In the current economic climate, the pharmaceutical industry may not seem like the likeliest place to forge a career, particularly in light of the recent spate of layoffs. While today's economic conditions are brutal, the forces shaping the drug industry's downsizing and restructuring have been at play for some time. By Virginia Gewin
"Pharmacometric analysis is key to understanding the relationship between exposure and response, and this critical knowledge will help companies avoid late-stage failures."
Out of the carnage come opportunities. Recent high profile drug failures—notably the withdrawals of the painkiller Vioxx and diabetes drug Avandia—have shed light once again on drug safety. As well, pharmaceutical companies are looking for new ways to address their two-pronged pipeline problem: many big-name blockbuster drugs will soon lose their patents and, perhaps more important, there is a dearth of new blockbusters headed to the market. In fact, over 90 percent of all candidate compounds that are developed never reach the market (Beresford AP, et al. Drug Discov Today 7(2):109–116, 2002).
Drug discovery and development has historically been a trial-and-error approach whereby companies screen thousands to millions of compounds. But regulatory bodies and a growing number of drug discovery companies want more predictive scientific approaches in place to lower the chance of failure in clinical trials. "The biggest resource demand in toxicology is time because many studies, such as carcinogenicity assays, require years to complete," says David Jacobson-Kram, associate director for pharmacology and toxicology at the US Food and Drug Administration's Center for Drug Evaluation and Research in Silver Spring, Maryland. Therefore, FDA is encouraging industry to use existing knowledge to do earlier stage simulations of the absorption, metabolism, and excretion of drug candidates. The search is on for new drug safety metrics, and the "virtual" trial may not be far behind.
As a result, an interesting mix of quantitative skills is in demand. Toxicology, the traditional tests for adverse drug effects, continues to drive early drug discovery efforts. But pharmacometrics—the modeling and statistical analysis of drug metabolism data—touts a more tailored, predictive approach to safety and efficacy. In simple terms, pharmacokinetics, PK, measures what the body does to the drug, while pharmacodynamics, PD, measures what the drug does to the body. Therefore, the simulation and modeling of PK/PD using in vitro data from human and animal models is becoming a more important means to better weed out potentially harmful drugs sooner in preclinical drug development. As such, their adoption in industry is gaining traction, and providing career opportunities.
"Pharmacometric analysis is key to understanding the relationship between exposure and response, and this critical knowledge will help companies avoid late-stage failures," says Jill fielder-Kelly, vice president of pharmacometric services and chief scientific officer of Buffalo, New York–based Cognigen Corporation, a contract research organization (CRO). In fact, she says, recruiters are actively looking for persons with pharmacometric skills, a group that seems fairly immune to the ongoing layoffs.
"The safety science space has been using the same models for 25 years, making the field ripe for innovation," says Joseph DeGeorge, worldwide head of safety assessment at Merck & Co, in West Point, Pennsylvania. Innovation includes developing metrics able to predict outcomes. Prediction requires robust models that can simulate biological systems. Predictive approaches to pharmacodynamics and pharmacokinetics in humans using nonclinical data sets, and developing safety biomarkers, are two areas ready for big scientific advances, says Jacobson-Kram.
DeGeorge says people who can cross disciplines are making the scientific breakthroughs. "It isn't good enough to be trained in pharmacokinetics or pharmacodynamics or gene expression," says DeGeorge. For example, he says, Merck recently combined traditional toxicology endpoints, genomics, and accessible protein biomarkers measured in humans and animals to develop a new biomarker signature that was qualified by the FDA in 2008 to indicate kidney toxicity.
Many see the development of new drug safety metrics as the field's future. Perhaps not surprisingly, that is because FDA is leading the charge. Jacobson-Kram expects to sustain at least some level of last year's 30 percent increase in staffing as FDA bolsters its own quantitative capacity.
"The most interesting and exciting area in early drug safety is to develop new systems for assessing toxicity," says Jacobson-Kram. "We're getting away from 'dose 'em and count 'em' studies and moving toward using computer models to predict toxicities, and in vitro assays to make predictions about organ toxicities—that's the future. In the next five to 10 years, we'll eliminate much of animal toxicology and move on to smarter systems using computer algorithms and modeling," he says. Combining new and old techniques—for example, fusing molecular biology skills, pharmacometrics, and classical toxicology—to find new safety biomarkers and surrogate endpoints represents a new pharmacological frontier.
While drug safety is an industrywide concern, the adoption and development of new drug safety metrics varies widely within the industry. Some companies are sticking with traditional toxicological approaches, while others, like Novartis, are actively exploring new directions. While Novartis is still recruiting traditional veterinary pathologists trained in toxicology, it is also looking for toxicologists with a background in biochemistry, molecular biology, and modern biology techniques who are interested in molecular-based risk assessment. "To better understand safety issues, we are making efforts to employ more modern investigational, molecular approaches to safety assessment—and increase our safety testing work force by about 10 percent," says Phil Bentley, vice president and global head of preclinical safety at Novartis Institutes for BioMedical Research in New York City.
Other top pharmaceutical and biotechnology companies—including Genentech, Amgen, Pfizer, and Merck—also have an ongoing need for pharmacometric skills to optimize drug efficacy and minimize potential toxicities. Genentech in San Francisco, for example, currently needs people with bioanalytical skills, with specific emphasis on biotransformation and drug transport, says Holly Butler, senior staffing manager for research.
Recruitment efforts, however, aren't typically designed to grow company size. Instead, companies are more focused on maintaining productivity while keeping the cost of drug discovery and development down. But that leads to tough choices. For example, so-called Phase 0 trials—preclinical trials designed to gather PK/PD data from subtherapeutic doses given directly to humans—will bypass animal experiments in this earliest stage in order to better understand and model the drug's action in humans. This information should not only identify any potential problems earlier but also help design more effective clinical trials. But, at this early stage, the benefits of virtual trials remain to be proven. "Everybody's idea of a Phase 0, or virtual, trial is different," says Amin Rostami-Hodjegan, pharmacologist at the University of Sheffield in the UK.
Low Supply, High Demand
Both the classical animal toxicology skills and futuristic pharmacometric skills are in huge demand but short supply. Indeed, industry professionals agree that the dearth of persons with doctorates in veterinary medicine (D.V.M.) degrees is a stumbling block to their ability to develop new methods. "D.V.M. pathologists are hard to recruit," says DeGeorge. Johan Gabrielsson, senior principal scientist at AstraZeneca in Mölndal, Sweden, agrees and says the ongoing shortage of trained people in the field of quantitative pharmacology and in vivo animal experimentation is in part the result of political pressure in the 1980s and '90s to replace, refine, or reduce the use of animal studies. "Unfortunately, the 'omics' sciences do not yet offer a complete solution. We have to rely on in vivo data when we go from animals to humans," says Gabrielsson. Seeing the continued demand, a coalition of the Society of Toxicological Pathology (STP) and the American College of Veterinary Pathologists (ACVP) jointly fund training opportunities.
Training in pharmacometrics is equally hard to come by. "All big pharma companies are searching for people with PK/PD skills, but the demand is much greater than the actual output from academia," says Gabrielsson. Meindert Danhof, head of the Leiden/Amsterdam Center for Drug Research in the Netherlands, agrees. "The bottleneck is the number of academic sites to learn these skills," he says. Indeed, graduate schools in the United States offering PK/PD programs are few—fewer still in Europe. And most graduates are lured into industry. "Few graduates stay in academia, which perpetuates its limited capacity to fulfill the increasing demand for these skills," says Danhof. Driven by the need for well-trained professionals, industry employees such as Gabrielsson and Fiedler-Kelly, serve as instructors of university courses.
Gabrielsson says the lack of quantitative pharmacology training opportunities is hurting the industry. "We lack people able to design, carry out, and analyze in vivo experiments, and then build models to extrapolate data across species and into man," he says. CROs play an increasingly important role in drug development—from routine clinical trial management to more specialized modeling and simulation work—and are also having trouble recruiting. Fiedler-Kelly says Cognigen needs both senior and junior level pharmacometricians, but finding people with five to six years of experience applying these methods to address real issues in the regulatory environment is most difficult. "We need people who use these methods, and can interpret and synthesize the results, to answer questions that come up during regulatory review and to provide the necessary feedback to FDA," she says.
Pharmaceutical company and CRO dollars fund many of the top programs, and offer valuable internship opportunities for students. Interestingly, many programs, such as the University of Sheffield's new M.Sc. in modeling and simulation in PK and PD, serve people who already work in the pharmaceutical industry. "Our students understand there is a future in quantitative pharmacology and would like to shift careers into this area," says Rostami-Hodjegan. Likewise, Gabrielsson's new M.Sc. course, called Integrated Quantitative Pharmacology, at the University of Göteborg in Sweden, has garnered enormous interest from people already in industry.
Without industry support, these programs wouldn't exist. "Universities aren't as interested in funding the applications of cutting-edge science, they'd rather fund its creation," says Rostami-Hodjegan. And industry is wary of funding the training of students that may join their competition.
But Rostami-Hodjegan points out that companies that fund training get first pick of the best talent. Until recently, most of the two annual postdoctoral fellows trained in clinical PK/PD at the University of North Carolina, Chapel Hill were hired by either of its two industry funders—Quintiles, a Research Triangle Park–based CRO, and GlaxoSmithKline (GSK). Danhof says the Leiden Center's funders, including Eli Lilly, Johnson & Johnson, and GSK, gobble up the five Ph.D.s they graduate each year. William Jusko chairs one of the most highly regarded US programs at the pharmaceutical sciences department at the State University of New York (SUNY) at Buffalo. He says over 85 percent of his students are actively recruited into the pharmaceutical industry, but those in the heaviest demand have focused on PK/PD. More would like to stay in academia, he says, but the current climate of support from NIH ensures that only the most highly motivated do so.
Jusko says that, among his students, the hottest topics of interest involve biotech—particularly protein therapeutic agents—because many future new drugs will likely be biologicals. Jusko says his faculty members are primarily supported by NIH grants and typically employ PK/PD in almost every project that studies the mechanistic basis of drugs in animals or people.
Yet the degree to which industry uses these techniques to develop new drug safety metrics is not well defined—leaving others in academia optimistic, but cautious, about career prospects.
Howard Lee, director of the University of San Francisco's Washington, D.C.–based Center for Drug Development Science, says that PK/PD modeling has so far been geared mostly toward drug efficacy, but that concern in the US Congress and among the public ensures that safety is going to become more important in the future. But he thinks the widespread adoption of pharmacometrics is an uphill battle that is just beginning, so he tempers his enthusiasm for the burgeoning field with the pragmatism that it may be some time before drug safety biomarkers or endpoint surrogates are sought by industry.
Advice from Industry
Understanding the myriad factors influencing such tough industry decisions is hard to do outside of a company. Since most jobs are in industry, gaining insider experience is essential. Butler advises graduate students interested in industry to do two things: complete an internship at a company and make sure that they attend critical conferences related to drug metabolism and pharmacokinetics—such as the International Society for the Study of Xenobiotics. Most big pharma companies offer internships as do some CROs, including Cognigen.
Beyond internships, Fielder-Kelly suggests that people interested in PK/PD careers should consider working for a CRO. "My big selling point is that a CRO can offer someone tons of experience analyzing the data from drugs used in a variety of therapeutic areas," she says.
Understanding the entire drug development process, however, necessitates experience in big pharma. Gabrielsson says it is important to start making contacts with industry early during graduate studies to secure internship opportunities. Given the dearth of formal training programs, the growing number of workshops and short courses is a good way to not only hone technical skills, but also network with the industry leaders eager to train students in the field. For the last 11 years, he's taught a weeklong workshop on advanced pharmacometric-pharmacodynamic data analysis at Cambridge University, which is an excellent networking opportunity.
"Come visit us, present a seminar on your research project, and aim for thesis advisers also from industry," says Gabrielsson. "I'm happy to mentor graduates and postdocs and am currently looking for a postdoc to work on both preclinical mechanistic data as well as patient biomarker and disease data," he adds.
Given the push toward Phase 0 trials, experience modeling all types of data is increasingly important. "Experience working with pharmacometric data from humans, or animals closest to humans, is highly sought after," says Ed Dupuis, pharmacologist at University of North Carolina, Chapel Hill. But the changing world of drug safety also requires that scientists adopt a systems biology approach. "We can't understand toxicity if we don't understand cellular processes and how molecules interact," says Bentley.
Putting that understanding into practice will provide endless career opportunities—especially if they help companies save money currently misspent chasing dead-end products.
Careers in Green Science — April 3
The Road to Diversity in Science: Are We There Yet? — April 24
Careers in Biotech and Pharma — May 8
This article was published as an advertising feature in the March 13, 2009 issue of Science.
View this article in PDF format.
Virginia Gewin is a freelance science journalist based in Portland, Oregon.