(Mis)Understanding Drug Discovery: It’s Much Harder Than Rocket Science


Xconomy Seattle — 

Developing new medicines is an amazingly difficult undertaking. The research portion alone is daunting, and for those of us who have actually attempted it, humbling. A recent article reminded me just how little many people understand about the drug discovery process. The basic premise of “Pharma Needs an Innovation Intervention” was that pharma should change its focus from “finding druggable targets” to “deliver a consumer-focused product and service and a business model that goes beyond the product itself.”

Unfortunately, the article doesn’t address two critical issues; one grounded in the past, the other the future. First off, why has Big Pharma, after being one of the most profitable industries for decades, suddenly become so remarkably unproductive in coming up with new medicines? Second, if Big Pharma walked away from the difficult task of “finding druggable targets,” then who would take over the job of creating new medicines? From a business point of view, one can see a clear need to alter a revenue model that may not be working for Big Pharma any longer. However, from a practical and societal perspective, there needs to be some way of innovating new medicines, not just new business models.

I wish I could provide a definitive reason why Big Pharma, as a group, has become so unproductive in recent years. These companies are not monolithic and have distinct styles for running their businesses. To paraphrase Tolstoy’s Anna Karenina “Productive drug companies are all alike; every unproductive drug company is unproductive in its own way.” It’s easy to congregate a lineup of the usual non-productivity suspects: excessive layers of bureaucracy, fear of making the wrong (or any) decisions, entrenched industrial group-thinking, failure to recognize and support an innovative culture, and new regulatory uncertainties. I suspect that all of these concerns contribute to the productivity problem, but it’s difficult to quantitate exactly how important each of these factors really is.

I think Big Pharma is still hung over after a difficult transition from a decades-long focus on screening chemical libraries to more of a recombinant DNA/genomic biology mindset. It may very well be that after a century or so of effort, all of the “low hanging fruit” really has been picked off of the medicinal tree. Finally, consider that Big Pharma’s recent attention was focused on developing blockbuster drugs that would bring in sufficient revenues to feed their bloated organizations. Genzyme’s marked success in treating rare diseases (along with its recent acquisition by Sanofi) illustrates how that thinking has changed.

So why is coming up with new drugs so difficult? The answer is actually pretty straightforward: because biology is amazingly complex. It’s not rocket science; it’s much harder. With all due respect to the people that design and build our space vehicles, uncovering the functional role of thousands of unique biological molecules is a significantly more complicated undertaking. Twelve years is about the average length of time it takes for a single drug to be discovered, developed, tested, and approved by the FDA. Twelve years also defines the time period between the start of the space age (the launch of Sputnik 1 in 1957) and humans landing on the moon in 1969.

Even more difficult than figuring out the function of biological molecules is coming up with a way to alter, in an appropriate way, the divergent activities of selected subsets of these molecules to treat various diseases. People can be afflicted with literally thousands of different ailments. As living organisms, we reflect eons of genetic diversification and are vulnerable to rogue viruses, bacteria, fungi, and environmental pollutants. No two of us are alike, not even identical twins. People also suffer from pain, are susceptible to psychological disorders including addictions, and are at risk of unintentional side effects brought on by numerous medications. The biological and rocket sciences do share one primary characteristic: they are both very expensive, high cost-of-entry businesses.

Mechanistically, how do medicines work? In general terms, most drugs act by either stimulating something (these are called agonists) or blocking something (these are called antagonists). These effects are generally directed against specific molecules, even if the exact target remains unidentified. Within these broad definitions, however, lies a great diversity of approaches that drug makers have taken to treat diseases. Let me share some examples:

Drugs can directly stimulate (e.g. morphine) or block (HIV protease inhibitors) enzymes. They can bind to and sequester molecules (TNF blockers for rheumatoid arthritis). Drugs can replace missing molecules (insulin, hemophilia) and alter the rate of movement of molecules into or out of cells (anti-arrhythmics like sodium channel blockers). Some drugs stimulate the immune system (Provenge, Yervoy), change the pH balance in the body (sodium bicarbonate for acidosis), or interfere with the assembly or function of intracellular structures (anti-cancer drugs like taxanes). Drugs can stimulate the release of stored molecules (epinephrine), or interfere with DNA synthesis (sulfa antibiotics). Drugs can perturb cell membranes (anesthetics), and effect the modification of proteins, thereby altering their function (histone deacetylase inhibitors). In gene therapy, the drug is often a replacement gene; anti-sense drugs block the formation of proteins by binding up specific mRNAs.

The above examples demonstrate the variety of approaches drug makers have taken in coming up with new medicines. Their goal: design in characteristics that enable the drugs to achieve the desired effects, while at the same time designing out their ability to bind to and effect secondary molecules. These secondary interactions (and sometimes primary ones as well) often lead to side effects that can be sufficiently serious to prevent a drug from ever being used in the marketplace. This particular aspect of designing small molecules is daunting. X-ray crystallography and other techniques often enable scientists to generate 3D images of the protein that wish to target. This information is extremely valuable in tailoring the design of a drug that is meant to bind to and modulate the function of this particular protein. Ideally, the drug only attaches itself to this one (or in some cases, a few closely related) target(s). Often, the primary challenge isn’t finding a chemical that can bind to the chosen protein, it’s identifying one that doesn’t bind well to the other 21,000 or so proteins that it might also interact with. Imagine trying to design a mask that will precisely fit your face, but won’t fit well on the faces of thousands of other individuals. This inherent difficulty has fueled the rise of biologics (protein-based drugs), where such discriminatory specificity is much easier to achieve because the appropriate molecules have already been selected for by the powerful forces of evolution.

For a number of medical conditions, the exact molecular defect that is responsible for the disease is now well understood. In a best-case scenario, the genetic defect that causes chronic myelogenous leukemia was identified and led to the development of an amazingly effective drug with a very high cure rate. Knowing exactly what causes a disease is useful, but this information doesn’t always translate into curative medicines. The mutated protein that causes cystic fibrosis was discovered after years of intensive research in 1989. In the two decades since this discovery was made, however, no one has developed a reliable method for replacing, correcting, or bypassing the single defective gene that encodes this key protein (although a recent trial looks somewhat promising for one specific subset of patients).

A recent study suggested that mutations in genes that are essential to survival are the cause of many rare diseases, whereas alterations in non-essential genes are the primary drivers of more common illnesses, such as heart disease. Many health problems, however, result not from a single defective gene, but from numerous mutations throughout the genome. It is estimated that the average tumor contains 15 or more altered genes that are responsible for its unchecked growth and metastasis. Other diseases may arise simply from a combination of what turns out to be an “unhealthy mix” of genetic variations acting in concert, or in combination with environmental factors. Finding effective treatments for these types of diseases is much more difficult than those caused by single genetic alteration (not that these are necessarily easy to treat either).

The various molecules that exist within our cells don’t function independently of each other. There are numerous interconnections between them, the study of which has spawned a recent approach referred to as systems biology. The complex nature of these interactions makes them extremely difficult to study. It’s obviously simpler to design a drug that affects a protein that operates within a single defined pathway than one that functions at the intersection of numerous metabolic junctions. Drugs that are directed at proteins that function at biological intersections may affect any and all of the pathways that lead to and away from the target.

It’s important to have a clear understanding of the drug discovery process if we hope to accelerate the pace of creating new medicines. Prior to the past 120 years or so, virtually all medicines were derived from naturally occurring plant materials. These eventually gave way to chemically synthesized molecules, which ruled a large share of the market for the better part of a century. Drugs these days may still come from these sources, but many of the newer medicines are biologics (purified recombinant proteins, including monoclonal antibodies). Insulin, the first protein drug, was originally purified from cadaver pigs, but bacteria growing in brewery-sized stainless steel tanks now generate most of it. Biology has usurped chemistry as the dominant force behind many of the newest drugs, even though most biologics are incredibly expensive compared to small, chemically synthesized drugs. Some of the newest types of medicines (most of which are still in development) are comprised of various types of DNA or RNA, and they may be packaged in viruses, liposomes, or as nanoparticles for administration to people. Various medicines require distinct routes for drug delivery: they can be swallowed, inhaled, sniffed, injected, or supplied via a skin patch.

This tremendous degree of biological diversity and complexity demands that pharma and biotech companies devise new and pioneering approaches in both research and development. This will take guts, which some feel many of these companies are lacking. Sequencing of the human genome was a great milestone in biology, but it only served to underscore just how much work remains to be done in solving biological problems. The “innovation intervention” approach mentioned at the beginning of this article was actually targeted at business models, not drug discovery. What is abundantly clear is that the industry needs to change its modus operandi, and several new initiatives have been launched.

One popular approach that a number of Big Pharma companies have adopted is to align themselves financially with university researchers to take advantage of the expertise contained within academia. If properly done, this tactic should undoubtedly be helpful. Many of the most innovative drug development papers that I’ve seen recently originated in academia. Another approach has been for Big Pharma to re-align and narrow the focus of their internal research programs; some companies have made significant cuts in their research and development spending. This can be successful if the aim is to focus on fewer research areas, but increased spending, not less, will be needed going forward to support the remaining avenues of investigation.

I believe that all of these changes are necessary, but I have doubts as to whether they will be sufficient to reinvigorate Big Pharma’s drug discovery efforts. As novelist Ellen Glasgow put it “All change is not growth, as all movement is not forward.” I have no doubt that “bushels of fruit” remain to be picked from the medicinal tree, although reaching those higher branches will require the industry to wisely adapt a long term perspective. They need to establish an innovative culture, correctly employ new technologies, reward smart thinking, and have a brave heart. Though many are loath to admit it, the truth is plain to see: there will be no shortcuts on the long and difficult journey towards the medicines of tomorrow.

Stewart Lyman is Owner and Manager of Lyman BioPharma Consulting LLC in Seattle. He provides strategic advice to clients on their research programs, collaboration management issues, as well as preclinical data reviews. Follow @

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

8 responses to “(Mis)Understanding Drug Discovery: It’s Much Harder Than Rocket Science”

  1. R. Jones says:

    There is a story today out about Martin Mackay, the new president of R&D at Astra Zeneca. In his PR photograph, Dr. Mackay is sitting in front of a laminar flow hood, inside of a laboratory.


    You’ll recall the Astra Zeneca not only sacked the R&D staff in Wilmington DE, they razed 450,000 square feet of lab space.

    Business as usual is taking an office based scientist and thinking that one person will bring about proper changes. This is the same cycle that has been going on for decades. Sack the staff, keep the decision makers. Photograph them inside a laboratory with a white lab coat.

  2. Pharmaceutical companies need to spend more of the time and money on developing bioinformatics and computer modelling in drug development so as to better understand the biological pathways associated with a particular disease and hence reduce developmental costs.

    However I think the main reason there is a lack of an adventurous spirit in the pharmaceutical industry is the real fear of costly litigation especially in the US when things go wrong. Perhaps the expansion in the US is largely due to the necessity to curry favour with US policy makers etc.

    Also one has to take into account the competition by pharmaceutical companies (usually in countries not worried about such things as patents) that don’t do the research but don’t won’t to pay any royalties.

  3. For some odd technical reason, Mark Minie was unable to post a comment on this article. Since he asked nicely, I’m pasting his comment below.

    From Mark Minie:

    In direct response to a column in Xconomy Stewart wrote back in January 2011 in which he lamented the lack of such access to such restricted online resources to consultants and researchers who are not part of the UW faculty, staff and student body-

    What Does Biotech Really Suffer From? Information Overload, or Underload?

    and in part to also address some of the issues he brings up in his post to Xconomy this morning (9/7/2011) I’ll be running an Online Journal Club through the UW this fall, and it will be on Monday evenings 6:30-7:30 PM. It is for 1 credit, and will cost $590. We plan on having this every quarter for the year 2011. It will also be pass/fail, and will be open for anyone with an interest to apply. The web page is at-

    Journal Research Seminar

    People can apply/register there, and we plan on limiting it to a total number of 20 people. My plan is to have people work in teams of 2 to give online presentations using the videoconferencing tool Adobe Connect Pro (with video and audio), and these talks will be scored and such by me and the other students-this will be done along the lines of a similar presentation format I have used successfully for my Molecular and Cellular Biology classes with the UW. Students will have FULL ACCESS to ALL UW online journals, and we will be using the free online reference management/science social networking tool MENDELEY (http://www.mendeley.com/ ). I will manage the class using a wiki (all talks will be recorded and posted on the wiki, and there will be an online discussion board for conversations about the talks asynchronously), and I will be very flexible about what papers and topics people want to use in their coursework.

    If this first version of the course is successful, it is possible that next year we will scale it up to include more people on multiple days and perhaps even specific focuses. One of the main purposes of this offering is to build a community of non-UW consultants, researchers and others who need access to the online holdings at the UW for their work and also want to keep up with the most up to date science through peer reviewed research journals and also want to network with others with similar interests.

    The Professional and Continuing Education unit at the UW (along with the UW BioE Department), which will be sponsoring this course, also recognizes that access to such resources is a continuing education issue for the biotech/pharma community and is a way for the UW to provide more support for those industries.

  4. Stewart correctly points out that the picture is not a simple one, but I think it’s fair to say that no industry with the wealth or importance of pharma has ever had more trouble controlling its environment. Public perception, regulatory framework, and the behavior of payors are all far from what pharma would like.

    I agree with Stewart that nature hasn’t been very cooperative lately, but given all the other things that pharma can’t seem to control, one is tempted to see the problem beginning and ending with pharma.

    However, while systematic organizational weaknesses and pervasive delusional thinking are factors, the industry has also had terrible luck. Just about the time pharma worked its way through all the easy wins from the last 5,000 years of medical experimentation, the economics of drug discovery and development got much worse, compounded (pun intended) by payment systems that (particularly in the US) make drug costs much more visible to consumers than are other medical costs. Oops.

    What to do? Drug companies are sales and marketing machines, and to a lesser extent manufacturing and distribution machines. Like Cisco Systems, they don’t have a particular talent for innovation. Cisco’s solution is to concede the obvious and buy innovation by acquiring startups. That’s working great. Or it was, up until recently. Now Cisco is shedding people as fast as pharma and exiting markets it never really understood. Oops again.

    The way pharma sells and markets drugs is somewhat broken, but it’ll do for now. The way new drugs are approved is broken, but earlier approval for more limited labeling combined with post-market surveillance would patch over the problem for now. The most broken thing is the way pharma decides what drug targets to invest in, and the way pharma decides what candidate molecules to take into trials. Concentrate your fire on those areas, and those who make progress will be richly and deservedly rewarded.

  5. Thanks to all of you for taking the time to comment.

    R. Jones shouldn’t be surprised that the President of Astra Zeneca had his photo taken in front of some scientific equipment; such pics are standard operating procedure in the industry, along with photos of pills and patients. See nearly any biotech/pharma annual report. I think your observation that the staff gets sacked while the decision makers keep their jobs applies to a large number of industries, such as automobiles; it is not unique to biopharma.

    Paul Whitesman is correct that bioinformatics and computer modeling may be helpful in drug development, but the same could be said for a number of other tools used in the scientific research process. Advanced imaging techniques, for example, may also be quite useful. The primary issue is whether or not companies are willing to make a long-term investment in these techniques at the expense of their short-term profits. It’s not clear to me why a fear of litigation would lead to a lack of an adventurous spirit. Even developing a drug that is closely related to another drug would not protect you from lawsuits, as the makers of Vioxx and Baycol can attest to. Intellectual property issues are always a concern, but they will carry the day in the major markets of the world (e.g. North America and Europe), so I don’t see that as too big a problem.

    Congratulation to Mark Minie for his success in putting together the UW Online Journal Club, which will help to solve the problem of a lack of journal access for a small number of local scientists. For those who are interested, I followed up on my previous Xconomy article that Mark referred to with an op-ed in Nature Biotechnology (http://www.nature.com/nbt/journal/v29/n7/full/nbt.1909.html) that put forth some possible solutions to the problem. I have been following up on both articles with a number of interested people in Seattle and beyond to try and develop a more wide spread solution to the problem. Anyone interested in helping with this effort should contact me at [email protected].

    Erik Nilsson correctly points out that there are a number of issues that pharma is struggling to deal with at present beyond the one that I wrote about here. Unfortunately, some of these problems are of their own making, such as the low opinion that they are held in by a large percentage of the population. These used to be highly admired companies in the not-to-distant past. Post market surveillance is something the industry has promised to do numerous times in the past, but then has failed to do, so I am not a big fan that this will efficiently fix any problems. My focus here was simply to point out that drug discovery is grounded in scientific inquiry and excellence, and that new business models are no substitute for this. The science is difficult, expensive, and time consuming, but is the only way forward in developing new medicines.

  6. Mark Minie says:

    Thanks to Stuart again for getting the issue of access to scientific literature on the radar screen as being a criticality one component of science driven businesses such as biotech and pharma.

    I do, however, want to clear up a few potential misperceptions in his post about my upcoming online journal club course. First, this course is completely Internet based, and therefore is open to people from any location on Earth (or even the ISS if any astronauts/cosmonauts doing biomedical research up there are interested  )…and in fact some of our early applicants are from outside of the Seattle area (we have one applicant from Boston, another from San Francisco and someone from Europe at this point). The course is not in any way restricted to those local to Seattle…it is a truly online course/resource! Second, while this initial course will be limited to 20 students (primarily for technical/financial/logistical reasons…which can be overcome easily if decisions to do so are made), if it is successful it will be easy to scale it up to make it available to larger numbers of people and it may also be possible to even have multiple instructors and specialized online journal clubs…this of course is dependant upon the success of this first experiment. Finally, this course is based on the continuing education missions of the UW Professional and Continuing Education Department and the UW Bioengineering Department…keeping up with the science is clearly an important part of that mission and this course could serve as an easy to implement template elsewhere…and indeed, we are already getting inquiries form other organizations about this.

    Clearly the online journal club is just one way to help address this need, and absolutely there will need to be other approaches as well…Stuart, count me in to help in any way I can with your efforts to find and build these other solutions!