Small molecules and large biologics navigating a body of 1013 cells, looking for the one protein they were designed to bind.
Every dose runs the same gauntlet. The fraction that reaches the target is often single digits.
Oral, IV, transdermal, inhaled. Bioavailability varies wildly — insulin orally is ~0%.
Bound to plasma proteins, partitioned by lipophilicity. The blood-brain barrier turns most drugs away.
The liver's CYP450 enzymes chop drugs into metabolites — sometimes activating, sometimes deactivating.
Kidneys (urine) and bile (feces) carry the remains. Half-life sets the dosing schedule.
Pharmacokinetics asks where the drug goes. Pharmacodynamics asks what happens when it gets there.
Almost every drug works by binding a protein — a receptor, an enzyme, an ion channel, a transporter — and changing what that protein does.
The relationship between concentration and effect is rarely linear. Doubling the dose rarely doubles the effect. Saturate the receptors and adding more accomplishes nothing… until you hit the toxicity ceiling.
Many of the best-selling drugs in history work by sitting in an enzyme's active site so the substrate cannot.
Block cholesterol synthesis in the liver. Atorvastatin, rosuvastatin. Among the most prescribed drugs on Earth.
Lower blood pressure by stopping the conversion of angiotensin I → II. Lisinopril, enalapril.
Shut off rogue growth signals in cancer cells. Imatinib (Gleevec) turned chronic myeloid leukemia into a managed condition.
Penicillins, cephalosporins, macrolides, fluoroquinolones. Selectively kill bacterial machinery our cells lack.
Ibuprofen, naproxen, aspirin. Inhibit COX enzymes to reduce prostaglandins.
Morphine, oxycodone, fentanyl. Bind μ-opioid receptors. Powerfully effective — powerfully addictive.
SSRIs (sertraline, fluoxetine), SNRIs, MAOIs. Modulate serotonin / norepinephrine over weeks.
Warfarin, heparin, DOACs (apixaban, rivaroxaban). Prevent strokes, treat DVT — bleeding is the price.
Each class targets a different receptor, enzyme, or transporter. The categories keep multiplying.
Every drug has a dose where it works and a dose where it kills you. The ratio is the therapeutic index.
For every approved drug, roughly 10,000 starting molecules were screened. Most of the cost is paying for the failures.
Phase III is where most candidates die — a drug that works in 200 people can fail in 2,000. The attrition rate from preclinical to approval is around 90%.
A US drug patent runs 20 years from filing — effectively 8–12 years on market before competitors can copy.
When the patent expires, generic manufacturers must only prove bioequivalence: the active ingredient, the same blood concentration. They skip the discovery, the trials, the failures.
Prices typically fall 80–90% within a year. Atorvastatin went from $5/pill to under 10¢.
No drug is selective enough to bind only its intended target. Even when it is, the target does more than one thing.
Stronger dose, stronger side effect. NSAIDs and ulcers, opioids and constipation, beta-blockers and fatigue.
Rare reactions specific to the patient — genetic variants, immune responses. Stevens-Johnson syndrome from carbamazepine.
One drug induces or inhibits the CYP enzymes that metabolize another. Grapefruit juice does this too.
Adverse drug events cause an estimated ~100,000 US deaths annually — many from interactions in patients on 5+ medications. Polypharmacy is the silent epidemic of modern medicine.
Drugs matched to a tumor's specific mutation, not its tissue of origin. Companion diagnostics required.
Antibody-drug conjugates: an antibody finds the cancer cell, the linked toxin kills it. Trastuzumab deruxtecan.
Deliver instructions, let the cell build the protein. COVID vaccines were the proof; cancer vaccines are next.
Generative models propose binders for a protein structure. AlphaFold made every target druggable in principle.
Plus: gene therapies (CRISPR, AAV), PROTACs that destroy proteins instead of inhibiting them, gut-microbiome modulators, psychedelics in psychiatry. The pipeline has never been more diverse.
Pharma is the only industry where 9 of every 10 products entering human trials never sell a single dose — and the model still works.
It works because the winners win enormously. Statins added years of life to entire populations. ART turned HIV from a death sentence into a chronic condition. Imatinib, GLP-1 agonists, hepatitis C cures, immunotherapy — each one rewrote a chapter of medicine.
The criticism is fair: prices are high, marketing is aggressive, the same molecule costs different amounts in different countries for opaque reasons. The replication crisis touches preclinical research too.
But the alternative — no industry willing to spend a decade and $2B on a 10% shot — is not better drugs. It is no new drugs.
A century ago, an infected cut could kill you. A bad heart meant the funeral home. Cancer meant months. The molecules in the bottle on your nightstand are the accumulated payoff of millions of failed experiments.
Pharmacology is humanity's best argument that biology is, in fact, engineerable.
Three centuries of trying to make molecules behave. Below, where to keep going.
End of deck. Press Home to restart, ← to step back.