A mathematical model can link plasma drug concentrations with a measured brain response without turning a plasma result into a direct measurement of that response. A 2019 study by van Gaalen and colleagues illustrates both the scientific value and the limits of that approach for dextroamphetamine.
The investigators developed a pharmacokinetic–pharmacodynamic model, commonly shortened to PK–PD. Pharmacokinetics described the drug’s distribution and elimination; pharmacodynamics described the changing dopamine response. The model integrated animal experiments and earlier published datasets. It was not a human impairment calculator.
What was measured, and in which specimens?
The new experiments used a single intravenous bolus of 0.1 mg/kg dextroamphetamine, expressed as free base, in male rats and male cynomolgus monkeys. The investigators measured dextroamphetamine in plasma, brain extracellular fluid and cerebrospinal fluid. They also measured dopamine in brain extracellular fluid using microdialysis.
These matrices have different meanings. Plasma is the liquid fraction obtained after centrifuging anticoagulated blood. Extracellular fluid is fluid outside brain cells sampled locally through a probe. Cerebrospinal fluid is another compartment, not a sample of every brain synapse. Dopamine was the response analyte; dextroamphetamine was the administered drug.
Keeping those distinctions intact prevents a common interpretive shortcut: calling all the results either a blood level or a brain level. A whole-blood forensic result is also not automatically the same quantity as the unbound plasma concentration used in the model.
How many animals supplied the new data?
Eleven Wistar rats contributed to the new low-dose experiments. All supplied blood samples, but the matrix-specific datasets were smaller: four rats supplied extracellular-fluid drug data and five supplied cerebrospinal-fluid drug data, with overlap between these groups. A separate group of five supplied extracellular dopamine data.
Three cynomolgus monkeys contributed overall, but that number does not describe three independent animals for every endpoint. Extracellular-fluid drug measurements came from one animal sampled in four brain regions on one dosing occasion. Dopamine measurements came from that same animal on three dosing occasions. Cerebrospinal-fluid samples came from two other animals.
Repeated dosing occasions and multiple probes increase observations, but they do not create additional independent animals. A model supported by these measurements must retain that distinction when its apparent precision is discussed.
Why did the model also use published studies?
The authors needed a wider dose range than the new 0.1 mg/kg experiments provided. They digitized mean concentration or dopamine-response curves from earlier publications. Rat datasets included intravenous and subcutaneous dosing; higher-dose primate data came from rhesus macaques rather than the cynomolgus monkeys used in the new experiments.
Consequently, the analysis combined different rat strains, primate species, routes and sampling schedules. It did not follow one uniform cohort through all doses. Population modeling allowed the investigators to integrate such data, but could not remove every difference between the contributing experiments.
Which assumptions connected plasma exposure with dopamine?
The investigators converted measured total plasma drug concentrations to unbound concentrations using an assumed unbound fraction of 0.6 in both species. Brain distribution parameters were assumed to remain independent of dose over the modeled ranges. Some volumes or clearance terms had to be fixed because the data could not estimate them separately with adequate precision.
The model also assumed that drug-transfer behavior could be applied across brain sampling regions. For the subcutaneous rat dataset lacking corresponding pharmacokinetic measurements, rapid and complete absorption was assumed. These were stated modeling choices, not directly observed facts for every animal and dose.
An estimated brain concentration therefore should be identified as a prediction conditional on those assumptions. It should not be presented as if a probe measured that concentration at every dopamine-response time point.
What did the dopamine time courses add?
The dopamine response did not simply mirror drug exposure throughout the observation period. Time-matched plots showed patterns consistent with acute tolerance, and the final model incorporated delayed negative feedback that limited dopamine input into the extracellular space.
This formulation fit the available data better than the alternative feedback structures tested. It supported a mechanistic account involving both amphetamine’s effects and the system’s response to them. A better fit among candidate models does not prove that every represented molecular step was independently measured or uniquely identified.
Are the model’s tolerance half-lives clinical recovery times?
No. Estimated half-lives governing development and loss of the modeled feedback were 9.6 hours in rats and approximately one hour in nonhuman primates. Those are parameters of a modeled dopamine-response process, not measured durations of intoxication, safe-driving intervals or times to reversal of dependence.
Precision was limited: relative standard errors for the feedback rate constant were 68.1% in rats and 33.9% in primates. Higher-dose primate sampling ended after two hours, substantially shorter than the reported 340-minute plasma elimination half-life following the low dose. The authors said longer sampling and additional higher-dose studies were needed to determine whether the apparent species difference was real.
What does this permit in a forensic evidence review?
The study supports examining how specimen, sampling site, exposure history and feedback assumptions shape an exposure–response model. It does not validate estimating an individual person’s synaptic dopamine concentration, mental state or driving ability from one amphetamine result.
Its strongest contribution is a framework for further experimental testing. The defensible description is an animal model integrating measured and predicted quantities, with translational potential and stated uncertainty. A case-specific human opinion would require evidence beyond the model’s successful fit to these animal datasets.
References
van Gaalen MM, Schlumbohm C, Folgering JH, et al. Development of a semimechanistic pharmacokinetic-pharmacodynamic model describing dextroamphetamine exposure and striatal dopamine response in rats and nonhuman primates following a single dose of dextroamphetamine. Journal of Pharmacology and Experimental Therapeutics. 2019;369(1):107–120. doi:10.1124/jpet.118.254508. Historical primary animal experiments and modeling using additional published data.
Leave a Reply