From Assay Inputs to Oral Exposure
Permeability and intestinal enzyme assays provide different pieces of an oral PBPK prediction. Permeability informs how much drug can leave the intestinal lumen and enter the gut wall; enzyme activity informs how much of that absorbed drug may be metabolized before reaching portal blood. A useful integrated case therefore tracks both fractions and then asks whether their combined prediction is consistent with observed oral exposure.
The calculations below use a deliberately simplified human model. It is useful for organizing evidence and diagnosing discrepancies, but it does not make permeability, enzyme activity, and systemic pharmacokinetics uniquely identifiable from one oral exposure measurement. The introductory material on effective permeability, intestinal enzyme scaling, and gut extraction supplies the underlying concepts; here they are combined in one quantitative workflow.
This product assumes the fractions are meaningful and approximately separable: absorption occurs first, gut extraction acts on absorbed drug, and hepatic extraction acts on drug entering the liver. It also assumes linear pharmacokinetics over the doses being compared. In particular, the product is not a substitute for a coupled enterocyte model when efflux, re-entry, and metabolism materially interact.
Estimating Absorption From Effective Permeability
For a simple intestinal segment, suppose the lumen is cylindrical with radius \(R\), effective permeability \(P_{\mathrm{eff}}\) is constant, and the drug remains in the segment for time \(\tau\). Assume the lumen is well mixed radially, the concentration available for permeation declines in proportion to the amount remaining, and dissolution or precipitation does not limit the dissolved concentration. The cylinder's surface-area-to-volume ratio is \(2/R\), so the first-order absorption rate constant is \(k_a=2P_{\mathrm{eff}}/R\). The units agree when \(P_{\mathrm{eff}}\) and \(R\) are expressed in the same length units.
The last expression follows by solving the first-order depletion equation: \(M_{\mathrm{lumen}}(t)/M_{\mathrm{lumen}}(0)=\exp(-k_at)\). The fraction removed from the lumen by time \(\tau\) is one minus the fraction remaining. In a regional model, segment-specific geometry, permeability, and residence times can be represented separately; the one-segment expression here is an illustrative reduction, not a claim that the human intestine is uniform.
Worked Example: Converting Permeability to Absorbed Fraction
Suppose a hypothetical human effective permeability estimate is \(1.50\times10^{-4}\) cm/s. Use a representative intestinal radius of 1.50 cm and a residence time of 3.00 h. Assume the drug is dissolved and permeability and geometry remain constant over this interval.
First convert permeability to cm/h: \(1.50\times10^{-4}\ \mathrm{cm/s}\times3{,}600\ \mathrm{s/h}=0.540\ \mathrm{cm/h}\). Then \(k_a=2(0.540\ \mathrm{cm/h})/(1.50\ \mathrm{cm})=0.720\ \mathrm{h}^{-1}\).
The estimated absorbed fraction is \(F_a=1-\exp[-(0.720\ \mathrm{h}^{-1})(3.00\ \mathrm{h})]=1-\exp(-2.16)=0.885\). Under these assumptions, about 88.5% of the dose is absorbed from the lumen. This is not yet the fraction reaching systemic circulation: intestinal and hepatic extraction have not been applied.
Scaling an Enzyme Assay to Gut Escape
An in vitro intrinsic clearance per unit microsomal protein must be scaled before it can be compared with intestinal blood flow. In this example, assume a microsomal assay reports intrinsic clearance in volume per time per milligram of microsomal protein. Scale it by microsomal protein per gram of intestine and total intestinal mass to obtain whole-gut intrinsic clearance. If \(f_{u,\mathrm{gut}}\) is the unbound fraction available to the enzyme in the gut, use \(f_{u,\mathrm{gut}}CL_{\mathrm{int,gut}}\) as the effective intrinsic clearance in the simplified well-stirred model.
The well-stirred model assumes constant intestinal blood flow \(Q_{\mathrm{gut}}\), linear metabolism, and rapid mixing such that the enzyme clearance and flow determine the outlet concentration. Its extraction ratio is effective intrinsic clearance divided by the sum of flow and effective intrinsic clearance. Gut escape is one minus that extraction ratio.
Worked Example: Scaling Enzyme Activity to Gut Escape
Assume a hypothetical assay gives \(CL_{\mathrm{int,assay}}=12.0\ \mathrm{\mu L/min/mg}\). Use 35.0 mg microsomal protein per gram of intestine, an intestinal mass of 250 g, \(f_{u,\mathrm{gut}}=0.500\), and intestinal blood flow \(Q_{\mathrm{gut}}=18.0\ \mathrm{L/h}\). Assume the assay clearance scales linearly with protein amount and the unbound fraction is known.
The total microsomal protein is \((35.0\ \mathrm{mg/g})(250\ \mathrm{g})=8{,}750\ \mathrm{mg}\). Therefore, \(CL_{\mathrm{int,gut}}=(12.0\ \mathrm{\mu L/min/mg})(8{,}750\ \mathrm{mg})=105{,}000\ \mathrm{\mu L/min}\). Since \(1{,}000{,}000\ \mathrm{\mu L}=1\ \mathrm{L}\), this is \(0.105\ \mathrm{L/min}\times60\ \mathrm{min/h}=6.30\ \mathrm{L/h}\).
The effective intrinsic clearance is \(f_{u,\mathrm{gut}}CL_{\mathrm{int,gut}}=(0.500)(6.30)=3.15\ \mathrm{L/h}\). Thus, \(F_g=18.0/(18.0+3.15)=0.851\). The model estimates that 85.1% of absorbed drug escapes gut metabolism. It does not say that 85.1% of the administered dose reaches portal blood; that fraction also depends on absorption.
Integrated Case: Predicting and Checking Exposure
Consider the two hypothetical input estimates above, together with a hepatic escape fraction \(F_h=0.750\). Suppose systemic clearance is \(8.00\ \mathrm{L/h}\). The model predicts \(F=F_aF_gF_h\). To compare that prediction with data, use dose-normalized oral and intravenous AUCs, assuming the same linear clearance applies to both routes and the observations are at comparable conditions.
Worked Example: Comparing Predicted and Observed Oral Exposure
For a hypothetical compound, a 20.0 mg intravenous dose produces \(AUC_{\mathrm{IV}}=2.50\ \mathrm{mg\cdot h/L}\), and a 100 mg oral dose produces \(AUC_{\mathrm{oral}}=5.00\ \mathrm{mg\cdot h/L}\). Use \(F_a=0.885\), \(F_g=0.851\), and \(F_h=0.750\) from the illustrative inputs. Assume dose-proportional pharmacokinetics, complete intravenous availability, and a common systemic clearance.
First, intravenous data give \(CL=D_{\mathrm{IV}}/AUC_{\mathrm{IV}}=(20.0\ \mathrm{mg})/(2.50\ \mathrm{mg\cdot h/L})=8.00\ \mathrm{L/h}\). The model's predicted bioavailability is \(F_{\mathrm{pred}}=(0.885)(0.851)(0.750)=0.565\).
The predicted oral exposure is \(AUC_{\mathrm{oral,pred}}=D_{\mathrm{oral}}F_{\mathrm{pred}}/CL=(100\ \mathrm{mg})(0.565)/(8.00\ \mathrm{L/h})=7.06\ \mathrm{mg\cdot h/L}\). The observed absolute bioavailability is estimated from dose-normalized AUCs: \(F_{\mathrm{obs}}=(AUC_{\mathrm{oral}}/D_{\mathrm{oral}})/(AUC_{\mathrm{IV}}/D_{\mathrm{IV}})=(5.00/100)/(2.50/20.0)=0.400\). The observed AUC is 2.06 mg·h/L below prediction, and observed bioavailability is about 29.2% below the predicted value: \((0.565-0.400)/0.565=0.292\).
This mismatch establishes that the combined prediction is too high under the stated assumptions; it does not establish which input is wrong. Possible explanations include an overestimated human permeability, lower in vivo enzyme activity or abundance than the assay scaling implies, a poor estimate of \(F_h\), or unrepresented loss of dissolved drug before absorption. The next step is to assess these explanations using independent evidence rather than adjusting several parameters to fit this one AUC.
Derivation and Interpretation of the Combined Prediction
The product form follows by conditioning on successive steps. Out of an administered dose \(D\), the amount absorbed is \(DF_a\). Of that amount, the amount escaping gut metabolism is \(DF_aF_g\). Of the amount escaping the gut, the amount escaping hepatic extraction is \(DF_aF_gF_h\). Under linear systemic disposition, AUC equals the amount reaching systemic circulation divided by clearance. Therefore \(AUC_{\mathrm{oral}}=DF_aF_gF_h/CL\). This argument depends on each fraction applying to the amount entering its corresponding step and on the absence of nonlinear dose dependence.
The intravenous comparison offers a practical check because, under the same linear-clearance assumption, \(AUC_{\mathrm{IV}}/D_{\mathrm{IV}}=1/CL\). Dividing oral dose-normalized AUC by this quantity cancels clearance and estimates \(F\). This comparison is especially useful when clearance is uncertain, but it cannot by itself separate \(F_a\), \(F_g\), and \(F_h\): many combinations can yield the same total bioavailability.
In the integrated example, permeability and assay scaling imply a predicted \(F\) of 0.565, while the route comparison implies 0.400. The size and direction of the discrepancy guide investigation, not a unique correction. If an independent measure supports \(F_h=0.750\), then the product \(F_aF_g\) required to match the observation is \(0.400/0.750=0.533\), compared with the input-based \(0.885\times0.851=0.753\). This localizes the mismatch to the combined pre-hepatic part of the model only if the hepatic estimate and the assumptions behind the route comparison are credible; it still does not distinguish absorption from gut escape.
Diagnostics, Alternatives, and Limitations
A mismatch should prompt checks of inputs, assumptions, and model structure before parameter fitting. In particular, verify unit conversions in the enzyme scaling, assay substrate and enzyme relevance, the protein-per-gram and gut-mass values, the unbound fraction, and the blood-flow units. For absorption, check that the permeability estimate applies to the human intestinal region and formulation being modeled, that the residence time is plausible, and that dissolved drug is available. A calculation that gives \(F_a>1\), \(F_g>1\), or negative clearance signals an implementation or unit error; fractions must lie between zero and one under this model.
Useful comparison approaches answer different questions. A total bioavailability estimate from oral and intravenous AUC is data-driven and can validate the overall product, but it cannot identify its component fractions. The permeability-plus-enzyme approach supplies a mechanistic decomposition, but introduces assay scaling and physiological assumptions. A coupled intestinal model can represent efflux, re-entry, regional metabolism, or dissolution-related input more explicitly; it requires more parameters and independent evidence to constrain them. Choosing the more complex model is justified when the simpler model fails in a systematic way and the added processes can be informed by data.
The simplified calculations have important limits. A single radius and residence time compress regional differences; a single effective permeability can conceal changes in transporter activity or intestinal conditions. The well-stirred gut model treats enzyme activity and blood flow as aggregate quantities and may not represent heterogeneous enterocyte exposure. If efflux returns drug to the lumen, an effective permeability estimate and a separate gut extraction term may not be independent. Counting that loss in both \(F_a\) and \(F_g\) would double count it. Likewise, using an assay-derived clearance that already incorporates binding or abundance adjustments and applying those adjustments again would bias \(F_g\).
A disciplined interpretation therefore reports the component estimates, their assumptions, and the observed-versus-predicted exposure difference. It avoids claiming that a mismatch proves high gut metabolism, low permeability, or any other single cause. Additional evidence—such as independent enzyme scaling information, region-specific permeability, or a relevant transporter assessment—can help distinguish competing explanations.
Check Your Understanding
Answer each question using the assumptions stated in the tutorial. Solutions follow the questions.
- If \(k_a=0.500\ \mathrm{h}^{-1}\) and residence time is 2.00 h, what is the predicted \(F_a\)?
- If \(Q_{\mathrm{gut}}=15.0\ \mathrm{L/h}\) and effective intrinsic clearance is \(5.00\ \mathrm{L/h}\), calculate \(F_g\).
- For \(F_a=0.800\), \(F_g=0.900\), and \(F_h=0.750\), calculate predicted absolute bioavailability.
- Why can an oral-versus-intravenous AUC comparison estimate total \(F\) but not separately identify \(F_a\), \(F_g\), and \(F_h\)?
- Name two checks to make before interpreting a predicted oral AUC that is higher than observed.
Solutions. 1. \(F_a=1-\exp[-(0.500)(2.00)]=1-\exp(-1.00)=0.632\), or 63.2%. 2. \(F_g=15.0/(15.0+5.00)=0.750\). 3. \(F=(0.800)(0.900)(0.750)=0.540\), or 54.0%. 4. The AUC ratio provides the product \(F_aF_gF_h\); different component values can have the same product, so the total alone does not determine each factor. 5. Examples include checking permeability and residence-time assumptions, enzyme assay scaling and unit conversions, unbound fraction and gut flow, and whether dissolution limits the available drug. A predicted-exposure excess does not by itself identify which input caused it.