Until now, drug research has focused on developing blockbusters to treat millions of people. Pharmacogenomics works from the inside out: Researchers use genetic information to interpret disease pathways and create drugs for small, likely-to-respond populations. How it affects payers and patients.
Abstract
Until now, drug research has focused on discovering blockbusters to treat millions of patients. Pharmacogenomics, a multidisciplinary effort arising from the Human Genome Project, strives to deliver “personalized medicine.” Researchers use genetic information to understand disease pathways and create drugs designed for small, likely-to-respond populations. The path from research to finished drugs is as logistically complex as landing a human on the moon, but don’t expect a giant leap; progress will come throughout the next couple of decades via incremental steps.
Aclara Biosciences, in Mountain View, Calif., looks at very small quantities of solid tumor slices to see if protein pathways are present in a tumor. If the protein pathway is not expressed and the protein complexes are not formed, the drug that acts on that pathway will not work in that patient, says Tom Klopack, CEO, right, standing in Aclara’s laboratory facilities.
In tomorrow’s personalized medicine, each person’s unique predisposition to disease will be more knowable, so that a disease could be predicted and treated before the patient is aware of its presence. Personalized, individualized, or targeted medicine — the terms are used interchangeably — refers to the ability to deliver the right drug to the right person at the right dose and time. Of course, another parameter must be added — at the right cost. Ultimately conserving precious healthcare dollars, the approach promises to be cost-effective because it focuses on patients with real medical problems and allows prevention and control.
The potential impact of pharmacogenomics is huge. As usual in bursts of creative development, there is much hype and a certain amount of misapplied energy. Yet, there is absolutely no doubt that medicine is moving in a pharmacogenomic direction. “Some oncologists believe that with diagnostics and targeted therapies of the future, cancer will become a disease that is chronic and controlled (as with HIV) rather than acute and often fatal,” says Tom Klopack, CEO of Aclara Biosciences in Mountain View, Calif.
The same fate could await heart disease, metabolic diseases, and mental illness. Research in all these fields could lead to drugs for specific genetic types, with fewer adverse reactions and increased certainty of drug response. Perhaps, some day, adverse events will be practically a thing of the past, thanks to pharmacogenomics. Already, it is estimated that the discovery of genetic variants of cytochrome P450 — a family of more than 60 enzymes that the body uses to break down toxins — can reduce adverse events by up to 20 percent for existing drugs.
Although pharmacogenomics has grown in gold-rush fashion, at times lacking coordination of effort, in 1999 private and public investigators established The SNP Consortium (TSC). Single nucleotide polymorphisms (SNPs), common DNA sequence variations among individuals, promise significant advances in the ability to understand and treat disease. The consortium has discovered and mapped nearly 1.8 million SNPs.
Pharmacogenomics encompasses several disciplines — genetics, proteomics, informatics, and others — that are identifying new biomarkers for disease, or resistance to disease, from a genetic standpoint. Researchers can better calculate drug response among individuals or groups by genetically analyzing tissues (including blood), tumors, or pathogens. Investigators compare samples from healthy with disease-associated protein pathways, RNA concentration and patterns, and cell processes. Companies design drugs or biologics that remedy subcellular malfunctions that may result from minute genetic mistakes, such as “typographical” errors in the nucleotides of DNA, that result in under- or over-expression of genes among other mechanisms.
Pharmacogenomics is not a new phenomenon, but the rate of new discoveries has quickened since the Human Genome Project’s completion.
PHARMACOGENOMICS TIMELINE
| 1865 | Genes are discovered. |
| 1902 | Mendelian genetics is proven in humans. |
| 1940s | Clinical observations are made of slow metabolizers (acetylators). |
| 1950s | Watson and Crick explain the structure of DNA. Individual responses between patients to drugs are noted in clinical studies, with speculation that inherited traits might explain individual differences in response to drugs and with respect to side effects. |
| 1980s | Pharmacogenomics is “born” with the identification of exaggerated adverse drug effects in some patients due to decreased oxidative metabolism. Understanding the biology of poor metabolizers is homozygous for a recessive allele of cytochrome P450 (CYP2D6). |
| 1999/2001 | Consortium to identify single nucleotide polymorphisms (SNPs) established between government and private industry researchers; 1 million human SNPs are identified. |
| 2000/2001 | Human Genome Project is completed and published. |
| Ongoing | New targets and SNPS are developed. |
| FUTURE | |
|
| |
| 2010 | Genetic testing will be commonplace for major illnesses. |
| 2020 | Physicians will be adjusting drugs and doses based on DNA variants. Targeted therapies developed by pharmacogenomics will be available in mental illness, diabetes, heart disease, and cancer. |
“I first became aware of the term pharmacogenomics in the work that was done with Cognex (tacrine) for Alzheimer’s disease,” says Stephen Lash, PharmD, managed care clinical manager at Genentech. “Testing for APOE4 was about finding a way to identify patients who were likely to respond to the drug and less likely to experience adverse events, and to do this using commonly available clinical tools.” APOE4 (apolipoprotein E4) is one form of a protein that shuttles lipids through the body, and it is believed to have a role in the development of Alzheimer’s disease.
COMPENSATING FOR AN INEXACT SCIENCE
Why do we need pharmacogenomics? The best answer is in an eyebrow-raising statement made by Allen D. Roses, senior vice president of genetics research at Glaxo-SmithKline and outspoken proponent of pharmacogenetic drug development: “Most of our drugs don’t work in most of the people most of the time.”
Many current drugs are effective in fewer than 40 percent of patients, even in potentially fatal diseases. Probability statistics suggest that benefits of drugs outweigh their risks for large populations, but some individuals still have serious and unpredictable side effects. An analysis published in JAMA in 1998 found that serious adverse drug reactions accounted for more than 2 million hospitalizations and 100,000 deaths in the United States in one year, ranking adverse drug reactions fifth among major causes of death in the United States. Presumably, that number can be reduced once we obtain more specific information on patient subgroups, such as ethnic groups or genetic haplotypes — groups that share a phenotype that makes the group react a certain way to a physical agent.
“In the future we may not call anyone an ‘asthmatic,’ but we may have different names for people who are allergic to dogs and cats and for people who have other, vastly different etiologies for their airway obstruction.”
— Stephen Lash, PharmD, Genentech
As part of the ever-growing influence of genetics on medicine since the 1980s, pharmacogenomics is not a new phenomenon, but the rate of new discoveries has quickened since the completion of the Human Genome Project. There are about 3 billion base pairs in human DNA, all of which are subject to the errors known as SNPs, which occur in every 100 to 300 base pairs in the human genome. One key element of pharmacogenomics is toxicogenomics, or studying the pharmacokinetics and dynamics of a given therapeutic agent within a specific genotype or haplotype, or even in an individual with a known SNP or a pattern of SNPs. There could be hundreds of undiscovered similar genetic predispositions to drug accumulation/toxicity.
“When treating a patient, we used to prescribe a 350 to 500 mg dose of a hypothetical drug to everyone,” says Jeff Kimmell, RPh, vice president for healthcare services and chief pharmacy officer at drug-store.com. “We may soon find that patient A needs only a 62.5 mg dose and patient B may need 75 mg of a different salt of the same drug, and that patient C needs to be treated for only six days, instead of ten.”
All humans share approximately 99.9 percent of DNA sequences; the other 0.1 percent accounts for individual variations. Some diseases such as hemophilia, Huntington’s disease, sickle cell anemia, and cystic fibrosis result from mutations in single genes; although most diseases result from a combination of genetic “misspellings” involving many genes.
“There are about 10,000 genes in a given cell that can be either turned on or off,” says John Ryan, MD, PhD, senior vice president for experimental medicine at Wyeth Research. “Complex diseases like cancer could involve hundreds of genes.”
PROFOUND IMPACT
It is difficult to overestimate how profoundly pharmacogenomics will affect medicine. Some of the ways medicine may change as a result include:
Reclassification of diseases such as hypertension, long known to have multiple etiologies, into genomic taxonomies
Prediction, early diagnosis, staging, and prognosis of diseases
Innovations in medication design
Dosing/duration of therapy
Foreknowledge of interindividual responses to medications
U.S. Food and Drug Administration approval process and approval time
New partnerships between managed care plans and medical providers to identify appropriate patients for particular therapies
“As genomic tools are sharpened, so will be our ability to dissect disease into its component parts,” predicts Geoffrey Ginsburg, MD, PhD, vice president of molecular medicine at Millennium Pharmaceuticals. “In cardiovascular disease, genetic heterogeneity has been identified in the Long QT syndrome, within at least four different ion channels. Clinical manifestations can range from no visible signs to sudden death. In another example of heterogeneity, it is known that familial cardiomyopathy results from more than 80 different mutations.” The clinical importance of subclassifying this disease and using this knowledge to design drugs that do not affect the QT interval are obvious.
Genentech’s Lash provides a third example affecting millions of patients:
“When we think of an asthmatic patient, we may think of of someone with a mold-spore sensitivity leading to restrictive airway disease. In the future we may not call anyone an ‘asthmatic,’ but we may have different names for people allergic to dogs and cats and for people who have other, vastly different etiologies for their airway obstruction.”
DIAGNOSTICS ARE KEY
While other technologies are under development (Illumina’s work with bead arrays is one example), the work engines of pharmacogenomics are often bioarrays. A variety of diagnostic methods already can identify SNPs in an individual that can affect disease susceptibility or resistance. This is where the abstract world of genomics meets real-world biology. Different companies are looking at different tissues at various stages in protein production and interaction cascades via micro-arrays (biochips) or other high-tech nanotechnologies.
Diagnostic tests based on genetics already are available online at «www.genetests.org», a federally funded Web site that is run by researchers at the University of Washington. “Even though genetic testing adds an extra step, this approach has been accepted by patients and physicians for nearly a decade with, for example, testing a breast tumor for estrogen receptivity before administering tamoxifen,” says Adam Hedgecoe, PhD, a lecturer in sociology at the University of Sussex, England. He is completing a Wellcome Trust postdoctoral fellowship exploring the impact of pharmacogenetics on clinical practice and is author of The Politics of Personalized Medicine (to be published this year by Cambridge University Press).
Five stops in the protein-development line, with examples of current research
| 1 |
Gene sequencing (identify SNPs) and mutations GlaxoSmithKline has compiled a database of DNA mutations since the late 1990s. |
| 2 |
Gene expression (transcription and translation) Currently used in the clinic for prescribing medications — e.g., predicting a black patient’s response to an angiotensin-converting enzyme inhibitor or a Japanese patient’s response to gefitinib. In preliminary research, Wyeth has been able to look at RNA in blood samples to compare individuals with diseases against those without disease and identify patients most likely to benefit from specific drugs, in the process developing databases of gene-expression information. |
| 3 |
Protein-to-protein signal (transduction) Determining protein interactions. |
| 4 |
Tissue genomics (as in blood, urine, or tumor samples) Examining signal pathways in tissue samples. Aclara looks at very small quantities of solid tumor slices to see if protein pathways are present in a tumor. If the protein pathway is not expressed and the protein complexes are not formed, the drug that acts on that pathway will not work in that patient. |
| 5 |
Tumor protein complexes and pathways Genitope, of Redwood City, Calif., is developing an antibody-type individualized cancer vaccine for follicular non-Hodgkin’s lymphoma, based on a biopsy of a tumor; scientists identify the tumor-derived expression factor, then expand the cell culture line, and purify the tumor-derived antibody to formulate an immunotherapy vaccine. The extent of influence on disease progression and longevity, as opposed to tumor response, is unknown. |
NEW TARGETS, NEW DRUGS
Another benefit of pharmacogenomics is its expansion of the drug discovery frontier. It opens a new page on targets for drug inventions, and could ultimately reduce development cost and time, as much early-stage research can be done in silico rather than in expensive human trials. Current research involves asthma, diabetes, and heart disease, and a number of drugs, including antiarrhythmics, beta blockers, antithrombotic, and lipid-lowering drugs. Research is also underway in hypertension, Alzheimer’s disease, and depression — the latter focusing on gene variants within the serotonin transporter and cytochrome P450 drug metabolizing enzymes.
Glossary
- Adverse drug reaction (ADR).
A negative side effect attributed to a drug and not to extraneous factors or the disease being targeted.
- Allele.
One of several alternative forms of a gene at a specific location (locus) on the chromosome that controls expression of the protein product in different ways. A single allele for each location is inherited from each parent (e.g., at a locus for eye color the allele might result in blue or brown eyes).
- APOE4.
APOE4, 3, and 2 are variants of the gene for Alzheimer’s disease that can be used to distinguish it from dementia.
- Biochips.
See Microarrays.
- Bioinformatics (computational biology).
Application of computer technology to managing biological information. Specifically, the science of developing computer databases and algorithms to expedite biological research, particularly in genomics. Computer programs process large amounts of data that enable important inferences to be made, such as the likely functions of novel proteins, or the existence of gene expression patterns that correlate with disease states.
- Expressed Sequence Tags (EST) E-tags.
Molecular tags used to label a gene, for example, to denote the gene’s function/protein. Functions of unknown human genes often are based initially on similarity to genes with known functions found in simpler organisms such as the worm C. elegans. For example, of the nearly 300 known disease-causing genes in the human genome, more than half have an analogous gene within the C. elegans genome.
- Gene.
An ordered sequence of nucleotides located in a particular position on a particular chromosome that encodes a specific functional product (i.e., a protein or RNA molecule). A unit of heredity.
- Gene expression.
Process by which a gene’s coded information is converted into the structures present and operating in a cell. Expressed genes include those that are transcribed into messenger RNA (mRNA) and then translated into protein and those that are transcribed into RNA but not translated into protein (e.g., transfer and ribosomal RNAs). Gene expression determines the protein(s) produced when a gene is either switched on or off (knocked out). In many cancers, genes are known to be amplified or overexpressed.
- Genetics.
The study of inheritance patterns of specific traits.
- Genome/Genomics.
A genome is all the genetic material in the chromosomes of a particular organism. Structural genomics examines where gene sequences are located within the genome, while functional genomics looks at what traits or functions result from which gene sequences, generally via protein expression.
- Genotype.
The genetic constitution of an organism, as distinguished from its physical, biochemical, or physiological characteristics — its phenotype. Genotyping is identifying specific alleles (or gene variants) within an individual.
- Haplotype.
A subgroup such as an ethnic group sharing a phenotype that makes the group react a certain way to a physical agent. An example is lactose intolerance, which in North America occurs in more than 70 percent of persons of African descent, but in fewer than 19 percent of Caucasians. Another example is that some drugs such as acetaminophen, aspirin, and diazepam are eliminated more slowly in women than in men.
- Microarrays (biochips).
Gene chip technology, used to monitor and measure changes in gene expression for each gene represented on the chip, measuring the extent to which specific genes are represented as mRNA “copies” within a cell. Typically, gene expression profiling is performed by comparing mRNA expressed in healthy and diseased samples to identify genes involved in particular disease processes. Microarrays measure hybridization in many thousands of gene products at the same time, but only for relatively small numbers of samples, one sample per microarray. In contrast, polymerase chain reaction analyzes large tissue sample sets, typically measuring only one to three gene products per sample.
- Nucleotide.
Structural units of a nucleic acid — DNA or RNA; also found in nonnuclear parts of the cell.
- Pharmacogenomics.
The study of how individual phenotypic and/or genetic makeup (genotype) influences drug response. The term pharmacogenomics, now used synonymously with pharmacogenetics, is debated among researchers. A more precise expression might be proteo-genetic SNP-based pharmacology.
- Polymorphism.
Difference in DNA sequence among individuals that may underlie differences in health. The word now is becoming virtually indistinguishable from a mutation — anything that changes the nucleotide bases in DNA.
- Proteome.
Proteins expressed by a cell or organ at a particular time and under specific conditions.
- Proteomics.
The study of an organism’s full set of proteins determined by their genome and the proteins’ role in an organism’s structure, growth, health, and disease (or resistance to disease). In protein interaction analysis, researchers study how newly discovered proteins interact with known proteins to infer the newly discovered protein’s function, sometimes comparing them with simpler species.
- RNA (mRNA).
A long, usually single-stranded nucleic-acid marker used in proteomics to determine gene expression. Contrast with DNA, the marker of inherited genes (genomics). RNA functions in a cell nucleus as a messenger (mRNA) for protein synthesis or that actively participates in synthesizing a protein within a cell.
- Single nucleotide polymorphisms (SNPs).
DNA sequence variations that occur when a single nucleotide (A, T, C, or G) in the genome sequence is altered. To be considered a SNP, the variation must occur in at least 1 percent of the population. Millions of SNPs have been identified in the human genome. Researchers are studying which ones are markers for disease susceptibility and progression.
- Toxicogenomics.
A branch of toxicology that deals with differences in organisms’ responses to toxins of as a result of genomes.
- Transcription.
The enzyme-catalyzed process of using genetic information in a strand of DNA as a template for mRNA. Transcriptional profiling uses microarray-based procedures for determining and comparing the expression levels of tens, hundreds, or even thousands of genes in parallel.
- Translation.
Unlike transcription, which uses the same language to decipher information, translation transforms information from RNA into protein (amino acid) language. Translation is how the genetic information in a mRNA molecule directs the order of incorporations of specific amino acids and the growth of polypeptides during protein synthesis.
The greatest advantage will be gained in chronic diseases with treatments having dangerous adverse drug reactions, low drug efficacy rates, or few treatment options. For now, cancer is the therapeutic area with the most intense pharmacogenomics investigation.
There are 1.1 million solid tumors diagnosed in the United States every year, Klopack points out. “Drugs are usually developed for specific tumor types (breast, lung, colon, prostate),” he says. “If you are diagnosed with a deadly disease, you want to get the most effective treatment as quickly as you can. If you don’t get the right drug at the correct time, you’re giving away your time advantage. Cancer is a very heterogeneous disease, so you have to quickly pick the drug that manages the disease characteristics of the tumor you have.”
TIMELINES
How fast individualized drugs reach routine use depends on many variables and on economic models of drug development that must be transformed from current models and assumptions. Methodically sifting through substantial genetic data and determining their functions in biological processes could take decades, if not longer.
“I think pharmacogenomics will yield 10, perhaps 20 new drugs within my lifetime,” surmises Ken Bauer, PhD, chief science officer at ChromaVision in San Juan Capistrano, Calif. “It won’t be a deluge of new drugs, and, fortunately for clinicians, the drugs and diagnostics will be introduced gradually.”
Beyond the technical challenges lie daunting economic realities. “In a way, the practice of individualized medicine is an impossible undertaking, a misnomer,” says Lash. “Would we ever have the resources to build a drug for Mr. Smith and then another drug for Mrs. Jones? Not unless Mr. Smith and Mrs. Jones are willing to pay hundreds of millions of dollars for ‘his’ or ‘her’ drug. Someone along the line has to pay for this research.”
One factor that could seriously slow research progress is a lack of financial resources. “There is not adequate funding for academic and commercially unbiased clinical research,” says Janice Kurth, MD, PhD, vice president for life sciences at Visualize Inc. in San Diego. “Healthcare organizations stand to gain from this research in the future. Therefore, they need to step forward to help steer the direction this research is taking and somehow encourage the entities doing research, which threatens the markets of bigger players.”
FIRST, A PRICE SPIKE
According to Hedgecoe, the potential gain of pharmacogenetics for healthcare administrators is obvious — reducing costs by avoiding drugs that cause serious or fatal reactions or have no therapeutic effect. Certainly, the newer drugs will cost more in the beginning, as their makers attempt to recoup development costs. That could change if the FDA overhauls the current approval structure and creates exemptions and rewards similar to those enjoyed by orphan drugs to encourage innovation and improve chances for financial success.
Long-established methods of clinical studies will need to change, thanks to forthcoming FDA mandates. Draft FDA guidance documents, issued last November, give pharmacogenomic research a push by requiring that manufacturers include genetic studies on inter-individual differences in populations in their investigational plans. Such studies could improve risk-benefit ratios. Final FDA guidelines for pharmacogenomic research submissions, are scheduled to appear this summer.
“Pharmacogenomics affects every payer, every employer, every pharmacy benefit manager, and every individual. Every one of us is going to wonder, ‘Will there be a drug for me?’”
— Stephen Lash, PharmD, Genentech
It’s unknown the extent to which patients and MCOs will be willing to pay for greater certainty of a drug’s effectiveness.
“Personalized medicine could be very cost-effective for payers in healthcare,” Hedgecoe says. “Nonetheless, any new drugs coming out of pharmacogenetics will have to fit into the clinical situation and current practices and systems. The clinical situation will affect how the drugs are used, rather than the reverse.”
While the interests of the biotech and pharmaceutical companies often coincide with those of diagnostic companies and payers, that’s not always the case.
“Unlike reducing adverse drug reactions, which is in the interests of patients, developers, and payers, this convergence of interests and incentives may not be true in the case of identifying or eliminating leading compounds or in efficacy matters,” Hedgecoe says. “More exact diagnostics may eliminate 50 percent of a potential consumer population, so it may not be in the interest of a pharmaceutical company to pursue or publicize them — and they are not likely to take cuts into their markets lying down.”
Current research also may create unforeseen economic advantages. “Redeploying older generic drugs to treat multiple-resistant microbes that work in some subset of patients but were formerly shelved may be possible,” Kimmell says. He uses the example of antibiotics, noting that new patterns of pathogen resistance or genetically linked adverse-effect information may encourage clinicians to find new uses for old medications.
There will be a considerable impact on financial structures of healthcare, and the formulary will look very different in 10 or 20 years because of this research. “Pharmacogenomics affects every payer, every employer, every pharmacy benefit manager, and every individual,” Lash says. “Every one of us is going to wonder ‘Will there be a drug for me?’”