesizes, transports, and catabolizes dietary and endogenous lipids are complex. Lipids are packaged in intestinal cells as a complex with various proteins known as apolipoproteins to form triglyceride-rich chylomicrons. These are secreted into the lymph and transported to the liver, where, in association with endogenous synthesis of triglyceride and cholesterol, they are packaged and secreted into the circulation as triglyceride-rich very low-density lipoproteins (VLDLs). VLDL is degraded to intermediate-density lipoprotein (IDL), which is further broken down into cholesterol-rich low-density lipoprotein (LDL). High-density lipoproteins (HDLs) are formed from lipoproteins secreted by the liver, chylomicrons, and VLDL remnants.

High levels of LDLs are associated with an increased risk of coronary artery disease. Conversely, high levels of HDLs are inversely correlated with a risk of coronary artery disease. Consequently, the LDL:HDL ratio has been used as a risk predictor for coronary artery disease and as an indicator for therapeutic intervention. Statins are effective drugs for lowering LDL cholesterol levels.

Family and Twin Studies

The risk to a first-degree relative of a person with premature coronary artery disease, defined as occurring before age 55 in males and age 65 in females, varies between 2 and 7 times that for the general population (Table 15.5). Twin studies of concordance for coronary artery disease vary from 15% to 25% for dizygotic twins and from 39% to 48% for monozygotic twins. Although these figures support the involvement of genetic factors, the low concordance rate for monozygotic twins clearly supports the importance of environmental factors.

Table 15.5 Recurrence Risks for Premature Coronary Artery Disease

Proband	Relative Risk
Male (<55 Years)
Brother	5
Sister	2.5
Female (<65 Years)
Siblings	7

Data from Slack J, Evans KA 1966 The increased risk of death from ischaemic heart disease in first degree relatives of 121 men and 96 women with ischaemic heart disease. J Med Genet 3:239–257

Single-Gene Disorders of Lipid Metabolism Leading to Coronary Artery Disease

Although there are a number of individually rare inherited disorders of specific lipoproteins, levels of the various lipoproteins and the hyperlipidemias are determined by a complex interaction of genetic and environmental factors. Family studies of some of the hyperlipidemias are, however, consistent with a single gene being a major factor determining genetic susceptibility.

Familial Hypercholesterolemia

The best-known disorder of lipid metabolism is familial hypercholesterolemia (FH) (p. 175). FH is associated with a significantly increased risk of early coronary artery disease and is inherited as an autosomal dominant disorder. It has been estimated that about 1 person in 500 in the general population, and about 1 in 20 persons presenting with early coronary artery disease, are heterozygous for a mutation in the LDLR (low-density lipoprotein receptor) gene. Molecular studies in FH have revealed that it is due to a variety of defects in the number, function, or processing of the LDL receptors on the cell surface (p. 176).

Susceptibility Genes

Since 2007, numerous large-scale GWA and follow-up replication studies have identified 12 susceptibility loci for coronary artery disease and myocardial infarction. The strongest association identified is on chromosome 9p21 (odds ratio per allele ≈ 1.3). The nearest genes, CDKN2A and CDKN2B, are over 100 kb away. Interestingly, the SNPs most strongly associated with coronary artery disease are only 10 kb away from those associated with type 2 diabetes (see Table 15.2). However, the two disease associations are independent and not in linkage disequilibrium with one another. Much work is already being done to investigate the role of ANRIL, a large, noncoding RNA which overlaps with the coronary artery disease–associated haplotype. It is expressed in tissues associated with atherosclerosis, and initial studies have shown correlations between expression of ANRIL transcripts and severity of atherosclerosis. However, additional evidence from large-scale association studies has shown that the same haplotype on 9p21 is associated with abdominal aortic aneurysm and intracranial aneurysm, suggesting that its role is not limited to atherosclerotic disease. Together with the other 11 loci, the locus at 9p21 only explains a small fraction of the heritability of coronary artery disease, and it is likely that many more loci will be identified.

Progress in uncovering susceptibility loci has also come from large GWA studies of lipid levels. Common variants at at least 30 loci are now robustly associated with circulating levels of lipids, with over one-third of these associated with LDL levels. Kathiresan and colleagues (2008) showed that individuals who inherit a higher number of LDL-raising alleles at these loci are more likely to have clinically high LDL cholesterol (>160 mg/dL) than those who inherit few alleles. The frequency of these LDL-raising alleles is higher in patients with coronary artery disease than in controls, indicating that they predispose to the disease via their primary effect on LDL levels. In many cases, the genes implicated by the loci are already associated with single-gene disorders. For example, PCSK9 harbours a full spectrum of LDL-altering alleles, from rare mutations which cause large differences in LDL (>100 mg/dL), through low-frequency variants with more modest effects (e.g., PCSK9 R46L has a 1% minor allele frequency and a 16 mg/dL effect size), to common variants at 20% minor allele frequency which change LDL levels by less than 5 mg/dL. The resequencing of further loci is likely to uncover rarer variants and mutations at lipid trait loci, which may further explain genetic susceptibility to coronary artery disease.

Schizophrenia

Schizophrenia is a serious psychotic illness with an onset usually in late adolescence or early adult life. It is characterized by grossly disorganized thought processes and behavior, together with a marked deterioration of social and occupational functioning, and can be accompanied by hallucinations and delusions.

Epidemiology

Schizophrenia is a principal cause of chronic mental illness. There is a 1% lifetime risk for a person to develop schizophrenia, and at any one time, approximately 0.2% of the population is affected. Schizophrenia occurs more commonly in individuals of poorer socioeconomic status and has an earlier age of onset and worse prognosis in males. There is an excess of winter births in schizophrenic individuals, which has suggested that environmental factors such as certain viral infections or nutritional factors could be contributory.

Evidence for Genetic Factors

The nature and extent of the genetic contribution to schizophrenia is unclear. This is partly because of past and continuing controversy concerning the definition of schizophrenia and the term schizoid. The latter term refers to the schizophrenia-like traits often seen in relatives of schizophrenics. The problem arises because clinical criteria to distinguish schizoid from normal personality are lacking. For the sake of simplicity, we can regard the term schizoid as referring to a person with the fundamental symptoms of schizophrenia but in a milder form. It has been estimated that roughly 4% of the general population have schizophrenia or a schizoid personality disorder.

Family and Twin Studies

The results of several studies of the prevalence of schizophrenia and schizoid disorder among the relatives of schizophrenics are summarized in Table 15.6. If only schizophrenia is considered, the concordance rate for identical twins is only 46%, suggesting the importance of environmental factors. If, however, schizophrenia and schizoid personality disorder are considered together, then almost 90% of identical co-twins are concordant.

Table 15.6 Proportions (%) of First-Degree Relatives of Individuals with Schizophrenia Who Are Similarly Affected or Have a Schizoid Disorder

Susceptibility Genes

Genome-wide association studies of copy number variations (CNVs) have identified large (>500 kb) deletions associated with the condition—for example on chromosomes 1q21.1, 15q13.3, and 22q11.2 (p. 282). These deletions are rare but penetrant: the odds ratio for the 15q13.3 deletion has been estimated at between 16 and 18 in two independent studies. A key observation is that these deletions are not only associated with schizophrenia. The 1q21.1 deletion (pp. 284, 287) has also been associated with autism, learning disability, and epilepsy. Thus, current clinically defined disease boundaries are not mirrored by the underlying genetics. While these deletions explain some of the genetic susceptibility to schizophrenia, they also explain susceptibility to other conditions. It is likely that a better understanding of the genetics will lead to better definition of clinical phenotypes.

Common genetic variants are also implicated in the aetiology of schizophrenia. Recent meta-analyses of GWA studies have identified associations with the HLA region on chromosome 6p21.3-6p22.1, suggesting an immune system component to the risk of disease. Robust associations have also been observed with variants near the NRGN gene and in the TCF4 gene, which implicate biological pathways involved in brain development, cognition, and memory. Analyses of existing GWA data conducted by the International Schizophrenia Consortium in 2009 suggested that there are likely to be thousands more common variants of small effect that collectively explain much of the heritability of schizophrenia.

Alzheimer Disease

Dementia is characterized by an irreversible and progressive global impairment of intellect, memory, social skills, and control of emotional reactions in the presence of normal consciousness. Dementia is aetiologically heterogeneous, occurring secondarily to both a variety of nongenetic causes such as vascular disease and infections such as AIDS, as well as genetic causes. Alzheimer disease (AD) is the most common cause of dementia in persons with either early-onset dementia (less than the age of 60 years, or presenile) or late onset (greater than age 60 years, or senile). The classic neuropathological finding in persons with AD is the presence at postmortem examination of amyloid deposits in neurofibrillary tangles and neuronal or senile plaques. In addition, individuals with Down syndrome have an increased risk of developing dementia (p. 273), which at postmortem has identical CNS findings to those seen in persons with typical AD.

Epidemiology

Limited numbers of studies of the incidence and prevalence of AD are available, owing to problems of ascertainment. However, the risk of developing AD clearly increases dramatically with age (Table 15.7).

Table 15.7 Estimates of Age-Specific Cumulative Prevalence of Dementia

Age Interval (Years)	Prevalence (%)
<70	1.3
70–74	2.3
75–79	6.4
80–84	15.3
85–89	23.7
90–94	42.9
>95	50.9

From Heston LL 1992 Alzheimer’s disease. Chapter 39 in King RA, Rotter JI, Motulsky AG eds The genetic basis of common diseases. New York: Oxford University Press

Twin and Family Studies

Differences in the age of onset of AD in identical twins are consistent with the importance of environmental factors, but there are difficulties with family studies in AD. Many studies are based on a clinical diagnosis. However, a significant proportion of persons with a clinical diagnosis of AD are found to have other causes at postmortem, such as cerebrovascular atherosclerotic disease. Attempts to confirm diagnoses in relatives who have died previously are often unsuccessful. Obviously, given the age of onset, it is generally neither practical nor possible to obtain funding for prospective studies of the risk to offspring. Therefore, family studies of the risk to siblings are the only practical type of family study to provide reliable data. Although there are numerous retrospective reports of families with AD that are consistent with autosomal dominant inheritance, recurrence risks in a number of studies for first-degree relatives are less than 10%. The risks are age related and greater the younger the age at diagnosis in the affected individual.

Biochemical Studies

The amyloid deposits in the neurofibrillary tangles and neuronal plaques have been shown to consist of the amyloid-β A4 precursor protein (APP). The major protein component of the neurofibrillary tangles has been shown to be derived from a microtubule-associated protein (MAP) called tau (τ). Along with other MAPs, it interacts with β-tubulin to stabilize microtubules.

Single-Gene Disorders

The identification of APP in the amyloid deposits of the neuronal plaques, its mapping in or near to the critical region of the distal part of chromosome 21q associated with the phenotypic features of Down syndrome (p. 273), and the increased risk of AD in persons with Down syndrome led to the suggestion that duplication of the APP gene could be a cause of AD. Evidence of linkage to the APP locus was found in studies of families with early-onset AD, and it is now known that mutations in the APP gene account for a small proportion of cases.

Evidence of linkage to early-onset AD was found for another locus on chromosome 14q. Mutations were identified in a proportion of affected individuals in one of a novel class of genes known as presenilin-1 (PSEN1), now known to be a component of the notch signaling pathway (p. 86). A large number of mutations in PSEN1 have now been identified and account for up to 70% of familial early-onset AD. A second gene, presenilin-2 (PSEN2), with homology to PSEN1, was mapped to chromosome 1q and has been shown to have mutations in a limited number of families with AD. PSEN1 and PSEN2 are integral membrane proteins containing multiple transmembrane domains that localize to the endoplasmic reticulum and the Golgi complex. All of the presenile dementias following autosomal dominant inheritance demonstrate high penetrance.

Susceptibility Genes

Polymorphisms in the apolipoprotein E (APOE) gene are the most important genetic risk factor identified for late-onset AD. The locus was initially identified in the early 1990s through linkage studies. The APOE gene has three major protein isoforms, ε2, ε3, and ε4. Numerous studies in various populations and ethnic groups have shown an increased frequency of the ε4 allele in persons with both sporadic and late-onset familial AD. In addition, the ε2 allele is associated with a decreased risk of the disease. The finding of apolipoprotein E in senile plaques and neurofibrillary tangles, along with its role in lipid transport, possibly in relation to the nerve injury and regeneration seen in AD, provides further evidence for a possible role in the acceleration of the neurodegenerative process in AD.

Although the APOE ε4 allele, found in up to 40% of cases, is a clearly important risk factor, the strongest association is with the age of onset rather than absolute risk of developing AD. The APOE ε4 allele is therefore neither necessary nor sufficient for the development of AD, emphasizing the importance of other genetic and environmental aetiological factors.

In 2009, large meta-analyses of GWA studies, involving up to 6000 cases and 10,000 controls, began to extend our knowledge of AD susceptibility loci. Robust evidence of association was found for common variants in the clusterin (CLU) gene (previously known as apolipoprotein J), which is found in amyloid plaques and is an excellent functional candidate. Associations were also confirmed at loci marked by the CR1 and PICALM genes. Further work will be needed to uncover the mechanisms underlying these associations, and it is likely that many more loci with more modest effects remain to be discovered.

Hemochromatosis

Hemochromatosis is a common disorder of iron metabolism that results in accumulation of iron. The liver is the most commonly damaged tissue, with iron deposition leading to cirrhosis and liver failure. Patients are at increased risk of hepatocellular carcinoma. Other organs that may be affected include the pancreas, heart, pituitary gland, skin, and joints. The iron overload is easily treated by venesection, and this is very effective at reducing morbidity and mortality. The ratio of affected males to females is 5 : 1, and the disease is underdiagnosed in the general population but overdiagnosed in patients with secondary iron overload.

Linkage and Gene Identification

In 1996, the HFE gene was discovered close to the HLA region on 6p21. Two variants were described, C282Y and H63D. Between 85% and 100% (depending on population) of affected individuals were found to be homozygous for C282Y, and the carrier frequency in Northern Europe is approximately 1 in 10. H63D is more common in the general population, and homozygosity for this variant is associated with only a modest increased (about fourfold) risk of hemochromatosis. Compound heterozygosity for C282Y and H63D is associated with reduced penetrance; only 1% are thought likely to develop symptoms. Homozygosity for C282Y was thought to confer a high risk of hemochromatosis, and it was suggested that population screening would be useful, since the iron overload is easily treated. However, population-based studies have suggested that the penetrance of hereditary hemochromatosis due to homozygosity for C282Y may be as low as 1%.

Genetic Testing

Genetic testing is now commonplace in families where an index case is found to be homozygous for C282Y or compoundly heterozygous for C282Y and H63D. Current strategy involves first testing the spouse for carrier status, as there is a 1 in 10 chance they will be heterozygous for C282Y. If the offspring have inherited a predisposing genotype, annual monitoring of serum ferritin is recommended to detect iron overload at an early stage and initiate prompt treatment.

Genetic Heterogeneity

Hemochromatosis is a genetically heterogeneous disorder (Table 15.8), with mutations also reported in the transferrin receptor 2 (TFR2) gene and the SLC40A1 gene which encodes ferroportin. In addition to the common recessive adult-onset form, there is a rare juvenile form with iron overload and organ failure before the age of 30 years, which is lethal if untreated. Neonatal hemochromatosis is a severe form of unknown aetiology.

Table 15.8 Genetic Heterogeneity in Hemochromatosis