Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

AZT and DDI are useful drugs for treating patients with AIDS (Yarchoan et al., 1986, 1990). It has been reported that 40 to 50% of adults and 70 to 80% of children with AIDS suffer from neurological dysfunction (Portegies et al., 1989; Petito, 1988). In the brain, the HIV infection is restricted, in most cases, to the brain parenchyma around capillary endothelial cells and brain macrophages (Wiley et al., 1986; Koenig et al., 1986). Therefore, for effective treatment of AIDS-dimentia complex or AIDS encephalopathy, efficient distribution of anti-HIV drugs into the CNS is essential (Price et al., 1988). However, only limited distribution of AZT, DDI and related nucleoside derivatives in the CNS has been demonstrated after systemic administration (Doshi et al., 1989; Anderson et al., 1990; Galinsky et al., 1991; Masereeuw et al., 1994; Wang and Sawchuk, 1995).

Because the BBB is well recognized to play an important role in regulating the entry of nutrients and drugs into the brain from the systemic circulation (Pardridge, 1977, 1983; Terasaki and Tsuji, 1994), the very low influx of AZT from the circulating blood into the brain across the BBB could be one possible explanation of the limited distribution into brain (Terasaki and Pardridge, 1988). However, recent pharmacokinetic studies, with use of a brain microdialysis technique, have reported that significant efflux transport from the brain to the circulating blood across the BBB may also contribute to the restricted distribution of AZT in the CNS (Dykstra et al., 1993; Wong et al., 1993; Wang and Sawchuk, 1995). In contrast to these reports, an in vitro transport study with primary cultured bovine brain capillary has failed to demonstrate the selective efflux transport of AZT across the BBB (Masereeuw et al., 1994). Moreover, it is also noteworthy that a probenecid-sensitive efflux transport system at the BCSFB may play an important role in reducing the distribution of DDI both in the CSF and brain parenchyma (Galinsky et al., 1991).

Accordingly, the primary purpose of the present study is to determine the presence of an efflux transport system for AZT and DDI from the brain interstitial fluid into the brain capillary lumen. To investigate the BBB efflux transport system directly, we used the in vivo intracerebral microinjection technique reported previously (Kakee et al., 1996).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. This study was approved by the University of Tokyo Animal Care Committee. Male Wistar rats (weight range, 230-270 g) were purchased from Nippon Ikagaku Doubutsu Ltd., Tokyo, Japan, and the animals had free access to food and water.

Chemicals. [³H]AZT (20 Ci/mmol) and [³H]DDI (42 Ci/mmol) were purchased from Moravek Biochemicals Inc. (Brea, CA). [Carboxyl-¹⁴C]inulin-carboxyl (2.6 mCi/g) and D-[1-¹⁴C]mannitol (56.7 mCi/mmol) were purchased from New England Nuclear (Boston, MA). Unlabeled AZT, DDI, thymidine, inosine, deoxyinosine and probenecid were purchased from Sigma Chemical Co. (St. Louis, MO). Hionic-fluor was used as a liquid scintillation cocktail and was purchased from Packard Instruments Corp. (Meriden, CT). Ketaral 50 (ketamine hydrochloride) was purchased from Sankyo Co., Ltd. (Tokyo, Japan). All other chemicals were commercially available compounds of reagent grade.

Intracerebral microinjection technique. The in vivo brain efflux experiments were carried out by the intracerebral microinjection technique reported previously (Kakee et al., 1996). Male rats were anesthetized with intramuscular doses of ketamine (235 mg/kg) and xylazine (2.3 mg/kg) and placed on a warm plate, the surface temperature of which was maintained at 37°C during the experiment by circulating hot water. After exposure of the skull, a 1.0-mm hole was made in the skull, 0.20 mm anterior and 5.5 mm lateral to the bregma by use of a dental drill (pen type grinder, Leutor Mini Gold, Nihonseimitsukikaikousaku, Hyogo, Japan). A stereotaxic frame (Narishige, Tokyo, Japan) was used to determine the coordinates of the rat brain coinciding with the Par2. The microinjection needle (350 µm o.d.; Seiseido Medical Industry, Tokyo, Japan) was inserted into the hole to a depth of 4.5 mm. Administered to the brain over 1 sec was 0.50 µl of drug solution containing 1 µM (20 µCi/ml) [³H]AZT or 1 µM (42 µCi/ml) [³H]DDI and 2 µCi/ml or 4 µCi/ml of [¹⁴C]inulin in physiological buffer (122 mM NaCl, 25 mM NaHCO₃, 10 mM glucose, 3 mM KCl, 1.4 mM CaCl₂, 1.2 mM MgSO₄, 0.4 mM K₂HPO₄, 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, saturated with 95%O₂-5%CO₂, pH 7.4). Subsequently, CSF was collected from the cisterna magna as reported previously (Suzuki et al., 1985, 1989). A hole of approximately 3 mm was made with a thermoknife KN-299 (Natsume, Tokyo, Japan) at sagittal midline through the suture between the interparietal and supraoccipital bones. A syringe (27.5 G; Terumo, Tokyo, Japan) was introduced into the hole to a depth of 6 mm. CSF (50-150 µl) was gently withdrawn at designated sampling times and mixed with liquid scintillation cocktail (10 ml). Immediately after collection of CSF, rats were decapitated and the ipsilateral (left) and contralateral (right) cerebrum and cerebellum were removed. Brain tissue samples were dissolved in 2.5 ml 2 N NaOH at 50°C for 3 hr. Fifteen milliliters of liquid scintillation cocktail was added and the radioactivity was measured by using a double-channel system for ³H- and ¹⁴C-containing samples with a liquid scintillation counter (LC 6000, Beckmann, Instruments, Inc., Fullerton, CA).

Determination of BEI from the brain. The BEI value was defined by equation 1 and determined by equation 2.

(1)

(2)

In the present study, [¹⁴C]inulin was used as an internal reference compound known to exhibit very limited permeation through the BBB (Takasato et al., 1984; Kakee et al., 1996). Because the percentage drug remaining in the brain is given by (100  BEI), K_eff from the brain was obtained from the slope of the semilogarithmic plot of (100  BEI) vs. time using a nonlinear least squares regression program. In addition to the BEI value, K_eff was also used as a kinetic parameter to characterize efflux transport of drug from the brain into the circulating blood across the BBB.

(3)

Intracerebroventricular injection technique. Intracerebroventricular injection was performed by the method reported previously (Suzuki et al., 1985, 1989). Under urethane anesthesia (1.5 g/kg), a needle (350 µm o.d.; Seiseido Medical Industry, Tokyo, Japan) connected to a silastic tube (0.3 i.d. × 0.7 o.d. × 100 mm) was introduced into the left lateral ventricle of rats. Subsequently, 10 µl of [³H]AZT (0.80 µCi) or [³H]DDI (0.80 µCi) and [¹⁴C]mannitol (0.018 µCi) dissolved in physiological buffer was administered via an inoculation needle. At an appropriate time after administration, CSF was collected as described above. The radioactivity was analyzed by HPLC followed by liquid scintillation counting to estimate the apparent efflux clearance across the BCSFB. Samples for AZT were dissolved in liquid scintillation cocktail and counted in a liquid scintillation counter.

HPLC assay. The amount of [³H]AZT and [³H]DDI metabolism in the brain and plasma after intracerebral microinjection was determined by HPLC. At 15 min after intracerebral microinjection of [³H]AZT or [³H]DDI (1 µCi/0.2 µl), blood was collected through the ipsilateral carotid vein and the rat was immediately decapitated. After centrifuging the blood, 350 µl of plasma was mixed well with the same volume of acetonitrile and kept at 4°C for 1 hr. After centrifugation for 5 min at 13,000 rpm with a Microfuge E (Beckman, CA), the supernatant was evaporated to dryness under a stream of N₂ and reconstituted with 150 µl of mobile phase. A 0.6-g aliquot of ipsilateral cerebrum was homogenized with 0.60 ml of physiological buffer by use of an ULTRA-TURRAX T25 (Janke & Kunkel, IKA-Labortechnik, Germany). Acetonitrile (2.1 ml) was added to deproteinize the brain homogenate after thorough mixing and kept at 4°C for 1 hr. After centrifugation for 15 min at 3,000 rpm, the supernatant was evaporated to dryness under N₂ and reconstituted with 180 µl of mobile phase. After filtration through a membrane filter (0.45 µm pore size; Millipore, Bedford, MA), 5 to 10 µl (brain) or 50 to 100 µl (plasma) was loaded onto a reversed-phase HPLC column TSK gel ODS-80^TM (15.0 cm × 4.6 mm i.d., Toso, Tokyo, Japan). For the analysis of the CSF, an aliquot (50-150 µl) of CSF was centrifuged for 0.5 min at 13,000 rpm with a Microfuge E, and 5 to 10 µl of the filtrate was directly injected onto the HPLC system. A constant flow solvent delivery system, model 655A-11 (Hitachi, Tokyo, Japan), was equipped with an injection system, consisting of a Rheodyne model 7125 injector and a 200-µl injection loop, a UV detector (model 638-41, Hitachi, Japan) and a fraction collector, model 2110 (Bio-Rad, Richmond, CA). A guard column (1.5 cm × 3.2 mm i.d., Toso, Tokyo, Japan) was placed between the injector and the analytical column. The mobile phase was methanol-0.01 M ammonium acetate (23:77 (v/v), AZT; 11:89 (v/v), DDI), pH 6.5 at a flow rate of 0.8 ml/min. The column and solvent were kept at ambient temperature. The absorbance of the eluent was monitored continuously at 254 nm. Fractions of the eluent were collected automatically and the radioactivity in each (0.60 ml for plasma samples or 0.30 ml for brain homogenate samples) was determined by a liquid scintillation counter.

Statistical analysis. Statistical analysis was performed by Student's t test except where a one-way analysis of variance was appropriate. Statistical significance was taken as P < .05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Time courses of [¹⁴C]inulin, [³H]AZT and [³H]DDI efflux from the brain. The percentage [¹⁴C]inulin remaining in the ipsilateral cerebrum was determined 2, 5, 10 and 20 min after microinjection into the cerebral cortex, Par2 region. No significant loss was observed with respect to the percentage [¹⁴C]inulin remaining relative to the injected dose during the investigation period following a one-way analysis of variance (P < .01). The average value of the percentage [¹⁴C]inulin remaining in the ipsilateral cerebrum was found to be 44.0 ± 3.4% (mean ± S.E.) after microinjection into the cerebral cortex. These data indicate that no significant efflux of inulin occurred during the period 2 min to 20 min after microinjection. By contrast, the time course of the percentage [³H]AZT remaining relative to the reference, [¹⁴C]inulin, in the brain obtained by use of equation 2 was monoexponential (fig. 1A) with an efflux rate constant of 0.0317 ± 0.0068 min¹ (mean ± S.D.). The efflux of [³H]DDI was similar but with an apparent efflux rate constant estimated to be 0.0253 ± 0.0037 min¹ (mean ± S.D.) (fig. 1B). During the course of the efflux study, a maximum of 0.5% [³H]AZT and [³H]DDI, relative to the administered dose, was found in the contralateral cerebrum, cerebellum and CSF, indicating very limited diffusion into the rest of CNS region from the injection site.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Time courses of percentage [3H]AZT and [3H]DDI remaining in ipsilateral cerebrum following intracerebral microinjection to the Par2 region. (A) [3H]AZT; (B) [3H]DDI. The solid line was obtained by the nonlinear least squares program. Each point represents the mean ± S.E. of three to eight determinations.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Typical HPLC chromatogram of ipsilateral cerebrum and carotid vein plasma after intracerebral administration of [3H]AZT and [3H]DDI. (A) [3H]AZT in ipsilateral cerebrum; (B) [3H]AZT in ipsilateral carotid vein plasma; (C) [3H]DDI in ipsilateral cerebrum; (D) [3H]DDI in ipsilateral carotid vein plasma. HPLC separation was performed at a flow rate of 0.8 ml/min. Each point represents radioactivity in the respective fraction.

Concentration dependence of [³H]AZT and [³H]DDI efflux from the brain. The concentration dependence of brain efflux of AZT was examined and the percentage [³H]AZT remaining in the brain was found to increase after the addition of unlabeled AZT to the injectate (table 2). As shown in figure 3, the efflux clearance of [³H]DDI fell with increasing unlabeled DDI concentration, also demonstrating saturable efflux of DDI from the brain. Based on equation 3, the following kinetic parameters were obtained: V_max = 85.9 ± 45.6 pmol/min/g brain; K_m = 11.3 ± 5.7 µM and K_non = 7.05 ± 0.54 µl/min/g brain (mean ± S.E.), respectively.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Concentration-dependent efflux of DDI from the brain across the BBB. The solid line was generated from the parameters obtained by the nonlinear least squares program with equation 3. Each point represents the mean ± S.E. of 3 to 12 determinations.

Effect of inhibitors on the efflux of [³H]AZT and [³H]DDI from the brain. The effects of several compounds on [³H]AZT efflux from the brain are summarized in table 2. Significant enhancement was observed with 1.7 mM DDI, 3.3 × 10² mM probenecid, 3.3 mM PAH, 3.3 mM PCG and 3.3 × 10² mM DIDS with respect to the percentage [³H]AZT remaining in brain. By contrast, no inhibitory effect with 3.3 mM thymidine was observed (table 2).

Contribution of efflux transport at the BCSFB to the apparent elimination of [³H]AZT and [³H]DDI from the brain. The amount of [³H]AZT and [³H]DDI remaining in the CSF was significantly inhibited by coadministration of 0.35 µmol probenecid after an icv bolus injection (table 4). These results indicate that a probenecid-sensitive transport system is responsible for part of the apparent elimination of AZT and DDI from the CSF. As shown in table 5, the administration of 0.35 µmol probenecid, 0.5 µmol AZT or 0.5 µmol DDI into the lateral ventricle did not cause any significant increase in percentage [³H]AZT and [³H]DDI remaining in brain after an intracerebral microinjection, which indicated that the probenecid-sensitive efflux transport system at the BCSFB is not responsible for the apparent elimination of [³H]AZT and [³H]DDI from the brain under the experimental condition.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Several methods have been developed to study the BBB efflux transport mechanism of drugs, e.g., pharmacokinetic analysis (Emanuelsson et al., 1987; Anderson et al., 1990; Galinsky et al., 1991; Adkison et al., 1994; Wang and Sawchuk, 1995), brain microdialysis (Terasaki et al., 1991; Dykstra et al., 1993; Wong et al., 1993), primary cultured brain capillary endothelial cells (Tsuji et al., 1992, 1993), in vivo perfusion into ATP-depleted brain (Sakata et al., 1994) and brain washout studies after carotid artery injection (Cornford et al., 1985). In the present study, characterization of the system responsible for the efflux transport of AZT and DDI from the brain directly was investigated by the intracerebral microinjection technique, previously developed as a useful technique to measure in vivo efflux across brain capillary endothelial cells from the abluminal side to the luminal side (Kakee et al., 1996). To validate this technique, the following are essential: 1) an internal reference compound is retained in the ipsilateral cerebrum over the entire experiment; 2) the apparent elimination rate of test drug from the ipsilateral cerebrum reflects the efflux transport across the BBB, in other words, both efflux transport from the BCSFB and diffusion into the contralateral cerebrum and/or cerebellum can be neglected during the experiment; 3) there is no significant metabolism in the ipsilateral brain. No significant decrease in [¹⁴C]inulin in the ipsilateral cerebrum was observed, which suggests that nonspecific permeation of [¹⁴C]inulin through the brain capillary is very limited for at least 20 min after cerebral microinjection into the Par2 region of rat brain. This result agrees well with a previous report that the apparent sucrose permeability through the brain capillary is also very limited in brain with implanted microdialysis fiber (Terasaki et al., 1991; Shimura et al., 1992). Moreover, no significant amount of [¹⁴C]inulin was found in the contralateral cerebrum, cerebellum and CSF after intracerebral microinjection. These results suggest that [¹⁴C]inulin can be used as an internal reference compound because it meets these criteria. Although the average recovery of [¹⁴C]inulin in the ipsilateral cerebrum during the experiments was only 44.0 ± 3.4%, a low recovery of reference compound does not affect the apparent efflux rate of test drugs (Kakee et al., 1996). A possible explanation for the low recovery after a 0.5-µl injection is that the injection solution may run up the injection needle and into the corpus callosum, as suggested previously (Cserr and Ostrach, 1974).

No significant amount of [³H]AZT and [³H]DDI was found in the contralateral cerebrum, cerebellum and CSF, which suggests that the microinjection technique used is satisfactory as far as the above-mentioned points are concerned. Although the direct implantation of microdialysis fiber has been reported to cause no significant effect on the transport of -aminoisobutyrate at the BBB (Benveniste et al., 1984), we could not rule out the possibility that the intracerebral microinjection technique may change the transport function of the BBB. In the present study, further characterization of the BBB function was not carried out and we assumed that this technique does not cause any significant change in the transport function of the brain capillary endothelial cells. Moreover, the apparent K_eff of AZT determined (0.0317 ± 0.0068 min¹) was in the same range as those reported values, i.e., 0.0528 min¹, calculated as the efflux clearance across the BBB (0.153 ± 0.029 ml/min/kg b.w.) divided by the distribution volume in the brain (2.9 ml/kg b.w.) obtained by a pharmacokinetic model analysis of AZT distribution in rabbits (Wang and Sawchuk, 1995), and 0.013 min¹, calculated as the whole-brain elimination constant (0.013 ml/min/g brain) divided by the distribution volume in the brain (1.02 ml/g brain) obtained in a brain microdialysis study in rats (Dykstra et al., 1993). These results also suggest that the brain microinjection technique allows determination of the drug efflux rate from the brain across the BBB. One of the advantages of this technique is that the efflux transport system from the brain is directly characterized by measuring the amount of test drug remaining in the ipsilateral cerebrum relative to that of a very slowly permeable reference compound injected simultaneously.

Because AZT and DDI are reported to be metabolized significantly in the body (De Miranda et al., 1990; Ray et al., 1990), it is possible that AZT and DDI are metabolized in the brain after intracerebral microinjection and the resulting metabolites are transported across the BBB. As shown in figure 2, there was no significant metabolism in the brain and most of the radioactivity in the ipsilateral carotid vein was accounted for by intact AZT and DDI (table 1), which demonstrates that AZT and DDI are transported from the brain into the circulating blood in their intact form.

Comparing the brain distribution of AZT and DDI in vivo reported previously (Galinsky et al., 1990; Anderson et al., 1990), the apparent brain-to-plasma concentration ratio (K_p,app) of AZT was approximately 2-fold greater than that of DDI. As the unbound fraction of AZT and DDI was determined to be 80% and 95% (Collins et al., 1988; Anderson et al., 1990), respectively, either tissue binding in the brain or the ratio of influx and efflux clearances across the BBB and/or the BCSFB are responsible for the differential distribution in the brain. As shown in figure 1, the apparent efflux rate constant of AZT was similar to that of DDI, which suggests that efflux transport is not a determining process for the differential distribution of AZT and DDI after systemic administration. Because the BBB is reported to lack a thymidine transport system and deoxynucleoside analogs have no significant affinity for the BBB nucleoside transport system (Cornford, 1975), carrier-mediated transport may not be responsible for influx from the circulating blood into the brain. Because the octanol-water partition coefficient (P_app) of AZT is reported to be 1.1, i.e., logP_app = 4.1 × 10² (Collins et al., 1988) and approximately 20-fold greater than that of DDI (P_app = 0.07, i.e., logP_app = 1.2; Ahluwalia et al., 1987), differential influx clearance by passive diffusion may be responsible for the differential in vivo distribution in the brain.

The efflux transport of [³H]DDI was reduced by unlabeled DDI (fig. 3) and probenecid (table 3), which suggests that DDI is transported from the brain into the circulating blood across the BBB via a carrier-mediated transport system, which would be shared by probenecid. Moreover, the efflux transport of [³H]AZT was inhibited by unlabeled AZT and probenecid, which indicates that AZT is also transported from the brain into the blood across the BBB via a probenecid-sensitive carrier system. Mutual inhibitory effects of AZT and DDI (tables 2 and 3) suggest that both drugs are transported by a probenecid-sensitive system across the BBB. It is believed that the BBB nucleoside influx transport system does not transport deoxy- and dideoxynucleoside (Cornford, 1975; Terasaki and Pardridge, 1988). Although we cannot rule out the possibility that such an efflux transport system for dideoxynucleoside and azido-deoxynucleoside analogs is present at the BBB and this is responsible for AZT and DDI efflux from brain into the circulating blood, the lack of any significant inhibitory effect of thymidine on the efflux of AZT and deoxyinosine and inosine on the efflux of DDI demonstrates that this transport system is not responsible for the efflux of AZT and DDI from the brain (tables 2 and 3). Comparing the maximal percentage AZT and DDI remaining in the brain, a significant difference was demonstrated, i.e., 69.3 ± 6.0% for AZT in the presence of probenecid (table 2) and 94.6 ± 3.0% for DDI in the presence of AZT (table 3). These differences for the noninhibitable component of the brain efflux process of AZT and DDI can be explained by the differential passive diffusion rates of AZT and DDI. This is because, as mentioned above, the P_app value of AZT reported was 20-fold greater than that of DDI (Collins et al., 1988; Ahluwalia et al., 1987).

In previous studies of the in vitro uptake of PCG by the isolated choroid plexus and icv injection technique (Suzuki et al., 1987, 1989), a probenecid-sensitive carrier-mediated transport system was proposed for the efflux transport of -lactam antibiotics at the BCSFB. Moreover, as shown in table 4, a probenecid-sensitive organic anion transport system at the BCSFB also participates in the transport of AZT and DDI from the CSF into the circulating blood. Therefore, we examined the contribution of a putative efflux transport system located at the BCSFB to the apparent elimination of AZT and DDI from the brain after intracerebral microinjection. The absence of an inhibitory effect of intracerebroventricularly administered unlabeled AZT and/or DDI and probenecid on the efflux of [³H]AZT or [³H]DDI from the brain (table 5) demonstrates that an efflux transport system located at BCSFB is not responsible for the efflux transport characterized by microinjection into Par2 of rat brain. However, in the case of systemic administration of AZT or DDI, it is noteworthy that we were unable to determine the contribution of the efflux transport system at the BBB on the apparent restricted distribution of AZT and DDI in the brain. This was because, after systemic administration, a drug which may be taken up by the brain parenchyma region close to the ventricle could diffuse into the CSF, and then, could be significantly transported from the CSF into the blood, causing a sink condition to the brain parenchyma. However, as discussed above, there would be no significant participation of BCSFB efflux transport in the apparent efflux from the brain after a cerebral microinjection, which would account for the distance of the Par2 region and the CSF being great enough to disregard diffusion through the parenchymal tissue (Kakee et al., 1996). It has been reported recently that a significant efflux of AZT from the brain into the circulating blood is essential to analyze the apparent brain, plasma and extracellular fluid concentrations of drug after systemic administration with pharmacokinetic models (Dykstra et al., 1993; Wang and Sawchuk, 1995). Accordingly, the efflux transport system characterized in the present study would be responsible for the restricted cerebral distribution of AZT and DDI.

In the development of anti-HIV drugs, several factors need to be considered to produce an antiviral effect in the CNS: 1) systemic clearance, 2) influx clearance from the capillary lumen into the brain extracellular fluid, and 3) efflux clearance from the brain into the circulating blood. Because drug diffusion through the brain parenchyma is known to be significantly limited (Blasberg et al., 1975; Fenstermacher and Kaye, 1988), it is very important to increase the net transport rate of drug across the BBB for effective anti-HIV drug therapy. Although there have been reports that a chemical delivery system is one successful way to increase the brain concentration of AZT and to retain AZT in the brain (Kawaguchi et al., 1990; Brewster et al., 1993), it will also be very important to decrease the affinity of anti-HIV drugs for the efflux transport carrier. To solve this problem, a strategy involving either suitable chemical modification of anti-HIV drugs or the discovery of a significant inhibitor of the efflux transport system is needed. The BEI method used, as shown in the present study, may help lead to important developments in the future.