Abstract
The P-glycoprotein (P-gp)-deficient mouse model is used to assess the influence of P-gp-mediated efflux on the central nervous system (CNS) distribution of drugs. The steady-state unbound plasma/unbound brain concentration ratio ([plasma],u/[brain],u) is an alternative method for assessing CNS distribution of drugs independent of the mechanism(s) involved. The objective of this study was to compare the degree of CNS distributional impairment determined from the in vivo P-gp efflux ratio with that determined from the [plasma],u/[brain],u ratio. CNS distribution of 34 drugs, including opioids, triptans, protease inhibitors, antihistamines, and other clinically relevant drugs with either poor CNS distribution or blood-brain barrier efflux, was studied. Plasma and brain unbound fractions were determined by equilibrium dialysis. Kp,brain and the P-gp efflux ratio were obtained from the literature or determined experimentally. The P-gp efflux ratio and the [plasma],u/[brain],u ratio were in concurrence (<3-fold difference) for 21 of the 34 drugs. However, the [plasma],u/[brain],u ratio exceeded the P-gp efflux ratio substantially (>4-fold) for 10 of the 34 drugs, suggesting that other, non-P-gp-mediated mechanism(s) may limit the CNS distribution of these drugs. The P-gp efflux ratio exceeded the [plasma],u/[brain],u ratio by more than 3-fold for three drugs, suggesting the presence of active uptake mechanism(s). These observations indicate that when mechanisms other than P-gp affect CNS distribution (non-P-gp-mediated efflux, poor passive permeability, cerebrospinal fluid bulk flow, metabolism, or active uptake), the P-gp efflux ratio may underestimate or overestimate CNS distributional impairment. The [plasma],u/[brain],u ratio provides a simple mechanism-independent alternative for assessing the CNS distribution of drugs.
The efflux transporter P-glycoprotein (P-gp) attenuates the central nervous system (CNS) distribution of many drugs, including opioids, triptans, protease inhibitors, and antihistamines. One method used to assess the influence of P-gp on the CNS distribution of compounds is the P-gp-deficient mouse model. The P-gp efflux ratio, calculated from the ratio of brain/plasma partition coefficient (Kp,brain) in P-gp-deficient (mdr1a–/–) mice to Kp,brain in P-gp-competent (mdr1a+/+) mice, reflects the degree to which P-gp-mediated efflux attenuates CNS distribution. However, when other processes influence CNS distribution, the P-gp efflux ratio may be a poor indicator of the degree to which CNS distribution of a compound is impaired.
Kp,brain is the most widely used in vivo parameter for assessing the extent of CNS distribution. A common assumption is that compounds with large Kp,brain values have more extensive CNS distribution than compounds with small Kp,brain values. For example, a Kp,brain ≥1is often used as an arbitrary cutoff to classify compounds as having “good” CNS distribution, whereas a Kp,brain ≪1 is used to classify compounds as having “poor” CNS distribution. Although this type of classification is common, it may be misleading. It is recognized that tissue partition coefficients such as Kp,brain are influenced by the relative binding affinity of a substrate for proteins in plasma versus the proteins in the tissue in question (Gillette, 1971; Kurz and Fichtl, 1983). For a compound that distributes solely by passive diffusion, at distribution equilibrium the unbound concentration in tissue will equal the unbound concentration in plasma, and the steady-state tissue partition coefficient is simply a function of the relative plasma and tissue unbound fractions (i.e., Kp,tissue = fu,plasma/fu,tissue). When brain and plasma unbound fractions are similar, then a Kp,brain ∼1 would be consistent with unrestricted distribution solely by passive processes. However, the Kp,brain value by itself provides little information without knowledge of brain and plasma unbound fractions. A Kp,brain value <1 could be the result of more extensive binding to plasma proteins than to proteins in brain tissue. Alternatively, a Kp,brain value <1 could reflect significant impairment in CNS distribution as a result of processes such as efflux transport at the blood-brain barrier (BBB).
Several recent literature reports have used the fu,plasma/fu,brain ratio to predict Kp,brain and to assess the CNS distribution of compounds. In one published account, the utility of the fu,plasma/fu,brain ratio to predict the Kp,brain of CNS discovery compounds was assessed (Kalvass and Maurer, 2002). As expected, the fu,plasma/fu,brain ratio predicted the Kp,brain for compounds that did not evidence active efflux at the BBB, and overpredicted Kp,brain when active efflux at the BBB limited brain uptake. The degree to which the fu,plasma/fu,brain ratio overpredicted Kp,brain was identical to the P-gp efflux ratio for the single member of the compound set for which the P-gp efflux ratio had been determined. Another study used the fu,plasma/fu,brain ratio to predict Kp,brain and to assess the CNS distribution of 33 marketed CNS drugs (Maurer et al., 2005). The fu,plasma/fu,brain ratio predicted the Kp,brain value for the majority of CNS drugs (25 of 33), indicating that most CNS drugs do not have impaired CNS distribution. In those cases for which the fu,plasma/fu,brain ratio did not predict Kp,brain, the discrepancy could be explained by active efflux or poor BBB permeability. More recently, for compounds subject to active efflux, the fu,plasma/fu,brain ratio combined with in vitro efflux data was shown to provide superior estimates of Kp,brain compared with the fu,plasma/fu,brain ratio alone (Summerfield et al., 2006).
When CNS distribution is impaired, the fu,plasma/fu,brain ratio overpredicts Kp,brain. The magnitude of the overprediction is reflective of the degree to which unbound plasma concentrations exceed unbound brain concentrations (eq. 1). As shown in eq. 1, the fu,plasma/fu,brain ratio and Kp,brain can be used to calculate the unbound plasma/unbound brain concentration ratio [plasma],u/[brain],u. The [plasma],u/[brain],u ratio represents the degree to which unbound plasma concentrations exceed unbound brain concentrations and is a meaningful expression of the degree of impairment in CNS distribution. A [plasma],u/[brain],u ratio of unity for a given compound is indicative of unimpaired CNS distribution (i.e., distribution consistent with passive diffusion; [plasma],u = [brain],u), whereas a [plasma],u/[brain],u ratio greater than unity indicates impairment in CNS distribution (i.e., efflux uptake or poor BBB permeability; [plasma],u > [brain],u). In contrast, a [plasma],u/[brain],u ratio value less than unity is consistent with enhanced CNS distribution (i.e., active uptake; [plasma],u < [brain],u).
The [plasma],u/[brain],u ratio is expected to be equal to the in vivo P-gp efflux if P-gp-mediated efflux is the only active process affecting brain disposition. Using this principle, the present study was conducted to compare the degree of CNS distributional impairment expressed as the [plasma],u/[brain],u ratio to the P-gp efflux ratio for 34 marketed drugs. Opioids, triptans, protease inhibitors, and antihistamines (n = 24 total) were included in this analysis because these classes of agents are known to include P-gp substrates, and the extent to which these compounds distribute into the CNS may have important implications regarding safety and efficacy. In addition, 10 marketed drugs from various drug classes with either poor CNS distribution or BBB efflux also were included as part of the analysis.
Materials and Methods
Materials. Sufentanil was obtained from Abbott Laboratories (North Chicago, IL). Amprenavir, indinavir, nelfinavir, ritonavir, and saquinavir were obtained through the National Institutes of Health AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health (Rockville, MD). Olanzapine was obtained from Pfizer Global Material Management (Groton, CT). Eletriptan was extracted from 40-mg tablets obtained from the Roerig Division of Pfizer (New York, NY), and the raw extract was used as a stock solution. Cetirizine, desloratadine, rizatriptan, and zolmitriptan were purchased from Sequoia Research (Oxford, UK). Alfentanil was obtained from Taylor Pharmaceuticals (Decatur, IL). Naratriptan and sumatriptan were purchased from U.S. Pharmacopoeia (Rockville, MD). All the other drugs were purchased from Sigma-Aldrich (St. Louis, MO). Solvents and other reagents were obtained from common sources and were of reagent grade or better.
Drug Selection. Twenty-four marketed drugs from four main drug classes (seven opioids, five triptans, five protease inhibitors, and seven antihistamines) were chosen for this study because drugs within each class exhibit varying degrees of interaction with BBB P-gp and because the extent of CNS distribution is known to be important to the efficacy and/or toxicity of these drug classes. In addition, 10 marketed drugs from other drug classes with either poor CNS distribution or BBB efflux also were chosen to assess the general utility of the approach with as diverse a dataset as possible.
Animals. Male CF-1 mdr1a(+/+) and mdr1a(–/–) mice (30–40 g; Charles River Laboratories, Inc., Wilmington, MA) were maintained on a 12-h light/dark cycle in a temperature- and humidity-controlled room with access to water and food ad libitum. All the procedures involving mice were approved by The Institutional Animal Care and Use Committee of the University of North Carolina and were conducted in accordance with Principles of Laboratory Animal Care (National Institutes of Health Publication 85-23, revised in 1985).
Determination ofKp,brain from Animal Studies.Kp,brain values in mdr1a(–/–) and mdr1a(+/+) mice were obtained from the literature or, when published values were not available, were determined experimentally. Separate pharmacokinetic studies were conducted in mdr1a(–/–) and mdr1a(+/+) mice to determine the Kp,brain values for alfentanil, fentanyl, loperamide, and methadone. Briefly, mice received an s.c. dose of the opioid, and at designated time points (nine time points, n = 4 animals per time point), mice were sacrificed, and trunk blood and brain tissue were collected. Plasma was harvested following centrifugation. Brain and plasma samples were stored at –20°C until analysis by high-performance liquid chromatography (HPLC) with tandem mass spectrometry (MS/MS) (see below). Kp,brain was determined from the ratio of brain to plasma area under the curve0-∞. Kp,brain values for cimetidine, meperidine, ranitidine, and sufentanil were determined under steady-state conditions in mdr1a(–/–) and mdr1a(+/+) mice. Briefly, osmotic mini-pumps (Alzet, Cupertino, CA) were implanted s.c. per the manufacturer's instructions. Mice were sacrificed, and trunk blood and brain tissue were collected (n = 3) 24 h later. Plasma was harvested following centrifugation, and brain and plasma samples were stored at –20°C until analysis by HPLC/MS/MS (see below). Kp,brain was calculated from the ratio of the 24-h brain and plasma concentrations. All the other Kp,brain values were obtained from the literature.
Equilibrium Dialysis Experiments. Plasma and brain unbound fractions were determined in a 96-well equilibrium dialysis apparatus (HTDialysis, Gales Ferry, CT) using a previously reported method (Kalvass and Maurer, 2002). Briefly, fresh CF-1 or FVB mouse plasma and brain tissue were obtained the day of the study. Spectra-Por 2 membranes obtained from Spectrum Laboratories Inc. (Rancho Dominguez, CA) were conditioned in HPLC-grade water for 15 min, followed by 30% ethanol for 15 min, and 100 mM sodium phosphate buffer (pH 7.4) for 15 min. Brain tissue was diluted 3-fold with 100 mM sodium phosphate buffer (pH 7.4) and homogenized with a sonic probe. The drug of interest was added to plasma and brain homogenate to achieve a final concentration of 3 and 1 μM, respectively; 150-μl aliquots (n = 6) were loaded into the 96-well equilibrium dialysis apparatus and dialyzed against an equal volume of 100 mM sodium phosphate (pH 7.4) buffer. The 96-well equilibrium dialysis apparatus was incubated for 4.5 h in a 155-rpm shaking water bath maintained at 37°C. Previous experience with the equilibrium dialysis apparatus indicated that equilibrium would be achieved by 4.5 h (data not shown). After 4.5 h, 10 μl of matrix sample (plasma or brain homogenate) and 50 μl of buffer sample were removed from the apparatus and added directly to HPLC vials containing 100 μl of methanol containing an appropriate internal standard. A 50-μl aliquot of control buffer was added to the brain homogenate and plasma samples, and either a 10-μl aliquot of control brain homogenate or control plasma was added to the buffer samples to yield identical sample composition between buffer and non-buffer samples. The samples were vortex-mixed, centrifuged, and the supernatant was analyzed by HPLC/MS/MS. Plasma unbound fraction was calculated from the ratio of concentrations determined from the plasma and buffer samples. Equation 2, a previously described approach to account for the effect of tissue dilution on unbound fraction (Kalvass and Maurer, 2002), was used to calculate the brain unbound fraction: where D represents the -fold dilution of brain tissue, and fu,measured is the ratio of concentrations determined from the buffer and brain homogenate samples.
HPLC/MS/MS Analysis of Samples. All samples were quantified using either a PE-Sciex API-3000 (Turbo Ionspray source, 500°C) or an API-4000 (Turbo V Ionspray source, 700°C, PerkinElmerSciex Instruments, Boston, MA) quadruple mass spectrometer as summarized in Table 1. Equilibrium dialysis samples were prepared as described in the equilibrium dialysis section. Plasma and brain samples from animal experiments were prepared as follows. Brain samples were homogenized in water (1:2 v/v) with a sonic probe. An aliquot of homogenate or serum (2–25 μl) was transferred to an HPLC vial, and protein was precipitated with 4 to 125 volumes of methanol containing an appropriate internal standard. The sample was vortex-mixed and centrifuged, and the supernatant was analyzed by HPLC/MS/MS. Samples were injected (2–10 μl; CTC Analytics autosampler, Zwingen, Switzerland) onto either a Phenomenex (Torrance, CA) 2.0 × 30 mm, 5-μm Gemini 110A or a Phenomenex 2.0 × 30 mm, 4-μm Synergi Max-RP column maintained at room temperature. The total run time was 3 min. Analytes were eluted with a linear gradient consisting of ammonium acetate (pH 6.8, 10 mM) (“A”), methanol (“B”), and acetonitrile (“C”) produced by three Shimadzu LC-10ADVP binary pumps. An initial condition (80–95% A) was ramped to an intermediate condition (5–25% A) over 2 min, held for 0.5 min at the intermediate condition, and then returned to initial conditions in a single step to reequilibrate the column (Table 1). During the run, the flow rate was increased from 750 to 1500 μl/min over the first 2 min, held at 1500 μl/min for 1 min, and then returned to the initial flow rate of 750 μl/min in a single step. For samples run on the API-3000, the flow rate was increased from 500 to 750 μl/min over the first 2 min, held at 750 μl/min for 1 min, and then returned to the initial flow rate of 500 μl/min in a single step. The entire column effluent was diverted from the source of the quadrupole mass spectrometer for the first 0.8 min and final 0.5 min of the run. Standards were prepared with either plasma or brain homogenate and were identical in final composition to corresponding samples. Accuracy of standards was within ±20%.
Data Analysis. The in vivo P-gp efflux ratio for each drug was calculated as the ratio of mdr1a(–/–) and mdr1a(+/+) Kp,brain values. The steady-state [plasma],u/[brain],u ratio was calculated for each drug according to eq. 3, where Kp,brain is the value from P-gp competent mice. The P-gp efflux ratio and [plasma],u/[brain],u ratio were used to assess the distributional behavior of each drug based on the graphical scheme in Fig. 1. The horizontal and vertical lines represent the point at which the P-gp efflux ratio and the [plasma],u/[brain],u ratio equal 3, respectively. A P-gp efflux ratio or [plasma],u/[brain],u ratio >3 was considered meaningful because such ratio is sufficiently different from unity and most CNS drugs (29 of 32) have been shown to have P-gp efflux ratios <3 (Doran et al., 2005). The figure was divided into four quadrants (I–IV) based on whether the P-gp efflux or [plasma],u/[brain],u ratio values were ≥3. The solid line passing through the origin represents the line of identity ± 3-fold (dashed lines). Drugs were assessed as follows: quadrant I, impaired CNS distribution as a result of P-gp-mediated efflux; subsections Ia and Ib, impaired CNS distribution as a result of P-gp with other active process(es) present; quadrant II, impaired CNS distribution as a result of non-P-gp mechanism; quadrant III, no impairment in CNS distribution; and quadrant IV, P-gp substrate but CNS distribution is not impaired as a result of the presence of a compensatory mechanism. The CNS distribution behavior of the protease inhibitors, opioids, antihistamines, and triptans was evaluated separately.
Results
The Kp,brain,fu,plasma, and fu,brain values for all the drugs included in this study are reported in Table 2. The Kp,brain and fu,brain values varied by more than 4 orders of magnitude, whereas the fu,plasma values varied by more than 3 orders of magnitude among the drugs studied. Vinblastine was unstable in mouse plasma, so the fu,plasma value was reported as equal to or greater than the fu,plasma value determined from the buffer concentrations, assuming complete mass balance.
The P-gp efflux ratio and the [plasma],u/[brain],u ratio were compared within each drug class (opioid, antihistamine, triptan, and protease inhibitor; Fig. 2, A, B, C, and D, respectively). The P-gp efflux ratio and the [plasma],u/[brain],u ratio for all 34 drugs examined in this study are compared in Fig. 3.
The P-gp efflux ratios varied between ∼1 and 50 for the examined drugs; 18 of the 34 drugs had a P-gp efflux ratio exceeding 3. The [plasma],u/[brain],u ratio varied between ∼1 and >1000, with 23 of 34 drugs having a [plasma],u/[brain],u ratio greater than 3. The P-gp efflux ratio and the [plasma],u/[brain],u ratio were in concurrence (<3-fold difference) for 21 of 34 drugs (quadrants I and III). However, the [plasma],u/[brain],u ratio exceeded the P-gp efflux ratio substantially (>4-fold) for cetirizine, cimetidine, dexamethasone, digoxin, doxorubicin, fexofenadine, ivermectin, ranitidine, sumatriptan, and zolmitriptan (quadrants Ib and II). The P-gp efflux ratio was more than 3-fold higher than the [plasma],u/[brain],u ratio for methadone, ritonavir, and saquinavir (quadrant IV).
Discussion
The P-gp efflux ratio and the [plasma],u/[brain],u ratio were in concurrence (<3-fold difference) for 21 of the 34 drugs studied. This concurrence indicates that for most of the drugs examined, there was little difference between the P-gp efflux ratio and the [plasma],u/ [brain],u ratio, and that any impairment in CNS disposition would be consistent with P-gp-mediated efflux. The [plasma],u/[brain],u ratio exceeded the P-gp efflux ratio substantially (>4-fold) for 10 of the 34 drugs studied (cetirizine, cimetidine, dexamethasone, digoxin, doxorubicin, fexofenadine, ivermectin, ranitidine, sumatriptan, and zolmitriptan), suggesting that other non-P-gp-mediated mechanism(s) may limit the CNS distribution of these drugs. The P-gp efflux ratio exceeded the [plasma],u/[brain],u ratio by more than 3-fold for three of the drugs examined (methadone, ritonavir, and saquinavir), suggesting the presence of active uptake mechanism(s). The results for the opioids, triptans, protease inhibitors, and antihistamines are discussed separately in the following paragraphs.
Opioids. Consistent with their clinical use as analgesics, there was minimal impairment in CNS distribution (<3-fold as assessed by both the P-gp efflux ratio and the [plasma],u/[brain],u ratio) for sufentanil, fentanyl, morphine, and meperidine. Alfentanil evidenced modest impairment (∼3-fold), consistent with P-gp-mediated efflux. Both methadone and loperamide were significant P-gp substrates (P-gp efflux ratios of 7 and 33, respectively). However, only loperamide had substantial impairment in CNS distribution as assessed by the [plasma],u/[brain],u ratio.
The discrepancy between the P-gp efflux ratio and the [plasma],u/ [brain],u ratio for methadone may indicate the presence of one or more compensatory mechanism(s) (e.g., active uptake) that may have negated the impact of P-gp-mediated efflux. Methadone is a substrate for active uptake in the lung (Chi and Dixit, 1977), so it is plausible methadone also may undergo active uptake across the BBB. The [plasma],u/[brain],u ratio predicted that loperamide alone would have substantially reduced central activity, whereas the P-gp efflux ratio predicted that both loperamide and methadone would evidence substantially reduced central activity. Clinically, only loperamide is associated with reduced central activity, so the [plasma],u/[brain],u ratio was better able to differentiate between opioids with and without reduced central activity than the P-gp efflux ratio.
Triptans. The [plasma],u/[brain],u ratio indicated that all the triptans examined have impaired CNS distribution. However, only rizatriptan and eletriptan showed significant impairment as a result of P-gp (>3-fold). Non-P-gp-mediated mechanism(s) may be responsible for impaired CNS distribution of sumatriptan and zolmitriptan (quadrant II). One non-P-gp-mediated mechanism that may impair the CNS distribution of sumatriptan and zolmitriptan is cerebrospinal fluid (CSF) bulk flow. When passive permeability and brain uptake clearance are very low, CSF bulk flow may represent a significant clearing mechanism from the CNS, ultimately resulting in reduced CNS exposure (Shen et al., 2004). Because sumatriptan and zolmitriptan both have very low passive permeability (less than the paracellular marker mannitol) (Mahar Doan et al., 2002), it is plausible that CSF bulk flow may limit sumatriptan and zolmitriptan CNS distribution.
If triptans possess a degree of CNS distributional impairment in humans similar to that indicated by the [plasma],u/[brain],u ratio, the impaired distribution may have important implications regarding mechanism of action and CNS side effect profile of triptans. There is debate as to whether the antimigraine action of triptans is solely through vascular-mediated events or whether antinociceptive activity within the brainstem trigeminal nuclei is partially responsible (Dodick and Martin, 2004). In addition, the incidence of CNS side effects varies between the different triptans. By understanding the interrelationships between in vivo efficacy, incidence of CNS side effects, and the extent of CNS distribution, the optimal CNS distributional characteristics of triptans may be deduced.
Protease Inhibitors. All the protease inhibitors examined undergo significant P-gp efflux (P-gp efflux ratio ≥7, quadrant I and IV). However, the [plasma],u/[brain],u ratio indicated that ritonavir and saquinavir, despite being P-gp substrates, do not have impaired CNS distribution (<3-fold). This observation may be explained if a compensatory mechanism (i.e., active uptake) negates the impact of P-gp-mediated efflux. This explanation is supported by reports showing that both ritonavir and saquinavir are substrates for uptake transporters (Anthonypillai et al., 2004; Su et al., 2004).
Even though it has been thought that ritonavir and saquinavir have poor CNS distribution because of significant P-gp-mediated efflux (in vivo P-gp efflux ratio >5) and low Kp,brain (0.17 and 0.13, respectively), this may not necessarily be the case. The [plasma],u/[brain],u ratio indicates that steady-state unbound concentrations in plasma and brain are approximately equal for these agents. Compensatory uptake mechanism(s) may overcome the efflux by P-gp, and the low Kp,brain values therefore would simply be a function of more extensive protein binding in plasma than in brain. The binding of ritonavir and saquinavir to plasma proteins is higher than that in the brain. Therefore, based on fu,plasma and fu,brain, the Kp,brain is expected to be 0.25 and 0.22, respectively, and the [plasma],u/[brain],u ratio indicates the BBB has no net effect on the CNS distribution of ritonavir and saquinavir. These observations suggest that compounds that evidence P-gp-mediated efflux together with low Kp,brain (Kp,brain ≪ 1) may not have impaired CNS distribution if active uptake counters the effects of P-gp efflux and if the low Kp,brain can be explained by binding in plasma that exceeds binding in brain. Overall, ritonavir and saquinavir may have better CNS distribution and thus may be more effective in combating HIV viral infection in the CNS than previously thought.
Antihistamines. Consistent with their central activity, the sedating antihistamines triprolidine, diphenhydramine, and hydroxyzine had minimal impairment in CNS distribution (quadrant III). The nonsedating histamines desloratadine and cetirizine fell within quadrants I and Ib, respectively, indicating substantial impairment in CNS distribution as a result of P-gp-mediated efflux. However, the P-gp efflux ratio did not indicate impairment in CNS distribution of the nonsedating antihistamine fexofenadine, whereas the [plasma],u/[brain],u ratio suggested significant impairment (quadrant II). Non-P-gp-mediated mechanism(s) may contribute to the impairment in CNS distribution of cetirizine and fexofenadine (quadrants Ib and II). Assuming that loratadine is a prodrug of desloratadine, only the [plasma],u/ [brain],u ratio correctly distinguished between the sedating and nonsedating antihistamines.
Other Drugs with Poor CNS Distribution or BBB Efflux. Other marketed drugs with poor CNS distribution or BBB efflux were examined along with the opioids, triptans, protease inhibitors, and antihistamines (Fig. 3). As expected, P-gp substrates such as quinidine, verapamil, and paclitaxel fell within quadrant I (consistent with P-gp being the only efflux mechanism), whereas efflux substrates for transporters other than P-gp such as ranitidine, digoxin, and doxorubicin fell within quadrants Ib and II (Fig. 3; Table 3).
Discrepancy between P-gp Efflux Ratio and the [Plasma],u/[Brain],u Ratio. The P-gp efflux ratio and the [plasma],u/[brain],u ratio differed by more than 3-fold for 13 of the 34 drugs examined. In most cases, the difference can be explained by the physiochemical properties and/or transport characteristics of the individual drugs (Table 3). Ten of the 13 drugs for which discrepancy was noted were located in either quadrant Ib or II, indicating more extensive impairment in CNS distribution than predicted by the P-gp efflux ratio. For these 10 drugs, six drugs (digoxin, doxorubicin, ivermectin, cimetidine, dexamethasone, and ranitidine) are known to be substrates for efflux transporters other than P-gp (Table 3), two (sumatriptan and zolmitriptan) have very low permeability values (less than mannitol), and two (cetirizine and fexofenadine) have reduced central activity relative to other drugs in the same class. The remaining 3 of the 13 drugs (methadone, ritonavir, and saquinavir) for which the P-gp efflux ratio and the [plasma],u/[brain],u ratio differed by more than 3-fold were classified in quadrant IV, indicating less impairment in CNS distribution than indicated by the P-gp efflux ratio. One possible explanation for such drugs is the presence of a compensatory active uptake mechanism. Consistent with this explanation, all three of these drugs are substrates of active uptake (Table 3). Although not the intent of this work, additional detailed studies on individual drugs would be useful to confirm whether the discrepancy between the P-gp efflux ratio and the [plasma],u/[brain],u ratio are real and are not experimental artifacts. For future studies examining the influence of non-P-gp-mediated mechanisms on CNS distribution, the 13 drugs identified as having discrepancies between the P-gp efflux ratio and the [plasma],u/ [brain],u ratio would be logical choices.
Accurate determination of steady-state Kp,brain is necessary for accurate determination of the P-gp efflux ratio and the [plasma],u/ [brain],u ratio. In addition, accurate experimental determination of fu,plasma and fu,brain is required for accurate assessment of the [plasma],u/[brain],u ratio. Any error in determining Kp,brain,fu,plasma,or fu,brain may lead to artificial discrepancies between the P-gp efflux ratio and the [plasma],u/[brain],u ratio, which may also lead to incorrect conclusions regarding the CNS distribution of a compound. The likelihood of this occurring can be minimized by examining Kp,brain over multiple time points to ensure steady-state Kp,brain has been achieved and by using a validated method for determining fu,plasma and fu,brain.
In summary, the P-gp efflux ratio and [plasma],u/[brain],u ratio were similar for most of the drugs examined, indicating P-gp-mediated efflux is the predominate mechanism limiting the CNS distribution of drugs in the selected compound set. The [plasma],u/ [brain],u ratio differentiated between sedating and nonsedating antihistamines and between opioids with and without reduced central activity, whereas the P-gp efflux ratio did not. Furthermore, when there were differences between the P-gp efflux ratio and the [plasma],u/[brain],u ratio, additional supporting evidence was consistent with the [plasma],u/[brain],u ratio. When mechanisms other than P-gp affect CNS distribution (non-P-gp-mediated efflux, poor passive permeability, CSF bulk flow, metabolism, or active uptake), the P-gp efflux ratio may underestimate or overestimate CNS distributional impairment. The [plasma],u/[brain],u ratio provides a simple alternative means for assessing the CNS distribution of drugs independent of the mechanism(s) involved.
Footnotes
-
This work was supported by Grant R01 GM61191 from the National Institutes of Health and Pfizer Inc. J.C.K. was supported by an Eli Lilly and Company Foundation Predoctoral Fellowship in Pharmacokinetics and Drug Disposition.
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.106.012294.
-
ABBREVIATIONS: P-gp, P-glycoprotein; CNS, central nervous system; BBB, blood-brain barrier; HPLC, high-performance liquid chromatography; MS/MS, tandem mass spectrometry; CSF, cerebrospinal fluid.
- Received August 1, 2006.
- Accepted January 16, 2007.
- The American Society for Pharmacology and Experimental Therapeutics