Endpoint explanations

The figure follows a drug from its dosage form, through absorption, into the systemic circulation, and shows how that exposure is quantified. On the plasma concentration–time curve, the peak concentration (Cmax) and the time at which it occurs (Tmax) index the rate of absorption, while the area under the curve (AUC) measures the extent of exposure — the total amount of active moiety that actually reaches the circulation.

Panel 2 defines absolute bioavailability. Because an intravenous dose enters the bloodstream completely, it is taken as the 100% reference, and any non-intravenous route (oral, subcutaneous, …) is compared against it after correcting for dose — so F is the dose-normalised ratio of the two AUCs. A value below 100% reflects drug that never reaches the circulation intact, through incomplete absorption across the gut wall and/or presystemic (first-pass) loss in the gut and liver.

Panel 3 contrasts this with relative bioavailability, which compares two non-intravenous formulations or routes with each other — for example a new tablet against a reference oral solution — rather than against IV. The same dose-corrected AUC ratio is used, so the result expresses how the test formulation performs against the chosen reference, not against the theoretical maximum.

The figure starts from the definition: AUC is the area enclosed beneath the plasma (or serum/whole-blood) concentration–time curve, written formally as the integral of concentration over time, ∫ C(t) dt. Geometrically it sums every moment of exposure, so it captures the total amount of drug the body is exposed to after a dose — not just the peak.

Panel 2 distinguishes the two forms in which AUC is reported. AUC0–t is measured directly from time zero to the last sample with a quantifiable concentration (tlast), typically by trapezoidal integration of the observed points. Because sampling stops while drug is still present, the remaining tail from tlast out to infinity is estimated from the terminal decline and added on: AUC0–∞ = AUC0–t + extrapolated tail. The dashed segment in the figure is exactly this extrapolated portion.

Panel 3 explains why AUC matters. Reported in units of concentration × time (e.g. ng·h/mL), it is the standard index of total systemic exposure, and together with the peak concentration Cmax it forms the pair of parameters on which bioavailability and bioequivalence assessments rest.

The figure reads Cmax straight off the concentration–time profile: it is the single highest drug concentration reached after dosing, marked at the apex of the curve, with the time at which it occurs being tmax. Because exposure is sampled at discrete time points, Cmax is the largest of those measured concentrations — an observed value, not a number interpolated or fitted between samples.

Panel 2 makes the determination explicit: Cmax = max[C(ti)] over the sampled times ti. The practical consequence, shown by the sampling dots, is that adequate sampling around the expected peak is needed — if the schedule is too sparse near the maximum, the observed Cmax (and tmax) can under-capture the true peak.

Panel 3 explains why it is reported. Cmax indexes peak systemic exposure and, as a hybrid parameter, reflects both the rate and the extent of absorption. Reported in concentration units (e.g. ng/mL), it is paired with AUC (total exposure) and tmax (timing) and, together with AUC, anchors bioavailability and bioequivalence assessment.

The figure locates Tmax on the concentration–time profile: it is the time after dosing at which the highest observed concentration (Cmax) is reached, read straight off the x-axis at the curve's apex. Because concentrations are only measured at discrete sampling times, Tmax is taken as the sampling time ti at which the observed Cmax occurs (Tmax = ti where C(ti) = Cmax) — an observed value, not one interpolated between samples. A schedule that is too sparse near the peak therefore blurs Tmax, which is why adequate sampling around the expected peak is needed.

Tmax serves as a simple index of the rate of absorption: an earlier Tmax generally signals that the drug appears in the systemic circulation faster, while a later Tmax reflects slower absorption (or, for modified-release products, deliberately delayed release).

Tmax is reported in units of time (hours or minutes) and is presented alongside Cmax and AUC. Unlike those two, however, it is generally not a formal bioequivalence acceptance criterion — regulators (e.g. EMA) do not require a statistical evaluation of Tmax, and compare it descriptively only when the rate of onset is clinically important.

The figure builds the volume of distribution up from its definition: Vd is the proportionality constant relating the total amount of drug in the body to the concentration measured in plasma. The dashed container and the “no-entry” marker stress that this is an apparent, theoretical volume — a scaling factor, not a real anatomical space the drug fills.

Panels 2 and 3 show why the same dose can imply very different volumes. When a drug stays largely within plasma and blood (limited tissue distribution), the measured plasma concentration is high, so the calculated volume is small — a low Vd. When a drug partitions extensively into tissues or binds tissue components, little remains in plasma; the plasma concentration is low and the apparent volume becomes large — a high Vd.

Because the volume is defined relative to plasma concentration rather than any physical space, highly tissue-bound compounds can have a Vd that exceeds total body water, sometimes by orders of magnitude. The value is reported in several forms — Vd, steady-state Vss, or terminal-phase Vz — in litres or normalised to body weight (L/kg).

The figure builds protein binding up from the reversible reaction at its core: a drug molecule and a plasma protein associate and dissociate continuously (drug + protein ⇌ protein–drug complex), so at any instant the drug exists as a bound pool and a free pool in dynamic equilibrium. The two complementary ways of quantifying this are shown as formulas — percent bound (100 × bound/total) and the fraction unbound, fu = Cunbound/Ctotal = free/total.

The central panel shows why the split matters. Only the unbound (free) drug is small and mobile enough to cross the capillary endothelium into tissue, engage its pharmacological target, and exert an effect; the protein-bound fraction is confined to plasma and acts as a circulating reservoir. As free drug leaves the plasma, the equilibrium shifts and bound drug is released to replenish it. The two carrier proteins differ in preference — albumin chiefly binds acidic and neutral drugs, while α1-acid glycoprotein chiefly binds basic drugs.

The consequence for exposure is summarised on the right: a highly bound compound leaves little free drug, so fu is low; a weakly bound compound leaves more free drug, so fu is high. Because it is fu that drives tissue distribution and target engagement, it is the unbound fraction — not the total concentration — that governs pharmacological exposure. Binding is reported as percent bound and/or fu, and is sensitive to species, matrix, and assay method (equilibrium dialysis vs ultrafiltration).

The left panel fixes the anatomy: the barrier itself is the brain capillary endothelium, whose cells are sealed by tight junctions, so a compound moving from blood/plasma into the brain interstitial fluid must cross those cells rather than slip between them. The central message — echoed in the definition — is that this crossing is not captured by one universal number; different assays quantify different aspects of brain entry and exposure.

The middle panel lays out the five parameters that are actually reported, and what each one measures. logBB is the log of the total brain-to-plasma concentration ratio, so it reflects total (binding-dependent) drug. Kp,uu,brain is the unbound brain-to-plasma partition coefficient (Cu,brain / Cu,plasma) — the pharmacologically meaningful free-drug measure, because it is the unbound concentration that drives a CNS effect. Papp/Pe is an in vitro permeability (cm/s); %ID/g is in vivo brain uptake (percent injected dose per gram tissue); and the Kin/PS product is the unidirectional influx clearance (permeability–surface-area product).

The right panel is the interpretation key: the parameters are placed by whether they are measured in vitro or in vivo and whether they capture total or unbound drug. The essential cautions follow — these metrics are distinct and must not be used interchangeably; total-drug metrics such as logBB do not equal free-drug exposure; Kp,uu,brain best reflects pharmacologically relevant unbound brain exposure; and one should always state the exact parameter, its units, and the assay context, because “BBB penetration” is not a single number.

The left panel defines the measurement. Permeability is the rate at which a compound crosses a membrane, captured by the apparent permeability coefficient Papp = (dQ/dt)/(A × C0) — the transport rate dQ/dt normalised to the membrane area A and the initial donor concentration C0, giving units of cm/s. The compound moves from the apical (donor) side, across a tight-junction-sealed cell monolayer (Caco-2 or MDCK) or a cell-free artificial membrane (PAMPA), into the basolateral (receiver) side.

Permeability is direction-resolved, so the assay is run both ways and reported separately: apical→basolateral (Papp A→B, the absorptive direction) and basolateral→apical (Papp B→A, the secretory direction). Cell-based monolayers (Caco-2, MDCK) carry transporters and tight junctions, whereas PAMPA is cell-free and reflects passive permeability only — one reason magnitudes are not comparable across assay systems or labs.

The two directions are combined into the efflux ratio = Papp(B→A) / Papp(A→B). A ratio near 1 indicates largely passive, symmetric transport; an elevated ratio flags an active efflux transporter — classically P-glycoprotein (P-gp), which sits in the apical membrane and pumps drug back, raising B→A while lowering A→B. A higher Papp simply means faster membrane crossing; because values are assay- and direction-dependent, the assay model and transport direction should always be reported.

The figure frames serum/plasma stability as a resistance measurement: a peptide is incubated in serum or plasma and its intact form is gradually cleaved by endogenous peptidases and proteases at body temperature (~37 °C) into degraded fragments. What is quantified is how much intact peptide survives — the percentage remaining over time and, from that decline, an apparent degradation half-life (t1/2).

The middle panel shows the standard workflow: a peptide stock is spiked into serum/plasma at 37 °C, aliquots are quenched at several time points to stop enzymatic activity, and the intact peptide is quantified (typically by LC–MS/MS). Plotting percent intact against time gives a decay curve, and the time at which it falls to 50% is the apparent half-life. Because the assay reports loss of the parent peptide, the analytical method must be specific for the intact molecule rather than its fragments.

This in vitro readout is a primary predictor of a peptide's circulating metabolic stability in vivo, since enzymatic (peptidase/protease) degradation is a main driver of peptide clearance; as a quantitative, reproducible number it lets candidates be rank-ordered and guides stability engineering and selection. Values are condition-dependent (species, matrix, dilution, temperature, timepoint), and serum and plasma are not interchangeable because their enzyme complements differ — so the exact conditions, and whether the value is % remaining (with its timepoint) or a half-life, should always be stated.

The figure defines protease stability as resistance to cleavage by one or more defined, purified proteases (e.g. trypsin, chymotrypsin, pepsin, pronase, carboxy-/aminopeptidases) under controlled buffer, pH, temperature, and enzyme-concentration conditions. The intact peptide is converted by those proteases into cleaved fragments, and what is measured is the percentage of intact peptide surviving over time and/or an apparent degradation half-life (t1/2).

The middle panel shows the workflow: the peptide is combined with the protease solution, incubated under defined conditions (commonly ~37 °C), aliquots are quenched at several time points to stop the reaction, and the remaining intact peptide is quantified by LC–MS or RP–HPLC. Plotting percent intact against time gives a decay curve whose 50% crossing is the apparent half-life.

Unlike whole serum/plasma, this assay isolates a defined enzyme, so it is mechanistic: it identifies which protease a peptide is vulnerable to and guides engineering to fix it (residue substitution, terminal modification, cyclization). Crucially, cleavage is enzyme-specific — trypsin cuts after Lys/Arg, chymotrypsin after Phe/Tyr/Trp — so a single-protease result does not generalize, and the protease(s), enzyme:substrate ratio, buffer/pH, temperature, and timepoint must all be reported alongside the value.

The figure starts from the operating definition: clearance is the volume of plasma (or blood) completely cleared of drug per unit time, with CL = (rate of elimination)/Cp. It is a volume-per-time measure of how efficiently the body removes the compound — a higher CL means more rapid elimination — not a physical volume.

The middle panel shows how it is obtained. Systemically, the dose enters the body, is distributed and eliminated, and the drug leaves via urine, bile/feces, expired air, etc. In practice systemic clearance is most often computed from intravenous data as CL = DoseIV / AUC0–∞, read off a concentration–time profile (a single-exponential straight line on the semi-log plot of an IV bolus). Organ-specific clearances follow the same form — renal clearance is the urinary excretion rate of unchanged drug divided by Cp, biliary clearance the biliary excretion rate divided by Cp — and total clearance is their sum.

Clearance matters because it is the primary determinant of steady-state exposure: for a given dosing rate R0, Css,avg = R0 / CL, so a lower clearance gives higher exposure (all else equal). It integrates hepatic (metabolism) and renal/biliary (excretion) contributions into one number, and is reported in standardized units (mL/min, L/h) — often normalized to body weight (mL/min/kg) — so values can be compared across drugs, species, and studies. Note that clearance derived from oral data is an apparent oral clearance (CL/F), confounded by bioavailability.

The figure starts from the single idea that unites the endpoint: half-life (t1/2) is the time for the amount or concentration of a compound to fall to one-half of its value. Reading down from the 50% level on the decay curve gives t1/2, and because the decline is exponential, that same halving time recurs for each successive 50% drop.

Panel 2 stresses that one name covers two distinct quantities. (A) An in vitro stability/degradation half-life is the time for half the peptide to be degraded (or for 50% to remain intact) in a defined matrix or protease system — serum, plasma, or a purified protease assay. (B) An in vivo pharmacokinetic elimination half-life is the time for the plasma concentration to fall by 50% during elimination; it is a derived parameter governed by both volume of distribution and clearance, t1/2 = 0.693 × Vd / CL.

These two are not interchangeable, and in ADMETatlas this endpoint is predominantly the former — most records are in vitro/ex vivo proteolytic stability half-lives — so the assay or matrix context must be read to know which meaning is intended. Report the matrix/protease/species/method for in vitro records, or the PK context for in vivo records, and remember that the PK half-life is a dependent parameter of Vd and CL, not an independent measure of elimination.

This endpoint is the potency-metric form of cytotoxicity — the top-left quadrant of the figure. Cultured mammalian cells are exposed to a serial dilution series of the peptide, viability is read by an assay such as MTT, XTT, resazurin, or LDH, and the resulting dose–response curve (viability as a percentage of an untreated control, falling as concentration rises) is fitted to extract a threshold concentration.

That threshold is most often IC50 / CC50 — the concentration reducing viability of the test cell line to 50% of control. It answers “what concentration produces a defined cytotoxic effect,” so it is meaningful only together with the cell line, exposure time, and assay chemistry (the same peptide gives different IC50/CC50 across them), and a curve-fitted threshold is distinct from the single raw % readout of the percent-observation endpoint.

This endpoint is the percent-observation form of cytotoxicity — the top-right quadrant of the figure. Rather than a full dilution series, the peptide is tested at a single stated concentration and the effect is read once: in the figure, 70% of the cells remain viable and 30% are killed.

Because % viability and % cytotoxicity are complementary (% viability ≈ 100% − % cytotoxicity), confirm which one is reported. It answers “how much effect at this concentration,” not “what concentration reaches 50%,” so it is interpretable only with its test concentration, cell line, and exposure time — and a single point cannot be converted into an IC50/CC50.

This endpoint is the potency-metric form of hemolysis — the bottom-left quadrant of the figure. Red blood cells are exposed to a serial dilution series of the peptide; lysis releases hemoglobin, and the dose–response curve (percent hemolysis rising with concentration) is fitted to extract a threshold concentration.

The usual threshold is HC50, the concentration that lyses 50% of the RBCs; a related metric is the minimum hemolytic concentration (MHC), which has no universal cutoff (published as 5%, 10%, or just-detectable lysis), so values compare only when the underlying threshold matches. Results also depend on the RBC source/species (human vs sheep/horse/rabbit) and the assay conditions.

This endpoint is the percent-observation form of hemolysis — the bottom-right quadrant of the figure. The peptide is tested at a single stated concentration: RBCs are incubated with it, the sample is centrifuged, and the hemoglobin released into the supernatant is read by absorbance (≈540 nm).

Percent hemolysis is normalised between two controls run with every assay — 0% (buffer, negative) and 100% (full lysis with Triton/water, positive) — i.e. % hemolysis = (A − A0)/(A_total − A0) × 100 (45% in the figure). As a single dose-point it is interpretable only with its concentration, RBC source/species, and normalization, and cannot be converted to HC50/MHC.

The figure frames lipophilicity as how a compound distributes between two immiscible phases — classically n-octanol (organic) and water (aqueous) — with higher values meaning greater hydrophobicity (a stronger preference for octanol). Two related but distinct coefficients are reported, and the figure separates them deliberately.

logP (Panel 2) is the IUPAC partition coefficient: the log10 ratio of a single neutral species between octanol and water, logP = log10([neutral]octanol / [neutral]water). Because it is defined for the neutral form only, it is pH-independent and reflects the compound's intrinsic lipophilicity.

logD (Panel 3) is the distribution coefficient: the log10 ratio of all species — neutral plus ionized — at a stated pH, logD(pH) = log10([all species]octanol / [all species]water). Since ionized forms partition preferentially into water, logD is pH-dependent and the measurement pH must always be reported (e.g. logD7.4). For non-ionizable compounds logD ≈ logP, but for ionizable compounds the two can differ substantially — and partition systems (octanol/water, alkane/water, RP-HPLC retention) measure distinct phenomena and are never merged.