💊 Drug-Induced Liver Injury (DILI) Risk Simulator

Drug Absorption & Hepatic First-Pass Metabolism

Every orally dosed drug that reaches the liver passes through hepatocytes before entering systemic circulation. This first-pass exposure is exactly why the liver — despite receiving only a fraction of a drug's eventual body burden — bears a disproportionate share of metabolic and toxicologic risk. Understanding the entry route is the first step in modeling drug-induced liver injury (DILI).

• ~78,000: Acetaminophen-related ED visits/yr (US) (CDC overdose surveillance)
• ~33,000: APAP-related hospitalizations/yr (intentional + unintentional)
• 40–50%: Hepatic first-pass extraction (typical for CYP-cleared drugs)
• ~1,500 mL/min: Portal venous flow to liver (≈75% of total hepatic blood flow)

The portal route and hepatocyte uptake

After intestinal absorption, drug-laden blood is routed directly into the portal vein rather than the systemic circulation, delivering the entire absorbed dose to the liver before dilution into the rest of the body. Hepatocytes lining the sinusoidal capillaries express organic anion and cation transporters (OATP1B1/1B3, OCT1) on their basolateral membrane that actively concentrate drug molecules intracellularly.

This anatomical arrangement is a double-edged sword: it enables the liver to efficiently clear xenobiotics before they reach the brain, heart, or kidneys, but it also concentrates any downstream toxic metabolite formation within a single organ and even within specific zones of the liver lobule.

Zonal heterogeneity of hepatocytes

Hepatocytes are not uniform. The liver lobule is organized into three functional zones along the sinusoid, from the portal triad (zone 1) to the central vein (zone 3):

• Zone 1 (periportal): oxygen-rich, high in gluconeogenic and oxidative enzymes • Zone 2 (mid-zonal): transitional metabolic profile • Zone 3 (centrilobular): oxygen-poor, enriched in CYP2E1 and CYP3A4 — the primary bioactivating enzymes for many hepatotoxic drugs

Because reactive metabolite formation is concentrated in zone 3, most classic DILI presentations (including acetaminophen toxicity) show centrilobular necrosis on histology — a direct histologic signature of where bioactivation occurs.

Dose, exposure, and the toxicologic threshold

At therapeutic doses, the fraction of drug shunted toward a reactive bioactivation pathway is typically small (5–10%), with the majority safely cleared via glucuronidation and sulfation — conjugation pathways that do not generate reactive intermediates. As dose rises, these high-capacity conjugation pathways saturate, forcing a larger relative fraction of the dose through the CYP450 oxidative route. This dose-dependent shift in metabolic routing is the pharmacokinetic basis for why DILI is frequently a threshold phenomenon rather than a linear dose-response — small increases above a critical dose can disproportionately increase reactive metabolite exposure.

CYP450 Bioactivation — Forming the Reactive Electrophile

Bioactivation is the pharmacological paradox at the heart of most idiosyncratic and dose-dependent DILI: the very enzymes that detoxify most xenobiotics occasionally convert an inert parent drug into a short-lived, highly reactive electrophilic metabolite capable of damaging cellular macromolecules within microseconds of formation.

• <1 sec: NAPQI reactive metabolite half-life (extremely short-lived electrophile)
• ~90%: CYP2E1 contribution (toxic APAP dose) (dominant bioactivating isoform)
• 5–10%: Bioactivated fraction (therapeutic dose) (minor oxidative route)
• >15%: Bioactivated fraction (overdose) (conjugation pathways saturate)

CYP2E1 and CYP3A4 — the bioactivating isoforms

CYP2E1 is inducible by ethanol, fasting, and obesity, and sits at the metabolic crossroads for many hepatotoxicants because its active site favors oxidation of small, hydrophilic substrates into unstable intermediates. CYP3A4, the most abundant hepatic CYP by mass, contributes to bioactivation for a broad range of larger, lipophilic drugs including several antiepileptics and macrolide antibiotics.

Both enzymes normally perform productive detoxification — hydroxylation reactions that render most xenobiotics more water-soluble for excretion. Bioactivation occurs when the oxidation reaction instead generates an electrophilic species: a quinone imine, epoxide, or acyl glucuronide capable of reacting with cellular nucleophiles rather than being excreted intact.

Genetic polymorphism and the metabolizer spectrum

CYP450 activity varies substantially across individuals due to genetic polymorphism, classifying patients along a metabolizer spectrum:

• Poor metabolizers: reduced-function alleles slow bioactivation, generally lowering DILI risk for CYP-dependent toxins but increasing risk for drugs whose parent compound itself is toxic • Intermediate metabolizers: partially reduced enzyme activity • Extensive (normal) metabolizers: population-typical enzyme kinetics • Rapid and ultra-rapid metabolizers: gene duplication or induction (e.g., chronic alcohol use inducing CYP2E1) accelerates bioactivation, concentrating reactive metabolite formation into a shorter time window and increasing the odds that detoxification capacity is overwhelmed

This genetic variability is one reason the same nominal drug dose produces wildly different hepatotoxic outcomes across a population — the same 4 g of acetaminophen is inert in one patient and dangerous in another.

Why the metabolite, not the drug, causes injury

A defining feature of bioactivation toxicity is that the parent drug itself is often essentially non-toxic — it is the transient oxidative intermediate that carries chemical reactivity. Because these metabolites typically persist for milliseconds to a few seconds before reacting with the nearest available nucleophile, their damage is confined almost entirely to the hepatocyte in which they were generated, rather than diffusing to affect neighboring cells. This localized reactivity explains the classic zonal (centrilobular) distribution of injury and is central to why bioactivation-driven DILI is a cell-autonomous, dose- and enzyme-activity-dependent process.

Hepatotoxic drug classes and their bioactivation pathways

Glutathione Conjugation & the Limits of Hepatic Defense

Glutathione (GSH) is the hepatocyte's frontline nucleophilic scavenger — a tripeptide present in millimolar concentrations whose sulfhydryl group intercepts electrophilic metabolites before they can react with proteins, lipids, or DNA. Detoxification succeeds as long as the rate of metabolite formation does not outpace the rate of GSH conjugation and regeneration.

• ~10 mmol/kg: Hepatic GSH pool (normal) (liver tissue concentration)
• ~70%: GSH depletion toxicity threshold (below this, injury accelerates sharply)
• hours: GSH resynthesis (glutathione synthetase) (cannot keep pace with acute overdose)
• GSH precursor: N-acetylcysteine (NAC) mechanism (clinical antidote, restores nucleophile supply)

Glutathione S-transferase-mediated conjugation

Reactive electrophiles are conjugated to glutathione both spontaneously and enzymatically via glutathione S-transferases (GSTs), forming a stable thioether adduct that is subsequently processed through the mercapturic acid pathway — sequential cleavage of glutamate and glycine followed by N-acetylation — and excreted in bile or urine as a mercapturate.

This pathway is remarkably efficient at physiological metabolite loads: at therapeutic drug doses, GSH conjugation clears essentially all of the reactive metabolite generated, and no measurable hepatocellular injury occurs. The system is designed with substantial reserve capacity, which is precisely why toxicity is a threshold phenomenon rather than a graded response to every dose.

The depletion curve — from reserve to exhaustion

GSH depletion follows a characteristic curve: initial reactive metabolite exposure is absorbed with little functional consequence because the hepatocyte draws down a substantial reserve pool. As depletion progresses past roughly 70% of baseline, the marginal protective capacity of remaining GSH falls sharply — each additional metabolite molecule is now much more likely to escape conjugation and react with a cellular target instead.

GSH resynthesis via glutathione synthetase and the rate-limiting enzyme glutamate-cysteine ligase is a comparatively slow process, measured in hours — far too slow to keep pace with the metabolite flux generated during an acute overdose, even though it is adequate to replenish the pool over a period of days during recovery.

Therapeutic exploitation — N-acetylcysteine as antidote

The clinical antidote for acetaminophen overdose, N-acetylcysteine (NAC), works by directly replenishing the cysteine substrate needed for hepatic GSH synthesis, and can itself act as an alternative nucleophile that reacts with reactive metabolite directly. NAC is most effective when administered within 8–10 hours of ingestion, before GSH depletion has progressed to the point of irreversible mitochondrial injury — a therapeutic window that is a direct clinical consequence of the depletion kinetics described above.

Covalent Adduct Formation & Mitochondrial Dysfunction

Once glutathione reserves are exhausted, reactive metabolites are no longer efficiently intercepted and instead bind covalently to cysteine residues on cytosolic and mitochondrial proteins. This adduction event is the molecular pivot point between reversible metabolic stress and irreversible hepatocellular injury.

• detectable: APAP-protein (APAP-CYS) adducts (serum biomarker of bioactivation)
• ~4–8 h: Mitochondrial permeability transition onset (post-overdose in severe cases)
• JNK / ASK1: Stress kinase activation (amplifies mitochondrial ROS burst)
• >1.0 nmol/mg: Adduct level associated with injury (protein-bound APAP, serum)

Covalent binding — from reactive metabolite to protein adduct

Unconjugated reactive metabolites react preferentially with nucleophilic cysteine thiols on cellular proteins, forming stable covalent adducts. Mitochondrial proteins are disproportionately affected because a significant fraction of bioactivation occurs in close proximity to mitochondria within the hepatocyte, and because mitochondrial proteins involved in the electron transport chain and ATP synthesis are particularly rich in accessible cysteine residues.

Adduction is not merely a passive biomarker of exposure — for several key mitochondrial proteins, covalent modification directly impairs enzymatic function, degrading the organelle's capacity to maintain membrane potential and produce ATP.

Oxidative stress and the JNK amplification loop

Protein adduction triggers a self-amplifying signaling cascade. Early mitochondrial dysfunction generates reactive oxygen species (ROS), which activate apoptosis signal-regulating kinase 1 (ASK1) and downstream c-Jun N-terminal kinase (JNK). Phosphorylated JNK translocates to mitochondria, where it further impairs electron transport and amplifies ROS production — a feed-forward loop that converts an initially localized chemical insult into progressive, organelle-wide mitochondrial failure.

This amplification explains why hepatocyte injury often continues to worsen for hours after the peak of reactive metabolite formation has already passed: the downstream signaling cascade, not the initiating chemistry, becomes rate-limiting.

Mitochondrial permeability transition — the point of no return

Sustained oxidative stress ultimately triggers the mitochondrial permeability transition (MPT): opening of a high-conductance pore in the inner mitochondrial membrane that collapses the proton gradient, halts ATP synthesis, and allows matrix swelling. Once MPT becomes widespread across a hepatocyte's mitochondrial population, ATP depletion becomes irreversible and the cell is committed to a necrotic (rather than apoptotic) death, since the ATP-dependent apoptotic machinery itself fails under these conditions. This mechanistic transition — from reversible oxidative stress to committed necrosis — is the key inflection point separating recoverable hepatocellular stress from clinically significant liver injury.

Hepatocyte Necrosis & Hy's Law Risk Stratification

The culmination of unchecked bioactivation is centrilobular hepatocyte necrosis, releasing intracellular enzymes into the bloodstream and producing the biochemical signature clinicians use to diagnose DILI: rising transaminases and, in the most severe cases, rising bilirubin. Hy's Law provides the clinical framework for distinguishing benign transaminase elevation from potentially fatal liver failure.

• ALT/AST >3× ULN + bilirubin >2× ULN: Hy's Law criteria (no cholestatic alternative explanation)
• ~10–50%: Hy's Law case mortality / transplant (historically observed rate)
• 1,000–10,000+ U/L: Peak ALT, severe APAP toxicity (vs. ULN ~40 U/L)
• ~46%: US acute liver failure from APAP (leading single cause)

From mitochondrial failure to centrilobular necrosis

As ATP stores collapse following widespread mitochondrial permeability transition, hepatocytes lose the ability to maintain ionic gradients and membrane integrity, undergoing necrotic cell death characterized by cell swelling, organelle breakdown, and rupture of the plasma membrane. Because bioactivation and mitochondrial injury are concentrated in the oxygen-poor, CYP-rich centrilobular (zone 3) region, necrosis classically forms a distinctive band around the central vein — a histologic pattern used diagnostically to implicate a bioactivation-driven mechanism over other causes of liver injury such as viral hepatitis or biliary obstruction.

Necrotic cell rupture releases intracellular contents, including alanine aminotransferase (ALT) and aspartate aminotransferase (AST), into the sinusoidal blood — the basis for using serum transaminases as a sensitive, if non-specific, marker of ongoing hepatocellular injury.

Hy's Law — from biomarker to risk classification

Hyman Zimmerman first observed that isolated transaminase elevation, however dramatic, rarely predicts severe outcomes — the liver has substantial functional reserve, and even markedly elevated ALT often resolves without progression to liver failure. The clinically decisive combination, now known as Hy's Law, requires:

1. ALT or AST elevated more than 3 times the upper limit of normal (ULN) 2. Total bilirubin elevated more than 2 times the ULN 3. No alternative explanation such as biliary obstruction, viral hepatitis, or pre-existing liver disease

When all three criteria are met, historical cohort data associate this presentation with a case fatality or transplant rate of roughly 10–50% — because concurrent hyperbilirubinemia indicates that hepatocyte loss has become extensive enough to compromise the liver's bilirubin-conjugating and excretory capacity, not merely its enzyme-release signaling.

Idiosyncratic versus intrinsic DILI

DILI is broadly classified into two mechanistic categories that this simulation bridges:

• Intrinsic (dose-dependent) DILI: predictable, reproducible in animal models, and directly proportional to dose and bioactivation capacity — acetaminophen is the archetypal example, and the mechanism modeled throughout this simulation • Idiosyncratic DILI: rare, not clearly dose-dependent, and typically involves an adaptive immune response to drug-modified (haptenized) proteins, often with a delayed onset of weeks and a strong association with specific HLA alleles

Many real-world hepatotoxic drugs (isoniazid, amoxicillin-clavulanate) sit at the intersection: bioactivation generates the initial protein adducts, but the resulting immune sensitization — rather than direct chemical toxicity alone — determines whether clinically significant injury ultimately develops.