The emerging science of pharmacogenomics in drug development

By L. Scott Clark, Ph.D., Chief Scientific Officer, Gentris Corporation

Doubters of pharmacogenomics beware. Personalized medicine has become a household word. Its exposure has long surpassed the day of the scientific journal and has now been mentioned in mainstream magazines such as Time (2001) and Newsweek (2005) as well as numerous newscasts and talk shows.

Pharmacogenomics, or the response of a patient to a drug based on their underlying genetics, defines the goals of personalized medicine in the drug development process.  The field of pharmacogenomics has grown in the past 25 years, including correlations such as dextromethorphan and the drug metabolizing enzyme CYP2D6 to newer correlations such as efficacy of cetiximab or panitumumab with mutations in KRAS.

Pharmacogenomics (PGx) has slowly but steadily been adopted by pharmaceutical companies.  Still some skeptics don’t support the drug/gene correlations mentioned above even though there are more response associations being revealed every year.  Some scientists believe that the industry has found the simple associations; that is, one gene that affects drug response.  Others feel that there are limited drug alternatives; that is, there may be only a few drugs on the market for treatment and pharmacogenomic testing may leave patients without any alternative treatment.  Educating the healthcare provider is yet another barrier. 

Pharmaceutical companies are reluctant to utilize pharmacogenomic testing because it may limit the market potential of the drug; however, just the opposite may be true. By NOT testing patients when there is an association, the risk of money loss to a pharmaceutical company can be great.  Recent news has demonstrated that some pharmaceutical companies have had significant legal payouts due to adverse drug events including death while patients were on their drug.  For those companies on the fence about implementing a PGx program, it is time to get on the bandwagon as PGx is here and is here to stay.  Regulatory agencies are recommending that a pharmacogenomic sample be drawn during clinical trials. By utilizing pharmacogenomics in clinical trials, adverse drug reactions can be prevented, allowing for quicker time to market, and cost savings for the trial.

A PGx program can include analysis of DNA, RNA, and protein.  Historically, PGx was centered on Phase I and II drug metabolism enzymes.  Before genetic analysis of pharmacogenomics was utilized, drug metabolism was examined phenotypically.  Poor metabolism has been well documented in trials based on their parent/metabolite ratio.  For example, some individuals would have a different dextromethorphan/ dextrorphan ratio than others in the same study. 

Eventually, it was determined that the enzyme CYP2D6 was responsible for the metabolism of the parent compound, detromethorphan to the metabolite compound, dextrorphan.  Genetic analysis determined that genetic variants in CYP2D6 can affect the metabolism of compounds that utilize this pathway.  To date there are over 70 polymorphisms associated with CYP2D6 that can have an effect on the activity of the compound which in turn determines the safety and efficacy of the compound.  In addition, some of these polymorphisms have ethnic associations.  As the field advanced, additional genes besides drug metabolism enzymes demonstrated response with polymorphisms within the gene. Polymorphisms in transporters such as MDR1 (aka ABCB1) or receptors such as FcγIIa, FcγIIb and FcγIIIa can affect transport or cell signaling and increase safety and efficacy concerns. TPMT or thiopurine methyltransferase is a purine salvage pathway.  Polymorphisms within this gene can cause life threatening myelotoxicity if treated with azothiopurine, thioguanine, or mercaptopurine.

Pharmacogenomic programs also include gene expression analysis to examine biomarkers of disease as well as treatment conditions to provide critical information on drug safety and efficacy.  For example, 5-fluorouracil has been shown to be involved in the thymidylate synthase pathway.  Genetic variations in this gene, such as TSVNTR (variable number of tandem repeats) can affect the expression of the gene as well as the activity of the gene.  An increase or decrease in expression is dependent upon the number of repeats in the TS gene as well as a single nucleotide change within each tandem repeat.  Another example of gene expression analysis used for PGx programs is PML/RARα.  PML/RARα is a biomarker used for the treatment of acute promyelocytic leukemia with tretinoin and may be dependent upon the presence of the fusion gene.  Individuals who do not express this gene do not seem to respond to tretinoin therapy.

The benefit of a PGx program is multi-factorial.  The PGx program can be used in discovery and lead to an increase in targets and diversity which can lead to new biomarkers.  In development, the use of pharmacogenomics in trials can reduce the trial size and reduce the time to market.  In clinical medicine, the use of PGx can be used in point of care decisions such as tailoring dose, reducing adverse drug reactions, and increase the efficacy of the compound by identifying the correct patient who should take the medicine.  Identifying individuals who will benefit and who will not benefit from the drug can determine the success or failure of a drug.

Several medications have been pulled from the market that may have benefited from PGx testing.  Rofecobix, terfenadine, and cervistatin are just a few medications to mention.  Many of these medications are metabolized by drug metabolism enzymes or are involved with receptors.  Terfenadine, an antihistamine, was pulled from the market for causing arrhythmias.  This medication is metabolized by CYP3A4.  Genetic variations in this pathway can cause an increase in toxicity.  Rofecobix, an anti-arthritic COX-2 inhibitor was pulled from the market due to an increased incidence of myocardial infarction.  This medication is metabolized by CYP2C9.  Lastly, cervistatin, an anticholesterol compound, was pulled from the market due to an increase in rhabdomyolysis.  Testing for genetic variations in these pathways may or may not have saved that medication from being pulled from the markets, but it may have identified patients with a predisposition to having a serious adverse event.

Additional associations with drug response and genes have helped strengthen the field.  Some notable associations are with abacavir, warfarin, clopidogrel, panitumumab, and cetuximab.   Individuals with polymorphisms in HLAB*5701 and are undergoing abacavir therapy, can develop severe skin hypersensitivity.  Individuals undergoing warfarin therapy may lack metabolism of warfarin if they have polymorphisms in CYP2C9 or have sensitivity to warfarin if they have polymorphisms in VKORC1.  Patients may have an increase in cardiac risk if they have polymorphisms in CYP2C19 and are taking clopidogrel, and patients that possess mutations in codons 12 or 13 in the oncogene KRAS may not respond to panitumumab and cetuximab.  Additional associations are continually being demonstrated.  Testing for these polymorphisms in these genes can save money by preventing adverse drug reactions by only putting individuals that can tolerate therapy on the medication.  Many of these medications have a black box warning such as Plavix (clopidogrel) and Epzicom (abacavir), while others have pharmacogenomic testing as an indication such as with Erbitux (cetiximab) as well as dosing adjustments for CYP2D6 poor metabolizers for Strattera (amoxetine).  This type of warning not only identifies who should take the medication, but those who shouldn’t.  This labeling, although, perceived as limiting the market to a certain population, is not only providing the medication to the appropriate population, but will also save money in costs associated with adverse drug reaction whether in the clinical trial or post-market.

An additional advantage of pharmacogenomic therapies is that they can help pharmaceutical and biotechnology companies understand their compounds better.  Adding pharmacogenomics program into clinical trials can help reduce adverse drug reactions which can save money as well as maintain the subject numbers as compared to a trial with attrition due to adverse reactions.  In this scenario, the patients enrolled in the trial are the patients that will benefit from the therapy, thus providing a better understanding of the compound and its therapeutic effect.  This activity, in turn, can accelerate drug development.  Using PGx in clinical trials can also aid in correlations between dose response and gene expression.  This can be useful in the timing and selection of chemotherapies.  Pharmacogenomics help in the understanding of trial findings in respect to individual genetic variations.

The addition of PGx to a trial can have tremendous cost savings.  By utilizing PGx in preclinical and Phase I or Phase II studies, one can determine not only who should and who shouldn’t take the drug, but can also help determine whether the drug should continue in the drug development pipeline.  While that is a difficult decision to make, it is better made early on instead of finding out the drug will not perform as expected in Phase II or Phase III trials. 
Once a trial starts and there is a known gene/drug response association, testing for genetic variants can reduce clinical trials costs.  For example, in a three arm Phase III study that was conducted in 2000, one group received test drug, one group placebo, and one group a comparator drug that was metabolized by CYP2D6.  Patients were genotyped for polymorphisms in CYP2D6 prior to entry into the trial.  Patients that genotyped as CYP2D6 poor metabolizers were excluded from the study.  Results consisted of discontinuation of therapeutic dose monitoring as there was no concern for monitoring because poor metabolizers were excluded from the study.  No patient attrition occurred from the study because there were no adverse reactions, and there was an additional cost saving as treatment for adverse events did not occur.  The resulting net savings for the trial was nearly $750,000 and included the cost of genotyping. This study occurred 10 years ago, so the average cost savings today would be greater.

Pharmacogenomics is currently being used by industry, as mentioned earlier, to exclude poor metabolizers.  Stratification of ultrarapid, extensive, intermediate, and poor metabolizer or responders and non-responders is another pharmacogenomic application.  By stratifying the population, the resulting data may help determine dosing strategies.  One of the difficult challenges for some pharmaceutical companies is the need to include poor metabolizers in their clinical trial.  Poor metabolizers or non-responders can be represented anywhere from 1-30% of the general population.  In an effort to not spend a tremendous amount of money searching for a poor metabolizer or non-responder, pre-screening for these specific patients is an attractive proposition.  Some pharma and contract research organizations are pre-screening subjects under research conditions first to populate their subject database.  When there is a clinical trial that requires a certain metabolizer, the pharma or contract research organization can pull those subjects and have them re-genotyped under regulatory guidelines.  Both time and money are saved in the clinical trial using this strategy.  Gentris has successfully worked with several companies performing just this task. Lastly, performing PGx may help explain outliers in pharmacokinetic data.

The advantages of PGx are multi-factorial, however, if no sample is drawn, none of the benefits can be utilized.  Collecting a pharmacogenomic sample and storing it is an intelligent decision.  No company wants to be in a Phase III clinical trial and show an unexpected result.  Without a pharmacogenomic sample, patients would have to be recalled and re-consented, taking time and money.  By collecting and storing a PGx sample at the initiation of the trial, companies can go back to specific patient samples and examine them for any variations that may have contributed to the unexpected results.  The money saved by avoiding the necessity of recalling patients, re-consenting, and reanalyzing samples is significant.

In summary, pharmacogenomics, while still viewed by many in its infancy or failing to deliver, is now part of the drug development process in the pharmaceutical industry.  Technologies such as microarray analysis and next generation sequencing are capable of examining rare variants and looking for associations of response among thousands of genes.  With these technologies, it is now possible to examine multi-gene and drug response associations. 

Pharmacogenomic testing will help the market and save money as compensation for adverse drug reactions and legal suites will be reduced.  Education of the health care professionals will always be a challenge.  It is up to the health care professional and the pharmaceutical industry to be abreast of the indications of medications.  Ultimately, these activities will help pharmaceutical companies with cost and time to market for their trials and more importantly help the patients get the appropriate drug the first time.