3.1. The Current Status of Genomic Drug Discovery

 pp. 24-27 in Bioethics and the Impact of Human Genome Research in the 21st Century

Author: Yoshiji Fujita (TARA, University of Tsukuba)

Editors: Norio Fujiki, Masakatu Sudo, and Darryl R. J. Macer
Eubios Ethics Institute

Copyright 2001, Eubios Ethics Institute All commercial rights reserved. This publication may be reproduced for limited educational or academic use, however please enquire with the author.
Drug Discovery in the Past

 Examination of the two major chemistry databases (CAS and Beilstein) reveals that during the over hundred year history of organic chemistry up to the end of 1997 approximately 16 million compounds had been synthesized (79% of the 17 million CAS registered compounds (13.43 million) are organic. 34% of the 7.05 million Beilstein registered compounds (2.4 million) are Beilstein specific. This total is increasing at a yearly rate of 0.7 million. Of these compounds only 1-2% (200-300 thousand) are utilized as reagents (under 150 thousand in CSCHEM, under 210 million in the now unavailable CHEMQUEST ). When it comes to medicines in the Merck Index, where synthetic method, biological activity, absorption, toxicity as well as clinical benefit are identified in terms of "Best-in-Class", only 10-20 thousand compounds remain. There are roughly 450 classes of medicine ( e.g. classing the "antihypertensive drugs" diuretics and B-antagonists separately ) suggesting that it is very common for medicines to be similar in effect. With the recent advances in genomic science it is thought that the number of known medicinal targets will increase to nearly 4,000 over the next 10 years. Will we be able to create compound libraries with enough variety to exploit these?

There are many examples where a drug still under development or already on sale reveals unexpected beneficial effects, for example Viagra originally targeted at circulatory disorders is being used in MED, a renin- inhibitor originally targeted at hypertension has been applied to HIV, Minoxidil originally intended as an anti-hypertensive drug is used to promote hair growth, Amantidine originally targeted at Parkinson's disease has been applied to Type A influenza etc, there are no end of examples. Thalidomide, once driven out from the market due to its toxicity, is making a come-back as a promising anti-cancer (anti-angiogenesis) drug. The frequency of these unexpected discoveries and applications is too high to say that " it's only due to serendipity", but rather shows the limitations and insufficiency of the information we have, as well as of our knowledge. In other words, drug-discovery has just about come to the limit of what can be achieved through dependence on the luck, intuition and experience of individual researchers.

Even if through good luck a compound is found active in screening, and after optimization goes into development, it only has a 10% chance of approval and then reaching the market. This success rate is expected to rise to 30% by 2008, but how is this to be achieved? When we analyze the reasons for compounds failing during development we find that 39% are due to ADME (absorption, metabolism or excretion problems) and 11% are due to toxicity. 30% are dropped as the effects seen in pre-clinical research in animals are not reflected in efficacy in clinical trials. Clinical studies have hitherto been evaluated on the statistical average data of populations multi-variant in age, sex, genetic background and environmental influences. For this reason, after its marketing in the large, general patient population differences in response to a drug appear in1/4 to 1/3 of the patients, and non-responsiveness and/or severe side-effects may be encountered. It has been reported that "for the 3 billion prescriptions made in 1994, 2 million patients were hospitalized and 100 thousand died. This is the fourth major cause of death in the USA and consumes 70 billion dollars in medical costs". There are then numerous challenges that genomic drug discovery has to face in the transition from "population to individual" medicine (i.e. custom-made medicine, personalized medicine) in the 21st century.

Innovation Crisis in Science and Technology

 As the phrase "Innovation Crisis" suggests, it is getting more and more difficult to select, evaluate and exploit what we are really looking for out of the vast information "Big-Bang" caused by the unbelievable pace of evolution in science and technology during the 1990's. Life science in the 21st Century must integrate professional knowledge from diverse areas such as chemistry, pharmaceutical science, genomics, proteomics, functional genomics, bioinformatics, robot system technology etc. This change has a great impact on the whole process of drug discovery, affecting its speed, scale, quality and cost aspects. Even global companies can no longer cope with the speed of technological innovation and information growth by themselves. This is why partnerships and alliances are more frequently utilized now than ever before. Numerous venture companies with new technologies arise almost every day, and takeovers, mergers and alliances take place among these companies themselves, making an overview very difficult to achieve. A skill at "technology selection" based on a wide-ranging vision is very necessary.

Still fresh in our memories is the attention attracted by combinatorial chemistry in the 1990's, when it ushered in the age of large-scale, high-throughput screening. One organic chemist can synthesize no more than 100 compounds per year, whereas one robot system can synthesize and screen ten to several hundred thousand compounds a day. It is no wonder that chemists became worried about their futures. With an increase in sample number increased sensitivity became necessary, and increasing well-number from 96 to 384, miniaturization, and lowering sample amount requirement became necessary to reduce costs. Gathering and analyzing a large volume of data is also effective for building a compound group possessing high activity towards specific targets such as 7-TM, kinases, proteases, and ion-channels (focused library). It is necessary to identify and eliminate compounds which might possess high activity in various screening tests, but also show low solubility, cellular toxicity, and poor absorption. By raising the quality of a library we can decrease the number of compounds screened by approximately ten-fold, while still increasing the probability of a hit by a factor of ten (a total increase in effectiveness of 100 times).

The profiling of the biological characters of a compound (efficacy, toxicity etc.) at the genetic expression level is also proceeding apace (genetic fingerprinting). Compounds can be re-designed to a safer form by identifying genes that relate to unwanted side effects. When disease-related genes are identified, compounds possessing the genetic fingerprint matching that gene group can be investigated and a new treatment may be found. The NCI (National Cancer Institute) has initiated a project to re-evaluate the classification of cancers from the standpoint of genetic information. This Cancer Genome Anatomy Project (CGAP) discloses 45,000 pieces of gene-related information on cancer as well as preparing databases on, and analyzing various cancers by using common array filters at the 24 array centers across the USA. It is absolutely certain that in the future the results will have a large impact on the diagnosis and treatment of cancers.

As described above the "Innovation Crisis" is a great challenge to our "knowledge management" capability. From the vast pool of information we need to effectively select and implement the "Data-Information-Knowledge" process.

The current status and future of genetic research

The draft sequencing of the 3 billion base pairs of the human genome is almost complete, and we are about to set about the identification and functional analysis of the approximately 100 thousand genes (the figure 40-60 thousand is gaining credence) from this dictionary of several hundred to a thousand weighty books written only with the letters A,T,G and C. Since gene-related genes may also become possible drug targets, fierce competition is growing throughout the world over intellectual property rights. The standards for gene patents differ between countries and therefore discussion aimed at drawing up common guidelines is taking place between Japan, the USA and the EU. Over 99.9% of the sequence is common to all people, and the remaining less than 0.1% contributes to individual differences towards disease susceptibility, drug efficacy, side effects etc. In order to edit, assemble and annotate the necessary information from the vast pool of the genomic database we need to utilize the power of bioinformatics.

In parallel with the draft sequencing of the human genome the identification and mapping of Single Nucleotide Polymorphisms (SNPs) are also progressing rapidly. It is said that over 700 thousand useful SNPs will be identified and enter the public domain by the end of this year. The millennium project aiming to identify Japanese population specific SNPs and carry out their frequency analysis has also started. Race and individual differences in drug metabolism enzymes are widely known, and can cause serious side effects when failure to metabolize results in raised blood concentrations. This is the background to the SNP project on the major metabolic enzymes in the Japanese population that is also about to start. The SNP itself is rarely in the disease-causing gene, but rather important as a genetic marker positioned close to the mutated site. They are more useful than the conventional markers such as RFLP or microsatellites because of their higher frequency level, one in every several hundred to one thousand base pairs distributed evenly throughout the genome. Most of them are distributed in the intron area, spliced in the process of post-transcriptional modification. For this reason SNPs in the coding and non-coding regions, cSNPs (coding) and rSNPs (non-coding) of genes correlating to disease-related genes, efficacy, side effects and metabolism are attracting attention. When there is a mutation in the coding region the presence or not of functional change in the product protein can be evaluated, and for mutations in the promoter region effects on the level of protein produced can be investigated. High-density SNP maps will enable us to investigate causative genes for common disease such as diabetes, asthma, heart disease and migraine by comparing SNPs for the healthy population with those for patients.

Once new genes have been identified, next comes their functional analysis. A key technology in this phase is proteonomics, and most of the major genomic companies in the West are shifting towards this area. Genomics deals with a combination of 4 different bases, whereas proteomonics faces a combination of 20 different amino acids, making information unbelievably complicated. Protein, the genomic product, undergoes post-translational modification such as phosphorylation, and through protein-protein interactions etc. affects cellular signal transduction and contributes to the expression of the body's functions. Modification in the coding region may lead to production of a different protein, and cause changes in the 3-D and sub-unit structures, resulting in complete or partial loss of function. Proteonomics is indeed the science of large-scale protein kinetics, possessing a wide range of application from basic research to clinical studies, through the investigation of disease mechanisms, identification of drug discovery targets and of biological markers of efficacy as well as of toxicity etc. Were we to have an index applicable commonly from study in cells, through evaluation, and on to clinical trials in humans, we could decrease the risk to volunteer patients participating in clinical trials, as well as the probability of a product failing during development.

Pharmacogenomics

 Research using genetic information in clinical applications (pharmacogenomics) is being actively pursued. Efforts are being made to obtain high clinical efficacy by identifying drug responders and non-responders, and/or reducing risks by identifying those likely to show severe side effects. Identifying patients who have a mutation in, or lack certain metabolic enzymes (Poor Metabolizers PM) enables control of dosage to ensure that their drug exposure does not rise to high due to high plasma concentrations. The following are several examples:

  1. Terfenadine (anti-histamine)

    2. This compound is mainly metabolized by two enzymes (CYP2D6, CYP3A4), but 10% of the Caucasian population lack CYP2D6 and therefore rely on metabolism by CYP3A4. However this enzyme is inhibited by co-administration of Ketoconazole or Erythromycin, as well as foods such as grapefruit. When metabolism cannot take place, blood levels rise and severe side effects such as ventricular arrhythmia occur. For this reason Terfenadine has been replaced by its active-metabolite Phexofenadine in the US and Western Europe. Lack of CYP2D6 is seen in less than 1% of the Japanese population, making it a good example of race difference.

  2. Uracil anti-cancer drugs

    3. 5-Fluorouracil may rise to very high plasma levels (which may result in death) even after normal dosing when there is a mutation in the gene coding the metabolic enzyme Dihydropyrimidine Dehydrogenase (DPD), a Val335Leu leading to marked decrease in activity and Glu386Ter leading to loss of activity. The past tragedy involving Sorbidine was due to the fact that although it itself did not inhibit DPD, one of its metabolites did, leading to rises in 5-FU concentration, an example of side effects due to drug-drug interaction. For this reason it is important to identify PM patients with a DPD mutation prior to dosing with Uracil anti-cancer drugs.

  3. Herceptin (breast cancer drug)

    4. Breast cancer in patients possessing normal HER2 genes progresses rather slowly over about 6-7 years, but patients with over-expression of the HER2 gene (25-30% of the total breast cancer population) show extremely rapid tumor growth. Administration of the monoclonal antibody Herceptin to these patients produces remarkable effects, and co-administration with another anti-cancer drug, Taxol, results in a shrinkage of over 50% in tumor size in 49% of patients, with some patients even surviving for more than 10-12 months. The high response and efficacy seen in this clinical study couldn't have been achieved without identification of the patients' genotype.

  4. Others

Pharmacogenomic studies are also being carried out for Alzheimer's disease (ApoE), arteriosclerosis (CETP), asthma (L-O promoter region and beta 2 adrenergic receptor), HIV/AIDS and antibiotics (drug resistance) etc.

On the other hand, pharmacogenomics also raises many issues. At present the cost of genotyping is too high to apply this technique to all diseases. So far application has been limited to cancers, HIV/AIDS, diabetes, asthma and other serious diseases urgently in need. Market fragmentation is unavoidable, as the patient population for each drug will become smaller. Building an environment in which pharmaceutical companies can recoup their high R&D costs is necessary. With the worldwide suppression of medical costs, how should we manage this high cost of advancing technology? Should we focus on medicines that bring the maximum efficacy to the maximum patient population? Or should we focus on decreasing risk factors even for small patient populations?

Conclusion

 With the rapid developments in genomic science the number of issues in various areas of society that need urgent discussion is increasing rapidly. Japan is still at the stage of building the framework for genetic research in the areas of bioethics, protection of privacy, and with respect to its legal and political aspects, and is clearly behind the USA and Europe. Unlike hitherto used diagnostic methods, genetic information not only reveals the individuals personal status, but also affects their whole family, and for this reason requires thorough consideration of the rights-and-wrongs of diagnosis, explanations to the patient and the obtention of informed consent, handling of results, counseling, mental care etc. However the number of physicians in Japan with a professional counseling permit recognized by a national society is only 430 (370 in the Society of Human Genetics and 60 in the Society of Clinical Genetics: source the Asahi newspaper), or only 0.2% of all the doctors in this country. A huge gap in awareness also exists between, on the one hand, large hospitals and national laboratory institutes at the forefront of technology, and, on the other, small and medium-sized hospitals. In the field of medical education as well, we have a definite shortage of paramedics able to cope with the new medical needs, though the need for training such staff has been stressed. Both the handling of genetic information, and its relation to insurance systems is attracting worldwide social concern, with examples also arising in Japan.

There can be no doubting that in the 21st century we are heading towards the era of personalized medicine, but for we, the patient, to participate actively in treatment, selecting our own medical procedures, requires the provision of adequate explanations, information disclosure, counseling and privacy protection. The recent trend has been to put too much weight on genotype research, but I would also like to emphasize the importance of linking this with clinical (phenotype) information. I would like to close my presentation by reminding you all that the principle of drug discovery is not "selecting the patient" that suits the drug, but rather "selecting the drug" that most patients can use without discomfort.

Please send comments to Email < asianbioethics@yahoo.co.nz >.

To contents page
To Eubios book list
To Eubios Ethics Institute home page