How New Drugs Are Developed: From Discovery to Pharmacy Shelf

Target Identification and Validation

Drug development begins with identifying a molecular target, typically a protein, that plays a causal role in a disease. Targets can be receptors, enzymes, ion channels, transport proteins, or even specific genes. The ideal target is essential for the disease process but dispensable for normal physiology, ensuring that drugs acting on it will produce therapeutic effects without widespread side effects.

Target identification draws on multiple sources. Genomic studies (genome-wide association studies, or GWAS) identify genes associated with disease risk. Proteomic and metabolomic analyses reveal proteins and metabolites that are altered in disease states. Animal models of disease show which pathways, when disrupted, reproduce or alleviate disease symptoms. Clinical observations sometimes provide the initial clue: the discovery that patients with certain genetic mutations are resistant to a disease can point directly to a druggable target. The identification of PCSK9 as a target for cholesterol-lowering drugs came from the observation that people with loss-of-function PCSK9 mutations have exceptionally low LDL cholesterol levels and are protected from heart disease.

Target validation confirms that modulating the target actually affects the disease. This is done through genetic knockdown or knockout experiments (using RNA interference or CRISPR to silence the target gene in cells or animals), chemical tool compounds (early-stage molecules that hit the target and produce the expected biological effect), and analysis of human genetic data linking target activity to disease outcomes. Many drug development programs fail because the target, while associated with the disease, does not actually drive it, making target validation one of the most critical and most difficult steps in the entire process.

Lead Discovery and Optimization

Once a validated target exists, the next step is finding chemical compounds that interact with it. High-throughput screening (HTS) tests large compound libraries (often containing 1 to 2 million molecules) against the target in automated assays, identifying "hits," compounds that show activity at the target. A typical HTS campaign yields hundreds to thousands of hits, which are then filtered based on potency, selectivity, and chemical properties to identify a smaller set of "lead" compounds worth further development.

Rational drug design uses knowledge of the target's three-dimensional structure (from X-ray crystallography or cryo-electron microscopy) to design molecules that fit into the active site or binding pocket. Computational docking simulations predict how candidate molecules would bind to the target, allowing chemists to prioritize the most promising structures before synthesizing them. The HIV protease inhibitors developed in the 1990s (saquinavir, ritonavir, indinavir) were among the first major successes of structure-based drug design.

Fragment-based drug discovery screens very small, simple molecules (typically under 300 daltons) that bind weakly to the target. These fragments are then grown or linked together to create larger, more potent molecules. This approach can explore chemical space more efficiently than HTS because a small fragment library (a few thousand compounds) can represent a much larger diversity of binding interactions.

Lead optimization is the iterative process of modifying the lead compound's chemical structure to improve its drug-like properties. Medicinal chemists make hundreds of analogs, testing each for potency at the target, selectivity against related targets, metabolic stability (resistance to breakdown by CYP enzymes), solubility, oral bioavailability, and absence of obvious toxicity. The goal is a compound that is potent and selective enough to work at achievable doses, stable enough to survive the gastrointestinal tract and liver, and safe enough to proceed to animal testing. This stage typically produces a "development candidate" suitable for preclinical studies.

Preclinical Testing

Before a drug can be tested in humans, extensive preclinical studies must establish its safety and pharmacological properties. Regulatory agencies require both in vitro (cell-based) and in vivo (animal) studies.

Pharmacokinetic studies in animals determine how the drug is absorbed, distributed, metabolized, and excreted in a living system. These studies establish the relationship between dose and plasma concentration, identify the major metabolites, and predict what doses might be needed in humans. Scaling pharmacokinetic data from animals to humans uses allometric methods that account for differences in body size, metabolic rate, and organ blood flow.

Toxicology studies evaluate the drug's potential to cause harm. Acute toxicity studies determine the dose that causes adverse effects after a single administration. Repeat-dose toxicity studies (lasting 2 weeks to 6 months) identify toxic effects from chronic exposure, including target organ toxicity (which organs are damaged first). Genotoxicity assays test whether the drug damages DNA, which could cause cancer. Reproductive toxicity studies assess effects on fertility and fetal development. Safety pharmacology studies evaluate effects on the cardiovascular system (especially QT prolongation, which can cause fatal cardiac arrhythmias), respiratory system, and central nervous system.

Efficacy studies in animal disease models provide evidence that the drug actually works against the disease in a living system. While no animal model perfectly replicates a human disease, these studies help establish proof of concept and guide dose selection for human trials.

All preclinical data is compiled into an Investigational New Drug (IND) application submitted to the FDA (or equivalent regulatory agency in other countries). The IND must demonstrate that the drug is reasonably safe to test in humans and that the proposed clinical trial design is adequate. If the FDA does not object within 30 days, clinical trials may begin.

Clinical Trials: Testing in Humans

Clinical trials proceed through four phases, each with distinct objectives, populations, and designs.

Phase I trials are the first human exposure. They typically enroll 20 to 100 healthy volunteers (or patients with serious diseases for which no treatment exists, such as cancer). The primary objectives are to determine the drug's safety, tolerability, pharmacokinetics, and pharmacodynamics in humans. Doses start low (usually one-tenth the no-observed-adverse-effect level from animal studies) and escalate gradually, with intensive monitoring for side effects. Phase I establishes the maximum tolerated dose (MTD) and the recommended dose range for Phase II.

Phase II trials are the first tests of efficacy. They enroll 100 to 300 patients with the target disease and compare the drug to placebo or existing treatment in a randomized, controlled design. Phase II has the highest failure rate of any clinical phase: approximately 60 to 70% of drugs that enter Phase II fail, usually because the drug does not produce a clinically meaningful improvement over placebo despite its activity in preclinical models. This gap between animal model efficacy and human disease efficacy is one of the most persistent challenges in drug development. Phase II also further characterizes side effects and refines the optimal dose.

Phase III trials are large, definitive, randomized controlled trials designed to confirm efficacy and monitor adverse reactions in a broader patient population. They typically enroll 1,000 to 3,000 patients (sometimes more) across multiple clinical sites and countries. Phase III trials must demonstrate that the drug provides a clinically significant benefit compared to the current standard of care, with an acceptable side effect profile. These trials are expensive, often costing $100 million to $300 million per trial, and take 2 to 4 years to complete. Approximately 40 to 50% of drugs that enter Phase III fail.

Regulatory review follows successful Phase III trials. The drug manufacturer submits a New Drug Application (NDA) to the FDA, containing all preclinical and clinical data, manufacturing information, and proposed labeling. FDA reviewers evaluate the data, and an advisory committee of independent experts may weigh in. The FDA targets a 10-month review period for standard applications and 6 months for priority review (drugs addressing serious conditions with unmet needs). Approval rates for drugs that reach NDA submission are approximately 85 to 90%.

Phase IV (post-marketing surveillance) continues after approval. It monitors long-term safety in the general population, which is broader and more diverse than clinical trial participants. Phase IV has detected serious adverse effects that were too rare to appear in clinical trials, leading to label changes, restricted distribution, or market withdrawal. Rofecoxib (Vioxx) was withdrawn in Phase IV after cardiovascular risks became apparent in long-term data. Troglitazone (Rezulin) was withdrawn after post-marketing reports of severe liver toxicity.

Accelerated Pathways and Orphan Drugs

Regulatory agencies have created several mechanisms to speed access to drugs for serious or life-threatening conditions. Fast Track designation provides more frequent FDA meetings and the possibility of rolling review (submitting sections of the NDA as they are completed rather than waiting for the entire package). Breakthrough Therapy designation, introduced in 2012, provides intensive FDA guidance throughout development for drugs that show substantial improvement over existing treatments based on preliminary clinical evidence.

Accelerated approval allows drugs to be approved based on a surrogate endpoint (a laboratory measurement or physical sign that is reasonably likely to predict clinical benefit) rather than waiting for proof of actual clinical benefit. HIV viral load, rather than AIDS-related death, was the surrogate endpoint that enabled several antiretroviral drugs to reach patients years earlier than traditional approval would have allowed. Drugs approved via accelerated approval must conduct confirmatory Phase IV trials to verify that the surrogate endpoint does indeed predict the clinical benefit.

Orphan Drug Act incentives (in the US since 1983, with similar programs in Europe and Japan) encourage development of drugs for rare diseases (affecting fewer than 200,000 people in the US). Incentives include 7 years of market exclusivity, tax credits for clinical research costs, reduced FDA fees, and eligibility for FDA research grants. Before the Orphan Drug Act, fewer than 40 drugs for rare diseases had been approved in the US. Since its passage, over 600 orphan drugs have reached the market.

The Economics and Future of Drug Development

The cost of drug development has risen steadily, a phenomenon sometimes called Eroom's Law (Moore's Law spelled backward): the number of new drugs approved per billion dollars of R&D spending has roughly halved every 9 years since 1950. Contributing factors include increasingly stringent regulatory requirements, the exhaustion of easily druggable targets ("low-hanging fruit"), rising clinical trial costs driven by larger and longer trials, and the increasing difficulty of demonstrating superiority over existing treatments.

Several approaches aim to improve development efficiency. Biomarker-driven trials use molecular measurements to select patients most likely to respond, enriching the trial population and increasing the signal-to-noise ratio. Adaptive trial designs allow modifications to the trial protocol (such as dropping ineffective dose arms or reallocating patients to more promising treatments) based on interim data, reducing the number of patients needed and shortening timelines. Real-world evidence from electronic health records and insurance claims databases supplements traditional clinical trial data, particularly for post-approval safety monitoring.

Artificial intelligence is increasingly used in early drug discovery to predict which compounds will have favorable drug-like properties, to identify new uses for existing drugs (drug repurposing), and to design novel molecules with desired properties. AI-designed drugs have begun entering clinical trials, and while the technology is still maturing, it has the potential to significantly reduce the time and cost of lead discovery and optimization.