Cytochrome P450 Drug Metabolism

How a P450 enzyme works

Cytochrome P450s are a superfamily of heme-thiolate enzymes embedded in the smooth endoplasmic reticulum of liver hepatocytes (and to a lesser extent, intestinal enterocytes, kidney, lung, and brain). The name comes from the absorption peak at 450 nanometers when their reduced iron binds carbon monoxide — first observed in the 1960s before anyone knew what these proteins did.

Each enzyme is a globular protein roughly 50 kDa with a substrate channel that ends at the active site. At the active site sits a heme b group — an iron atom held in a porphyrin ring, with a fifth coordination position bound to a cysteine sulfur from the protein. That thiolate ligand is what makes P450 chemistry possible.

The catalytic cycle:

1. Drug binding. The substrate slips into the active site, displaces water from the iron, and twists the spin state to high-spin. This raises the iron's reduction potential.
2. First electron transfer. Cytochrome P450 reductase (CPR), using NADPH, hands one electron to the iron, reducing it from Fe(III) to Fe(II).
3. Oxygen binding. Molecular O₂ binds the iron, forming Fe(III)-O-O⁻.
4. Second electron + proton. A second electron arrives (sometimes via cytochrome b5), then protonation forms the iron-hydroperoxo species.
5. O-O bond cleavage. Another proton, then heterolytic cleavage of the O-O bond, releases water and produces Compound I — formally Fe(IV)=O with a porphyrin radical cation.
6. Hydrogen abstraction + rebound. Compound I rips a hydrogen atom off the drug, producing a substrate radical and Fe(IV)-OH. The hydroxyl then rebounds onto the substrate radical. The drug now carries a new OH group.
7. Product release. The metabolite leaves; water re-coordinates the iron; the enzyme is reset.

Net reaction: drug + O₂ + NADPH + H⁺ → drug-OH + H₂O + NADP⁺. The whole cycle takes milliseconds and produces a more polar metabolite — ready for phase II conjugation (glucuronidation, sulfation, glutathione) or direct renal excretion.

The clinically important CYPs

• CYP3A4 (with closely related CYP3A5) is the dominant drug-metabolizing enzyme — it handles approximately 50% of all marketed drugs. Substrates include most statins, midazolam, cyclosporine, tacrolimus, many opioids, calcium channel blockers, and protease inhibitors. CYP3A4 has an unusually large, flexible active site that accommodates a vast variety of substrates.
• CYP2D6 metabolizes about 20% of drugs and is the most clinically important polymorphic CYP. Substrates: codeine (→ morphine), tramadol, tamoxifen (→ endoxifen), many SSRIs (fluoxetine, paroxetine), tricyclic antidepressants, antipsychotics (risperidone, haloperidol), beta-blockers (metoprolol, carvedilol), atomoxetine.
• CYP2C9. Warfarin (S-isomer), phenytoin, many NSAIDs, sulfonylureas. Combined with VKORC1 genotyping, CYP2C9 variants explain roughly a third of warfarin dose variance.
• CYP2C19. Clopidogrel activation, omeprazole and other PPIs, citalopram/escitalopram, voriconazole. Common loss-of-function alleles (*2, *3) are very prevalent in East Asian populations.
• CYP1A2. Caffeine, theophylline, clozapine, olanzapine, tizanidine. Induced by smoking; smokers need higher clozapine doses, and quitting risks toxicity.
• CYP2E1. Alcohol (minor pathway, induced by chronic use), acetaminophen activation to NAPQI (drives hepatotoxicity at overdose).
• CYP2B6. Efavirenz, methadone, bupropion, ketamine, propofol.

The codeine story — pharmacogenomics in real life

Codeine itself is a weak opioid. About 10% of an oral dose is O-demethylated by CYP2D6 into morphine — the molecule that actually drives analgesia. People with no functional CYP2D6 ("poor metabolizers", roughly 7% of white Europeans) get almost no morphine from codeine and complain that it does nothing. People with multiple functional copies of CYP2D6 — gene duplications producing ultrarapid metabolizers — convert codeine to morphine far faster than expected. Their blood morphine levels can spike to several times the predicted concentration at standard doses.

Ultrarapid metabolizer frequencies vary dramatically: about 1–3% in white Europeans, but up to 29% in some Ethiopian populations, 21% in Saudi Arabians, and 11% in Algerians. In 2007, a breastfed neonate died from morphine toxicity after the mother — an ultrarapid metabolizer — took codeine post-Caesarean. In 2009, three pediatric tonsillectomy deaths in ultrarapid metabolizers prompted the FDA to issue a black-box warning. In 2017 the FDA contraindicated codeine in children under 12 and in breastfeeding women.

The same story plays out across pharmacogenomics: CPIC (Clinical Pharmacogenetics Implementation Consortium) now issues genotype-guided dosing guidelines for over 100 drug-gene pairs. Major academic centers routinely panel test CYP2D6, CYP2C19, CYP2C9, CYP3A5, TPMT, DPYD, and HLA-B*57:01 at the start of complex therapy. The cost is now under $200 — far cheaper than the adverse drug reactions it prevents.

Drug-drug interactions through CYPs

Inhibition is fast — within hours — and raises levels of co-administered substrates. The classic dangerous pairs:

• Strong CYP3A4 inhibitors (clarithromycin, erythromycin, ketoconazole, itraconazole, ritonavir, grapefruit juice for intestinal CYP3A4) + simvastatin/lovastatin → rhabdomyolysis.
• Fluoxetine, paroxetine (strong CYP2D6 inhibitors) + tamoxifen → reduced endoxifen, reduced antitumor effect in breast cancer.
• Fluconazole + warfarin (CYP2C9) → bleeding.
• Omeprazole + clopidogrel (CYP2C19) → reduced antiplatelet effect (controversial in practice).

Induction is slow — days to weeks, because it requires transcription of new enzyme — and lowers levels of co-administered substrates. The classic dangerous pairs:

• Rifampin + oral contraceptives → contraceptive failure.
• Carbamazepine, phenytoin, phenobarbital + warfarin, immunosuppressants → subtherapeutic levels.
• St. John's wort + cyclosporine, indinavir → transplant rejection, HIV breakthrough.

The FDA requires every new drug application to characterize CYP inhibition, induction, and metabolism. Drugs that depend on a single CYP are flagged. Drugs that strongly inhibit or induce CYPs get explicit labels.

How CYPs were discovered

In 1958, Martin Klingenberg and David Garfinkel independently noticed a strange absorption peak at 450 nm in liver microsomes treated with carbon monoxide. Ronald Estabrook later identified the "P-450 pigment" as a hemoprotein, in 1962. Over the 1960s and 1970s, Tsuneo Omura, Ryo Sato, and Minor "Jud" Coon worked out that this was an oxygenase, used NADPH and O₂, and was responsible for hepatic xenobiotic metabolism.

The first crystal structure of a P450 (P450cam from Pseudomonas putida) came from Tom Poulos in 1985, revealing the heme-thiolate active site. Human P450 crystal structures followed in the 2000s — CYP2C9 by Eric Johnson's group, CYP3A4 shortly after. The 1990s sequencing era catalogued the human CYPome: 57 genes, 18 families, of which roughly 10–15 are responsible for the majority of drug metabolism.

Pharmacogenomics built on top of this. CYP2D6 polymorphisms were first recognized through clinical observations of variable debrisoquine metabolism by Robert Smith and Bob Mahgoub in 1977. The molecular basis — null alleles, gene duplications — was worked out through the 1990s. Today's genotype panels descend directly from this work.

Common misconceptions

• "Metabolism always inactivates the drug." Not for prodrugs. Codeine, tramadol, tamoxifen, clopidogrel, and many others must be metabolized to their active form. CYP loss-of-function reduces efficacy of these drugs.
• "Grapefruit juice affects all drugs the same way." Only CYP3A4 substrates (especially those with poor oral bioavailability like felodipine, simvastatin). The bergamottin and furanocoumarins selectively inhibit intestinal CYP3A4, not the hepatic enzyme.
• "P450 enzymes are unique to humans." They are ancient — present in bacteria, fungi, plants, and all animals. CYP-mediated detoxification is an old evolutionary arms race against plant secondary metabolites.
• "Phase II is downstream of phase I." Not always. Many drugs go directly to phase II (e.g., morphine glucuronidation, lorazepam glucuronidation). Phase II conjugation enzymes (UGT, SULT, GST, NAT) also have polymorphisms with clinical impact (UGT1A1*28 → irinotecan toxicity).
• "Smokers handle drugs the same as nonsmokers." Tobacco smoke induces CYP1A2 by a factor of ~1.5. Clozapine, theophylline, olanzapine, and caffeine all show meaningful differences. Quitting smoking can cause toxicity at unchanged doses.
• "Liver failure means all drug metabolism stops." The liver has substantial reserve. CYP-mediated phase I metabolism declines roughly in proportion to Child-Pugh score, but UGT-mediated phase II is relatively preserved until late cirrhosis.

Clinically important CYP enzymes

Enzyme	% of drugs metabolized	Key substrates	Key inhibitors	Key inducers	Polymorphism impact
CYP3A4	~50%	Statins, midazolam, tacrolimus, opioids, CCBs	Clarithromycin, ketoconazole, ritonavir, grapefruit	Rifampin, carbamazepine, phenytoin, St. John's wort	Low — limited PM phenotype
CYP2D6	~20%	Codeine, tamoxifen, SSRIs, β-blockers, antipsychotics	Fluoxetine, paroxetine, bupropion, quinidine	Not significantly inducible	Very high — PM, IM, EM, UM
CYP2C9	~10%	Warfarin (S), phenytoin, NSAIDs, sulfonylureas	Fluconazole, amiodarone, sulfamethoxazole	Rifampin	High — *2, *3 alleles
CYP2C19	~10%	Clopidogrel, PPIs, citalopram, voriconazole	Fluvoxamine, omeprazole	Rifampin	High — *2, *3 common in East Asians
CYP1A2	~5%	Caffeine, theophylline, clozapine	Ciprofloxacin, fluvoxamine	Smoking, charbroiled meat	Moderate
CYP2E1	~5%	Alcohol, acetaminophen (toxicity pathway)	Disulfiram	Ethanol, isoniazid	Modest