Skip to Main Content
WELCOME TO CREX, THE COLLABORATIVE RESEARCH EXCHANGE FOR THE NIH INTRAMURAL RESEARCH PROGRAM

Go to Main Navigation

CNS Drug Discovery: The Role of Cerebrospinal Fluid (CSF) in the In Vivo Assessment of Drug and Biomarker Brain Exposure

This blog post was written by Transpharmation, a global leader in translational biology.

The Blood-Brain, Blood-CSF and Blood-Arachnoid Barriers

For drugs acting in the central nervous system (CNS), it is assumed that the unbound drug in the brain interstitial fluid (ISF) is available to interact with the target site in the CNS. Estimating or measuring the unbound brain ISF concentration is not straightforward. Unbound plasma exposure of drugs in most cases is not suitable to describe the correlation between pharmacokinetics and pharmacodynamics, mainly due to the restrictive role of the blood brain barrier (BBB), which has highly developed tight junctions between adjacent endothelial cells and active efflux transporters, such as P-glycoprotein (P-gp), breast cancer resistance protein (BCRP) and multidrug resistance-associated protein 4 (MRP4), limiting drug penetration into the CNS. In addition to the BBB, there are also the blood-CSF (BCSFB; comprised of choroid plexus epithelial cells) and blood-arachnoid (BAB; comprised of arachnoid epithelial cells) barriers. Both barriers express transporter proteins, including active efflux transporters (Uchida et al. Drug Metab Dispos (2020) 48:135).

A few methods to estimate brain ISF exposure are available. Intracerebral microdialysis is the method that can assess brain ISF exposure most directly; however, because it is technically challenging and has low throughput, it is not routinely used. Another approach uses the product of the total brain concentration in vivo and the unbound fraction in brain, determined in vitro using brain homogenate (determined by equilibrium dialysis) or brain slices. The third approach uses CSF drug concentrations as a surrogate for brain ISF concentrations. CSF drug concentrations can be assessed by collecting CSF samples from the cisterna magna in rodents, dogs and non-human primates. An advantage of CSF sampling in non-clinical animal species is that the technique is also applicable to human subjects in clinical studies (CSF is collected by lumbar puncture), allowing for a direct comparison between animal and human CSF concentrations. In addition, administration of poorly brain penetrant compounds into the CSF, e.g., by intrathecal dosing, is used both in animals and humans to circumvent the three barriers. Biomarkers can also be measured in CSF, further providing high translational value, especially for neurodegenerative diseases.

A pharmacokinetics parameter that is especially useful in assessing brain ISF exposure of a drug is the ISF to unbound plasma partition coefficient, Kp,uu. When brain exposure of a drug is dependent on plasma exposure, a Kp,uu value >1 indicates net influx of a drug into the brain, <1 indicates net efflux or poor permeation of the drug and =1 indicates net diffusion. The Kp,uu is determined by the unbound drug concentration ratio in the ISF to plasma at steady-state. However, it can also be calculated from the ratio of the area under the concentration versus time curves for ISF/plasma following a single dose. The CSF/unbound plasma or unbound brain to unbound plasma Kp,uu can be similarly determined.

Serial CSF and Plasma Collection

Using the CSF drug concentration as a surrogate for brain ISF concentrations is based on the assumption that there is a rapid equilibrium between the ISF and CSF compartments across the ependymal layer. Additionally, CSF has a very low protein concentration, so the CSF concentration is considered to be essentially unbound.

Transpharmation offers serial collection of CSF and plasma from conscious rats and dogs. The advantage of this study design is that blood and CSF are collected longitudinally from conscious animals, significantly reducing the number of animals required in a study, as well as avoiding the use of inhalant anesthetics (e.g., isoflurane), which have been shown to reversibly open the BBB (Spieth et al. Neuro Oncology Adv (2021) 3:1). Figure 1 below depicts a typical profile of a drug administered subcutaneously (SC) to a group of 4 carotid artery (for blood plasma collection) and cisterna magna (for CSF collection) catheterized rats. The drug was quantified by LC-MS/MS and the unbound fraction in plasma was determined by in vitro equilibrium dialysis. In this case the terminal half-life of the drug appears to be slightly longer in the CSF than in plasma (plasma concentrations decline faster than CSF concentrations) and the estimated CSF/plasma Kp,uu is 0.7, indicating that the drug is not freely permeable across the BBB.

Figure 1. Mean (± SD, n=4) plasma and CSF concentrations of a drug administered SC to conscious rats.

Terminal CSF, Plasma and Brain Collection

Transpharmation also offers terminal plasma, CSF and brain collection studies in rats and mice. Figure 2 depicts the unbound plasma, CSF and brain concentrations of loperamide over time. Blood plasma, CSF and brain were collected from 3 rats/time-point following 1 mg/kg intravenous (IV) administration of loperamide (Imodium®), which is a non-brain penetrant opioid as a result of it being a substrate for P-gp. Samples were analyzed by LC-MS/MS and the fraction of loperamide unbound in plasma and brain homogenate was determined in vitro by equilibrium dialysis. The calculated CSF/plasma Kp,uu is 0.010, whereas the brain/plasma Kp,uu is 0.014, indicating that CSF is a good surrogate for unbound brain concentrations of loperamide. As expected, loperamide does not permeate the BBB to a great extent. Note also that plasma, CSF and brain concentrations decline in parallel (terminal half-lives are the same), since loperamide crosses the BBB, BCSFB and BAB from the systemic circulation.

Figure 2. Mean (+ SD, n=3) unbound plasma, CSF and brain concentrations of loperamide following IV dosing to rats.

Delivery of Non-Brain Penetrant Compounds by Intrathecal or ICV Drug Administration

Transpharmation further offers the administration of small molecules and other modalities into the CSF of rats and mice, either by intrathecal (IT) administration or by intracerebroventricular (ICV) administration. IT administration is less invasive as we perform direct injection into the spine between L5 and L6 of an anaesthetized animal. The IT formulation can include lidocaine, as this consistently results in transient paralysis (which resolves within 10 min in mouse and 30 min in rat), ensuring that each IT injection was successful. Figure 3 depicts the unbound concentrations of loperamide in plasma, CSF and brain following 0.2 mg/kg IT administration of loperamide. Interestingly, loperamide is rapidly cleared from the CSF, likely by P-gp efflux into the systemic circulation and only low concentrations are measurable in brain tissue. The calculated CSF/plasma Kp,uu is 9.2, whereas that for brain is 0.013. The latter Kp,uu is essentially unchanged from that following IV dosing of loperamide. Although the half-life of loperamide in CSF is very rapid, plasma and brain concentrations decline in parallel (have the same half-life).

Figure 3. Mean (+ SD, n=3) unbound plasma, CSF and brain concentrations of loperamide following IT dosing to rats.

ICV injection in rats and mice entails drilling a hole in the skull of an anesthetized animal above the lateral ventricle, and thus is more invasive and takes more time/injection compared to IT administration. Figure 4 depicts the unbound concentrations of loperamide in plasma, CSF and brain following 0.2 mg/kg ICV administration of loperamide. Again, loperamide is initially rapidly cleared from the CSF into the systemic circulation (plasma); however, unbound concentrations in the brain are higher than plasma. The calculated Kp,uu is 34.1, whereas that for brain is 43.3. In this case, both values are much higher compared to those following IV dosing of loperamide. Although the half-life of loperamide in CSF is initially very rapid, plasma and brain and eventually CSF concentrations decline in parallel. For loperamide, ICV injection resulted in much higher unbound brain and CSF concentrations when compared to IT injection. In addition, plasma exposure was almost 2-fold lower following ICV compared to IT administration. Relative to CSF, plasma and brain exposures of loperamide were very low (plasma to CSF and brain to CSF Kp,uu values are, respectively, 0.0053 and 0.033 following ICV dosing and 0.11 and 0.0014 following IT dosing). Similar comparisons for other molecules that are substrates for drug transporters as well as those that are not would be of interest.

Figure 4. Mean (+ SD, n=3) unbound plasma, CSF and brain concentrations of loperamide following ICV dosing to rats.

Contact Transpharmation today to see how we can help you with your CNS drug discovery!