Molecular Informatics: Sharpening Drug Design's Cutting Edge

Darren R. Flower

EDWARD JENNER INSTITUTE FOR VACCINE RESEARCH, COMPTON, BERKSHIRE RG20 7NN, UK

1 Introduction

The word 'drug', which derives from the Middle English word 'drogge', first appears in the English language during the 14th century and it has, at least during the last century, become, arguably, one of the most used, and misused, of words, becoming tainted by connotations of misuse and abuse. The dictionary definition of a drug is: 'a substance used medicinally or in the preparation of a medicine. A substance described by an official formulary or pharmacopoeia. A substance used in the diagnosis, treatment, mitigation, cure, or other prevention of disease. A nonfood substance used to affect bodily function or structure.' Even within the pharmaceutical industry, possessed, as it is, by a great concentration of intellectual focus, the word has come, in a discipline-dependent way, to mean different things to different people. To a chemist a drug is a substance with a defined molecular structure and attributed activity in a biological screen or set of screens. To a pharmacologist a drug is primarily an agent of action, within a biological system, but typically without a structural identity. To a patent lawyer it is an object of litigious disputation. To a marketing manager it is foremost a way to make money. To a patient – the pharmaceutical industry's ultimate end-user – a drug is possibly the difference between life and death.

Unmet medical need is, then, a constant stimulus to the discovery of new medicines, be they small molecule drugs, therapeutic antibodies, or vaccines. This unmet need has many diverse sources, including both life-threatening conditions – such as arise from infectious, genetic, or autoimmune disease – and other conditions that impinge deleteriously upon quality of life. The division between the causes of disease is seldom clear cut. Genetic diseases, for example, can be roughly divided between those resulting from Mendelian and multifactorial inheritance. In a Mendelian condition, changes in the observed phenotype arise from mutations in a single dominant copy of a gene or in both recessive copies. Multifactorial inheritance arises from mutations in many different genes, often with a significant environmental contribution. The search for genes causing Mendelian disorders has often been spectacularly successful. Multifactorial diseases, on the other hand, have rarely yielded identifiable susceptibility genes. The identification of NOD2 as causative component for Crohn's disease has been hailed as a major technical breakthrough, leading, or so it is hoped, to a flood of susceptibility genes for multifactorial diseases. Unfortunately, the mode of inheritance in many multifactorial diseases is probably so complex that the subtle interplay of genes, modifier genes, and causative multiple mutations, which may be required for an altered phenotype to be observed, will, for some time yet, defy straightforward deduction.

Heart disease, diabetes, and asthma are all good examples of multifactorial disroders. Asthma, in particular, is, arguably, one of the best exemplars of the complex influence of environmental factors on personal wellbeing. It is a major health care problem affecting all ages, although it is not clear if the disease is a single clinical entity or a grouping of separate clinical syndromes. Asthma is a type I, or atopic, allergic disease, as contrasted with type II (cytotoxic), type III (complex immune), or type IV (delayed type). The word 'asthma', like the word 'drug', first appears in English during the 14th century. It derives from the Middle English word asma: a Medieval borrowing from Latin and Greek originals, although the incidence of allergic disease has been known since ancient times. It is a condition marked by paroxysmal or laboured breathing accompanied by wheezing, by constriction of the chest, and attacks of gasping or coughing. It is generally agreed, that, over the past half-century, the prevalence of asthma, and type I allergies in general, particularly in western countries, has increased significantly. The reasons for this are complex, and not yet fully understood. Clearly, improvements in detection will have made a significant contribution to the increased apparent incidence of asthma, and other allergies, as is seen in many other kinds of condition, although this will only make a partial contribution to the overall increase. Other causative factors include genetic susceptibility; increased allergen exposure and environmental pollution; underlying disease; decreased stimulation of the immune system (the so-called hygiene or jungle hypothesis); and complex psycho-social influences. This final class includes a rich and interesting mix of diverse suggested causes, including the increasing age of first time parents, decreased family size, increased psychological stress, the increase in smoking amongst young women, decreases in the activity of the young, and changes in house design. The last of these, which includes increased use of secondary or double glazing, central heating, and fitted carpets has led to a concomitant increase in the population of house dust mites such as Dermatophagoides farinae and Dermatophagoides pteronyssinus, which are believed to be key sources of indoor inhaled aero-allergens.

Amongst the rich, developed countries of the first world – the pharmaceutical industry's principal target population – some of the most pressing medical needs are, or would seem to be, a consequential by-product of our increasingly technologized, increasingly urbanized personal lifestyles. These include diseases of addiction or over-consumption, those that characterize the West's ageing population, and those contingent upon subtle changes in our physical environment.

Certain diseases have increased in prevalence, while the major killers of preceding centuries – infectious diseases – have greatly diminished in the face of antibiotics, mass vaccination strategies, and improvements in hygiene and public health. In 1900, the primary causes of human mortality were influenza, enteritis, diarrhoea, and pneumonia, accounting between them for over 30% of deaths. Together, cancer and heart disease were responsible for only 12% of deaths. Today, the picture is radically different, with infectious disease accounting for a nugatory fraction of total mortality. Chronic diseases – the so-called 'civilization diseases' – account, by contrast, for over 60% of all deaths.

Many of these diseases, and indeed many other diseases per se, are preventable, and the development of long-term prophylactics, which may be taken over decades by otherwise healthy individuals, is a major avenue for future pharmaceutical exploration. Hand in hand with the newly emergent discipline of pharmacogenetics, the development of prophylactics offers many exciting opportunities for the active prevention of future disease. As Benjamin Franklin inscribed in Poor Richard in 1735: 'An ounce of prevention is worth a pound of cure'. However, for drugs of this type, problems common in extant drugs will be greatly magnified. 'Show me a drug without side effects and you are showing me a placebo,' a former chair of the UK's committee on drug safety once commented. As pharmaceutical products, of which Viagra is the clearest example, are treated more and more as part of a patient's lifestyle, the importance of side effects is likely to grow. A recent study concluded that over 2 million Americans become seriously ill every year, and over 100,000 actually die, because of adverse reactions to prescribed medications. A serious side effect in an ill patient is one thing, but one in a healthy person is potentially catastrophic in an increasingly competitive market place. If the industry is able to convince large sections of the population that it has products capable of preventing or significantly delaying the onset of disease, then financially, at least, the potential market is huge. Whether such persuasion is possible, and who would bear the cost of this endeavour, only time will tell.

Important amongst civilization diseases are examples that arise from addiction and over-consumption. While obesity undoubtedly has a genetic component, it also results from a social phenomenon, with a significant voluntary component, related in part to improvements in the quality and availability of food. Likewise, diseases relating to the addiction to drugs of misuse (tobacco, alcohol, and other illegal drugs, such as heroin or cocaine) give rise to both direct effects – the addiction itself – and dependent pathological impairment, such as lung cancer or heart disease. There is a need to intervene both to address and to mitigate the behaviour itself, primarily through direct drug treatment, with or without psychological counselling, and to address its resulting harmful physiological by-products. Caring for these consequent phenomena has now becoming a major burden on health services worldwide. As individuals, people find dieting difficult and giving-up strongly addictive substances, such as tobacco, even more difficult; pharmaceutical companies are now beginning to invest heavily in the development of anti-obesity drugs and nicotine patches, inter alia, as an aid to this endeavour. For example, the appetite supressant anti-obesity drug Reductil or Sibutramine – a serotonin, norepinephrine, a dopamine reuptake inhibitor – has recently been licensed by the National Institute for Clinical Excellence in the UK. Vaccines are also being developed to alter the behavioural effects of addictive drugs such as nicotine and cocaine. Xenova's therapeutic vaccine TA-NIC, a treatment for nicotine addiction, has recently entered Phase I clinical trials to test the safety, tolerability and immunogenicity of the vaccine in both smokers and non-smokers. TA-NIC is thought to be the first anti-nicotine addiction vaccine to be clinically tested. Other therapies for nicotine addiction include skin patch nicotine replacement, nicotine inhalers or chewing gum, or treatment with the nicotine-free drug Bupropion. A Xenova anti-cocaine addiction vaccine, TA-CD, is currently in Phase II clinical development. We shall see, as time passes, that this type of direct pharmaceutical intervention, targeting the process of addiction rather than just treating its outcome, will doubtlessly increase in prevalence.

The ageing population apparent in western countries is, amongst other causes, a by-product of the increased physical safety of our evermore comfortable, urbanized, post-industrial environment. Together with decades of enhanced nutrition and the effects of direct medical advancement in both medicines and treatment regimes, this has allowed many more people to exploit their individual genetic predisposition to long life. Estimates based on extant demographic changes would suggest that by 2050 the number of the super-old, i.e. those living in excess of 100 years, would, within the USA, be well in excess of 100,000. In terms of its implications for drug discovery, this has led to a refocusing of the attention of pharmaceutical companies onto gerantopharmacology and the diseases of old age. Examples of these include hitherto rare, or poorly understood, neurodegenerative diseases, such as Parkinson's disease, or those conditions acting via protein misfolding mechanisms, which proportionally affect the old more, such as Alzheimer's disease. The prevalence of stroke is also increasing: approximately 60,000 people die as the result of a stroke annually in England and Wales and approximately 100,000 suffer a non-fatal first stroke. However, the relative proportion of young people suffering a stroke has also increased. Here, 'young' refers to anyone under 65, but stroke is not unknown in people very much younger, including infants and children. Indeed, 250 children a year suffer a stroke in the United Kingdom. This disquieting phenomenon may, in the era of routine MRI scans, simply reflect the greater ease of successful detection amongst the young as well as the old.

Looking more globally – though the danger is still real in developed western countries – new or re-emergent infectious diseases, such as AIDS or tuberculosis, pose a growing threat, not least from those microbes exhibiting drug and antibiotic resistance. As the world appears to warm, with weather patterns altering and growing more unpredictable, the geographical spread of many tropical infectious diseases is also changing, expanding to include many areas previously too temperate to sustain these diseases. The threat from infectious disease, which we have seen has been largely absent for the last 50 years, is poised to return, bringing with it the need to develop powerful new approaches to the process of anti-microbial drug discovery.

From the foregoing discussion, we can identify a large array of new, or returning, causes of human disease, which combine to generate many accelerating and diversifying causes of medical need. These come from infectious disease, which have evolved, with or without help from human society, to exhibit pathogenicity, but also from diseases of our own creation, such as those resulting directly, or indirectly, from addiction or substance abuse, to other disease conditions, which have not previously been recognized, or have not been sufficiently prevalent, due to our ageing population or changing economic demographics. Patterns of disease have changed over the past hundred years and will change again in the next hundred. Some of these changes will be predictable, others not. Medical need is ever changing and is always at least one step ahead of us. Thus the challenge to medicine, and particularly the pharmaceutical industry, has never been greater, yet neither has the array of advanced technology available to confront this challenge. Part of this is experimental: genomics, proteomics, high throughput screening (HTS), etc., and part is based on informatics: molecular modelling, bioinformatics, cheminformatics, and knowledge management.

2 Finding the Drugs. Finding the Targets

Within the pharmaceutical industry, the discovery of novel marketable drugs is the ultimate fountainhead of sustainable profitability. The discovery of candidate drugs has typically begun with initial lead compounds and then progresses through a process of optimization familiar from many decades of medicinal chemistry. But before a new drug can be developed, one needs to find the targets of drug action, be that a cell-surface receptor, enzyme, binding protein, or other kind of protein or nucleic acid. This is the province of bioinformatics.

2.1 Bioinformatics

Bioinformatics, as a word if not as a discipline, has been around for about a decade, and as a word it tends to mean very different things in different contexts. A simple, straightforward definition for the discipline is not readily forthcoming. It seeks to develop computer databases and algorithms for the purpose of speeding up, simplifying, and generally enhancing research in molecular biology, but within this the type and nature of different bioinformatic activity varies widely. Operating at the level of protein and nucleic acid primary sequences, bioinformatics is a branch of information science handling medical, genomic and biological information for support of both clinical and more basic research. It deals with the similarity between macromolecular sequences, allowing for the identification of genes descended from a common ancestor, which share a corresponding structural and functional propinquity.

Within the drug discovery arena bioinformatics equates to the discovery of novel drug targets from genomic and proteomic information. Part of this comes from gene finding: the relatively straightforward searching, at least conceptually if not always practically, of sequence databases for homologous sequences with, hopefully, similar functions and roles in disease states. Another, and increasingly important, role of bioinformatics is managing the information generated by micro-array experiments and proteomics, and drawing from it data on the gene products implicated in disease states. The key role of bioinformatics is, then, to transform large, if not vast, reservoirs of information into useful, and useable, information.