- Email: [email protected]
Transport of alovudine (3′-fluorothymidine) into the brain and the cerebrospinal fluid of the rat, studied by microdialysis
Life Sciences, Vol. 66, No. 19, pp. 180%1816.2000 Copyright 0 2000 Ekvin Science Inc. Printed in the USA. All rights reserved 0024-3205/00/$-see front matter
PI1 SO0243205(00)00504-X
TRANSPORT OF ALOVUDINE (3’-FLUORO THYMIDINE) INTO THE BRAIN AND THE CEREBROSPINAL FLUID OF THE RAT, STUDIED BY MICRODIALYSIS Lars St&le and Natalia Borg. Department
of Clinical Pharmacology, Huddinge University Huddinge, Sweden.
Hospital, SE-14186,
(Received in final form December 7, 1999) Summary
Extracellular unbound concentrations of alovudine were sampled by microdialysis in order to study the transport of alovudine between the blood and the brain and the cerebrospinal fluid (CSF) in the rat. The AUC (area under the curve) ratio CSF/blood was higher than the brain/blood ratio after iv. infusion of alovudine 25mg/kg/hr after a loading dose of 25 mgkg in 5 minutes (n=4). Neither i.v. infusion of thymidine (25 mg/kgAu, n=5; 100 mg/kg/hr, n=2) nor acetazolamide (50 mgkg i.p. bolus followed by 25 mgkg i.p. every second hour, n=3) infhtenced the brain/blood AUC ratio after alovudine 25 mgkg S.C. injection compared to controls (n=5). Finally, perfusion through the microdialysis probe with thymidine (1000 pM, n=3) had also no effect on the brain/blood AUC ratio after alovudine 25 mgkg S.C. Because neither thymidine nor acetazolamide has significant influence on the ability of alovudine to penetrate the blood-brain barrier in the rat, neither thymidine transport nor carboanhydrase dependent CSF production appear to be major dete rminants of the blood-brain concentration gradient. Thus, it is concluded that alovudine reaches the extracellular fluid of the brain not by cerebrospinal fluid, but via the cerebral capillaries and that the existence of a concentration gradient over both blood-brain and CSF-brain barrier can probably be explained by the presence of an active process pumping alovudine out from the brain. Key Words:
alovudine, blood-brain barrier, thymidine, acetazolamide, microdialysis, rat
A major clinical problem in untreated HIV infected patients is the presence of neuropsychological symptoms in a large proportion of the patients (1). It is therefore important that anti-HIV agents are able to penetrate into the brain, not only because of the CNS (central nervous system) symptoms but also because the brain is one of the potential sanctuaries of the virus (2). Microdialysis, a method for continuous monitoring of unbound extracellular drug concentrations in sot? tissues, has previously been used to investigate the distribution of a range of nucleosides analogues into the brain (3, 4, 5, 6). The ability of alovudine and zidovudine to cross the blood-brain barrier (BBB) has previously been studied after S.C. injection in rats (5) and monkeys (4) by microdialysis. In these studies, it was shown that a blood-brain concentration gradient is actively maintained. Recently it was shown that the size of this gradient is a function of the electronic distribution properties of 5-substituents, such as CORRESPONDING AUTHOR: Natalia Borg, Dept. Clinical Pharmacology, Huddinge University Hospital, SE-141 86 Huddinge, Sweden. Tel: +46 8 58581068. Fax: +46 8 58581070. E-mail : [email protected].
1806
Alovudine Transport into CNS and CSF
Vol.
66, No. 19,200O
electronic parameters q,, and crp, rather than the lipophilicity of a series of thymidine analogues (3). This suggests that a specific transport system, rather than passive lipophilicity dependent diffusion, is a major determinant of the transport across the BBB of thymidine analogues. There are two different routes by which substances reach the central nervous system (7,8). The first is by direct transport across the cerebral capillaries into the brain. The special structural features of the ccrebrovascular endothelium, such as absence of fenestrations between endothelial cells and presence of tight junctions, a large number of mitochondria and a special role of astrocytes, are all responsible for the selective transport from the blood to the brain. The other possibility to reach CNS is via CSF, which fills the ventricular and subarachnoidal spaces in the brain. CSF is a product of active secretion processes mainly at the choroid plexus. A substantial proportion of the CSF is, however, formed at the brain capillaries and is subject to bulk transport to the ventricular and subarachhnoid spaces (9). The blood-CSF barrier is not as restrictive for drug transport as BBB. There is a free exchange of fluid between CSF and the extracellular space of the brain. At the choroid plexus the production of CSF can be reduced or inhibited by carbonic anhydrase inhibitors such as acetazolamide. It has previously been shown that acetazolamide reduces production of CSF by 55% in rats (10). Thus, treatment with acetazolamide is expected to reduce the brain concentration of drugs using the CSF to reach the brain and increase the brain concentration of drugs eliminated through the CSF. The aim of this study was to further investigate the uptake of alovudine into the brain and the CSF after a single S.C. dose and during continuous i.v. infusion. The intluence of coadministration of acetazolamide and thymidine on the blood-brain concentration gradient of alovudine was also a subject to investigation in order to elucidate possible transport mechanisms of alovudine over the BBB. Alovudine, an anti-HIV drug which was developed into phase II cinical trials, but has been stopped because of its toxicity, has been chosen because of its low protein binding, linear kinetics and absence of metabolism in the rat. Methods Subjects: Male Sprague-Dawley rats (weight, 180-550 g) were used throughout (B&K Universal AB, Sollentuna, Sweden). The rats had free access to tap water and standard lab chow and were housed 5-6 per cage. Each rat was used once only. These experiments were approved by the Local Ethical Committee of the Huddinge University Hospital. Antiviral drugs: Alovudine was a generous gift from Medivir AB, Huddinge, Sweden. Acetazolamide (Diamox@, Lederle Paranterals, Carolina, Puerto Rico, USA) was obtained from the hospital pharmacy. Thymidine was obtained from Sigma-Aldrich Sweden AB, Stockholm, Sweden. The compounds were dissolved in a Ringer solution (Pharmacia & Upjohn AB, Stockholm, Sweden) to a concentration of 25 mg/ml (alovudine and thymidine) and 50 mg/ml (acetazolamide). Ringer’s solution, containing 8.6 g NaCl, 300 mg KC1 and 330 mg CaCli2H,O per liter sterile water, was obtained from the hospital pharmacy. Microdialysis: Detailed accounts of the microdialysis method can be found elsewhere (11, 12, 13, 14). In the present study, rats were anaesthetised by halothane during the whole experiment and were placed in a David Kopf stereotaxic instrument with the bite bar 2.5 mm below the interaureal line. Dialysis probes (0.50 mm diameter, 4.0 or 10.0 mm long membrane) of the concentric type (CMA Microdialysis, Stockholm, Sweden) were implanted into the corpus striatum (4.0 mm long membrane) on one side of the brain (stereotaxic coordinates 2.2 mm lateral and 1.3 mm anterior to the bregma and 7.5 mm ventral to the brain surface), in the jugular vein (4.0 or 10.0 mm long) and into the cistema magna (4.0 mm long membrane) by
Vol. 66, No. 19,200O
Alovudine Transport into CNS and CSF
1807
insertion parallel to the occipital bone behind the cerebellum. Each dialysis probe was perfused with a Ringer solution at a rate of 1 ul/min or 2 pVmin samples of 20 pl or 40 pl were collected at 20 min intervals. The in vivo recovery of each dialysis probe was determined in the following way. After implantation and a 30 min washout period the perfusion medium was switched to a 10 pM alovudine solution. Three samples were collected and the perfusion medium was switched back to the Ringer solution. After a second washout period of 40-60 min no drug was detectable in the dialysates. The proportion of drug lost over the dialysis membrane was taken as an estimate of the in vivo recovery over the dialysis membrane (12, 13). After the second washout period alovudine was administered systemically and samples were collected postinjection for 180 min. The concentration measured in the dialysate was converted to estimates of the unbound extracellular concentration by means of the in vivo recovery. The different designs (e.g. different perfusion flow, fraction volume, length of the microdialysis membrane) of the microdialysis experiments in this study is due to the prolonged period of time over which this study was carried out, seven years. Thus, over time the probe design was changing to improve the reproducibility and the HPLC analysis of alovudine has been optimised. Analysis of alovudine: Analysis was made by isocratic HPLC separation and W-detection (254 nm). Two different columns were used in the study: a 100 x 4.0 mm 5 micron particle size C,, column (BAS Technical Ltd, Stockport, UK) and a 75 x 3.2 mm 3.5 micron particle size C,, column (Zorbax@, Hewlett Packard, USA). The mobile phase consisted of 0.05 M ammonium phosphate at pH=6.0, containing 20% (v/v) methanol (15). Alovudine was easily separated from endogenous compounds in the brain, blood and in the CSF dialysates. Under the conditions described, the lower level of detection for alovudine was 0.3 pM. Calibration curves, done in duplicate, were processed over the concentration range OS-100 PM. Linear regression analysis of the peak areas of alovudine gave a correlation coefficient of 0.99. Experimental design and statistical analysis: The following experiments were performed: Experiment 1: Constant i.v. infusion of alovudine 25 mg/kg/hr after a loading dose of 25 mg/kg in 5 minutes (n=4). Experiment 2: S.C. injection of alovudine 25 mg/kg (n=5). Experiment 3: S.C. injection of alovudine 25 mg/kg in rats during an iv. infusion of thymidine 25 mg/kg/hr after a loading dose of 25 mg/kg in 5 minutes (n=5) which was started prior to recovery determination. Experiment 4: SC. injection of alovudine 25 mg/kg in rats during an iv. infusion of_thymidine 100 mg/l&hr after a loading dose of 100 mg/kg in 5 minutes (n=2) which was started prior to recovery determination. Experiment 5: SC. injection of alovudine 25 mg/kg during perfusion through the microdialysis probe with 1000 pM thymidine (n=3). In experiment 5, two microdialysis probes were implanted in the brain in each animal (symmetrically). One probe was perfused by the thymidine solution, the other by Ringer’s solution serving as a control. Experiment 6: S.C. injection of alovudine 25 mg/kg in rats treated with acetazolamide 50 m&g i-p. bolus and 25 mg/kg i.p. every second hour during the whole experiment starting acetazolamide injections 30 min prior to recovery determination (n=3). Data Analysis and Statistics: The microdialysis samples are collected in 20 min fractions which can be represented in the graphs as the midpoint of the sampling period (i.e. 10,
1808
Alovudine Transport into CNS and CSF
Vol. 66, No. 19,200O
(area under the time versus concentration 30 ,... 170 min) (16, 17, 18, 19). The AU&.,,, curve) was calculated by summing the products between the sampling interval and the sample concentration (corrected for recovery). The AUC ratio was calculated as AU&_,,, (brain)/AUC,,,W (blood) or as AUC,,,tin (CSF)/AUC,,,,, (blood). The half-life was calculated as lfi/slope where the slope of the elimination phase was determined by linear regression. The clearance was calculated as dose/AUC, for blood and the volume of distribution was calculated as CL/slope. Data are presented as meanskstandard errors of means or as geometric mean when appropriate. Two-sided Student’s t-test for dependent samples was used to compare AUC ratio brain/blood and CSFIblood, Cmax in the brain and CSF, half-lives in the brain and CSF. For this test the half-life was log-transformed to get symmetrically distributed data. The AUC ratio brain/blood and CSFIblood and the difference between C, in the brain and the CSF were compared among experiments by one-way ANOVA. Hypothesis testing was performed at the 5% level. Results Experiment 1: Constant i.v. infusion of alovudine. Steady state levels of alovudine were rapidly attained in blood, cerebrospinal fluid and the brain after the loading dose of 25 mgkg in 5 min and the start of continuous i.v. infusion. The mean time-concentration curves are given in Fig. 1. The mean in vivu recoveries of alovudine at a flow rate of 1 pL/min were 0.187 (0.079), 0.222 (0.078) and 0.458 (0.088) in the brain, CSF and blood respectively. The dialysis membranes used here was 4 mm. The AUC brain/blood and CSF/blood were found to be 0.164f0.03 1 and 0.402M.082 respectively. The difference in AUC ratio brain/blood and AUC ratio CSF/blood is statistically significant.
0
20
40
00
80
100
120
140
160
180
Tlme,min
Fig. 1 Time-log(concentration) curves (geomean+SEM) for CNS (circles), CSF (squares) and blood (triangles) from rats treated by iv. infusion of alovudine (25 mgkg/hr after a loading dose of 25 mgkg in 5 minutes) (n=4). The concentration is free (unbound) extracellular concentratioa, corrected for in viva recovery at 1 @/min.
Alovudine Transport into CNS and CSF
Vol. 66, No. 19,200O
1809
Experiment 2: S.C.injection of alovudine. The mean time-concentration curves for alovudine are given in Fig. 2. The mean in vivo recoveries of alovudine at a flow rate of 2 pL/min were 0.034 (0.003), 0.054 (0.005) and 0.137 (0.033) in the brain, CSF and blood respectively. The membrane length used in this experiment was 4 mm. The C, was 17.0k1.8 pM for the brain, 25.9f3.7 uM for the CSF and 104.2f27.6 pM for blood. The T, was found to be from 30 to 50 min for the brain and 30 min for the CSF and blood. Alovudine was eliminated with half-lives of 1.7+0.5h from the brain, 1.0&O.1h from the CSF and 0.8+0.lh from blood. CL calculated for blood concentration data was 0.76+0.13L/hkg. Vd &as 0.85fO.l4L/kg. The AUC ratios (O-180 min) between brain and blood and between CSF and blood were 0.285kO.058 and 0.327fo.051 respectively. The analysis of the pooled data from experiments 2,3,4 showed that the difference in C, between the brain and the CSF and the difference in half-lives between the brain and the CSF are statistically significant, while the difference between AUC ratio brain/blood and AUC ratio CSF/blood is close to significant (p=O.O56).
loo-
_
Lh
5
~lo_ 1
! 0
I
20
40
60
1
,
80
100
120
140
160
180
Time, min
Fig. 2 Timelog(concentration) curves (geomean&SEM) for CNS (circles), CSF (squares) and blood (triangles) from rats treated with alovudine (25 mgkg S.C. injection) (n=5). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 2 uL/min.
Experiment 3 and 4: S.C.injection of alovudine during i.v. infusion of thymidine. The mean time-concentration curves for alovudine during i.v. infusion of thymidine 25 mg/kg/h are given in Fig. 3. The mean in vivo recoveries of alovudine at a flow rate of 2 uL/min were 0.043 (0.004), 0.106 (0.025) and 0.268 (0.021) in the brain, CSF and blood respectively. The membrane length used in this experiment was 4 mm. The C, was 9.5f2.1 PM for the brain, 17.M3.6 uM for the CSF and 58.1f2.9 FM for blood. The T_ was 50 min for the brain and 30 mm for CSF and blood. Alovudine was eliminated with half-lives of 1. HO. 1 h, 0.9&O.1 h and 0.7&O.1 f?om the brain, CSF and blood respectively. CL calculated for
Vol. 66, No. 19,200O
Alovudine Transport into CNS and CSF
1810
Fig. 3 Time-log(concentration) curves (geomeanLSEM) for CNS (circles), CSF (squares) and blood (triangles) from rats treated with alovudine (25 mgkg S.C. injection) and thymidine (25 mgkgh i.v. afkr a loading dose of 25 mg/kg in 5 minutes) (n=5). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 2 pL/min.
1
1
I
0
I
20
40
60
60 Time,
100
120
140
160
180
min
Fig. 4 Time-log(concentration) curves (geomean+SEM) for CNS (circles), CSF (squares) and blood (triangles) from rats treated with alovudine (25 mgkg S.C. injection) and thymidine (100 mg/kg/h i.v. after a loading dose of 100 mgkg in 5 minutes) (n=2). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 2 pL/min.
Alovudine Transport into CNS and CSF
Vol. 66, No. 19,200O
1811
blood concentration data was 1.22fO. 12 L/h/kg. Vd was 1.23fO. 1 L/kg. The AUC ratios (O-l 80 min) brain to blood and CSF to blood were 0.241rtO.059 and 0.348f0.046 respectively. The mean time-concentration curves for alovudine during i.v. infusion of thymidine 100 mgkg /h are given in Fig. 4. The mean in vivo recoveries of alovudine at a flow rate of 2 uL/min were 0.058 (0.012), 0.055 (0.029) and 0.249 (0.049) in the brain, CSF and blood respectively. The membrane length used in this experiment were 4 mm. The C, was 12.2k5.4 pM for the brain, 19.7k12.1 pM for the CSF and 64.2k14.1 ).tM for blood. The T, was 50 min for the brain and 30 min for CSF and blood. Alovudine was eliminated with half-lives of l.OkO.2 h, 0.7f0.02 h and 0.7kO.l h from the brain, CSF and blood respectively. CL calculated for blood concentration data was 1.18M.16 LMcg. Vd was 1.13kO.25 L/kg. The AUC ratios (O-180 min) between brain and blood and between CSF and blood were 0.232kO.063 and 0.34M0.170 respectively. The difference in C,,, the brain/blood AUC ratio and the CSF/blood AUC ratio between the CSF and the brain, were compared between experiments 2,3 and 4, and no statistically significant difference was found.
0
20
40
60
60
100
120
140
160
180
Time, min
Fig. 5. Time-log(concentration) curves (geomeaniSEM) for blood (triangles) and for CNS with (squares) and without (circles) perfusion through the microdialysis probe with thymidine 1000 pM (n=3) from rats treated with alovudine (25 mgkg S.C. injection). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 1 uL/min.
Vol. 66, No. 19,200O
Alovudine Transport into CNS and CSF
1812
Experiment 5: S.C.injection of alovudine during perfusion through the microdialysis probe with thymidine. The mean time-concentration curves for alovudine are given in Fig. 5. The mean in vivo recoveries of alovudine at a flow rate of 1 uL/min were 0.127 (0.009) and 0.142 (0.015) in the brain (probes perfused with Ringer solution and probes perfused with thymidine solution, respectively) and 0.275 (0.085) in blood. The membrane length used in this experiment was 4 mm. The C, obtained from the probes perfused by thymidine solution was 20.4k1.7 pM and in the dialysate obtained from the probes perfused by Ringer’s solution the C_ were 17.9k2.5 PM for the brain and 125.Ok4.7 PM for blood. The T, was 30 to 70 min for the brain (both probes) and 30 min for blood. Alovudine was eliminated with half-lives of 3.OkO.8 h (probes perfused with thymidine solution), 2.OzkO.l h (probes permsed with Ringer’s solution) and 1.2fO.l h from the blood. CL calculated for blood concentration data was 0.43f0.06 L/h/kg. Vd was 0.69zkO.05 L/kg. The AUC ratio (O-180 min) between brain and blood was 0.218kO.018 (control probes) and 0.248f0.014 (probes perfused by thymidine solution). The difference in AUC ratios with respect to the perfusion medium was not significant.
1
1 0
!
20
40
60
60
100
120
140
160
180
Time,min
Fig. 6 Time-log(concentration) curves (geomean+SEM) for CNS (circles) and blood (squares) from rats treated with alovudine (25 mgikg S.C. injection) and acetazolamide (50 mgkg i.p. bolus and 25 mgikg i.p. every second hour during the whole experiment (n=3). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 1 uL/min.
Experiment 6: S.C.injection of alovudine during acetazolamide treatment. The mean time-concentration curves for alovudine are given in Fig. 6. The mean in vivo recoveries of alovudine at a flow rate of 1 uL/min were 0.115 (0.024) and 0.648 (0.058) in the brain and blood respectively. The membrane lengths used in this experiment were 4 (brain) and 10 (blood) mm. The C,, was 17.2k4.5 uM for the brain and 107.7k12.5 PM for blood. The T_ was 70 min for the brain and from 30 to 50 min for blood. Alovudine was eliminated with
Vol.66, No.
19,200O
AlovudineTransport intoCNSandCSF
1813
half-lives of 3.2f0.5 h from the brain and 1.5f0.3 h from blood. CL calculated for blood concentration data was 0.47kO.01 L/h/kg. Vd was 1.56f0.62 L/kg. The AUC ratio (O-180 min) between brain and blood was 0.179kO.032. The differences in brain/blood AUC ratio, C, and half-life between this experiment and experiment 2 (control group) are not statistically significant.
Discussion
The ability of anti-HIV drugs to penetrate the BBB is necessary to suppress virus replication in the brain tissue. Recent clinical data suggest that the concentrations attained in CSF of zidovudine (20) and indinavir (21) are sufficient to reduce HIV-RNA in this compartment (22). To understand the mechanisms of nucleoside analogue transport across the blood-brain barrier is thus of considerable interest. The anatomical and physiological properties of the BBB are responsible for the restricted transport of biologically active compounds, including drugs (7, 8). Many pharmacologically active compounds use the transport systems normally used for endogenous substances and nutrients (10, 23, 24, 25, 26, 27). As mentioned above, there are two routes by which a drug can enter and/or leave the brain: via the endothelium of cerebral capillaries and via the CSF. If a substance crosses the BBB via CSF, it should result in higher concentration of the drug in the CSF compared to extracellular concentration in the brain, which is what was found in this study. In experiment 2, when alovudine was given subcutaneously, the levels are different in the brain and CSF. The C, of alovudine in the CSF was significantly higher than the C, in the brain extracellular fluid in the analysis of the pooled experiments 2, 3 and 4. However, the half-life of alovudine in the CSF is significantly shorter than that in the brain making the difference in AUC ratio brain/blood V.S.CSF/blood only borderline significant. The concentration gradient between the extracellular fluid in the brain and the CSF found in experiment 1, remained constant during the intravenous infksion of alovudine. Therefore, slow transport by passive diffusion over capillary membranes or the choroid plexus can not alone explain the observation that brain concentrations of thymidine analogues decline during the elimination phase despite higher concentrations in blood. Had transport from CSF to the brain tissue been the major mode of transport, the concentration gradient would decline with time. Thus it is necessary to postulate the existence of an active process which maintains the concentration gradient. Similar results have been obtained for zidovudine in the rabbit (28). In that study the brain/plasma and CSF/plasma concentration ratios (0.19 and 0.29 respectively) suggest the existence of active transport of zidovudine out from the brain. However, intracerebroventricular infhsion of zidovudine was accompanied by a much higher CSF/brain concentration ratio suggesting that transport from CSF to the brain is of minor importance when the drugs are systemically administered (2). The brain regions studied were different in our study and in the rabbit study, where the thalamus was sampled, while we inserted microdialysis probes into the basal ganglia. The similarity of the results obtained suggests that different gray matter regions are similar with ‘respect to thymidine analogue transport. This is important because the basal ganglia are relevant from the virological point of view being a main location of HIV infection in the brain (1). In the present study, the influence of CSF production on the penetration of alovudine was also investigated. By administration of acetazolamide, production of cerebrospinal fluid can be
1814
Alovudine Transport into CNS and CSF
Vol. 66, No. 19,200O
reduced by 55% (lo), which should reasonably influence the concentration levels in the brain in case a substance utilises transport via CSF into or out from the brain tissue. However, the distribution to the brain of alovudine, estimated by the ratio of the AUC between brain and blood was 0.179 in the presence of acetazolamide and 0.285 in the control group. This difference is not statistically significant even if the control data are pooled with data from thymidine treated rats. Neither was C,, in the brain different from the control group. However, the elimination half-life of alovudine from the brain in the presence of acetazolamide was significantly slower @=0.003) compared to pooled data from experiments 2, 3 and 4. One explanation to this finding could be that alovudine is, partly, transported out from the brain via the flow of CSF as has been suggested for theophylline (13). Nevertheless, the major route of transport of alovudine over the blood-brain barrier appears to be via the cerebral capillaries. We have previously found that there is no correlation between lipophilic properties and the blood brain concentration gradient for uridine analogoues in the rat suggesting that the mechanism maintaining the gradient is not primarily dependent upon passive diffision over a lipophilic barrier. Instead, the concentration gradient was related to the electronic distribution in the substituents of these compounds (3). This tinding indirectly suggests that the blood brain concentration gradient depends on the affinity to an energy requiring transport protein. The existence of many active transport systems has been reported in the literature. Transport systems for purines (adenine and guanine) have been described and pyrimidines (thymidine, cy-tidine and utidine) have been reported to utilise a specific nucleoside carrier (23, 24, 25, 26, 27). An energy-dependent, saturable high-affinity transport system for thymidine has been found in the rabbit brain (23). It was hypothesized that the thymidine analogues use the same transporter for penetration over the BBB, but the result from experiments number 3,4,5 do not support that hypothesis. In these experiments we have used thymidine in an attempt to saturate the thymidine transporter from either side of the brain capillaries and thereby inhibit alovudine transport into or out from the brain. Thus, both i.v. inmsion of thymidine in different doses (25 mg/kg/h and 100 mg/kg/hr) and perfusion with thymidine 1000 uM through the probe was used. The doses chosen were based on the basal thymidine concentration being less than 0.1 uM both in the brain extracellular fluid and in the CSF and around 0.2 pM in plasma in humans (29). The AUC ratios brain/blood and CSF/blood in the presence of thymidine showed no tendency to a change compared to the corresponding AUC ratios in the control group (experiment 2). Neither did local administration of thymidine by its inclusion in the perfusion medium of the microdialysis probe have an effect on the AUC ratio. Thus, applying thymidine on either side of the blood-brain barrier in high concentrations had no effect on alovudine transport. The pharmacokinetic properties of alovudine were found to be in the range expected for thymidine analogues and are stmunarised in Table 1. No differences in plasma pharmacokinetic parameters of alovudine among experiment groups were found suggesting that peripheral pharmacokinetic changes do not influence the interpretation of the experiments. A corrollary of the result from this study with respect to the concentration gradient of alovudine between the extracellular fluid in the brain and the cerebrospinal fluid, is that clinical data on CSF concentration can not be interpreted as equal to the brain concentration. The rabbit data on zidovudine (2, 28) suggests that the difference between CSF and the brain holds for thymidine analogs and it might be valid for a wide range of drugs. However, brain extracellular levels of zidovudine and alovudine are higher relative to plasma levels in cynomolgus monkeys than in rats (4, 5). Therefore, species differences can make the difference between CNS and CSF less important in some species.
Alovudine
Vol. 66, No. 19,200O
Transport
into CNS and CSF
1815
Thus, we conclude that alovudine reaches the extracellular fluid of the brain not via the cerebrospinal fluid, but via the cerebral capillaries. The blood-brain and CSF-brain concentration gradients may be maintained by an active capillary process pumping alovudine out from the brain. This pump was not inhibited by thymidine or by inhibiting CSF production.
TABLE 1. Experiment Control
CNS
Cmax, pM 17.0
Half-life, h 1.7
CL, L/kg/h
CSF
25.9
1.0
Blood
104.2
0.8
CNS
9.5
1.1
0.241
CSF
17.0
0.9
0.348
Blood
58.1
0.7
CNS
12.2
1.0
0.232
CSF
19.7
0.7
0.340
Blood
64.2
0.7
CNS
17.9
3.0
0.218
CNS*
20.4
2.0
0.248
Blood
125.0
1.2
CNS
17.2
3.2
Blood
107.7
1.5
Yd, L/kg
AUC ratio 0.285
group
Uovudine S.C. +THY 25 mg/kg*h
4lovudine S.C. +THY 100 mg/kg*h
Alovudine s.c.+lOOOuM THY in perfusion medium (*)
Alovudine s.c.+ ketaxolamid e i.p.
Summary of pharmacokinetic
0.327 0.76
1.22
1.18
0.43
0.85
1.23
1.13
0.69 0.179
0.47
1.56
properties (mean) of alovudine.
1816
Alovudine
Transport
into CNS and CSF
Vol. 66, No. 19,ZOOO
Acknowledgement The present study was supported by the Karolinska Institute, the Swedish Medical Research Council (grants no. 09069 and 12590), Swedish Physicians Against AIDS and Medivir AB. The expert technical assistance of Ewa Guzenda is gratefully acknowledged.
References 1. C. WILEY. Psychiatric Clinics ofNorth America. 17 1-15 (1994). 2. Y. WANG and R.J. SAWCHUK, J. Pharm. Sciences. 84 871-6 (1995). 3. N. BORG, X-X. ZHOU, N-G. JOHANSSON, B. dBERG and L. STAI-ILE, Antivir. Chem. & Chemother 8 47-53 (1997). 4. E. LJUNGDAHL-STAHLE, E. GUSENDA, D. BOTTIGER, B. WAHREN, B. OBERG and L. STAI-ILE, Antimicrob. Agents & Chemother. 36 2418-2422 (1992). 5. L. ST.&I-ILE, E. GUZENDA and E. LJUNGDAHL-ST&ILE, J. AIDS. 6 435-439 (1993). 6. L. STAI-ILE and B.oBERG, Antimicrob. Agents & Chemother. 36 339-342 (1992). 7. M. BRADBURY, The concept of a blood-brain barrier. London:Wiley. 1979. 8. J. B. M. M. VAN BREE, Thesis from the Division of Pharmacology of the Center BioPharmaceutical Science, Leiden University, The Netherlands (1990). 9. T.H. MILDHORAT, M.K. HAMMOCK, J.D. FENSTERMACHER and V.A. LEVIN, Science 173 330-332 (1971). 1O.B. VOGH and D.R. GODMAN, Canadian J. Physiol. & Pharmacol. 63 340-3 (1985). 11 .L. STAI-ILE, S. SEGERVARD and U. UNGERSTEDT, Eur. J. Pharmacol. 185 187-193 (1990). 12. L. ST&ILE, S. SEGERVARD and U. UNGERSTEDT, J. Pharmacol. Meth. 25 41-52 (1991). 13. L. ST&ILE, S. SEGERV.&RD and U. UNGERSTEDT, Life Sciences 49 1843-1852 (1991). 14. U. UNGERSTEDT, Measurement of neurotransmitter release by intracranial dialysis. In: Marsden CA (ed) Measurement of neurotransmitter release in vivo. 8 I-105 Wiley, London, (1984). 15. L. STAHLE, A. MIAN and N. BORG, J. Pharmaceut. & Biomed. Analysis 13 369-376 (1995). 16. L. STAI-ILE, Eur. J. Clin. Pharmacol. 43 289-294 (1992). 17. L. STAHLE, Eur. J. Drug Metab. Pharmacokin. 18 89-96 (1993). 18. L. STAHLE, Use of microdialysis in pharmacokinetics and pharmacodynamics. In: Robinson T Justice J (eds) Methods in Neuroscience, 155- 174 Elsevier, Amsterdam (199 1). 19. L. ST&ILE, Eur. J. Clin. Phatmacol. 45 477-481 (1993). 20. T.A. TARTAGLIONE, A.C. COLLIER, R.W. COOMBS, K.E. OPHEIM, D.K. CUMMINGS, S.R. MACKAY, J. BENEDETTI and L. CREY, Arch. Neurol. 48 695-699 (1991). 21. L. ST&ILE, C. MARTIN, J.O. SVENSSON and A. SilNNERBORG, Lancet 350 20. (1997). 22. M. GISSLEN, G. NORKRANS, B. SVENNERHOLM and L. HAGBERG, J. Infect. Diseas. 175 434-437 (1997). 23. R. SPECTOR and W.G. BERLINGER, J. Neurochem. 39 837-841 (1982). 24. R. SPECTOR, J. Neurochem. 35 1092-1098 (1980). 25. R. SPECTOR, J. Neurochem. 45 1411-1418 (1985). 26. R. SPECTOR and S. HUNTOON, J. Neurochem. 40 1474-1480 (1983). 27. R. SPECTOR and S. HUNTOON, J. Neurochem. 41 1131-l 136 (1983). 28. Y. WANG, Y. WE1 and R.J. SAWCHUK, J. Pharm. Sciences. 86 1484-1490 (1997). 29. J.T. EELLS and R. SPECTOR, Neurochem. Research. 8 1451-1457 (1983).
PI1 SO0243205(00)00504-X
TRANSPORT OF ALOVUDINE (3’-FLUORO THYMIDINE) INTO THE BRAIN AND THE CEREBROSPINAL FLUID OF THE RAT, STUDIED BY MICRODIALYSIS Lars St&le and Natalia Borg. Department
of Clinical Pharmacology, Huddinge University Huddinge, Sweden.
Hospital, SE-14186,
(Received in final form December 7, 1999) Summary
Extracellular unbound concentrations of alovudine were sampled by microdialysis in order to study the transport of alovudine between the blood and the brain and the cerebrospinal fluid (CSF) in the rat. The AUC (area under the curve) ratio CSF/blood was higher than the brain/blood ratio after iv. infusion of alovudine 25mg/kg/hr after a loading dose of 25 mgkg in 5 minutes (n=4). Neither i.v. infusion of thymidine (25 mg/kgAu, n=5; 100 mg/kg/hr, n=2) nor acetazolamide (50 mgkg i.p. bolus followed by 25 mgkg i.p. every second hour, n=3) infhtenced the brain/blood AUC ratio after alovudine 25 mgkg S.C. injection compared to controls (n=5). Finally, perfusion through the microdialysis probe with thymidine (1000 pM, n=3) had also no effect on the brain/blood AUC ratio after alovudine 25 mgkg S.C. Because neither thymidine nor acetazolamide has significant influence on the ability of alovudine to penetrate the blood-brain barrier in the rat, neither thymidine transport nor carboanhydrase dependent CSF production appear to be major dete rminants of the blood-brain concentration gradient. Thus, it is concluded that alovudine reaches the extracellular fluid of the brain not by cerebrospinal fluid, but via the cerebral capillaries and that the existence of a concentration gradient over both blood-brain and CSF-brain barrier can probably be explained by the presence of an active process pumping alovudine out from the brain. Key Words:
alovudine, blood-brain barrier, thymidine, acetazolamide, microdialysis, rat
A major clinical problem in untreated HIV infected patients is the presence of neuropsychological symptoms in a large proportion of the patients (1). It is therefore important that anti-HIV agents are able to penetrate into the brain, not only because of the CNS (central nervous system) symptoms but also because the brain is one of the potential sanctuaries of the virus (2). Microdialysis, a method for continuous monitoring of unbound extracellular drug concentrations in sot? tissues, has previously been used to investigate the distribution of a range of nucleosides analogues into the brain (3, 4, 5, 6). The ability of alovudine and zidovudine to cross the blood-brain barrier (BBB) has previously been studied after S.C. injection in rats (5) and monkeys (4) by microdialysis. In these studies, it was shown that a blood-brain concentration gradient is actively maintained. Recently it was shown that the size of this gradient is a function of the electronic distribution properties of 5-substituents, such as CORRESPONDING AUTHOR: Natalia Borg, Dept. Clinical Pharmacology, Huddinge University Hospital, SE-141 86 Huddinge, Sweden. Tel: +46 8 58581068. Fax: +46 8 58581070. E-mail : [email protected].
1806
Alovudine Transport into CNS and CSF
Vol.
66, No. 19,200O
electronic parameters q,, and crp, rather than the lipophilicity of a series of thymidine analogues (3). This suggests that a specific transport system, rather than passive lipophilicity dependent diffusion, is a major determinant of the transport across the BBB of thymidine analogues. There are two different routes by which substances reach the central nervous system (7,8). The first is by direct transport across the cerebral capillaries into the brain. The special structural features of the ccrebrovascular endothelium, such as absence of fenestrations between endothelial cells and presence of tight junctions, a large number of mitochondria and a special role of astrocytes, are all responsible for the selective transport from the blood to the brain. The other possibility to reach CNS is via CSF, which fills the ventricular and subarachnoidal spaces in the brain. CSF is a product of active secretion processes mainly at the choroid plexus. A substantial proportion of the CSF is, however, formed at the brain capillaries and is subject to bulk transport to the ventricular and subarachhnoid spaces (9). The blood-CSF barrier is not as restrictive for drug transport as BBB. There is a free exchange of fluid between CSF and the extracellular space of the brain. At the choroid plexus the production of CSF can be reduced or inhibited by carbonic anhydrase inhibitors such as acetazolamide. It has previously been shown that acetazolamide reduces production of CSF by 55% in rats (10). Thus, treatment with acetazolamide is expected to reduce the brain concentration of drugs using the CSF to reach the brain and increase the brain concentration of drugs eliminated through the CSF. The aim of this study was to further investigate the uptake of alovudine into the brain and the CSF after a single S.C. dose and during continuous i.v. infusion. The intluence of coadministration of acetazolamide and thymidine on the blood-brain concentration gradient of alovudine was also a subject to investigation in order to elucidate possible transport mechanisms of alovudine over the BBB. Alovudine, an anti-HIV drug which was developed into phase II cinical trials, but has been stopped because of its toxicity, has been chosen because of its low protein binding, linear kinetics and absence of metabolism in the rat. Methods Subjects: Male Sprague-Dawley rats (weight, 180-550 g) were used throughout (B&K Universal AB, Sollentuna, Sweden). The rats had free access to tap water and standard lab chow and were housed 5-6 per cage. Each rat was used once only. These experiments were approved by the Local Ethical Committee of the Huddinge University Hospital. Antiviral drugs: Alovudine was a generous gift from Medivir AB, Huddinge, Sweden. Acetazolamide (Diamox@, Lederle Paranterals, Carolina, Puerto Rico, USA) was obtained from the hospital pharmacy. Thymidine was obtained from Sigma-Aldrich Sweden AB, Stockholm, Sweden. The compounds were dissolved in a Ringer solution (Pharmacia & Upjohn AB, Stockholm, Sweden) to a concentration of 25 mg/ml (alovudine and thymidine) and 50 mg/ml (acetazolamide). Ringer’s solution, containing 8.6 g NaCl, 300 mg KC1 and 330 mg CaCli2H,O per liter sterile water, was obtained from the hospital pharmacy. Microdialysis: Detailed accounts of the microdialysis method can be found elsewhere (11, 12, 13, 14). In the present study, rats were anaesthetised by halothane during the whole experiment and were placed in a David Kopf stereotaxic instrument with the bite bar 2.5 mm below the interaureal line. Dialysis probes (0.50 mm diameter, 4.0 or 10.0 mm long membrane) of the concentric type (CMA Microdialysis, Stockholm, Sweden) were implanted into the corpus striatum (4.0 mm long membrane) on one side of the brain (stereotaxic coordinates 2.2 mm lateral and 1.3 mm anterior to the bregma and 7.5 mm ventral to the brain surface), in the jugular vein (4.0 or 10.0 mm long) and into the cistema magna (4.0 mm long membrane) by
Vol. 66, No. 19,200O
Alovudine Transport into CNS and CSF
1807
insertion parallel to the occipital bone behind the cerebellum. Each dialysis probe was perfused with a Ringer solution at a rate of 1 ul/min or 2 pVmin samples of 20 pl or 40 pl were collected at 20 min intervals. The in vivo recovery of each dialysis probe was determined in the following way. After implantation and a 30 min washout period the perfusion medium was switched to a 10 pM alovudine solution. Three samples were collected and the perfusion medium was switched back to the Ringer solution. After a second washout period of 40-60 min no drug was detectable in the dialysates. The proportion of drug lost over the dialysis membrane was taken as an estimate of the in vivo recovery over the dialysis membrane (12, 13). After the second washout period alovudine was administered systemically and samples were collected postinjection for 180 min. The concentration measured in the dialysate was converted to estimates of the unbound extracellular concentration by means of the in vivo recovery. The different designs (e.g. different perfusion flow, fraction volume, length of the microdialysis membrane) of the microdialysis experiments in this study is due to the prolonged period of time over which this study was carried out, seven years. Thus, over time the probe design was changing to improve the reproducibility and the HPLC analysis of alovudine has been optimised. Analysis of alovudine: Analysis was made by isocratic HPLC separation and W-detection (254 nm). Two different columns were used in the study: a 100 x 4.0 mm 5 micron particle size C,, column (BAS Technical Ltd, Stockport, UK) and a 75 x 3.2 mm 3.5 micron particle size C,, column (Zorbax@, Hewlett Packard, USA). The mobile phase consisted of 0.05 M ammonium phosphate at pH=6.0, containing 20% (v/v) methanol (15). Alovudine was easily separated from endogenous compounds in the brain, blood and in the CSF dialysates. Under the conditions described, the lower level of detection for alovudine was 0.3 pM. Calibration curves, done in duplicate, were processed over the concentration range OS-100 PM. Linear regression analysis of the peak areas of alovudine gave a correlation coefficient of 0.99. Experimental design and statistical analysis: The following experiments were performed: Experiment 1: Constant i.v. infusion of alovudine 25 mg/kg/hr after a loading dose of 25 mg/kg in 5 minutes (n=4). Experiment 2: S.C. injection of alovudine 25 mg/kg (n=5). Experiment 3: S.C. injection of alovudine 25 mg/kg in rats during an iv. infusion of thymidine 25 mg/kg/hr after a loading dose of 25 mg/kg in 5 minutes (n=5) which was started prior to recovery determination. Experiment 4: SC. injection of alovudine 25 mg/kg in rats during an iv. infusion of_thymidine 100 mg/l&hr after a loading dose of 100 mg/kg in 5 minutes (n=2) which was started prior to recovery determination. Experiment 5: SC. injection of alovudine 25 mg/kg during perfusion through the microdialysis probe with 1000 pM thymidine (n=3). In experiment 5, two microdialysis probes were implanted in the brain in each animal (symmetrically). One probe was perfused by the thymidine solution, the other by Ringer’s solution serving as a control. Experiment 6: S.C. injection of alovudine 25 mg/kg in rats treated with acetazolamide 50 m&g i-p. bolus and 25 mg/kg i.p. every second hour during the whole experiment starting acetazolamide injections 30 min prior to recovery determination (n=3). Data Analysis and Statistics: The microdialysis samples are collected in 20 min fractions which can be represented in the graphs as the midpoint of the sampling period (i.e. 10,
1808
Alovudine Transport into CNS and CSF
Vol. 66, No. 19,200O
(area under the time versus concentration 30 ,... 170 min) (16, 17, 18, 19). The AU&.,,, curve) was calculated by summing the products between the sampling interval and the sample concentration (corrected for recovery). The AUC ratio was calculated as AU&_,,, (brain)/AUC,,,W (blood) or as AUC,,,tin (CSF)/AUC,,,,, (blood). The half-life was calculated as lfi/slope where the slope of the elimination phase was determined by linear regression. The clearance was calculated as dose/AUC, for blood and the volume of distribution was calculated as CL/slope. Data are presented as meanskstandard errors of means or as geometric mean when appropriate. Two-sided Student’s t-test for dependent samples was used to compare AUC ratio brain/blood and CSFIblood, Cmax in the brain and CSF, half-lives in the brain and CSF. For this test the half-life was log-transformed to get symmetrically distributed data. The AUC ratio brain/blood and CSFIblood and the difference between C, in the brain and the CSF were compared among experiments by one-way ANOVA. Hypothesis testing was performed at the 5% level. Results Experiment 1: Constant i.v. infusion of alovudine. Steady state levels of alovudine were rapidly attained in blood, cerebrospinal fluid and the brain after the loading dose of 25 mgkg in 5 min and the start of continuous i.v. infusion. The mean time-concentration curves are given in Fig. 1. The mean in vivu recoveries of alovudine at a flow rate of 1 pL/min were 0.187 (0.079), 0.222 (0.078) and 0.458 (0.088) in the brain, CSF and blood respectively. The dialysis membranes used here was 4 mm. The AUC brain/blood and CSF/blood were found to be 0.164f0.03 1 and 0.402M.082 respectively. The difference in AUC ratio brain/blood and AUC ratio CSF/blood is statistically significant.
0
20
40
00
80
100
120
140
160
180
Tlme,min
Fig. 1 Time-log(concentration) curves (geomean+SEM) for CNS (circles), CSF (squares) and blood (triangles) from rats treated by iv. infusion of alovudine (25 mgkg/hr after a loading dose of 25 mgkg in 5 minutes) (n=4). The concentration is free (unbound) extracellular concentratioa, corrected for in viva recovery at 1 @/min.
Alovudine Transport into CNS and CSF
Vol. 66, No. 19,200O
1809
Experiment 2: S.C.injection of alovudine. The mean time-concentration curves for alovudine are given in Fig. 2. The mean in vivo recoveries of alovudine at a flow rate of 2 pL/min were 0.034 (0.003), 0.054 (0.005) and 0.137 (0.033) in the brain, CSF and blood respectively. The membrane length used in this experiment was 4 mm. The C, was 17.0k1.8 pM for the brain, 25.9f3.7 uM for the CSF and 104.2f27.6 pM for blood. The T, was found to be from 30 to 50 min for the brain and 30 min for the CSF and blood. Alovudine was eliminated with half-lives of 1.7+0.5h from the brain, 1.0&O.1h from the CSF and 0.8+0.lh from blood. CL calculated for blood concentration data was 0.76+0.13L/hkg. Vd &as 0.85fO.l4L/kg. The AUC ratios (O-180 min) between brain and blood and between CSF and blood were 0.285kO.058 and 0.327fo.051 respectively. The analysis of the pooled data from experiments 2,3,4 showed that the difference in C, between the brain and the CSF and the difference in half-lives between the brain and the CSF are statistically significant, while the difference between AUC ratio brain/blood and AUC ratio CSF/blood is close to significant (p=O.O56).
loo-
_
Lh
5
~lo_ 1
! 0
I
20
40
60
1
,
80
100
120
140
160
180
Time, min
Fig. 2 Timelog(concentration) curves (geomean&SEM) for CNS (circles), CSF (squares) and blood (triangles) from rats treated with alovudine (25 mgkg S.C. injection) (n=5). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 2 uL/min.
Experiment 3 and 4: S.C.injection of alovudine during i.v. infusion of thymidine. The mean time-concentration curves for alovudine during i.v. infusion of thymidine 25 mg/kg/h are given in Fig. 3. The mean in vivo recoveries of alovudine at a flow rate of 2 uL/min were 0.043 (0.004), 0.106 (0.025) and 0.268 (0.021) in the brain, CSF and blood respectively. The membrane length used in this experiment was 4 mm. The C, was 9.5f2.1 PM for the brain, 17.M3.6 uM for the CSF and 58.1f2.9 FM for blood. The T_ was 50 min for the brain and 30 mm for CSF and blood. Alovudine was eliminated with half-lives of 1. HO. 1 h, 0.9&O.1 h and 0.7&O.1 f?om the brain, CSF and blood respectively. CL calculated for
Vol. 66, No. 19,200O
Alovudine Transport into CNS and CSF
1810
Fig. 3 Time-log(concentration) curves (geomeanLSEM) for CNS (circles), CSF (squares) and blood (triangles) from rats treated with alovudine (25 mgkg S.C. injection) and thymidine (25 mgkgh i.v. afkr a loading dose of 25 mg/kg in 5 minutes) (n=5). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 2 pL/min.
1
1
I
0
I
20
40
60
60 Time,
100
120
140
160
180
min
Fig. 4 Time-log(concentration) curves (geomean+SEM) for CNS (circles), CSF (squares) and blood (triangles) from rats treated with alovudine (25 mgkg S.C. injection) and thymidine (100 mg/kg/h i.v. after a loading dose of 100 mgkg in 5 minutes) (n=2). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 2 pL/min.
Alovudine Transport into CNS and CSF
Vol. 66, No. 19,200O
1811
blood concentration data was 1.22fO. 12 L/h/kg. Vd was 1.23fO. 1 L/kg. The AUC ratios (O-l 80 min) brain to blood and CSF to blood were 0.241rtO.059 and 0.348f0.046 respectively. The mean time-concentration curves for alovudine during i.v. infusion of thymidine 100 mgkg /h are given in Fig. 4. The mean in vivo recoveries of alovudine at a flow rate of 2 uL/min were 0.058 (0.012), 0.055 (0.029) and 0.249 (0.049) in the brain, CSF and blood respectively. The membrane length used in this experiment were 4 mm. The C, was 12.2k5.4 pM for the brain, 19.7k12.1 pM for the CSF and 64.2k14.1 ).tM for blood. The T, was 50 min for the brain and 30 min for CSF and blood. Alovudine was eliminated with half-lives of l.OkO.2 h, 0.7f0.02 h and 0.7kO.l h from the brain, CSF and blood respectively. CL calculated for blood concentration data was 1.18M.16 LMcg. Vd was 1.13kO.25 L/kg. The AUC ratios (O-180 min) between brain and blood and between CSF and blood were 0.232kO.063 and 0.34M0.170 respectively. The difference in C,,, the brain/blood AUC ratio and the CSF/blood AUC ratio between the CSF and the brain, were compared between experiments 2,3 and 4, and no statistically significant difference was found.
0
20
40
60
60
100
120
140
160
180
Time, min
Fig. 5. Time-log(concentration) curves (geomeaniSEM) for blood (triangles) and for CNS with (squares) and without (circles) perfusion through the microdialysis probe with thymidine 1000 pM (n=3) from rats treated with alovudine (25 mgkg S.C. injection). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 1 uL/min.
Vol. 66, No. 19,200O
Alovudine Transport into CNS and CSF
1812
Experiment 5: S.C.injection of alovudine during perfusion through the microdialysis probe with thymidine. The mean time-concentration curves for alovudine are given in Fig. 5. The mean in vivo recoveries of alovudine at a flow rate of 1 uL/min were 0.127 (0.009) and 0.142 (0.015) in the brain (probes perfused with Ringer solution and probes perfused with thymidine solution, respectively) and 0.275 (0.085) in blood. The membrane length used in this experiment was 4 mm. The C, obtained from the probes perfused by thymidine solution was 20.4k1.7 pM and in the dialysate obtained from the probes perfused by Ringer’s solution the C_ were 17.9k2.5 PM for the brain and 125.Ok4.7 PM for blood. The T, was 30 to 70 min for the brain (both probes) and 30 min for blood. Alovudine was eliminated with half-lives of 3.OkO.8 h (probes perfused with thymidine solution), 2.OzkO.l h (probes permsed with Ringer’s solution) and 1.2fO.l h from the blood. CL calculated for blood concentration data was 0.43f0.06 L/h/kg. Vd was 0.69zkO.05 L/kg. The AUC ratio (O-180 min) between brain and blood was 0.218kO.018 (control probes) and 0.248f0.014 (probes perfused by thymidine solution). The difference in AUC ratios with respect to the perfusion medium was not significant.
1
1 0
!
20
40
60
60
100
120
140
160
180
Time,min
Fig. 6 Time-log(concentration) curves (geomean+SEM) for CNS (circles) and blood (squares) from rats treated with alovudine (25 mgikg S.C. injection) and acetazolamide (50 mgkg i.p. bolus and 25 mgikg i.p. every second hour during the whole experiment (n=3). The concentration is free (unbound) extracellular concentration, corrected for in vivo recovery at 1 uL/min.
Experiment 6: S.C.injection of alovudine during acetazolamide treatment. The mean time-concentration curves for alovudine are given in Fig. 6. The mean in vivo recoveries of alovudine at a flow rate of 1 uL/min were 0.115 (0.024) and 0.648 (0.058) in the brain and blood respectively. The membrane lengths used in this experiment were 4 (brain) and 10 (blood) mm. The C,, was 17.2k4.5 uM for the brain and 107.7k12.5 PM for blood. The T_ was 70 min for the brain and from 30 to 50 min for blood. Alovudine was eliminated with
Vol.66, No.
19,200O
AlovudineTransport intoCNSandCSF
1813
half-lives of 3.2f0.5 h from the brain and 1.5f0.3 h from blood. CL calculated for blood concentration data was 0.47kO.01 L/h/kg. Vd was 1.56f0.62 L/kg. The AUC ratio (O-180 min) between brain and blood was 0.179kO.032. The differences in brain/blood AUC ratio, C, and half-life between this experiment and experiment 2 (control group) are not statistically significant.
Discussion
The ability of anti-HIV drugs to penetrate the BBB is necessary to suppress virus replication in the brain tissue. Recent clinical data suggest that the concentrations attained in CSF of zidovudine (20) and indinavir (21) are sufficient to reduce HIV-RNA in this compartment (22). To understand the mechanisms of nucleoside analogue transport across the blood-brain barrier is thus of considerable interest. The anatomical and physiological properties of the BBB are responsible for the restricted transport of biologically active compounds, including drugs (7, 8). Many pharmacologically active compounds use the transport systems normally used for endogenous substances and nutrients (10, 23, 24, 25, 26, 27). As mentioned above, there are two routes by which a drug can enter and/or leave the brain: via the endothelium of cerebral capillaries and via the CSF. If a substance crosses the BBB via CSF, it should result in higher concentration of the drug in the CSF compared to extracellular concentration in the brain, which is what was found in this study. In experiment 2, when alovudine was given subcutaneously, the levels are different in the brain and CSF. The C, of alovudine in the CSF was significantly higher than the C, in the brain extracellular fluid in the analysis of the pooled experiments 2, 3 and 4. However, the half-life of alovudine in the CSF is significantly shorter than that in the brain making the difference in AUC ratio brain/blood V.S.CSF/blood only borderline significant. The concentration gradient between the extracellular fluid in the brain and the CSF found in experiment 1, remained constant during the intravenous infksion of alovudine. Therefore, slow transport by passive diffusion over capillary membranes or the choroid plexus can not alone explain the observation that brain concentrations of thymidine analogues decline during the elimination phase despite higher concentrations in blood. Had transport from CSF to the brain tissue been the major mode of transport, the concentration gradient would decline with time. Thus it is necessary to postulate the existence of an active process which maintains the concentration gradient. Similar results have been obtained for zidovudine in the rabbit (28). In that study the brain/plasma and CSF/plasma concentration ratios (0.19 and 0.29 respectively) suggest the existence of active transport of zidovudine out from the brain. However, intracerebroventricular infhsion of zidovudine was accompanied by a much higher CSF/brain concentration ratio suggesting that transport from CSF to the brain is of minor importance when the drugs are systemically administered (2). The brain regions studied were different in our study and in the rabbit study, where the thalamus was sampled, while we inserted microdialysis probes into the basal ganglia. The similarity of the results obtained suggests that different gray matter regions are similar with ‘respect to thymidine analogue transport. This is important because the basal ganglia are relevant from the virological point of view being a main location of HIV infection in the brain (1). In the present study, the influence of CSF production on the penetration of alovudine was also investigated. By administration of acetazolamide, production of cerebrospinal fluid can be
1814
Alovudine Transport into CNS and CSF
Vol. 66, No. 19,200O
reduced by 55% (lo), which should reasonably influence the concentration levels in the brain in case a substance utilises transport via CSF into or out from the brain tissue. However, the distribution to the brain of alovudine, estimated by the ratio of the AUC between brain and blood was 0.179 in the presence of acetazolamide and 0.285 in the control group. This difference is not statistically significant even if the control data are pooled with data from thymidine treated rats. Neither was C,, in the brain different from the control group. However, the elimination half-life of alovudine from the brain in the presence of acetazolamide was significantly slower @=0.003) compared to pooled data from experiments 2, 3 and 4. One explanation to this finding could be that alovudine is, partly, transported out from the brain via the flow of CSF as has been suggested for theophylline (13). Nevertheless, the major route of transport of alovudine over the blood-brain barrier appears to be via the cerebral capillaries. We have previously found that there is no correlation between lipophilic properties and the blood brain concentration gradient for uridine analogoues in the rat suggesting that the mechanism maintaining the gradient is not primarily dependent upon passive diffision over a lipophilic barrier. Instead, the concentration gradient was related to the electronic distribution in the substituents of these compounds (3). This tinding indirectly suggests that the blood brain concentration gradient depends on the affinity to an energy requiring transport protein. The existence of many active transport systems has been reported in the literature. Transport systems for purines (adenine and guanine) have been described and pyrimidines (thymidine, cy-tidine and utidine) have been reported to utilise a specific nucleoside carrier (23, 24, 25, 26, 27). An energy-dependent, saturable high-affinity transport system for thymidine has been found in the rabbit brain (23). It was hypothesized that the thymidine analogues use the same transporter for penetration over the BBB, but the result from experiments number 3,4,5 do not support that hypothesis. In these experiments we have used thymidine in an attempt to saturate the thymidine transporter from either side of the brain capillaries and thereby inhibit alovudine transport into or out from the brain. Thus, both i.v. inmsion of thymidine in different doses (25 mg/kg/h and 100 mg/kg/hr) and perfusion with thymidine 1000 uM through the probe was used. The doses chosen were based on the basal thymidine concentration being less than 0.1 uM both in the brain extracellular fluid and in the CSF and around 0.2 pM in plasma in humans (29). The AUC ratios brain/blood and CSF/blood in the presence of thymidine showed no tendency to a change compared to the corresponding AUC ratios in the control group (experiment 2). Neither did local administration of thymidine by its inclusion in the perfusion medium of the microdialysis probe have an effect on the AUC ratio. Thus, applying thymidine on either side of the blood-brain barrier in high concentrations had no effect on alovudine transport. The pharmacokinetic properties of alovudine were found to be in the range expected for thymidine analogues and are stmunarised in Table 1. No differences in plasma pharmacokinetic parameters of alovudine among experiment groups were found suggesting that peripheral pharmacokinetic changes do not influence the interpretation of the experiments. A corrollary of the result from this study with respect to the concentration gradient of alovudine between the extracellular fluid in the brain and the cerebrospinal fluid, is that clinical data on CSF concentration can not be interpreted as equal to the brain concentration. The rabbit data on zidovudine (2, 28) suggests that the difference between CSF and the brain holds for thymidine analogs and it might be valid for a wide range of drugs. However, brain extracellular levels of zidovudine and alovudine are higher relative to plasma levels in cynomolgus monkeys than in rats (4, 5). Therefore, species differences can make the difference between CNS and CSF less important in some species.
Alovudine
Vol. 66, No. 19,200O
Transport
into CNS and CSF
1815
Thus, we conclude that alovudine reaches the extracellular fluid of the brain not via the cerebrospinal fluid, but via the cerebral capillaries. The blood-brain and CSF-brain concentration gradients may be maintained by an active capillary process pumping alovudine out from the brain. This pump was not inhibited by thymidine or by inhibiting CSF production.
TABLE 1. Experiment Control
CNS
Cmax, pM 17.0
Half-life, h 1.7
CL, L/kg/h
CSF
25.9
1.0
Blood
104.2
0.8
CNS
9.5
1.1
0.241
CSF
17.0
0.9
0.348
Blood
58.1
0.7
CNS
12.2
1.0
0.232
CSF
19.7
0.7
0.340
Blood
64.2
0.7
CNS
17.9
3.0
0.218
CNS*
20.4
2.0
0.248
Blood
125.0
1.2
CNS
17.2
3.2
Blood
107.7
1.5
Yd, L/kg
AUC ratio 0.285
group
Uovudine S.C. +THY 25 mg/kg*h
4lovudine S.C. +THY 100 mg/kg*h
Alovudine s.c.+lOOOuM THY in perfusion medium (*)
Alovudine s.c.+ ketaxolamid e i.p.
Summary of pharmacokinetic
0.327 0.76
1.22
1.18
0.43
0.85
1.23
1.13
0.69 0.179
0.47
1.56
properties (mean) of alovudine.
1816
Alovudine
Transport
into CNS and CSF
Vol. 66, No. 19,ZOOO
Acknowledgement The present study was supported by the Karolinska Institute, the Swedish Medical Research Council (grants no. 09069 and 12590), Swedish Physicians Against AIDS and Medivir AB. The expert technical assistance of Ewa Guzenda is gratefully acknowledged.
References 1. C. WILEY. Psychiatric Clinics ofNorth America. 17 1-15 (1994). 2. Y. WANG and R.J. SAWCHUK, J. Pharm. Sciences. 84 871-6 (1995). 3. N. BORG, X-X. ZHOU, N-G. JOHANSSON, B. dBERG and L. STAI-ILE, Antivir. Chem. & Chemother 8 47-53 (1997). 4. E. LJUNGDAHL-STAHLE, E. GUSENDA, D. BOTTIGER, B. WAHREN, B. OBERG and L. STAI-ILE, Antimicrob. Agents & Chemother. 36 2418-2422 (1992). 5. L. ST.&I-ILE, E. GUZENDA and E. LJUNGDAHL-ST&ILE, J. AIDS. 6 435-439 (1993). 6. L. STAI-ILE and B.oBERG, Antimicrob. Agents & Chemother. 36 339-342 (1992). 7. M. BRADBURY, The concept of a blood-brain barrier. London:Wiley. 1979. 8. J. B. M. M. VAN BREE, Thesis from the Division of Pharmacology of the Center BioPharmaceutical Science, Leiden University, The Netherlands (1990). 9. T.H. MILDHORAT, M.K. HAMMOCK, J.D. FENSTERMACHER and V.A. LEVIN, Science 173 330-332 (1971). 1O.B. VOGH and D.R. GODMAN, Canadian J. Physiol. & Pharmacol. 63 340-3 (1985). 11 .L. STAI-ILE, S. SEGERVARD and U. UNGERSTEDT, Eur. J. Pharmacol. 185 187-193 (1990). 12. L. ST&ILE, S. SEGERVARD and U. UNGERSTEDT, J. Pharmacol. Meth. 25 41-52 (1991). 13. L. ST&ILE, S. SEGERV.&RD and U. UNGERSTEDT, Life Sciences 49 1843-1852 (1991). 14. U. UNGERSTEDT, Measurement of neurotransmitter release by intracranial dialysis. In: Marsden CA (ed) Measurement of neurotransmitter release in vivo. 8 I-105 Wiley, London, (1984). 15. L. STAHLE, A. MIAN and N. BORG, J. Pharmaceut. & Biomed. Analysis 13 369-376 (1995). 16. L. STAI-ILE, Eur. J. Clin. Pharmacol. 43 289-294 (1992). 17. L. STAHLE, Eur. J. Drug Metab. Pharmacokin. 18 89-96 (1993). 18. L. STAHLE, Use of microdialysis in pharmacokinetics and pharmacodynamics. In: Robinson T Justice J (eds) Methods in Neuroscience, 155- 174 Elsevier, Amsterdam (199 1). 19. L. ST&ILE, Eur. J. Clin. Phatmacol. 45 477-481 (1993). 20. T.A. TARTAGLIONE, A.C. COLLIER, R.W. COOMBS, K.E. OPHEIM, D.K. CUMMINGS, S.R. MACKAY, J. BENEDETTI and L. CREY, Arch. Neurol. 48 695-699 (1991). 21. L. ST&ILE, C. MARTIN, J.O. SVENSSON and A. SilNNERBORG, Lancet 350 20. (1997). 22. M. GISSLEN, G. NORKRANS, B. SVENNERHOLM and L. HAGBERG, J. Infect. Diseas. 175 434-437 (1997). 23. R. SPECTOR and W.G. BERLINGER, J. Neurochem. 39 837-841 (1982). 24. R. SPECTOR, J. Neurochem. 35 1092-1098 (1980). 25. R. SPECTOR, J. Neurochem. 45 1411-1418 (1985). 26. R. SPECTOR and S. HUNTOON, J. Neurochem. 40 1474-1480 (1983). 27. R. SPECTOR and S. HUNTOON, J. Neurochem. 41 1131-l 136 (1983). 28. Y. WANG, Y. WE1 and R.J. SAWCHUK, J. Pharm. Sciences. 86 1484-1490 (1997). 29. J.T. EELLS and R. SPECTOR, Neurochem. Research. 8 1451-1457 (1983).