The completion of the human genome project has heralded a new era in biology. Undoubtedly, knowledge of the genetic blueprint will expedite the search for genes responsible for specific medical disorders, simplify the search for mammalian homologues of crucial genes in other biological systems and assist in the prediction of the variety of gene products found in each cell. It can also assist in determining the small but potentially significant genetic variations between individuals. However, sequence information alone is of limited value without a description of the function and, importantly, of the regulation of the gene products. Our bodies consist of hundreds of different cell types, each designed to perform a specific role that contributes to the overall functioning of the organism. Every one of these cells contains the same 20 000 to 30 000 genes that we are estimated to possess. The remarkable diversity in cell specialization is achieved through the tightly controlled expression and regulation of a precise subset of these genes in each cell lineage. Further regulation of these gene products is required in the response of our cells to physiological and environmental cues. Most impressive perhaps is how a tightly controlled program of gene expression guides the development of a fertilised oocyte into a full-grown adult organism. The human genome has been called our genetic blueprint, but it is the process of gene expression that truly brings the genome to life.

Introduction

There is now a large body of evidence indicating that genetic factors can influence treatment response in a variety of human diseases. Pharmacogenetics, defined as the study of the genetic basis for individuality in response to drugs (Vogel, 1959) or use of genetic analysis to predict drug response, efficacy and toxicity (Roses, 2004) has seen an almost exponential rise in peer-reviewed publications in the last 10 years (Figure 34.1). The closely related field of pharmacogenomics can be defined as the use of genetic approaches in drug discovery and use and encompasses the study of all genes which may influence drug response although, as illustrated in Figure 34.1, the terms pharmacogenetics and pharmacogenomics are often used interchangeably.

Inter-individual differences in response to commonly prescribed drugs are increasingly recognized. Serious adverse drug reactions (ADRs) have been estimated to occur in more than 2 million patients/year in the USA and result in more than 100 000 fatalities; meta-analysis of a number of US-based prospective studies suggest that ADRs are the fourth leading cause of death, after heart disease, cancer and stroke (Lazarou et al., 1998). A recent UK-based prospective study reported that up to 6.5% of all hospital admissions in the UK are related to adverse drug reactions, with a projected annual cost to the National Health Service of £466 million (Pirmohamed et al., 2004).

Individuality in drug response can be both genetically and environmentally determined.

Introduction

The current rate of advances in genetic technology and statistical methods makes it difficult to discuss study design in mapping complex disease traits in a way that will have value beyond a relatively short time horizon. This chapter considers how knowledge about the nature of complex diseases and traits can inform study design and confines itself to genomic (rather than proteomic or metabonomic) approaches.

Genetic influences on complex traits can be considered in terms of susceptibility to disease (clinical and pre-clinical), susceptibility to differences in natural history of disease (severity, complications and prognosis), susceptibility to different therapeutic responses (efficacy and adverse effects) or in terms of the genetic determinants of normal phenotypic variation in health.

The choices between approaches depend not only on the context of the study, but also on the relative costs of ascertaining families, measuring phenotypes and genotyping. The costs of genotyping have been falling rapidly over the last decade and the trend is for genotyping to be done in a few automated high-throughput centres to maximize efficiency. In contrast, more stringent ethical and data protection legislation requirements have tended to increase unit recruitment costs, since ascertainment and recruitment procedures become more demanding and remain very labor intensive. It is likely therefore that the requirements for very large sample sizes and for large collaborative studies will increasingly involve research groups from countries of intermediate development which can assure high fidelity phenotyping, but at much lower cost than is possible in most industrialized nations.

The MHC, the region of the genome widely believed to be associated with disease resistance, is in fact linked with more disease susceptibility than any other region of the human genome (Price et al., 1999). One explanation for this paradox is that there is a net cost in providing resistance to infection. In other words, improved resistance to infection, manifest as a more effective immune response, results in a greater propensity to autoimmune disease. In this article we will explore this proposal by examining the main features of the MHC, the functions of the genes it contains and its role in disease (Lechler and Warrens, 2000; Marsh et al., 2000).

Features of the MHC

The human MHC is a gene-dense region that contains genes for the classical class I molecules HLA -A, -B and -C as well as class II molecules DP, DQ and DR, all of which are highly polymorphic (Figure 9.1). The ∼4Mbp complex is located on chromosome 6p21.3 and contains over 220 genes. The main known function of MHC class I and class II molecules is to present peptide fragments from pathogens on the cell surface for recognition by T cells. This guides killing of virus-infected cells, activation of phagocytes, or production of specific antibodies. Pathogens are known to employ a variety of different strategies in their attempts to evade presentation (Vossen et al., 2002) but features of MHC class I and class II genes help to counteract evasion.

Prevalence of hearing impairment

Hearing impairment is undoubtedly a common disease. Around 1.06 per 1000 children are born with a significant, permanent hearing impairment (40 dB or greater increase in threshold in their better hearing ear), and by the age of nine years, this number has risen to around 1.65 per 1000 (Fortnum et al., 2001). The prevalence of hearing impairment continues to increase with each decade of life, until 40% of the 71–80 years age group and 80% of the 80+ age group have a hearing loss of 35 dB or more (Davis, 1989; Davis and Moorjani, 2002). In total, approximately 20% of all adults over 18 in the UK suffer some form of hearing impairment (25 dB or greater hearing loss in at least one ear), and the proportions for other countries are very similar (Davis and Moorjani, 2002). The increase at various impairment levels is illustrated in Figure 33.1. However, thresholds are a crude reflection of the impairment, because it is not just the amplitude but the clarity of hearing that is affected. Our ability to distinguish speech sounds, to focus on specific sound sources such as one speaker in a noisy room, and to localize sounds in the environment all require accurate frequency and temporal discrimination, features that are disproportionately affected by hearing impairment. Much hearing loss with age affects sensitivity to high frequencies first, although some types of hearing loss have a more even effect across the frequency spectrum.

Introduction

The striking success in mapping Mendelian disease genes, coupled with the rapid development of genomic methodologies have generated an initial wave of enthusiasm that progress in understanding the genetics of common diseases might experience similar success. Over the past five years however, this optimism has diminished sharply as the difficulty of the task has become apparent. This difficulty is highlighted by the fact that, to date, very few genes have convincingly been shown to harbor variants that influence predisposition to a common disease (Glazier et al., 2002; Lohmueller et al., 2003).

However, recent progress has been relatively rapid in a number of relevant methodologies for genetic association studies. In particular, there has been considerable progress in making use of linkage disequilibrium (LD) to more efficiently and comprehensively represent human genetic variation in association studies. There now appears reason to be cautiously optimistic that the field may begin to progress more rapidly.

This chapter focuses primarily on factors influencing patterns of human genetic variation, and the implications of the patterns of variation in medical genetics. A central theme is the role of the HapMap project in facilitating genetic association studies. We consider both the use of the HapMap to identify single nucleotide polymorphisms (SNPs) that efficiently represent other SNPs (that is, tagging SNPs), and also the use of the HapMap data to help optimise methods for identifying the causal variants that underlie genotype–phenotype correlations.

Introduction

What is “gene mapping” and why is it useful?

One of the goals of human genetics research is to understand genetic variation between people in their susceptibility to disease. From twin and family studies, and the study of Mendelian disease, it is clear that some traits and diseases “run in families” and that the reason for the increased disease risk of relatives of affected individuals is, at least in part, because of their genetic predisposition. Genetic variation in populations is caused by mutations that cause differences in DNA sequence and by other genome events in the germline, for example insertions, deletions, duplications and translocations of stretches of DNA. If these mutation events have an effect on a phenotype of the carrier, for example an increased risk of disease or an effect on a continuously varying phenotype (such as blood pressure or body mass index), then there will be an association between the genotype and the phenotype. Gene mapping aims to identify locations on the genome that are responsible for genetic variation and, ultimately, to identify which specific variants cause the observed effect. Gene mapping is useful because it leads to an understanding of the nature of genetic variation and the identification of variants and biological pathways that cause or predispose to disease. This knowledge can be used to develop drug targets or other treatments and in the future may be used for disease diagnosis or the assessment of susceptibility to disease.

Coronary heart disease (CHD) is the single commonest cause of death in the developed world. One in four men and one in six women die from CHD. In the UK, around 15% of these deaths occur under the age of 65, and 35% under the age of 75 (www.heartstats.org). CHD frequency varies between populations, with the highest age-adjusted rates of 600–1000 deaths per 100 000 found in countries of the former Soviet Union and the lowest at around 60 per 100 000 in Japan (World Health Organization 2002, www.who.ch/). Age-adjusted rates in the UK and USA are around 200–300 per 100 000 of population. Age-adjusted CHD prevalence in the UK and USA has fallen by around 40% in the past 30 years, although this reflects more a postponement in age of CHD-related death by about 10 years rather than an absolute reduction in numbers of deaths (Fuster, 1999). CHD is predicted to remain the commonest single cause of death in developed countries over the next 20 years and will increase in frequency to become the commonest cause of disease-related disability in both developed and developing countries by the year 2020 (Murray and Lopez, 1997).

Genetic and environmental contributions to CHD pathogenesis

The significant changes in CHD incidence and mortality over the past 20 years can be attributed at least in part to variation in known environmental risk factors (Table 24.1).

An understanding of the complex processes that underlie the transition from zygote to newborn infant remains one of the major unsolved challenges in human biology. Failure of key steps in early embryogenesis leads to arrested development and embryonic wastage in a substantial proportion of conceptions (Wilcox et al., 1999). Interference with later developmental pathways which mediate the processes of morphogenesis and organogenesis can also lead to fetal demise but equally can produce a phenotypic effect evident at term. This chapter discusses, with selected examples, our current understanding of the influence that genetic and environmental factors have on these complex developmental processes in humans.

The medical significance of developmental disorders

Developmental disorders in humans are diverse in nature and individually relatively rare, but as a group constitute a “common disease”. Improvements in their recognition and pathogenesis, both as isolated entities and as components of syndromes, have been greatly aided by advances in the clinical speciality of dysmorphology and the construction of clinical databases which catalogue rare associations of phenotypic features (Donnai and Read, 2003).

The overall birth prevalence of disorders which are primarily considered to be due to defective morphogenesis is estimated to be between 2 and 3% (Kalter and Warnaky, 1983). If malformations associated with still births and abnormalities which do not present a requirement for significant medical intervention are included in this estimate, the figure rises to ∼ 5%.

Introduction

Since its discovery over 25 years ago, the TP53 gene is one of the “stars” of molecular cancer research. The p53 protein acts as an all-round regulator of many interconnected functions associated with cell cycle regulation, apoptosis, DNA repair, differentiation, senescence and development. Activation of p53 prevents DNA replication and cell proliferation when cells are subjected to stress that may perturb genetic or genomic integrity. Thus, TP53 acts as a “master suppressor gene” by exerting simultaneous, many-fingered control of several components of the molecular mechanisms of carcinogenesis. Mutations in TP53 result in loss of these suppressor functions. In some instances, it has been suggested that mutations may also exert gain-of-function effects that may explain the persistence of p53 mutant protein in cancer cells. TP53 is emerging as an important target for improving cancer detection, prognosis and treatment. However, forms of mutant p53 differ from each other and this may affect cancer development in an organ, tissue and context-specific manner. Addressing this diversity is essential for developing cancer management strategies using p53 as a target.

Cancer progression is characterized by acquisition of multiple genetic and epigenetic alterations in genes involved in interrelated processes controlling cell cycle, apoptosis, differentiation, replicative senescence, cell motility and migratory capacity (Hanahan and Weinberg, 2000). There are many ways in which cells develop defects in these processes, often in a cell-type, tissue- and context-specific manner. However, a small number of genes are commonly altered in many different cancers, irrespective of their histology or site.

Introduction

The ability of even the most complex protein molecules to fold to their biologically functional states is perhaps the most fundamental example of biological self-assembly, one of the defining characteristics of living systems (Vendruscolo et al., 2003). The process of protein folding in the cellular environment can, in principle, begin whilst the nascent chain is still attached to the ribosome, and there is clear evidence that some proteins do fold at least partially in such a co-translational manner (Hardesty and Kramer, 2001). Other proteins are known to undergo the major part of their folding only after release from the ribosome, whilst yet others fold in specific cellular compartments such as the endoplasmic reticulum (ER) or mitochondria following translocation through membranes (Hartl and Hayer-Hartl, 2002). Although the fundamental principles underlying the mechanism of folding are unlikely to differ in any significant manner from those elucidated from in vitro studies, many details of the way in which individual proteins fold will undoubtedly depend on the environment in which they are located. In particular, as incompletely folded chains inevitably expose to the outside world many regions of the polypeptide molecule that are buried in the native state, such species are prone to inappropriate interactions with other molecules within the complex and crowded cellular environment. Such interactions can result both in the disruption of normal cellular processes and in self-association or aggregation.

Introduction

Hemoglobin, the oxygen transporting molecule that makes up over 95% of the protein content of the red blood cell, has been at the forefront of research into the genetic causes of disease since the inception of such studies in the mid-twentieth century. In this short space of time, the normal structure and function of the molecule has been described in exquisite detail and accurate structure–function relationships have been derived from analyses of structurally abnormal hemoglobins. The thalassemias demonstrated a different class of disease-causing mutations, those resulting in defects in the synthesis of the globin polypeptides. Globin genes were among the first to be cloned and sequenced, providing an extensive list of mutations while functional studies of these genes have contributed enormously to our understanding of gene regulation. Application of this knowledge through prenatal diagnosis has made a major clinical impact on a growing number of populations and progress is being made in the development of gene therapy for the severe hemoglobin disorders. All in all, hemoglobin has provided a model system for the study of human genetic diseases.

Normal human adult hemoglobin (HbA, α2β2) is a tetrameric molecule consisting of two α-globin chains (141 amino acids) attached to two β-globin chains (146 amino acids). Each chain carries a prosthetic heme group that reversibly binds oxygen. The three dimensional structure of hemoglobin has been determined to a resolution of 1.7Å (Fermi et al., 1984), demonstrating a compact elliptical shape composed largely of α-helices.

Introduction

Since the inception of the current “genetic age” and its characterization as having created a “risk society”, the ethical and legal debate surrounding the regulation of genomics (genetic research and related biotechnology and clinical genetics) has largely been informed by evaluations of risk. These risks pertain to individuals, populations, future generations and the environment. Cutting across these cohorts, the not unsubstantial risks of genetic relativism and determinism have excited concerns about genetic discrimination, triggered debates about the place and form of justice in the modern (genomic) health context, and challenged the practical realization and protection of human rights, individually, internationally and inter-generationally.

As evidenced by the papers which make up this volume, genomic knowledge is accumulating at an impressive, if not unparalleled rate. What is also evident, however, is that our comprehension of genomics remains somewhat rudimentary, particularly when it comes to understanding the interactions that occur between genes, and between genes and the environment. Further, while numerous genetic tests have been developed, we have been less successful at devising effective genetic therapies and treatments. That is not to say that genomics is a purely “frontier” science. On the contrary, it is becoming more and more a part of regular clinical practice, most particularly in the developed world. It has been claimed that there is:

Introduction

Type 2 diabetes accounts for the overwhelming majority of diabetes worldwide (Zimmet et al., 2001) and represents a major and growing challenge to biomedical care. In contrast to many other complex traits, the environmental exposures which contribute to the development of this condition are well characterized: but as they are so pervasive, efforts to reduce the prevalence of this condition through environmental and behavioral manipulation have had only limited impact. Personal risk of developing type 2 diabetes results from the interaction between these pervasive exposures and our individual portfolios of susceptibility and protective genomic variants. Over the past decade, more and more of these variants have been identified and characterized. The challenge for the next decade is to understand how these variants interact with each other and with environment, and to use this information to target preventive and therapeutic interventions to maximize their effect.

Type 2 diabetes: the next global epidemic?

Definitions

In contrast to type 1 diabetes, which is known to result from autoimmune destruction of the insulin-secreting beta-cells of the pancreas, leading to lifelong dependence on exogenous insulin, the etiology of type 2 diabetes is poorly understood (Kahn, 2003). Whilst type 1 diabetes is typically diagnosed in childhood or early adulthood, type 2 diabetes classically presents in later life. These clinical distinctions lie behind previous disease classifications in which type 2 diabetes was known originally as maturity-onset diabetes, and subsequently, as non-insulin-dependent diabetes mellitus (World Health Organization Study Group, 1985).