Why Binding and Partition Inputs Must Be Checked Together
Plasma binding, blood-to-plasma concentration ratios, tissue binding, and tissue partition coefficients describe related but different features of drug distribution. A PBPK model can contain individually plausible values that nevertheless do not form a coherent set. For example, a tissue-to-plasma total concentration ratio may imply a large tissue-to-plasma unbound concentration ratio when combined with the measured fractions unbound. That implication might reflect real transport or other biology, but it could also arise from a mismatched matrix, concentration basis, or assay condition.
The purpose of a consistency check is not to force every tissue to have the same partition coefficient, nor to require that all unbound concentrations be equal in every circumstance. It is to make explicit what a set of inputs implies under stated assumptions, then decide whether the implication is biologically credible and supported by the data.
This relationship follows from \(C_{u,t}=f_{u,t}C_t\) and \(C_{u,p}=f_{u,p}C_p\). It is an accounting identity for the stated concentration and binding definitions; interpreting \(K_{p,uu,t:p}\) as approximately 1 requires an additional biological assumption, such as passive equilibration without meaningful active transport or persistent disequilibrium. Binding inputs alone do not establish that assumption.
Deriving a Useful Partition Check
Rearranging the definition gives a direct check on what the supplied inputs imply about unbound distribution:
If passive equilibrium is a reasonable working expectation for a tissue, then a value far from 1 calls for investigation. It is not, by itself, proof that an input is wrong: transport, pH gradients, sequestration, time-dependent distribution, and measurement limitations can all matter. Conversely, if a model assumes unbound equilibration but its inputs imply a large departure from it, that assumption and the inputs cannot both be accepted without explanation.
The same relationship can be used in the other direction. If \(K_{p,uu,t:p}=1\) is a justified approximation, then \(K_{p,t:p}=f_{u,t}/f_{u,p}\). This is a diagnostic expectation, not a universal equation for predicting tissue partitioning. Tissue composition, transport, and other processes may cause observed partitioning to differ.
Worked Example: Interpreting a Tissue Partition Coefficient
Suppose a hypothetical data set gives \(f_{u,p}=0.020\), \(f_{u,t}=0.0050\), and \(K_{p,t:p}=4.0\). Assume the measurements refer to the same species and relevant concentration range. Calculate the implied unbound ratio and assess the result.
Using the rearranged relationship:
The input set implies an unbound tissue concentration 16 times the unbound plasma concentration under the defined equilibrium comparison. If this tissue is expected to show passive equilibration without substantial transport or other gradients, that is a strong inconsistency to investigate. It does not establish which input is wrong, and it does not rule out a real biological explanation.
For comparison, if unbound equilibration were assumed, the expected total partition coefficient would be \(f_{u,t}/f_{u,p}=0.0050/0.020=0.25\), not 4.0. The gap between 0.25 and the supplied 4.0 makes the assumption behind a simple passive-equilibrium interpretation especially important to review.
Checking Blood-to-Plasma Inputs
The blood-to-plasma ratio \(R_b\) is defined as total whole-blood concentration divided by total plasma concentration. It is not interchangeable with a tissue-to-plasma partition coefficient. With hematocrit \(H\), and assuming blood consists of plasma and red-cell volumes represented by those fractions, the ratio can be written as:
Here \(K_{RBC:p}=C_{RBC}/C_p\) is the red-cell-to-plasma concentration ratio. This relation provides a basic check: the red-cell ratio implied by the reported \(R_b\) and hematocrit is \(K_{RBC:p}=[R_b-(1-H)]/H\), for \(H>0\). A negative implied ratio is incompatible with nonnegative concentrations under these definitions. A positive value is not automatically plausible; it should be compared with the drug's red-cell partitioning evidence and the conditions of measurement.
If unbound drug concentration is equivalent in plasma and blood under the conditions being considered, the blood fraction unbound is \(f_{u,b}=f_{u,p}/R_b\). This conversion depends on that unbound-concentration assumption and consistent concentration definitions. A reported \(f_{u,b}\) should therefore not be combined mechanically with \(f_{u,p}\) and \(R_b\) when the assays use different matrices, conditions, or definitions.
Worked Example: Auditing a Blood-to-Plasma Ratio
Consider hypothetical values \(H=0.45\), \(R_b=1.30\), and \(f_{u,p}=0.020\). Calculate the implied red-cell-to-plasma ratio and, under the stated unbound-equivalence assumption, the blood fraction unbound.
First, the plasma contribution to whole-blood concentration relative to plasma concentration is \(1-H=0.55\). Therefore:
The reported values imply a red-cell concentration about 1.667 times the plasma concentration. This is positive and algebraically consistent with the mixture relationship. Whether it is biologically credible still depends on the drug and the measurement conditions.
With equivalent unbound concentrations in plasma and blood, the implied blood fraction unbound is:
If a separate hypothetical assay reported \(f_{u,b}=0.030\), the three values would not agree under the stated assumption: \(0.030\) differs from \(0.01538\). Before changing a model input, check whether the assays used comparable species, matrices, concentrations, and definitions, and whether unbound concentrations can reasonably be treated as equivalent.
Using Tissue Binding Measurements Correctly
Tissue fraction unbound is often estimated using a tissue homogenate. Homogenization dilutes endogenous binding sites and can make the measured homogenate fraction unbound higher than the fraction in intact tissue. If \(f_{u,hom}\) is measured in homogenate and \(D\) is the dilution factor, the standard correction is:
This calculation assumes the dilution is correctly characterized and that the binding behavior represented by the assay can be adjusted using this dilution model. The corrected value should be used in a partition check only when it refers to the same tissue, drug, and relevant conditions as the tissue partition input. An uncorrected homogenate value can substantially distort the implied \(K_{p,uu,t:p}\).
Worked Example: Correcting Tissue Binding Before Comparing Inputs
Suppose a hypothetical tissue assay reports \(f_{u,hom}=0.080\) at a dilution factor of \(D=5\). A matched plasma assay gives \(f_{u,p}=0.040\), and the supplied tissue-to-plasma partition coefficient is \(K_{p,t:p}=2.0\). Correct the tissue value, then calculate the implied unbound ratio.
The correction is:
Using this corrected tissue fraction unbound:
The input set implies an unbound tissue-to-plasma ratio of about 4.681. Under a justified passive-equilibration expectation, this merits investigation. If the uncorrected value \(0.080\) were used instead, the calculated ratio would be \(2.0(0.040/0.080)=1.0\). That apparent agreement would result from using the diluted homogenate value as if it were the intact-tissue value, so the correction matters to the interpretation.
A Practical Audit and Its Failure Modes
An effective audit proceeds from definitions to comparisons rather than treating each parameter as an independent number. First establish whether each partition ratio uses plasma or whole blood as its reference. Then check whether fractions unbound are reported for the matching matrices and whether concentration-dependent binding measurements correspond to the concentration range relevant to the model. For tissue assays, verify the dilution correction and identify whether the value is an estimate for intact tissue or a direct homogenate result.
Record whether each ratio is tissue-to-plasma, tissue-to-blood, blood-to-plasma, or red-cell-to-plasma. Confirm compatible concentration units and matrices.
Fractions unbound should be positive and no greater than 1 under the usual definition. Use hematocrit and \(R_b\) to calculate the implied red-cell ratio, and use \(K_{p,t:p}\), \(f_{u,p}\), and \(f_{u,t}\) to calculate \(K_{p,uu,t:p}\).
Ask whether the implied unbound ratio agrees with the model's distribution assumptions. A discrepancy calls for investigation of data, conditions, and biology, not an automatic adjustment to force agreement.
A useful alternative is to estimate tissue partitioning directly from in vivo data rather than infer it from binding values. Direct estimates can be closer to the modeled quantity but may still be affected by sampling time, disequilibrium, and measurement error. A mechanistic estimate based on tissue composition and binding can provide another comparison, but it relies on its own assumptions. These approaches are complementary checks, not interchangeable measurements of exactly the same thing.
Common failure modes include using a blood-based partition coefficient with plasma-based binding values as if the reference concentrations matched; comparing assays from different species or concentration ranges; applying no homogenate dilution correction; and treating \(K_{p,uu,t:p}=1\) as a universal law. If a fraction unbound is near an assay's quantification limit, ratios involving it can be highly sensitive to small measurement changes. Report the underlying values and their conditions, and treat a striking ratio as a prompt for review rather than as a precise mechanistic conclusion.
Check Your Understanding
Work through these checks using the relationships in the tutorial.
- For \(f_{u,p}=0.10\), \(f_{u,t}=0.020\), and \(K_{p,t:p}=0.40\), calculate \(K_{p,uu,t:p}\). What does the value say under a passive-equilibration assumption?
- With \(H=0.40\) and \(R_b=0.70\), calculate the implied \(K_{RBC:p}\). Is it compatible with nonnegative red-cell concentration under the mixture relationship?
- Why should an uncorrected homogenate fraction unbound not automatically be substituted for \(f_{u,t}\)?
- A tissue-to-blood partition coefficient and a tissue-to-plasma fraction unbound are available. What must be checked before using them together in the tissue-to-plasma identity?
Solutions. (1) \(K_{p,uu,t:p}=0.40(0.10/0.020)=2.0\). The inputs imply an unbound tissue concentration twice the unbound plasma concentration; this differs from the simple passive-equilibration expectation of 1 and should be investigated, not automatically rejected. (2) \(K_{RBC:p}=[0.70-(1-0.40)]/0.40=(0.70-0.60)/0.40=0.25\). It is nonnegative and algebraically compatible with the mixture relationship. (3) Homogenization dilutes binding components and can raise the measured fraction unbound; apply the dilution correction when its assumptions and dilution factor are appropriate. (4) The partition coefficient must be converted to the same reference basis as the binding inputs, or the identity must be reformulated on a blood basis; also verify that species, conditions, and concentration definitions match.