How to discover a new drug (and why it's so difficult)

Unlike other scientific pursuits, drug hunting is a task not guided by any known equations or formulae



Donald R. KirschOgi Ogas
October 13, 2018 6:00PM (UTC)

Excerpted with permission from The Drug Hunters: The Improbable Quest to Discover New Medicines by Donald R. Kirsch and Ogi Ogas. Copyright 2018 by Arcade Publishing, an imprint of Skyhorse Publishing, Inc.

After spending my entire career searching for new medicines, I’ve learned that the only sure thing in the drug hunting business is that you almost never end up with the exact drug you started stalking. The vast majority of my colleagues, all educated at top-flight research universities and working at posh laboratories festooned with high-tech gear, have spent their entire careers groping through the labyrinth of bio-active molecules without ever finding a new compound that safely and effectively improves human health.


The professor who trained me in pharmacology, an MD, once told me that 95 percent of the time a patient visits a physician he will not actually be helped by the doctor. In most cases, either the patient’s body will heal itself without needing the doctor’s intervention or the disease will be untreatable, rendering the physician powerless. In his view, the physician has the ability to make a meaningful difference to his patients only 5 percent of the time. While that may seem low, those are fantastic odds compared to the ones faced by the drug hunter.

Only about 5 percent of a scientist’s ideas for a drug discovery project get funded by management. Of these, only 2 percent end up producing an FDA-approved medicine. That means a drug hunting scientist can only expect to make a difference about one tenth of 1 percent of the time. Finding new drugs is so challenging, in fact, that it has led to a crisis in the pharmaceutical industry. Big Pharma companies are becoming increasingly frustrated with the massive research expenditures necessary to come up with new drugs—an average of about $1.5 billion and fourteen years for each FDA-approved drug—and the exasperating fact that the vast majority of their endeavors don’t produce a usable drug. Executives at Pfizer recently told me they are thinking about getting out of the drug-discovery industry entirely. Instead, they want to be in the drug acquisition industry: they would prefer just to buy drugs that other people have invented. Think about that. Finding new drugs is so formidable that one of the oldest, most talented, and wealthiest drug makers—the largest drug maker in the world, in fact—would rather let other people deal with the problem.

So why is the “degree of difficulty” of finding new drugs so much greater than, say, landing a man on the moon or designing an atomic bomb? The moon shot and the Manhattan Project employed well-established scientific equations, engineering principles, and mathematical formulas. They were formidable and grueling endeavors, to be sure, but at least the researchers possessed clear scientific road maps and mathematical compasses to guide them. The moon shot engineers knew with certainty the distance from the Earth to the moon, and how much fuel was needed to get there. The Manhattan Project scientists knew that matter could be turned into city-obliterating energy according to E=mc2.

In contrast, the core challenge of designing a new drug — the trial-and-error screening of immense numbers of candidate compounds — is a task not guided by any known equations or formulas. While an engineer knows if his bridge will bear weight before he ever lays a girder down, a drug hunter has no clear idea how a particular drug will work until a human subject actually ingests it.

In the mid-nineties, chemists at Ciba-Geigy (now a part of Novartis) calculated the total number of possible drug compounds in our universe: 3 x 1062. When it comes to characterizing the size of a number, some numbers are big, some are enormous, and some are so incomprehensibly, inconceivably large that they might as well be infinity. The number 3 x 1062 falls into that third category. If you were able to test one thousand compounds every second to see if any of them could serve as an effective remedy for a particular malady—say, breast cancer—by the time our sun burned out you would still have not made a measurable dent in the total number of breast cancer-fighting drug possibilities.

There is a story by the blind Argentinean author Jorge Luis Borges that I think perfectly captures the central challenge of drug hunting. In “The Library of Babel,” Borges imagines that the universe is a library consisting of an infinite number of hexagonal rooms that extend forever in every direction. Each room is filled with books. Each book contains a random arrangement of letters, and no two books are the same. Once in a while, purely by chance, a book contains an entire readable sentence, such as “The gold is in the mountain.” But, as Borges puts it, “For every rational line or forthright statement there are leagues of senseless cacophony, verbal nonsense, and incoherency.”


Nevertheless, the library must contain books that, purely by chance, are filled with legible life-changing wisdom. Such books are known as “Vindications.” In Borges’s fantasy, solitary searchers known as librarians wander endlessly through the library, hoping to find these Vindications. Most librarians wander through the endless hexagons in vain, spending their life coming across nothing but nonsense. But Borges notes that there are librarians who, through good fortune or sustained force of will, have managed to discover a Vindication.

Similarly, every possible drug is contained somewhere in the vast theoretical library of chemical compounds. There is a molecular configuration that will safely destroy ovarian cancer, another that will halt the corrosive advance of Alzheimer’s, another that will cure AIDS—or maybe they do not exist at all. There is no way to know for sure. Modern drug hunters are like Borges’s Babelian librarians, forever questing for life-changing compounds and always suppressing their secret fear that the vindicating medicines may never be found.

The problem, ultimately, is the human body. Our physiological activity is not a closed, well-defined system like rocket propulsion or nuclear fission. It is an open and unfathomably arcane molecular system with innumerable undefined relationships among its components, rendered even more abstruse by the fact that each person’s body has their own idiosyncratic structure and dynamics. We only understand a tiny fraction of these physiological relationships and have not yet deciphered how most of our body’s basic molecular components actually work. Complicating matters still further is the fact that each individual has her own idiosyncratic genetic and physiological architecture, so that each person’s body operates slightly (or extremely) differently. Even more daunting, despite tremendous advances in our understanding of cells and tissues and organs, we simply cannot precisely predict in advance how a given chemical compound will interact with a given molecule in a living body. In fact, it is impossible to know with certainty if a particular disease possesses what pharmacologists call a “druggable protein” or a “druggable target”—some specific protein associated with a pathology that can be influenced by a chemical agent.

Designing an effective drug requires two things: the right compound (the drug) and the right target (the druggable protein). The drug is like a key that turns the protein lock to start the ignition on a physiological engine. If a scientist wants to intentionally influence a person’s health in a specific way—to reduce depression, relieve itching, treat food poisoning, or produce any other health benefit—she must first identify a target protein that influences the relevant physiological processes in the human body or that, conversely, interferes with the physiological processes of a pathogen.


For example, Lipitor acts upon HMG-CoA reductase, the protein controlling the rate of the body’s synthesis of cholesterol. Penicillin, in contrast, shuts down peptidoglycan transpeptidase, a protein required to synthesize the (essential) cell wall of a bacterium. But finding the drug key that will turn a protein lock . . . As Hamlet would say: Ay, There’s the rub! This is the daunting challenge for the drug hunter. Despite the humbling odds, some drug hunters, such as Suren Sehgal, through unwavering resolution or outrageous fortune, through individual genius or far-flung collaborations, have stumbled upon their Vindications.

The term drug hunters have bestowed upon the process of systematically searching through a library of compounds is screening. The prehistoric screening method consisted of snatching every new berry or leaf you came across and snorting it, smearing it, or swallowing it. After unknowable eons of our ancestors randomly sampling the natural landscape, in 1847 the first drug was discovered using a reasonably scientific method of screening. At the time, physicians were using ether as a surgical anesthetic, prompting them to reason that there could be other chemical compounds similar to ether that might work even better. Ether had a few obvious shortcomings—it irritated patients’ lungs and had an unfortunate tendency to explode—so physicians knew there would great clinical value for a new anesthetic without these issues.

Since ether was a volatile organic liquid, the Scottish physician James Young Simpson and two of his colleagues decided to test every volatile organic liquid they could get their hands on. Their screening process was simple: they opened a bottle of a given test liquid and inhaled its vapors. If nothing happened, they labeled the sample inactive. If they woke up on the floor, they labeled the sample active.


This screening protocol, of course, would never meet contemporary standards for laboratory safety. Benzene, for instance, is a volatile organic liquid that was widely available at the time and was almost certainly one of the compounds that Simpson screened. We now know that benzene is carcinogenic, and inhaling its vapors can cause long-term damage to your ovaries or testes.

Despite the recklessness of their screening method, on the evening of November 4, 1847, Dr. Simpson and his colleagues tested chloroform. When the three men inhaled the chemical it produced a mood of cheer and good humor—followed by collapse and unconsciousness. When they awoke hours later, Simpson knew they had identified an active sample.

Hoping to confirm their findings, Simpson insisted that his niece, Miss Petrie, inhale the chloroform while he watched. The girl blacked out. It is fortunate she woke again, since we now know that chloroform is a powerful cardiovascular depressant, producing a high incidence of death when used as a surgical anesthetic. Despite these dangers, by sniffing chemical after chemical in his living room, Simpson had discovered one of the blockbuster drugs of the nineteenth century—a pharmaceutical origin story unlikely to be duplicated today. But you never know. In the 1980s, I tried to find new drugs in the back of a Volkswagen microbus.


If you are thinking that I must have been indulging in tie-dyed psychedelic experimentation—after all, why else would anyone be diverting himself with unknown drugs in the back of a lime-green VW bus?—you would be wrong. One of my first paid jobs was working as a drug hunter in an antibiotics discovery group. A common method for searching for new antibiotics is to screen microorganisms living in the soil. I always kept an eye out for new kinds of soil that might hide a pharmaceutical payoff—and a commercial payoff. I was literally looking for pay dirt.

One weekend, I volunteered to travel to the Delmarva Peninsula to screen soil samples from the Chesapeake Bay side of the peninsula. I took our “mobile laboratory”—the microbus, which we had equipped with a sink and a Bunsen burner. Since my group had recently discovered a new type of antibiotic called monobactams, we christened our mobile laboratory the “Monobacvan.”

I somehow managed to recruit my wife to come with me with promises of sunbathing on the beach, but then conscripted her into driving the Monobacvan around the tight curves of the rural shoreline as I hunkered down in the back, abruptly ordering her to stop so I could dash out and fill bags full of dirt. During those moments when we were not driving or scooping up the dank, smelly Chesapeake earth, I was diluting the samples and slapping them on Petri plates. My wife was not pleased. The weekend was a bust for both of us, since when I got back to the lab on Monday and we tested my samples, each one turned out to be inactive. My wife informed me that if I did not want my marriage labeled inactive, our next road trip needed much more sunbathing and absolutely no more screening.

When people learn that I am a drug hunter they usually ask me at least one of the following three questions—usually expressed with some well-founded cynicism:


Why are my drugs so expensive?

Why do my drugs have such unpleasant side effects?

Why is there no medicine for the malady that afflicts me or someone I love?

The truth of the matter is that the answers to all three are tied to the fact that drug hunting—at least, thus far—is dismayingly difficult because every contemporary method of drug development relies, at some crucial juncture, on trial-and-error screening, just as it did when Neanderthals roamed the wilds. We still do not possess adequate knowledge of human biology to provide us with theories and principles that could rationally guide us to the salubrious molecules we so fiercely desire.


# # #

Excerpted with permission from The Drug Hunters: The Improbable Quest to Discover New Medicines by Donald R. Kirsch and Ogi Ogas.

Donald R. Kirsch

MORE FROM Donald R. Kirsch

Ogi Ogas



Read Now, Pay Later - no upfront
registration for 1-Hour Access

Click Here
7-Day Access and Monthly
Subscriptions also available
No tracking or personal data collection
beyond name and email address


Fearless journalism
in your inbox every day

Sign up for our free newsletter

• • •