Arachidonic acid to prostaglandins & leukotrienes — the COX/LOX branch point and where NSAIDs intervene
Every eicosanoid — every prostaglandin, thromboxane, and leukotriene — begins with a single hydrolysis event: cytosolic phospholipase A2 (cPLA2) cleaving arachidonic acid free from the membrane phospholipid pool. This is not a passive leak; it is a tightly Ca²⁺-gated, tissue-specific decision that determines whether a cell will fuel a local inflammatory response, and it is the step that sets the ceiling on how much downstream prostaglandin and leukotriene synthesis is even possible.
cPLA2-α is a 749-residue, 85 kDa cytosolic enzyme that is catalytically silent in resting cells despite being abundantly expressed. Its activation follows a strict two-signal logic:
1. Ca²⁺ mobilization: agonists (thrombin via PAR receptors, bradykinin via B2 receptors, ATP via P2Y, or complement C5a) trigger phospholipase C-mediated IP3 production, releasing ER Ca²⁺ stores and raising cytosolic [Ca²⁺] from ~100 nM resting to 300–800 nM. The C2 domain of cPLA2 binds Ca²⁺ directly, which — without catalysis — drives translocation from cytosol to the perinuclear envelope, Golgi, and ER membranes rich in arachidonate-containing phospholipids.
2. MAPK phosphorylation: parallel activation of p38 MAPK and ERK1/2 phosphorylates Ser505 of cPLA2, roughly doubling catalytic turnover once the enzyme is already membrane-bound. This dual requirement (Ca²⁺ for localization, phosphorylation for activity) creates an AND-gate that prevents spurious eicosanoid production from transient Ca²⁺ noise.
Substrate selectivity: cPLA2 shows strong preference for phospholipids bearing arachidonic acid specifically at the sn-2 position — phosphatidylcholine and phosphatidylethanolamine are the dominant donors. This channels the enzyme toward the ω-6 eicosanoid precursor pool rather than other sn-2 fatty acids (oleate, linoleate), even though those are far more abundant in the bulk membrane.
Rate-limiting significance: • cPLA2 knockout mice are essentially unable to mount acute inflammatory prostaglandin/leukotriene responses • Free AA concentration, not COX or LOX enzyme abundance, is usually the rate-limiting factor for total eicosanoid output in stimulated cells • Secretory PLA2 (sPLA2, group IIA) acts extracellularly and amplifies the response in chronic/exudative inflammation (synovial fluid concentrations reach µg/mL in rheumatoid arthritis)
Once liberated, free AA sits at a metabolic crossroads. The Lands cycle continuously re-esterifies AA back into membrane phospholipids via acyl-CoA synthetase and lysophospholipid acyltransferase (LPCAT), so under basal conditions >95% of released AA is simply recycled with no eicosanoid output.
During active inflammatory signaling this balance shifts: cPLA2 output rate transiently exceeds reacylation capacity, and the resulting free AA pool (low micromolar) becomes available to COX and 5-LOX, both of which have Km values in the same low-micromolar range — meaning enzyme access to substrate, not enzyme saturation, governs flux split between the two pathways.
Dietary and membrane composition modulate the whole cascade upstream of any enzyme: diets enriched in ω-3 fatty acids (EPA, DHA) partially displace AA from the sn-2 position, so cPLA2 releases EPA instead — EPA-derived eicosanoids (3-series prostaglandins, 5-series leukotrienes) are markedly less inflammatory than their AA-derived (2-series/4-series) counterparts. This is the mechanistic basis for fish-oil anti-inflammatory effects.
Cyclooxygenase (prostaglandin H synthase, PGHS) performs one of the most consequential reactions in human pharmacology: converting arachidonic acid into PGH2, the shared precursor for every downstream prostaglandin and thromboxane. Two isoforms — COX-1 constitutive and COX-2 inducible — share 60% sequence identity and near-identical catalytic mechanisms, yet their differential expression is the entire rationale for the NSAID and coxib drug classes.
COX is a homodimeric, heme-containing membrane protein anchored to the luminal face of the ER and inner nuclear membrane, with two coupled catalytic activities in a single polypeptide:
Cyclooxygenase activity (bis-dioxygenation): • AA enters a long hydrophobic channel; Tyr385, activated by the heme-Fe(IV)=O radical generated at the peroxidase site, abstracts the 13-pro-S hydrogen from AA • Molecular O2 is added twice in a stereospecific radical cascade, forming the bicyclic endoperoxide with a 15-hydroperoxy group: PGG2 • This step defines the "cyclo" in cyclooxygenase — the two new 5-membered and 6-membered rings created here are conserved in every downstream prostaglandin
Peroxidase activity: • A second, spatially distinct heme site reduces the 15-hydroperoxide of PGG2 to a 15-hydroxyl, yielding PGH2 • This reduction consumes two electrons, typically donated by cellular reducing cofactors (glutathione peroxidase substrates) or by co-oxidized xenobiotics
Kinetics: Km for AA is ~5–10 µM for both isoforms; kcat is roughly 1,000 catalytic cycles/min, but the enzyme undergoes suicide inactivation after only ~1,400 turnovers due to heme/radical-mediated self-oxidation — COX activity is inherently self-limiting even without any pharmacological inhibitor, providing a built-in brake on prostaglandin production.
COX-1 (gene PTGS1, chromosome 9): expressed constitutively at stable levels in nearly all tissues — gastric mucosa, platelets, kidney, vascular endothelium. It supplies baseline "housekeeping" prostaglandins: gastric PGE2/PGI2 that stimulate mucus and bicarbonate secretion and maintain mucosal blood flow, and platelet TXA2 that supports normal hemostasis.
COX-2 (gene PTGS2, chromosome 1): essentially undetectable in most resting tissues, but transcriptionally induced 10–80 fold within 1–3 hours of exposure to LPS, IL-1β, TNF-α, or growth factors, via NF-κB and NFAT response elements in its promoter. COX-2 is the dominant source of prostaglandins at sites of active inflammation, and is also constitutively expressed in kidney (macula densa, supporting renal blood flow autoregulation) and in the brain.
Structural basis for isoform-selective inhibition: COX-2's substrate channel has a slightly larger side pocket (Val523 instead of the bulkier Ile523 found in COX-1), which selective coxibs (celecoxib, rofecoxib) exploit — their bulky sulfonamide/sulfone side groups fit the COX-2 pocket but sterically clash with COX-1's narrower channel, giving 300–400 fold selectivity ratios.
While cyclooxygenase dominates prostaglandin pharmacology, roughly half of arachidonic acid liberated during inflammation is instead routed through 5-lipoxygenase to leukotrienes — potent chemoattractants and bronchoconstrictors entirely untouched by aspirin or ibuprofen. This parallel pathway explains both why NSAIDs sometimes worsen asthma (aspirin-exacerbated respiratory disease) and why a distinct drug class (leukotriene modifiers) exists alongside NSAIDs rather than replacing them.
5-lipoxygenase is a 78 kDa, non-heme iron dioxygenase found in myeloid lineage cells — neutrophils, eosinophils, monocytes/macrophages, mast cells — but critically, not in platelets or vascular endothelium, giving the leukotriene pathway a much more restricted cellular distribution than COX.
Activation sequence: 1. Resting 5-LOX is cytosolic/nucleoplasmic and catalytically inactive 2. Ca²⁺ influx (paralleling cPLA2 activation) drives 5-LOX translocation to the nuclear membrane 3. FLAP (5-lipoxygenase-activating protein), an integral 18 kDa nuclear membrane protein, binds free AA and presents it to 5-LOX — without FLAP, 5-LOX cannot efficiently access membrane-derived AA even though it is catalytically competent in vitro 4. 5-LOX abstracts the pro-S hydrogen at C7 of AA and inserts molecular O2 at C5, forming 5-HPETE (5-hydroperoxyeicosatetraenoic acid) 5. The same active site then performs a second reaction — dehydration of 5-HPETE — producing the unstable allylic epoxide LTA4 (half-life seconds at 37°C)
Branch point at LTA4: • In neutrophils: LTA4 hydrolase (a bifunctional zinc metalloenzyme with separate aminopeptidase activity) hydrolyzes LTA4 to LTB4 — a dihydroxy acid • In eosinophils, mast cells, basophils: LTC4 synthase, a membrane-bound glutathione-S-transferase, conjugates LTA4 with reduced glutathione to form LTC4, subsequently metabolized extracellularly to LTD4 and LTE4 (collectively "cysteinyl leukotrienes," historically termed slow-reacting substance of anaphylaxis)
LTB4 acts through two receptors: high-affinity BLT1 (Kd~1nM) on neutrophils driving chemotaxis, integrin activation, and degranulation — it is among the most potent neutrophil chemoattractants known — and lower-affinity BLT2, more broadly expressed.
Cysteinyl leukotrienes (LTC4/D4/E4) act primarily through CysLT1 receptors on airway smooth muscle, causing bronchoconstriction roughly 1,000-fold more potent than histamine on a molar basis, plus increased vascular permeability and mucus secretion — the core pathophysiology of asthma exacerbation.
Aspirin-exacerbated respiratory disease (AERD, Samter's triad): in ~10% of adult asthmatics, COX-1 inhibition by NSAIDs removes a tonic PGE2-mediated brake on 5-LOX (PGE2 normally suppresses leukotriene synthesis via the EP2 receptor). Blocking COX-1 without touching 5-LOX shunts AA preferentially into the leukotriene pathway, triggering severe bronchospasm — this is precisely why leukotriene-pathway drugs occupy a therapeutic niche NSAIDs cannot fill.
Pharmacological targeting: zileuton directly inhibits 5-LOX catalysis; montelukast and zafirlukast are CysLT1 receptor antagonists used as asthma controller therapy — none of these act anywhere on the COX pathway, underscoring that COX and 5-LOX are genuinely independent drug targets sharing only their AA substrate.
PGH2 itself has no biological activity — it is a universal, unstable intermediate whose fate is decided entirely by which terminal synthase a given cell expresses. This single branch point explains why the "same" arachidonic acid cascade produces opposite physiological outcomes in different tissues: vasodilation in the endothelium, vasoconstriction and clotting in platelets, and fever in the hypothalamus, all downstream of an identical precursor molecule.
Four principal terminal synthases compete for the shared PGH2 pool, each restricted to particular cell types:
• mPGES-1 (microsomal PGE synthase-1): induced in parallel with COX-2 during inflammation; converts PGH2 to PGE2, the most abundant prostaglandin in the body and the dominant mediator of inflammatory pain sensitization, vasodilation, and fever • Prostacyclin synthase (PGIS): highly expressed in vascular endothelium; converts PGH2 to PGI2 (prostacyclin), the principal endogenous inhibitor of platelet aggregation and a potent vasodilator • Thromboxane synthase (TXAS): highly expressed in platelets; converts PGH2 to TXA2, a potent vasoconstrictor and platelet activator — platelets, being anucleate, express only COX-1 and cannot upregulate COX-2, which is the pharmacological basis of low-dose aspirin • PGD synthase: expressed in mast cells and CNS; produces PGD2, involved in allergic responses and sleep regulation
The vascular endothelium/platelet axis is the clearest illustration of this logic: endothelial cells (COX-1 and inducible COX-2, both feeding PGIS) continuously release antithrombotic PGI2, while circulating platelets (COX-1 only, feeding TXAS) release prothrombotic TXA2 upon activation. Healthy hemostasis depends on this PGI2/TXA2 balance — pharmacological perturbation of one side without the other (as with COX-2-selective coxibs, which spare platelet TXA2 but reduce endothelial PGI2) tilts the system toward thrombosis.
Eicosanoids act as autacoids — "local hormones" — rather than classical endocrine messengers, and their receptor pharmacology reflects this:
• EP1: Gq-coupled, increases IP3/Ca²⁺; mediates GI smooth muscle contraction and part of pain sensitization • EP2/EP4: Gs-coupled, increase cAMP; mediate vasodilation, bronchodilation, and — critically — peripheral and central sensitization of pain fibers via PKA phosphorylation of TRPV1 and voltage-gated Na⁺ channels • EP3: Gi-coupled, decreases cAMP; mediates fever generation in the hypothalamus (PGE2 crossing a leaky region of the blood-brain barrier near the OVLT) and gastric mucus/bicarbonate secretion • IP (PGI2 receptor): Gs-coupled, vasodilation and inhibition of platelet activation via elevated platelet cAMP • TP (TXA2 receptor): Gq-coupled, vasoconstriction and platelet aggregation via Ca²⁺ mobilization and integrin αIIbβ3 activation
Because TXA2 and PGI2 both have half-lives measured in seconds to a few minutes (non-enzymatic hydrolysis to inactive TXB2 and 6-keto-PGF1α respectively), their signaling radius is limited to tens of micrometers around the cell of origin — this is why local tissue injury produces intensely local inflammation and pain rather than systemic effects, and why circulating eicosanoid levels measured in blood mostly reflect stable downstream metabolites rather than the active species themselves.
Every non-steroidal anti-inflammatory drug in clinical use — from 81 mg aspirin to 200 mg celecoxib — works by the same fundamental act: physically occluding the hydrophobic substrate channel of cyclooxygenase so arachidonic acid cannot reach Tyr385. The differences between drugs in this class are entirely a matter of which isoform they block, how tightly, and for how long — differences with direct, measurable consequences for GI bleeding risk, cardiovascular risk, and analgesic durability.
All clinically used COX inhibitors act at the same site — the hydrophobic AA-binding channel — but with three mechanistically distinct binding modes:
1. Aspirin (acetylsalicylic acid) — irreversible, covalent: aspirin's acetyl group covalently transfers to Ser530 in the COX-1 channel (Ser516 in COX-2), permanently blocking AA access for the remaining lifetime of that enzyme molecule. Because platelets are anucleate and cannot synthesize new COX-1 protein, a single low dose (75–100 mg) irreversibly suppresses platelet TXA2 production for the entire 8–10 day platelet lifespan — this is the pharmacological basis of daily low-dose aspirin for cardiovascular prophylaxis. Nucleated endothelial cells, by contrast, resynthesize COX within hours, so PGI2 production recovers quickly — a favorable asymmetry for the antithrombotic effect.
2. Traditional non-selective NSAIDs (ibuprofen, naproxen, diclofenac, indomethacin) — reversible, competitive: these compete directly with AA for the channel with IC50 values in the low micromolar range for both isoforms roughly equally. Because inhibition is non-covalent and reversible, effect duration tracks plasma drug concentration and requires repeat dosing (ibuprofen t½≈2h; naproxen t½≈14h).
3. COX-2-selective coxibs (celecoxib, etoricoxib; rofecoxib and valdecoxib withdrawn) — reversible, time-dependent, selective: bulky sulfonamide or methylsulfone side chains occupy the extra side pocket unique to COX-2's larger channel (Val523 vs. COX-1's Ile523), producing 300–400× selectivity for COX-2 over COX-1 and largely sparing gastric and platelet COX-1.
The clinical trade-off is now well quantified: non-selective NSAIDs cause serious GI bleeding in roughly 1–2% of chronic users per year (via COX-1 loss of gastroprotective PGE2/PGI2), while COX-2-selective coxibs cut that GI risk by roughly half but shift the vascular PGI2/TXA2 balance toward thrombosis — rofecoxib (Vioxx) was withdrawn in 2004 after the APPROVe trial showed roughly double the risk of myocardial infarction versus placebo, and the VIGOR trial had shown elevated MI risk versus naproxen years earlier. No COX inhibitor is risk-free; drug choice is a quantitative trade between GI and cardiovascular risk tailored to individual patient factors.
NSAID clinical effects separate by required occupancy of the target enzyme, which is why the same drug class treats fever, pain, and inflammation at overlapping but distinct dose ranges:
• Antipyretic/analgesic effect: achieved at partial (~50–80%) COX inhibition, sufficient to blunt hypothalamic PGE2 (fever) and peripheral pain-sensitizing PGE2 (EP2/EP4-mediated nociceptor sensitization) without full blockade • Anti-inflammatory effect: typically requires higher, sustained (>90%) COX-2 inhibition to meaningfully suppress the induced, high-output inflammatory prostaglandin pool — this is why anti-inflammatory NSAID doses (e.g., ibuprofen 600–800 mg) exceed simple analgesic doses (200–400 mg) • Antithrombotic effect: requires only near-complete, sustained COX-1 inhibition in platelets specifically — achieved at very low aspirin doses (75–100 mg) precisely because of irreversible, cumulative inhibition across the platelet's entire circulating lifespan; higher aspirin doses do not improve antithrombotic effect and instead increase GI risk
This dose-separated pharmacology is the reason aspirin is dosed completely differently for cardioprotection (75–100 mg/day) versus analgesia (325–650 mg every 4–6h) — the same molecule, same target, different required occupancy and different dosing kinetics entirely.