Drugs don’t create new functions — only alter pace of ongoing activity (exception: gene-based therapy).
Types of action: Stimulation (↑activity, e.g. adrenaline→heart; excess→rebound depression) · Depression (↓activity, e.g. barbiturates→CNS) · Irritation (nonselective, noxious; low-grade = counterirritant, e.g. eucalyptus oil — reflex ↑local circulation + blocks pain conduction) · Replacement (deficiency states — insulin, levodopa, thyroxine) · Cytotoxic (selective toxicity to pathogen/cancer — penicillin, cyclophosphamide).
Non-receptor (minority): physical (bulk laxatives, charcoal-adsorption, mannitol-osmotic, radioactivity) · chemical (antacids, KMnO4-oxidation, chelators, cholestyramine).
Target-based (majority) — 4 categories:
Agonist = affinity + full intrinsic activity. Antagonist = affinity + ZERO intrinsic activity, blocks agonist. Partial agonist = affinity + submaximal intrinsic activity — can be MORE potent than full agonist (buprenorphine 25× morphine) yet lower ceiling; also antagonizes full agonist (precipitates withdrawal in high-dependence, substitutes in low-dependence). Inverse agonist = affinity + NEGATIVE intrinsic activity — only matters for constitutively active receptors (BZD receptor; β-carbolines cause anxiety/convulsions).
Two-state model: receptor ⇌ Ri(inactive)/Ra(active). Agonist→stabilizes Ra. Competitive antagonist→binds both equally, no shift, just blocks agonist access. Partial agonist→mild Ra preference. Inverse agonist→stabilizes Ri (only visible if baseline constitutive activity exists).
| Family | Mechanism | Example | Speed |
|---|---|---|---|
| Ligand-gated ion channel | Direct pore opening | Nicotinic, GABA-A | Milliseconds (fastest) |
| GPCR | G-protein (Gs↑cAMP/Gi↓cAMP/Gq→IP3-DAG) | Muscarinic, adrenergic | Seconds |
| Enzyme-linked | Tyrosine kinase / JAK-STAT | Insulin, EGF, cytokine receptors | Minutes-hours |
| Nuclear | Ligand crosses membrane → DNA binding → transcription | Steroid, thyroid, vit A/D | Hours (slowest) |
Receptor regulation: Downregulation = chronic AGONIST use → ↓number/sensitivity → ↓drug effect (salbutamol tolerance). Upregulation = chronic ANTAGONIST use → ↑number/sensitivity → sudden stop → exaggerated response (propranolol withdrawal → rebound tachycardia/angina/HTN — hence taper, never stop abruptly).
Log dose → sigmoid curve, linear in 30-70% zone.
Potency = position on dose axis (EC50) — leftward=more potent. Efficacy = ceiling/Emax — independent of potency!
Efficacy = more clinically decisive. Steep DRC (hydralazine) = needs individualized dosing; flat DRC (HCTZ) = standard dose works for most.
Therapeutic index = LD50/ED50 (animal). Clinically: therapeutic window = dose range between min therapeutic effect and max acceptable adverse effect. Individual variation → hard to define universally. High-efficacy drug may still be dose-limited by toxicity (prednisolone in asthma).
Selectivity: separation between DRCs for different effects. Salbutamol (bronchodilation vs cardiac stim curves far apart = selective) vs isoprenaline (curves overlap = non-selective). Depends on: single vs multiple targets (omeprazole=1 target=narrow action; chlorpromazine=multiple receptors=broad action) AND target distribution (dexamethasone=1 receptor type but ubiquitous=widespread effects).
Synergism (combined > individual, same direction):
Antagonism (combined < individual):
Naloxone works because morphine antagonism is competitive/surmountable — displaceable, may need redosing since naloxone’s duration < morphine’s. Phenoxybenzamine-type irreversible block can’t be reversed by more agonist — recovery needs new receptor synthesis. β-blocker withdrawal danger = receptor upregulation, same logic as antagonist tolerance — always taper.
Pharmacodynamics is what a drug does to the body — the complete action-effect sequence, from the initial molecular interaction to the observed clinical response, along with how one drug’s action can be modified by another.
A fundamental principle worth stating before anything else: drugs (with the exception of gene-based therapies) do not create new functions in a cell, organ, or system. They only alter the pace of processes that are already ongoing — speeding something up, slowing it down, or blocking it. That said, this kind of modulation alone accounts for the entire range of therapeutic and toxic effects seen in medicine.
Non-receptor mechanisms account for only a small minority of drugs, acting through simple physical or chemical properties rather than any specific biological target:
The great majority of drugs act by interacting with a discrete target biomolecule, almost always a protein, which is what confers selectivity of action. These targets fall into four functional categories, beyond which only a few drugs act on other structures entirely (colchicine and vinca alkaloids on the structural protein tubulin, alkylating agents directly on nucleic acids):
Enzymes. Drugs mostly inhibit rather than stimulate enzymes, because physiological enzyme activity is usually already optimally set — true stimulation by a foreign drug is unusual (an exception: pyridoxine as a genuine cofactor for decarboxylase). Inhibition is either competitive (the drug resembles the substrate, competes reversibly for the catalytic site, raises the apparent Km but leaves Vmax unchanged, and can be overcome by enough substrate — physostigmine on cholinesterase, allopurinol on xanthine oxidase, captopril on ACE) or non-competitive (the drug binds an allosteric site, distorting the enzyme so Vmax falls while Km stays the same — aspirin on cyclooxygenase, omeprazole on H+/K+-ATPase, lovastatin on HMG-CoA reductase). A separate category — non-equilibrium/irreversible inhibition — occurs when the drug forms a strong covalent bond or has such high affinity that substrate cannot displace it at all (organophosphates on cholinesterase, methotrexate on dihydrofolate reductase, binding it 50,000 times more avidly than the natural substrate).
Ion channels. Drugs can directly block voltage-gated channels (local anaesthetics on neuronal Na+ channels, quinidine on cardiac Na+ channels, nifedipine on L-type Ca2+ channels, dofetilide on the cardiac delayed rectifier K+ channel) or modulate channels that are themselves a form of receptor (sulfonylureas closing pancreatic ATP-sensitive K+ channels, nicorandil opening them).
Transporters. Many drugs act by directly inhibiting a physiological carrier protein rather than a classical receptor — SSRIs blocking the serotonin transporter (SERT), tricyclics and cocaine blocking the noradrenaline transporter (NET), furosemide blocking the Na-K-2Cl symporter in the loop of Henle, thiazides blocking the Na-Cl symporter in the distal tubule, probenecid blocking the renal organic anion transporter.
Receptors. The largest category, and the one true pharmacology reserves the word “receptor” for specifically: a regulatory macromolecule (usually on the cell surface, sometimes intracellular) whose sole biological function is to recognize a signal molecule and initiate a response — it has no independent catalytic or structural role of its own. This distinction matters because using “receptor” loosely for every drug target (calling xanthine oxidase “the receptor” for allopurinol, for instance) dilutes a term that carries specific pharmacological meaning.
Evidence that drugs act through receptors: structural specificity (small changes in a molecule’s configuration can abolish or redirect its activity — isopropyl substitution on sympathomimetic amines confers β-selectivity; a 3-carbon vs 2-carbon side-chain separation in phenothiazines determines antipsychotic vs antihistaminic character), and stereospecificity (levo-noradrenaline is roughly 10 times more potent than its dextro isomer at the same receptor, despite identical chemical formula). Both point to a cell-surface structure that recognizes precise 3-dimensional shape — a receptor.
Key terms:
The two-state receptor model offers the mechanistic explanation: a receptor exists in equilibrium between an inactive (Ri) and an active (Ra) conformation. An agonist preferentially binds and stabilizes Ra, shifting the equilibrium toward the active state; a competitive antagonist binds both states with equal affinity, leaving the resting equilibrium unchanged (hence no effect of its own, but it occupies receptors an agonist would otherwise use); a partial agonist has only a modest preference for Ra, so even at saturating concentration the shift toward Ra is incomplete; an inverse agonist preferentially stabilizes Ri, and — only for receptors with baseline constitutive activity — this produces an effect opposite to the agonist’s.
| Family | Effector | Example receptors | Time to response |
|---|---|---|---|
| Ligand-gated ion channels (ionotropic) | Ion channel opens directly | Nicotinic (NM), GABA-A, glutamate | Milliseconds |
| G protein-coupled receptors (GPCRs, metabotropic) | Coupled channel or enzyme via a G protein | Muscarinic, adrenergic | Seconds |
| Enzyme-linked receptors | Intrinsic or associated enzyme (usually tyrosine kinase) | Insulin, epidermal growth factor, cytokine (JAK-STAT) receptors | Minutes to hours |
| Nuclear receptors | Regulate gene transcription directly | Steroid, thyroid hormone, vitamin A/D receptors | Hours (slowest) |
Ligand-gated ion channels are themselves the ion channel: agonist binding directly opens the pore, letting ions (Na+, K+, Ca2+, Cl-) flow and immediately depolarizing or hyperpolarizing the cell. The nicotinic receptor is the structural prototype — five subunits arranged around a central channel that opens when two acetylcholine molecules bind, admitting partially hydrated Na+ ions. Because there is no intermediate signalling step, this family produces the fastest onset of any receptor type.
G protein-coupled receptors span the membrane seven times and couple, via a heterotrimeric G protein (α, β, γ subunits), to an effector enzyme or channel. Agonist binding triggers the α-subunit to exchange GDP for GTP and dissociate, after which it (or the released βγ dimer) activates or inhibits its downstream target. Different G protein subtypes couple to different effectors and produce opposite outcomes even for chemically related receptors: Gs activates adenylyl cyclase (raising cAMP — β-adrenergic receptors), Gi inhibits it (lowering cAMP — α2-adrenergic receptors), and Gq activates phospholipase C (generating IP3 and DAG — M1 muscarinic receptors). The agonist here is the “first messenger”; cAMP, cGMP, IP3/DAG, and Ca2+ act as second messengers that carry the signal onward inside the cell.
Enzyme-linked receptors have catalytic activity built into (or immediately coupled to) their intracellular domain. Receptor tyrosine kinases (insulin, EGF receptors) dimerize on agonist binding and phosphorylate tyrosine residues on themselves and downstream proteins, triggering intracellular signalling cascades and altered gene transcription. JAK-STAT receptors (cytokine and growth hormone receptors) lack their own catalytic domain but recruit cytoplasmic JAK kinases on dimerization, which phosphorylate STAT proteins that then travel to the nucleus and alter transcription directly.
Nuclear receptors are intracellular (cytoplasmic or nuclear) rather than membrane-bound. A lipophilic ligand (steroid, thyroid hormone) diffuses across the cell membrane, binds its receptor, and the ligand-receptor complex migrates to (or is already in) the nucleus, binds specific DNA sequences, and regulates protein synthesis directly. This mechanism is inherently the slowest of the four, since it depends on new protein being made rather than modifying an existing one.
Receptor number and sensitivity are not fixed — they adapt to sustained drug exposure:
A dose-response curve has two conceptually separate stages: dose → plasma concentration (a pharmacokinetic question) and plasma concentration → effect (the pharmacodynamic question proper, and the one usually meant by “dose-response relationship”). At the receptor level, response follows a rectangular hyperbola against concentration (law of mass action), which becomes sigmoid — and usefully linear in its middle 30–70% portion — when dose is plotted on a logarithmic scale. Log-dose plotting also compresses a wide dose range onto one graph and makes comparisons between agonists and antagonists far easier to read.
Potency vs efficacy — genuinely different properties, and exam-favourite territory precisely because they’re so often confused:
They vary independently: aspirin is both less potent and less efficacious than morphine as an analgesic; pethidine is less potent than morphine but equally efficacious; furosemide is less potent than metolazone as a diuretic but more efficacious; diazepam is more potent than pentobarbitone as a CNS depressant but less efficacious. Efficacy is the more clinically decisive property — a highly potent drug that simply cannot achieve the needed effect at any dose is less useful than a less potent one that can. A steep dose-response curve (hydralazine) means small dose changes cause large response changes, so doses need individualization; a flat one (hydrochlorothiazide) means a standard dose suits most patients without much titration.
Therapeutic efficacy (a distinct, more clinical notion again) is expressed either as graded benefit within the recommended dose range (levodopa-carbidopa gives markedly more relief of parkinsonian symptoms than trihexyphenidyl can) or as a quantal success rate against a defined endpoint (a drug curing 95% of gonorrhoea cases is more therapeutically effective than one curing 75%, independent of either drug’s pharmacological potency).
Therapeutic index and therapeutic window. In animal studies, therapeutic index is classically expressed as LD50 ÷ ED50 (median lethal dose over median effective dose) — the larger this ratio, the wider the safety margin. Clinically, the more useful concept is the therapeutic window (or therapeutic range): the span between the dose producing minimal therapeutic effect and the dose producing maximal acceptable adverse effect. Because individual variability is real, a dose therapeutic for one patient may be toxic for another, which is exactly why defining a workable therapeutic range for many drugs is genuinely difficult in practice — and why a drug capable of a higher ceiling response (higher efficacy) may still be used at a lower, safer dose because intolerable adverse effects preclude going further (prednisolone in asthma is the standard example).
Drug selectivity reflects how widely a drug’s dose-response curves for different effects are separated. Salbutamol’s curve for bronchodilation is well separated from its curve for cardiac stimulation, making it a selective bronchodilator; isoprenaline’s two curves nearly overlap, so bronchodilating doses also stimulate the heart substantially. Selectivity itself depends on whether a drug acts on one target or several (omeprazole hits a single, tissue-restricted target — the gastric parietal cell proton pump — and so has one dominant action; chlorpromazine antagonizes dopamine D2, α-adrenergic, muscarinic, histamine H1, and some 5-HT receptors simultaneously, producing a correspondingly broad action profile) and on how widely that target itself is distributed in the body (dexamethasone acts on a single receptor type, but because glucocorticoid receptors exist in nearly every cell, its effects are still widespread).
When two drugs act together, the outcome is additive, supra-additive, indifferent, or antagonistic — and this interaction can happen at either a pharmacokinetic or a pharmacodynamic level (this section concerns the pharmacodynamic kind specifically).
Synergism — combined effect exceeds either drug’s individual effect:
Antagonism — combined effect is less than expected:
The competitive-vs-non-competitive antagonism distinction is not academic trivia — it predicts what happens at the bedside when an agonist overdose meets an antagonist. Naloxone reversing morphine overdose works because the antagonism is competitive: enough naloxone displaces morphine and restores the possibility of full agonist response if morphine is later reintroduced or naloxone wears off first (hence repeat dosing/monitoring is needed, since naloxone’s duration is often shorter than the opioid’s). By contrast, an irreversible receptor block (phenoxybenzamine) cannot be out-competed by simply giving more agonist — recovery depends on synthesis of new receptor protein, not on drug concentration at all. The same logic underlies why abrupt β-blocker withdrawal is dangerous (upregulated receptors, described above) and why potency and efficacy have to be evaluated as genuinely separate questions before calling one drug “better” than another for a given indication.
What to draw: A set of 3–4 sigmoid log dose-response curves on the same axes (log dose on x, % response on y), positioned to show one pair differing in potency only (parallel curves, different horizontal position, same ceiling) and one pair differing in efficacy only (same or different horizontal position, different ceiling).
Labelling requirements: mark EC50 for at least one curve explicitly with a dashed line from the 50%-response point down to the dose axis — this is the actual definition students are tested on, not just “the curve is to the left.” Label which curve is “more potent” and which is “more efficacious” directly on the diagram rather than leaving it to a legend.
Common exam-marking mistakes:
What to draw: Two side-by-side sets of log dose-response curves for an agonist alone and the same agonist with increasing concentrations of an antagonist. Left set (competitive): curves shift progressively rightward but all reach the same maximum. Right set (non-competitive): curves flatten, with progressively lower maxima, without a clean rightward shift.
Labelling requirements: explicitly annotate “same maximum reached — surmountable” under the competitive set and “maximum reduced — unsurmountable” under the non-competitive set; this contrast is the entire teaching point and is easy to lose if the curves aren’t directly compared side by side.
Common exam-marking mistakes:
Not rendered as a diagram — a 4-column comparison (family × effector × example receptors × time to response) communicates more clearly as the table already in notes.md than as a flowchart, and a rendered version would simply duplicate that table. See notes.md for the comparison and the mechanism detail (GPCR subunit dissociation, nuclear receptor membrane-crossing requirement) in prose.
Personal revision notes, mnemonics and reminders.
