AIT-082, a Cognitive Enhancer, Is Transported into Brain by a Nonsaturable Influx Mechanism and out of Brain by a Saturable Efflux Mechanism1

Materials and Methods

Animals.

Male Swiss-Webster CFW and C57BL/6 mice were supplied by Charles River Laboratories (Hollister, CA) and all experiments were conducted according to the National Institutes of Health Guide on Care and Use of Laboratory Animals. Young male mice were 2 to 3 months old at the time of use. Aged male mice were 13 to 25 months old. This large range in age was due to a supply shortage.

AIT-082 and Other Reagents.

AIT-082 (>99.5% pure) was synthesized by Eprova (Schaffhausen, Switzerland). The free acid of AIT-082 was produced by preparing a concentrated solution of AIT-082, adjusting to pH 1 to 2 with HCl, filtering the solution, washing and then drying the precipitate. The precipitate was then dissolved in PBS containing 160 mM NaOH. [¹⁴C]AIT-082 (51.5 mCi/mmol; 98% pure) was synthesized by Chemsyn Laboratories (Lenexa, KS). ¹²⁵I-BSA (300 Ci/mmol) was supplied by NEN (Boston, MA) and [³H]sucrose (5–15 Ci/mmol) was supplied by Amersham (Arlington Heights, IL).

Octanol:PBS Partition Coefficients.

A 400-μl aliquot of PBS (pH 7.4) was combined with 400 μl of octanol and 5 μl of [¹⁴C]AIT-082 (0.1 mg/ml) and mixed for 24 h. After 24 h, radioactivity was measured in 100 μl of the octanol layer and 100 μl of the aqueous layer. The partition coefficient (P) was calculated as disintegrations per minute in octanol/disintegrations per minute in PBS and log P was determined.

Influx Experiments.

These experiments were conducted as described in Blasberg et al. (1983), Patlak et al. (1983), and Banks and Kastin (1993). Under urethane anesthesia (2.25 g/kg i.p.), the right jugular vein and the left carotid artery were exposed. [¹⁴C]AIT-082 (1 μCi/animal) and¹²⁵I-albumin (0.25 μCi/animal) were coadministered via the jugular vein in a volume of 160 μl of PBS. At various times thereafter, brain and blood from the carotid artery were collected. Radioactivity in the weighed brain and 50 μl of serum was measured with an LS6500 liquid scintillation counter (Beckman Instruments, Fullerton, CA). The brain/serum ratios for [¹⁴C]AIT-082 and¹²⁵I-albumin were calculated and plotted against exposure time (Blasberg et al., 1983; Patlak et al., 1983; Banks and Kastin, 1993). Exposure time is the integral value for disintegrations per minute per milliliter of serum versus time from time 0 to timet divided by disintegrations per minute per milliliter at time t. The [¹⁴C]AIT-082 in brain was then corrected for blood volume (¹⁴C brain/blood ratio minus ¹²⁵I brain/blood ratio) and this corrected measurement was plotted against exposure time. An initial linear portion of the curve was identified and the influx constant (K_i) was calculated from the slope of the line (Banks and Kastin, 1993). To investigate the effect of excess unlabeled AIT-082 on influx of [¹⁴C]AIT-082 into brain, 110- or 650-fold molar excess unlabeled AIT-082 was included in the injection mixture. AIT-082 is commonly used as a K⁺ salt. However, due to the toxicity of the high plasma K⁺ levels, the free acid of AIT-082 was used as the unlabeled competitor in these experiments.

In Vivo Stability of [¹⁴C]AIT-082 in Brain and Serum.

Mice were anesthetized with 2.25 g/kg urethane. [¹⁴C]AIT-082 in PBS (10 μCi/mouse) was injected into the jugular vein in a volume of 160 μl. At various times after injection, blood was collected from the inferior vena cava and the animal was perfused through the left ventricle with 30 ml of ice-cold PBS. Thereafter, the brain was removed. Whole brain was homogenized in 300 μl of ice-cold water and the homogenate was diluted with a further 600 μl of water before centrifugation for 15 min at 12,000g at 4°C; the supernatant was retained. A 100-μl aliquot of serum was diluted with 300 μl of ice-cold water. To 500 μl of brain supernatant and 400 μl of diluted serum, 1000 and 800 μl of 2:1 chloroform:methanol was added, respectively. The samples were vortexed for 15 s and then centrifuged for 10 min at room temperature. The upper aqueous phase was lyophilized, resuspended in 160 μl of water, and filtered before reversed phase HPLC analysis.

[¹⁴C]AIT-082 in serum and brain was resolved with a 250- × 4.6-mm, 5-μm ODS reversed phase column (Phenomenex, Torrance, CA) coupled to a 30- × 4.6-mm guard column. The solvent system used was 0.2% o-phosphoric acid in water and acetonitrile. [¹⁴C]AIT-082 and metabolites were eluted from the column with a 10 to 12% gradient over 20 min followed by a 12 to 65% gradient over 5 min. With this method, both AIT-082 and [¹⁴C]AIT-082 elute at 19 to 20 min. Fractions were collected at 1-min intervals for 25 min and the radioactivity in each fraction measured.

Processing controls were obtained by adding 3 × 10⁴ dpm [¹⁴C]AIT-082 to brain or serum samples in vitro and processing as described above. These controls were used to correct in vivo data for degradation that occurred during processing.

Capillary Depletion.

These experiments were conducted with the method of Triguero et al. (1990) as modified by Gutierrez et al. (1993). [¹⁴C]AIT-082 (10 μCi/animal) and¹²⁵I-albumin (1 μCi/animal) were administered as described for influx experiments. At various times after administration, blood was collected from the inferior vena cava and the animal was perfused through the left ventricle with 30 ml of ice-cold PBS. The mouse was decapitated immediately, and the brain was dissected on ice to collect the cerebral cortex. Cortex was weighed and homogenized in 0.7 ml of ice-cold capillary buffer (10 mM HEPES, 141 mM NaCl, 4 mM KCl, 2.8 mM CaCl₂, 1 mM NaH₂PO₄, 1 mM MgSO₄, and 10 mM d-glucose, pH 7.4). Ice-cold 26% dextran (1.7 ml) was added and the homogenate centrifuged at 5400g for 15 min at 4°C in a Beckman TL-100 centrifuge with a TLS-55 swinging bucket rotor. The pellet, containing the capillaries, and the supernatant, containing brain parenchyma, were separated and radioactivity in each compartment was measured with a liquid scintillation counter (LS6500; Beckman, Fullerton, CA).

Pharmacokinetic Analysis.

AIT-082 at 60 mg/kg was administered i.p. to 2- to 3-month-old mice. At various times thereafter, trunk blood was collected into heparinized tubes and plasma was prepared. Plasma AIT-082 concentration was measured by HPLC as described above.

Microdialysis.

The surgical procedures were divided into two parts. Initially, mice were anesthetized with a ketamine/xylazine cocktail (ketamine 150 mg/kg; xylazine 10 mg/kg, 10 ml/kg i.p.) and placed in a stereotaxic apparatus (ASI Instruments, Eugene, OR). A saggital incision was made over the skull and a small hole drilled to expose the dura on the right side to allow for implantation of a guide cannula for the subsequent insertion of a precalibrated concentric microdialysis probe (CMA, Acton, MA) into the right cortex (coordinates: AP, 1.94 mm; LR, 2.00 mm relative to bregma; DV, 1.00 mm relative to the dura surface). Two additional burr holes were made for skull screws and the guide cannula (CMA) was secured using epoxy glue. After surgery, animals were returned to a home cage and allowed to recover overnight.

After recovery, mice were removed from their home cage and placed into a Plexiglas rodent dialysis box. A concentric dialysis probe was inserted unilaterally into the cortex, to a depth of 2 mm below the end of a guide cannula. The probe was 0.24 mm in diameter and had 2 mm of active membrane exposed at the tip. The probe was perfused with artificial cerebrospinal fluid [145 mM NaCl, 2.7 mM KCl, 1.0 mM MgCl₂(6H₂O), 1.2 mM CaCl₂(2H₂O), 2.0 mM NaHPO₄, pH 7.4, with 85% H₃PO₄], at a rate of 1.5 μl/min with a Harvard infusion pump (Harvard Apparatus, Holliston, MA), for ∼2 h to allow stabilization of injury-mediated neurotransmitter release (Benveniste and Huttemeier, 1990). Two 20-min baseline samples were then collected and AIT-082 (60 mg/kg, dissolved in 0.9% saline) was then administered i.p. and microdialysate collected every 20 min for 120 min. The microdialysate was immediately frozen at −80°C and subsequently the AIT-082 concentration in each sample was measured with liquid chromatography-tandem mass spectrometry (LC-MS-MS) as described below.

To assess in vitro recovery of AIT-082 by the microdialysis probe, the probe was placed in a 1 mg/ml solution of AIT-082 and microdialysis was performed. The AIT-082 concentration was measured in collected samples with the method described below.

The AIT-082 concentration in both in vivo and in vitro microdialysate samples was measured by LC-MS-MS with an LC-ABZ column (2.1 × 100 mm) fitted with a Supelcosil LC-ABZ guard column and a Micromass Quattro LC triple quadrupole spectrometer.

Efflux Experiments.

These experiments were conducted according to the method of Banks et al. (1997) with minor modifications. [¹⁴C]AIT-082 or [³H]sucrose in PBS was injected i.c.v. into mice at 1 mm posterior, 1 mm lateral to Bregma and 3.5 mm in depth relative to the skull surface, in a volume of 1 μl. After injection and on withdrawing the needle, there was often back flux of fluid; this was collected. At various times thereafter, the amount of radioactivity in brain, back flux, and a 10-μl aliquot of injection mixture was determined. The amount of radioactivity in brain was corrected for the back flux and the log of this corrected value (disintegrations per minute) was plotted against time. Thet_1/2 was calculated as described previously (Banks et al., 1997). To investigate the saturability of [¹⁴C]AIT-082 efflux from brain, 100- or 600-fold excess unlabeled K⁺-AIT-082 was included in the injection mixture.

Data Analysis.

To analyze whether there was a significant effect of age, mouse strain, or excess unlabeled AIT-082 on influx or efflux of [¹⁴C]AIT-082, one-way ANOVA was used with a Scheffe's post hoc analysis.

Results

Figure 1 shows the structure of AIT-082. It has a molecular weight of 365 and an octanol:PBS partition coefficient of 0.0029 (log P is −2.5). Figure2 shows a representative experiment demonstrating the influx of [¹⁴C]AIT-082 into brain of young Swiss-Webster mice. Clearly, the brain/blood ratio of [¹⁴C]AIT-082 was increased above that for¹²⁵I-albumin (Fig. 2A). The brain/blood ratio for¹²⁵I-albumin showed little change over time, whereas the brain/blood ratio of [¹⁴C]AIT-082 increased over time. When [¹⁴C]AIT-082 uptake into brain was corrected for blood content of brain, it was evident that the levels of [¹⁴C]AIT-082 in brain increased rapidly over the first 5 to 10 min of exposure time, thereafter reaching a plateau (Fig. 2B). The variability was large in the influx experiments and this was presumably due to the low influx rate. However, there was convincing evidence that uptake across the blood-brain barrier occurred and that a linear uptake phase existed. For consistency, the linear portion of the curve was identified with time points up to and including 5 min of experimental time in all experiments. Due to the large variability in the data, ther² value for many of the fitted lines was very low. To assess whether there was a statistically significant linear uptake phase, the data in each treatment group were averaged and a one-sample t test performed to assess whether the mean slope was significantly different from zero. The mean slope was significantly different from zero in all animal and treatment groups (P < .05). The K_i for uptake of [¹⁴C]AIT-082 into brain of Swiss-Webster mice was 0.6 ± 0.2 μl g⁻¹min⁻¹ (Table 1). Figures 3 and4 show that the influx of [¹⁴C]AIT-082 into brain of both young (Fig. 3) and aged (Fig. 4) C57BL/6 mice was qualitatively similar to that in young Swiss-Webster mice. The K_i for uptake was 0.35 ± 0.06 and 0.33 ± 0.02 μl g⁻¹ min⁻¹ for young and aged mice, respectively (Table 1). There were no statistically significant differences between the K_iin Swiss-Webster mice and C57BL/6 or between theK_i in young and aged C57BL/6 mice.

The saturability of transfer across the blood-brain barrier was examined by including 110- or 650-fold excess unlabeled AIT-082 in the i.v. administration of [¹⁴C]AIT-082. There was no significant effect of 110- or 650-fold excess unlabeled AIT-082 on the K_i in either Swiss-Webster or C57BL/6 mice (Table 1).

To confirm that the uptake data represented uptake of [¹⁴C]AIT-082 and not uptake of a¹⁴C-metabolite, the in vivo stability of [¹⁴C]AIT-082 was examined. Figure5 shows that the proportion of¹⁴C that was intact AIT-082 was >90% in serum at all time points examined. It decreased slightly from 100 ± 0.4% at 2 min to 94 ± 2.6% at 60 min. Similarly, the percentage of ¹⁴C in brain that was intact [¹⁴C]AIT-082 decreased slowly with time from 97 ± 2.9% at 2 min to 88 ± 11.8% at 60 min.

The above-mentioned influx data revealed that AIT-082 is taken up from the circulation by the brain. However, it did not describe whether the AIT-082 was simply sequestered by brain capillary endothelial cells or passed through them to enter the brain parenchyma. Figure6A demonstrates that at all time points examined, ≥90% of the [¹⁴C]AIT-082 detected in brain after perfusion passed into the brain parenchyma, whereas <10% remained in the capillaries. The level of [¹⁴C]AIT-082 in brain parenchyma increased over time to a peak at 30 min, after which it declined (Fig. 6B).

The passage of AIT-082 into brain parenchyma was further confirmed with microdialysis. Figure 7 shows the plasma and brain levels of AIT-082 after i.p. administration of 60 mg/kg. The peak in plasma levels (63.4 μg/ml) occurred at 10 min (Fig. 7A). The concentration of AIT-082 in microdialysate of cortical extracellular fluid increased with time to a peak of 111.4 ± 70.4 ng/ml at 40 ± 3.3 min and then declined (Fig. 7B). Based on the in vitro recovery of the microdialysis probe (9.5 ± 3.2%), the peak concentration of AIT-082 in vivo was 1173.1 ± 740.7 ng/ml or 3.2 ± 2.2 μM. However, caution must be exercised when extrapolating the in vivo concentration from in vitro recovery data because a number of confounding factors might influence in vivo recovery.

Figure 8 shows data from a representative experiment in which the efflux of [¹⁴C]AIT-082 and [³H]sucrose was examined after i.c.v. administration. [¹⁴C]AIT-082 and [³H]sucrose were cleared from brain exponentially with a t_1/2 of 20.0 ± 1.0 and 65.9± 16.4 min, respectively. The efflux of [¹⁴C]AIT-082 from brain was not inhibited by 100-fold molar excess unlabeled AIT-082; however, 600-fold molar excess unlabeled AIT-082 significantly inhibited [¹⁴C]AIT-082 efflux (Fig.9; P < .01). In the presence of 600-fold excess unlabeled AIT-082, thet_1/2 increased to 35.5 ± 3.6 min. Given that AIT-082 is a potassium salt, the effect of 600-fold excess K⁺ was examined. There was no significant effect of excess K⁺ on the efflux of [¹⁴C]AIT-082 from brain (Fig. 9).

Discussion

The design of drugs that are targeted at the central nervous system is commonly based on pharmacological approaches to disease mechanisms. However, as stated by others, such drug development programs are at a high risk of failure if the drug does not cross the blood-brain barrier in useful quantities (Pardridge, 1998). As part of a program to develop AIT-082 as a therapy for Alzheimer's disease, we investigated the transport of this purine derivative across the blood-brain barrier. Herein, we present data to support the transfer of AIT-082 across the blood-brain barrier in useful quantities.

Using multiple-time regression analysis, we have demonstrated that AIT-082 crosses the blood-brain barrier and that this transport is not inhibited by up to a 650-fold excess of unlabeled AIT-082. These experiments indicate that AIT-082 crosses the blood-brain barrier by a nonsaturable mechanism. In the absence of a specific, saturable transport mechanism, compounds may cross the blood-brain barrier by leakage or by simple diffusion (Banks and Kastin, 1993). Leakage is demonstrated by the entry of albumin into the brain at a very slow rate of ∼1.5 × 10⁻⁵ ml g⁻¹ min⁻¹ (Banks and Kastin, 1993). The rate of diffusion of a molecule across the blood-brain barrier is determined by its lipid solubility, hydrogen bonding potential, and molecular weight (Audus et al., 1992). If diffusion is the main mechanism by which AIT-082 enters the brain, the rate of entry would be similar to that of a hydrophilic molecule, of similar molecular weight, that is not transported by any saturable mechanism. In experiments not presented herein, we examined the passage of sucrose (mol. wt. = 350) across the blood-brain barrier and showed that the rate of transfer was similar to AIT-082 (10⁻⁴ μl g⁻¹min⁻¹). This supports the suggestion that [¹⁴C]AIT-082 crosses the blood-brain barrier by simple diffusion.

The specific activity of [¹⁴C]AIT-082 was such that 1 μCi of radioactivity represented 7 μg of compound. Given this high level, there is a possibility that self-inhibition of a transport system may have occurred in the experiments presented. Self-inhibition would effectively reduce theK_i and preclude demonstration of a saturable mechanism by inclusion of excess unlabeled AIT-082. This possibility was investigated with [³H]AIT-082 that had a 12.5-fold higher specific activity (4 Ci/mmol) than [¹⁴C]AIT-082; therefore, administration of 1 μCi/animal was equivalent to 90 ng/animal. There was no change in theK_i (data not shown). Thus, there is no evidence that AIT-082 crosses the blood-brain barrier via a saturable mechanism.

It has been demonstrated that the purine hypoxanthine is transported across the blood-brain barrier via a high-capacity, low-affinity transporter in rats (Cornford and Olendorf, 1975; Betz, 1985; Spector, 1987, 1988). The rate of transport of hypoxanthine was ∼100-fold faster than for the hypoxanthine derivative AIT-082 and was significantly diminished by increasing doses of unlabeled hypoxanthine, adenine, theophylline, and uracil (Spector, 1987). Given that [¹⁴C]AIT-082 influx is slower than that of hypoxanthine and is not saturable, it is unlikely that AIT-082 is transported across the blood-brain barrier by the hypoxanthine transporter.

There was no significant difference in either the quantitative or qualitative uptake of AIT-082 into brain between an outbred strain of mouse (Swiss-Webster) and an inbred strain of mouse (C57BL/6), both of which have been commonly used in AIT-082 studies. Nor was there a significant effect of age on the transfer of AIT-082 across the blood-brain barrier. These results are consistent with the literature. Although age induces significant histological and biochemical changes in the blood-brain barrier, including loss of endothelial cells and decreased capillary diameter in rats, there is no apparent change in blood-brain barrier permeability to hydrophilic and high molecular-weight substances in healthy, aged animals (Shah and Mooradian, 1997; Banks et al., 1999). In contrast, there may be a change in blood-brain barrier permeability associated with Alzheimer's disease. A study of sixty-five 85-year-old subjects revealed that in those that had dementia, there was an increase in the cerebral spinal fluid/plasma protein ratio suggestive of an increase in blood-brain barrier permeability (Skoog et al., 1998). This may have implications for the design and administration of compounds, such as AIT-082, which are targeted at the treatment of Alzheimer's disease and other neurodegenerative disorders.

All of these studies might be subject to complication by the possibility that [¹⁴C]AIT-082 is metabolized and that ¹⁴C measurements may represent measurement of a ¹⁴C-metabolite. However, the results presented demonstrate that the metabolism of [¹⁴C]AIT-082 is minimal in serum and brain of Swiss-Webster mice in vivo and thus confirm that the data discussed above do indeed relate to intact [¹⁴C]AIT-082 and not to a ¹⁴C-metabolite. Although degradation/metabolism was minimal in brain, it was slightly increased compared with serum and so suggests that there may be some brain-specific degradation/metabolism of [¹⁴C]AIT-082. It is known that the brain endothelium is enriched in degradative enzymes, including drug-metabolizing and purine-metabolizing enzymes (Audus et al., 1992). Specifically, adenine deaminase and purine nucleotide phosphorylase are highly enriched in brain capillary endothelial cells (Johnson and Anderson, 1996). It is, however, unlikely that these enzymes are responsible for the slight increase in degradation in brain because it has recently been shown that AIT-082 inhibits the action of these enzymes (Caciagli et al., 1999).

The influx data discussed above demonstrate that [¹⁴C]AIT-082 crosses the blood-brain barrier. To confirm the transport of [¹⁴C]AIT-082 into brain parenchyma, we used the capillary depletion technique and demonstrated that >80% of the [¹⁴C]AIT-082 that remained in brain after perfusion was present in the brain parenchyma. This clearly demonstrates that [¹⁴C]AIT-082 is not sequestered by endothelial cells and passes into the brain parenchyma where it is presumably available to interact with neurons, astrocytes, and other cells of the central nervous system. [¹⁴C]AIT-082 levels in brain parenchyma peaked at 30 min and from these data, the maximal concentration of AIT-082 in cortical extracellular fluid was estimated to be in the low micromolar range. To further confirm the passage of AIT-082 into brain interstitial fluid and to quantify the concentration more accurately, microdialysis experiments were performed. AIT-082 was readily detectable in cortical extracellular fluid and the concentration increased over time to a peak of ∼3 μM at 40 min. This confirmed the previous estimate. When comparing the data obtained with the two techniques, it is apparent that there was a shift in the time at which peak levels of AIT-082 were detected in parenchyma. Peak levels of [¹⁴C]AIT-082 were detected at 30 min with the capillary depletion technique, whereas peak levels of AIT-082 were detected at 40 min with the microdialysis technique. This shift is likely due to the shift in peak plasma levels from 0 min, when [¹⁴C]AIT-082 was given i.v. (capillary depletion), to 10 min, when AIT-082 was administered i.p. (microdialysis; Fig. 7). The apparent parallel between plasma and brain AIT-082 concentrations further supports the suggestion that AIT-082 crosses into brain by a nonsaturable mechanism.

The concentration of AIT-082 that has been used in in vitro studies ranges from 1 to 100 μM. At these concentrations, AIT-082 stimulates neurite outgrowth and an increase in extracellular neurotrophic factors (Middlemiss et al., 1995; Caciagli et al., 1998, 1999; Juurlink and Rathbone, 1998; Caciagli et al., 1998, 1999; Bintner et al., 1999;Ciccarelli et al., 1999). In vivo, 60 mg/kg AIT-082, administered i.p., ameliorates NMDA-induced (Caciagli et al., 1998; Di Iorio et al., 1999) and kainic acid-induced neurodegeneration in striatum and hippocampus (Di Iorio et al., 1999), and the neurodegeneration and motor deficits caused by spinal cord crush injury (Middlemiss et al., 1999). Given that the concentration of AIT-082 in brain reaches low micromolar concentrations when 60 mg/kg AIT-082 is administered i.p., these data suggest that there is parity between the in vitro and in vivo studies of AIT-082 neurotrophic properties.

The fact that uptake of AIT-082 into brain rapidly reached a plateau (Fig. 2-4) suggests that equilibrium was reached between the nonsaturable influx of AIT-082 and efflux of AIT-082 from brain to blood. Indeed, after i.c.v. administration, [¹⁴C]AIT-082 was transported out of brain with a t_1/2 of ∼20 min. In contrast [³H]sucrose was transported out of brain with at_1/2 of >60 min. This is similar to the efflux of albumin (Banks et al., 1997) and suggests that sucrose is transported out of brain via passive mechanisms, including cerebral spinal fluid reabsorption. The faster rate of efflux of [¹⁴C]AIT-082 from brain, compared with sucrose, and the fact that efflux was inhibited by 600-fold excess unlabeled AIT-082 indicate that AIT-082 is transported out of brain by an active mechanism.

In conclusion, AIT-082, a hypoxanthine derivative, crosses the blood-brain barrier and is available to astrocytes and neurons at significant concentrations. Together with the demonstration of an active efflux mechanism, the results suggest that there are two pools of AIT-082; one that is transported into brain parenchyma and one that is rapidly returned to the circulation via a specific, saturable efflux mechanism.