Targeting the Cerebrovascular Large Neutral Amino Acid Transporter (LAT1) Isoform Using a Novel Disulfide-Based Brain Drug Delivery System

Abstract

We describe a novel strategy to achieve high affinity recognition for the specific, cerebrovascular large neutral amino acid transporter (LAT1) isoform by covalent coupling of small molecules to the amino acid, L-cysteine (L-Cys). L-Cys (as the carrier) was covalently attached via a disulfide bond to either 6-mercaptopurine or 2-methyl-1-propanethiol (IBM) to form the brain-targeted drug delivery systems (BTDS). BTDS were designed for high affinity recognition by LAT1 at the cerebrovasculature. Using an in situ rat brain perfusion technique, competition between BTDS and the radiotracer [¹⁴C]L-Leu demonstrated significant inhibition of [¹⁴C]L-Leu brain uptake. BTDS possess affinity for cerebrovascular LAT1 in many distinct brain compartments, and the recognition of BTDS by LAT1 is influenced by hydrophobicity of the side-chain in BTDS. Thus, the BTDS strategy may be utilized for rapid shuttling of various neuropharmaceuticals into brain.

Keywords :

• Amino Acids
• BBB Targeting
• Brain Targeted Drug Delivery
• BTDS
• Cerebrovasculature
• Disulfides
• LAT1
• LAT2
• Neuropharmaceuticals

The brain capillary endothelial cells comprising the blood-brain barrier (BBB) greatly prohibit solute transport from blood to brain. This is due to tight intercellular junctions that, limit diffusion of relatively small as well as large molecular weight substances from blood to brain [1, 2, 3]. Thus, paracellular and transcellular permeability of many therapeutic agents into brain is minimal. Additionally, high brain capillary metabolism, low pinocytic vesicular trafficking, and efficient efflux mechanisms at the BBB significantly decreases net brain uptake of solutes [4, 5]. In contrast, carrier-mediated solute transport into brain can be rapid and is readily available at the BBB [6]. One such transporter, the large neutral amino acid transporter (LAT), is present at the BBB and efficiently transports L-amino acids into the brain [7]. Recently, we have provided evidence for the predominant functional expression of the specific LAT1 isoform at the cerebrovasculature [8].

Recent reports have described isolation, cloning, and structural identification of two LAT isoforms, LAT1 and LAT2 [9, 10, 11]. Until recently, the existence and functional predominance of these isoforms at the BBB has been controversial [8, 9, 12]. Although both LAT isoforms have affinity for L-neutral amino acids with large bulky side-chains (e.g., phenylalanine and leucine), LAT2 has noticeably broad substrate recognition. In general, LAT2 (L-Leu K_m = 119 μM) [9] has a much lower affinity for amino acids than LAT1 (L-Leu K_m = 18.1 μM) [10], as shown by L-Leu affinity. The LAT system at the cerebrovasculature (i.e., LAT1) is a high affinity transport system with typical K_m values of 10 − 50 μM [7]. Comparatively, LAT in peripheral tissues has higher K_m values (lower affinity) in the 1–10 mM range [13]. Cerebrovascular LAT1 recognizes a wide range of amino acid-based drugs [14, 15, 16]. Baclofen (an antispasmodic) [17], α-methyldopa (a centrally acting, antihypertensive) [18], and L-dopa (anti-Parkinson's agent) [19] are clinically used drugs that have been shown to cross the BBB via LAT1-mediated transport.

In this study, we employed an amino acid-based strategy to create a novel brain targeted drug delivery system (BTDS) displaying high affinity for the cerebrovascular LAT1. The carrier portion in the BTDS is the amino acid, L-cysteine (L-Cys). L-Cys possesses minimal affinity (K_m = 480 μM)⁷ for LAT1 at the BBB. Covalent attachment of “R” groups to L-Cys provide novel BTDS whose affinity for LAT1 at the BBB can be significantly improved and further modulated by R group properties (Figure 1). Thus, an R group with an appropriate “reaction handle” for attachment to the thiol group of L-Cys can be used in this strategy, which provide considerable flexibility in designing new BTDS design. We have attached an anticancer agent, 6-mercaptopurine (MP) and a model agent that mimics the side chain of L-Leu, i.e., 2-methyl-1-propanethiol (isobutyl mercaptan; IBM), to L-Cys to form either MP-BTDS or IBM-BTDS, respectively. The BTDS were designed to demonstrate versatility in attachment of R (i.e., drug portion) to L-Cys in the BTDS structure. The flexibility of the BTDS strategy could be useful for brain delivery of a variety of therapeutic agents that are required in the treatment of various central nervous system disorders (e.g., Parkinson's, Alzheimer's, AIDS, brain tumors). Thus, the BTDS approach converts molecules with minimal LAT1 affinity to create novel molecules with significant cerebrovascular LAT1 affinity. BTDS could be used to effectively shuttle central nervous system active therapeutics across the BBB.

FIG. 1 Scheme for the synthesis of L-cysteine-thiourea [1], MP-BTDS [2], and IBM-BTDS [3].

MATERIALS AND METHODS

[¹⁴C]L-leucine (323 mCi/mmole) and [¹⁴C]urea (56 mCi/mmol) were obtained from ICN Biochemicals (Irvine, CA, USA). [³H]Methoxyinulin (323 mCi/mmol) and [³H]diazepam (83.5 Ci/mmol) were obtained from NEN (DuPont Co., Wilmington, DE, USA). [³H]Methoxyinulin was purified prior to use using Sephadex G-25 chromatography. All other reagents were obtained from Aldrich Chemical (Milwaukee, WI, USA), unless stated otherwise. ¹H-NMR spectra were obtained using a 300 MHz spectrometer and was processed by MacNMR 5.5, Tecmag Zodiac. Mass spectral data (FAB) was obtained from Mass Spectral Analysis Department, University of Maryland (College Park, MD, USA). Mass spectrometric analysis was also perfomed using a LCQ-MS/MS instrument (Finnigan, San Jose, CA, USA) with a dual octapole–ion trap configuration using electrospray ionization (ESI), including MS/MS analysis.

Synthesis

2-Amino-3-Thioureadisulfanyl-Propionic Acid (L-cysteine-Thiourea) [1]

L-Cys (1.5 mmol) was activated using thiourea (1.8 mmol) in ethanol according to the report of Shirakawa et al [20]. Concentrated HCl (600 μ L) and H₂O (600 μ L) were added to form a clear solution. The mixture was cooled to 0–5° C. Hydrogen peroxide (30% solution; 190 μ L) was added immediately after cooling and the reaction mixture was stirred for 2 hr at 0–5° C. The reaction mixture was either rotary-evaporated and purified (as described below) or the crude reaction mix was coupled directly with isobutyl mercaptan or mercaptopurine, described below. Methanol and chloroform (3:1) was used as the mobile phase to purify the product [1] (R_f∼ 0.7) using silica-based preparative thin layer chromatography (TLC): Yield 70–80%.

2-Amino-3-(7H-Purin-6-yldisulfanyl)-Propionic Acid (MP-BTDS) [2]

Mercaptopurine (0.311 mmol in methanol) was dissolved in 2N sodium hydroxide (200 μ L). L-cysteine-thiourea [1] (0.306 mmol) was dissolved in methanol (500 μ L) and H₂O (50 μ L) and the solution of L-cysteine-thiourea was added dropwise to the solution of MP within 5 min. The reaction mixture was stirred for 1 hr at room temperature. The reaction conditions were based on previous report from Saneyoshi et al. (21). The reaction mixture was filtered and the filtrate was rotary-evaporated. The residue was purified using silica-based, preparative TLC and methanol: chloroform: ammonium hydroxide (6:4:1) as the eluting solvent (R_f∼ 0.2). The HPLC retention time of MP-BTDS [2] on a C18 column using a 5% (v/v) methanol in water as the mobile phase was 5.9 min. Purity was confirmed using TLC and HPLC. ¹H-NMR (D₂O), δ ppm; 3.2–3.5 (2H, m, CH₂), 4.0 (1H, m, CH), 8.4 (1H, s, CH), 8.8 (1H, s, CH). M.S. (FAB): M^{+ 1} = 272.

2-Amino-3-Isobutyldisulfanyl-Propionic Acid (IBM-BTDS) [3]

2-Methyl-1-propanethiol [isobutyl mercaptan (IBM); 390 μ L; 3.6 mM] was added to 2N sodium hydroxide (1 mL), H₂O (1 mL), and methanol (4 mL). The solution of IBM was added dropwise to the crude reaction mixture of L-cysteine-thiourea within 5 min. The reaction mixture was stirred for 2 hr at room temperature, filtered, and the filtrate was rotary-evaporated. The residue was purified using silica-based, preparative TLC and methanol: chloroform: ammonium hydroxide (6:4:1) as the eluting solvent (R_f∼ 0.7). The HPLC retention time of IBM-BTDS [3] on a C18 column using a 5% (v/v) methanol in water as the mobile phase was 13.3 min. Purity was confirmed by TLC and HPLC. ¹H-NMR (D₂O), δ ppm; 0.8–0.9 (6H, m, CH₃), 1.2–1.4 (1H, m, CH), 2.3–2.4 (2H, m, CH₂), 3.5–3.6 λ (2H, m, CH₂), 4.1 (1H, m, CH). M.S. (FAB): M^{+ 1} = 210.

HPLC Analysis

Analysis by reverse phase, HPLC was conducted using a C-18 column (Bondclone 300 × 3.9 mm, 10 μ m particle size, Phenomenex, St. Torrance, USA). The Shimadzu HPLC system consisted of 2 Shimadzu LC-10AS pumps, LC-10A controller, SIL-10A autoinjector, CR-501 chromatopac integrator, and a SPD-10A UV-vis detector. A 5% methanol in water (isocratic) mobile phase was used for the separation of BTDS [2] and [3] as well as the S-benzyl-L-cysteine. The following HPLC parameters were kept constant throughout analysis; flow rate (1 mL/min), attenuation (0), range (0.005 AUFS), λ (260 nm), and injection volume (10 μ L).

In Situ Brain Perfusion

An in situ brain perfusion technique in anesthetized rats was used to demonstrate affinity of the synthesized BTDS for cerebrovascular LAT1. Adult male Sprague-Dawley rats (250–300 g; Harlan, Indianapolis, IN, USA) were treated in accordance with institutional guidelines. Our method was modified from the original procedure [15, 23] and it was similar to recently reported, modified methods [24, 25]. The right common carotid artery was exposed. The right external carotid artery was ligated, the occipital artery was coagulated and cut, and the right common carotid artery was cannulated for normal grade perfusion. The brain perfusate was a pH 7.4, bicarbonate buffered physiological saline (128 mM NaCl, 24mM NaHCO₃, 4.2 mM KCl, 2.4 mM NaH₂PO4, 1.5 mM CaCl₂, 0.9 mM MgCl₂, and 9 mM D-glucose). The radiotracer [¹⁴C]L-leucine (L-Leu; 0.2 μ Ci/mL) was perfused alone (as control) to determine 100% capillary permeability-surface area (PA) product for L-Leu.

In the competition assays, excess concentration of either L-Leu, L-Phe, S-benzyl-L-cysteine, L-cysteine-thiourea, L-Cys (alone), a mixture of L-Cys and MP, IBM-BTDS, or MP-BTDS was coperfused with radiotracer [¹⁴C]L-Leu at the BBB to determine percent inhibition of [¹⁴C]L-Leu brain uptake. The in situ rat brain perfusion was initiated immediately after the heart-cut at a flow rate of 9 mL/min for 30 sec. The animal was then decapitated and the head was chilled on ice. The skull was opened and the brain was placed over buffer-saturated filter paper on ice.

Brain samples (15–25 mg) were taken from the right hemisphere (frontal, parietal, occipital, hippocampus, caudate nucleus, thalamus) and left hemisphere (parietal) as control. After overnight digestion of the samples (in 10% piperidine at 50° C), 10 mL of scintillation cocktail (Econo-Safe, Research Products International Corp., Mount Prospect, USA) was added and radioactivity in each brain sample was counted on a dual-label scintillation spectrometer (Beckman LS 5801).

BBB Permeability Calculations

Brain uptake calculations for the in situ rat brain perfusion technique have been described in detail previously [7, 15, 22, 23, 25]. Briefly, the polar carbohydrate polymer, [³H]methoxyinulin, does not penetrate the BBB because of its high molecular weight and high polarity. Thus, methoxyinulin was used as an ideal marker to determine the brain vascular volume (V_v) in the rat. However, due to rapid uptake process mediated by LAT1 at the BBB, a significant concentration of radiotracer [¹⁴C]L-Leu was not found intravascularly. Since the exact calculation of V_v is not significant in calculating intravascular concentration for highly permeable solutes, the V_v (0.77%) was determined using a separate set of experiments [23]. The amount of [¹⁴C]L-Leu in each sample was subtracted from the amount of residual [¹⁴C]L-Leu remaining in the vasculature to provide the unidirectional blood-to-brain transport constant, as follows [26]: Where q_tot is the amount of total [¹⁴C]L-Leu measured in the right parietal brain region (dpm/g). C_pf is the amount of [¹⁴C]L-Leu measured in the perfusate (dpm/mL). The net perfusion time is represented by T (in seconds). The transformation of K_in to the PA product is performed using the Crone-Renkin model [27] Where F is the flow rate (0.0491 mL/sec/g) determined from a separate set of experiments involving [³H]diazepam (a lipophilic solute used as a marker for cerebral blood flow). For determination of affinity, the Michaelis-Menten equation was used as follows: The kinetic variables of PA_i, PA₀, C_i, and K_i were substituted into the Michaelis-Menten equation as performed previously [7], thus yielding: Where C_i (μ mol/ml) is the concentration of the inhibitory analyte, PA₀ (ml/s/g) is the control PA product for [¹⁴C]L-Leu alone, and PA_i (ml/s/g) is the observed PA product of [¹⁴C]L-Leu upon concomitant perfusion with competitor. Modified Michealis-Menton was rearranged and solved for K_i (value for 50% inhibition of [¹⁴C]L-Leu transport), as follows: This analysis assumes that K_i values are equal to K_m values for cerebrovascular LAT1, which has been reported to be a reasonable assumption as shown previously [28]. The Lineweaver-Burk plot was determined by rearranging equation [D] to yield: Permeability (P) and flux (J) were determined from the PA product using the following relationships: Where the brain capillary surface area is reported to be 130 cm²/g [29]. Flux was ultimately substituted for PA value in plotting the Lineweaver-Burk equation [G] [30]. Student's t-test (two-sided) was used in all cases to compare two means. The criteria used for statistical significance was p < 0.05.

RESULTS AND DISCUSSION

BTDS Synthesis

The disulfide-based BTDS were synthesized using a 1-pot reaction scheme as shown in Figure 1. L-Cys was reacted with thiourea under acidic conditions using hydrogen peroxide as the oxidizing agent. The mechanism for this reaction has been suggested to be oxidation of the thiol on L-Cys to a sulfenic acid followed by reaction with thiourea to form the activated intermediate, L-Cys-thiourea [20]. The activated product, L-Cys-thiourea, was then reacted in situ with either MP or IBM under basic conditions, to yield BTDS. Thiourea serves as the leaving group on the activated intermediate L-Cys-thiourea to allow formation of BTDS. The reaction yield for both MP-BTDS and IBM-BTDS was 20%. Another disulfide product, isobutyl-thiourea, was observed during the IBM-BTDS reaction by mass spectrometry (LCQ MS/MS using electrospray ionization), thus providing one explanation to the low yield for IBM-BTDS. Another explanation could be that the use of unprotected L-Cys can lead to side reactions. The reaction scheme for BTDS was adapted from literature after minor modifications [20, 21].

Both IBM- and MP-BTDS were purified using silica-based preparative TLC using methanol: chloroform: ammonium hydroxide (6:4:1). BTDS was structurally identified using FAB mass spectrometry and ¹H-NMR. Stock solution for IBM-BTDS (5 mM) was prepared in methanol and stored at 2°C. IBM-BTDS was noted to be stable in methanol for at least a period of 6 months at 2°C as confirmed by UV-based HPLC. MP-BTDS and IBM-BTDS retention times were 4.2 and 13.3 min, respectively, using a C18 column and a 5% (v/v) methanol in water under isocratic conditions. UV detection was at 260nm. MP-BTDS was noticeably unstable in methanol as determined by HPLC. Therefore, MP- BTDS was added to brain perfusate immediately prior to brain perfusion. The activated synthetic intermediate L-Cys-thiourea and the BTDS were probed for cerebrovascular LAT1 affinity as discussed below.

Brain Perfusion

The in situ rat brain perfusion technique was validated using the following radiotracers: [³H]methoxyinulin, [³H]diazepam, [¹⁴C]urea, and [¹⁴C]L-Leu. Methoxyinulin is a high molecular weight carbohydrate polymer that does not cross the BBB. By using [³H]methoxyinulin (1.0–1.5 μ Ci/mL) as a vascular marker, the integrity of the BBB can be determined. Our vascular volume determination (0.77 ± 0.12%) did not differ significantly from previous reports of 0.7%–1.0% [7, 14, 15]. On the other hand, diazepam is a very lipophilic compound that undergoes rapid transmembrane diffusion. The K_in of [³H]diazepam (0.2 μ Ci/mL) was used in determining the cerebral blood flow (F value) at the BBB. Our F value (5.19 ± 0.23 × 10⁻² ml/s/g) is in close agreement with a previous report (4.91 × 10^{− 2} ml/s/g) [15]. Urea is a small, polar compound that enters the brain relatively slowly, via diffusion. It exhibited a PA product of 3.38 ± 0.95 × 10⁻⁴ ml/s/g) [23]. L-Leu is an endogenous amino acid with high affinity for LAT1 at the BBB [7, 8]. The validation of radiotracer [¹⁴C]L-Leu (0.2 μ Ci/mL) brain uptake and its competitive self-inhibition demonstrated functional expression of LAT1 at the cerebrovasculature in the rat. The PA product obtained for [¹⁴C]L-Leu was 5.04 ± 0.26 × 10⁻² ml/s/g, which is not significantly different from reported data [7, 8]. By using these four molecules exhibiting significantly different BBB transport properties, the in situ brain perfusion method was validated.

In competition assays, [¹⁴C]L-Leu (0.2 μ Ci/mL) was coperfused with excess concentration of various competing molecules (0.1–2.0 mM) (Table 1). Thus, the relative affinity of the competing substrate for LAT1 at the BBB can be determined. Validation of the competition process for cerebrovascular LAT1 was performed by competing the radiotracer [¹⁴C]L-Leu with excess concentration of two known LAT1 substrates, either L-Leu or L-Phe. This competition demonstrated almost complete inhibition (89–93%) of [¹⁴C]L-Leu brain uptake, thereby validating cerebrovascular LAT1 functional expression. S-benzyl-L-cysteine (100 μ M), which is a stable sulfur-containing amino acid analog (resembling L-Phe), was also competed with radiotracer [¹⁴C]L-Leu. This competition assay for [¹⁴C]L-Leu brain uptake demonstrated that the thioether compound, S-benzyl-L-cysteine has significant affinity for LAT1, since 88% of control [¹⁴C]L-Leu transport was inhibited that was similar to competition with excess unlabeled L-Leu (89% inhibition). Therefore, the introduction of a “sulfur” into R group of an amino acid does not disrupt cerebrovascular LAT1 affinity.

TABLE 1 Influence of 0.1 mM concentration of various competing molecules on the brain capillary permeability-surface area (PA) product of tracer concentration of [¹⁴C]L-Leu in parietal cortex of rat brain

Addition (100 μM)	PA× 10^3 (mL/sec/g)	Percent inhibition
None (control)	50.4 ± 2.6	0
L-cysteine-thiourea	41.2 ± 5.7	18.3 ± 11.3
L-cysteine + MP^a	32.7 ± 0.7	35.1 ± 1.5^b
L-cysteine	32.2 ± 0.6	36.1 ± 1.3^b
MP-BTDS	18.2 ± 5.3	63.9 ± 10.9^b
Benzyl-L-cysteine	5.85 ± 0.2	88.4 ± 0.4^b
L-leucine	5.42 ± 1.8	89.2 ± 3.6^b
IBM-BTDS	3.94 ± 1.1	92.2 ± 2.3^b
L-phenylalanine^c	3.51 ± 0.91	93.0 ± 1.8^b

Each determination is the mean ± SE for 3–6 animals.

^aL-cysteine and MP were coperfused at 100 μM each.

^bSignificantly different from control (p < 0.05).

^c L-phenylalanine was used at 2 mM concentration.

To determine if BTDS possess affinity for cerebrovascular LAT1, BTDS (100 μ M) were coperfused with the radiotracer [¹⁴C]L-Leu. Concomitant IBM-BTDS and [¹⁴C]L-Leu perfusion resulted in 92% inhibition of [¹⁴C]L-Leu brain uptake. These data demonstrate that IBM-BTDS has potent affinity for cerebrovascular LAT1, similar to that of L-Leu and S-benzyl-L-cysteine. MP-BTDS also was separately coperfused with [¹⁴C]L-Leu and it demonstrated significant inhibition of [¹⁴C]L-Leu brain uptake (63%). To ensure that the competition for the radiotracer [¹⁴C]L-Leu with BTDS was not due to instability of BTDS during perfusion conditions, competition between L-Cys and [¹⁴C]L-Leu also was performed. L-Cys inhibited [¹⁴C]L-Leu brain uptake by only 35%, which was significantly lower than the inhibition by MP-BTDS and IBM-BTDS (63% and 92%, respectively). In control studies, the combination of L-Cys and MP also was competed with [¹⁴C]L-Leu that showed similar inhibition (34%) as with L-Cys alone (35%). This demonstrated that MP does not interfere with the LAT1-mediated, [¹⁴C]L-Leu brain uptake and that BTDS can inhibit [¹⁴C]L-Leu brain uptake significantly much greater than L-Cys alone. It was thus confirmed that BTDSs are potent substrates for the cerebrovascular LAT1.

To determine affinity of BTDS for cerebrovascular LAT1, the concentration of BTDS that inhibits 50% of [¹⁴C]L-Leu brain uptake (i.e., K_i) was determined as shown in Figure 2. The K_i for IBM-BTDS was calculated using [¹⁴C]L-Leu brain uptake at various concentrations of competing IBM-BTDS. A modified Michaelis-Menten equation was used to calculate K_i in 6 different brain regions: frontal cortex (9.4 ± 2.7 μM), parietal cortex (11.8 ± 1.3 μM), occipital cortex (8.1 ± 2.5 μM), hippocampus (12.8 ± 4.1 μM), caudate nucleus (7.0 ± 2.1 μM), and thalamus (6.9 ± 2.5 μM). These data suggest no significant difference in K_i values for IBM-BTDS in different brain regions. Therefore, functional expression of the cerebrovascular LAT1 is relatively similar in various brain regions. The inset in Figure 2 depicts a Lineweaver-Burk plot (of 1/J vs. 1/C) for [¹⁴C]-L-Leu brain uptake in presence of increasing concentration of the competing IBM-BTDS in parietal cortex of rat brain. The x-intercept (−1/K_i) of the Lineweaver-Burk for IBM-BTDS yielded a K_i of 12.3 μ M, which is not significantly different from parietal cortex K_i calculated from the modified Michaelis-Menten equation (Figure 3; 11.3 ± 2.8 μM). The K_i for IBM-BTDS demonstrates LAT1 affinity similar to that of the endogenous amino acids L-Phe (K_m = 11 μ M) [6] and L-Leu (K_m = 29 μ M) [7] using the rat brain perfusion technique.

FIG. 2 Influence of IBM-BTDS concentration on the brain capillary-permeability surface area (PA) product for [¹⁴C]L-Leu in various brain regions as determined using an in situ rat brain perfusion technique. Inset depicts Lineweaver-Burk plot (1/J versus 1/C) of [¹⁴C]L-Leu flux in presence of increasing concentration of IBM-BTDS, to yield a calculated K_i of 12.3 μ M for parietal cortex.

FIG. 3 [¹⁴C]-L-Leu parietal cortex brain uptake in presence of increasing concentration of IBM-BTDS. K_i was determined using the modified Michaelis-Menten equation (K_i = 11.3 ± 2.8 μM).

Similar affinity studies for cerebrovascular LAT1 using amino acid-based, anticancer agents such as melphalan (K_i = 90 μ M)¹⁶ and acivicin (K_i = 290 μ M¹⁶, K_m = 667 μ M¹⁵) yielded much lower affinity than BTDS. A rigid, hydrophobic analog of melphalan, DL-2-amino-7-bis[(2-chloroethyl)amino]-1,2,3,4-tetrahydro-2-napthoic acid (DL-NAM), exhibited remarkably high affinity for cerebrovascular LAT1 (K_m∼ K_i∼ 0.2 μ M) [14, 22]. The only note able differences between DL-NAM and melphalan are that DL-NAM possesses a marked increase in lipophilicity (∼ 10-fold) and steric constraint.²² However, brain uptake of DL-NAM was not as high as expected from its K_m or K_i value determinations due to a markedly reduced V_max, ∼ 1/20th that of melphalan [14, 22]. These data demonstrate that cerebrovascular LAT1 affinity does not always predict extent of brain uptake of the substrate.

The novel, synthesized BTDS possess significant affinity for cerebrovascular LAT1. The use of MP and IBM as BTDS R groups demonstrates flexibility of BTDS design. IBM (a hydrophobic, branched alkane) mimics the R group of L-Leu. MP (a purine base) is a model, highly polar, heterocyclic anticancer agent. Both IBM-BTDS and MP-BTDS exhibit significant cerebrovascular LAT1 affinity, despite significantly different physicochemical properties of the R group. MP-BTDS, the more hydrophilic analog, displays lower affinity which is consistent with previous findings that cerebrovascular LAT1 affinity decreases with side-chain hydrophilicity [7]. It appears that differences in amino acid R groups, such as polarity and length, can significantly alter cerebrovascular LAT1 affinity. Although cerebrovascular LAT1 has been shown to possess higher affinity for amino acids with lipophilic R groups [7], amino acids with small hydrophobic R groups (e.g., L-Ala) possess low LAT1 affinity. Importantly, the high LAT1 affinity exhibited by S-benzyl-L-cysteine and by BTDS demonstrates that sulfur introduction in the amino acid R group, either as a thioether or disulfide, does not compromise cerebrovascular LAT1 affinity.

CONCLUSION

Overall, BTDS may prove useful in improving affinity and specific recognition of various important drug and drug candidate molecules for the cerebrovascular LAT1. The BTDS strategy is flexible as it allows incorporation of various R groups in the BTDS structure. Although IBM-BTDS and MP-BTDS contain a disulfide linkage, a monosulfur (thioether) linkage as in S-benzyl-L-cytsteine also can be an effective linker of an R group to an amino acid. Based on the above data, it appears that the R group will be the major determinant of BTDS affinity for cerebrovascular LAT1. Previous attempts to make an extremely high affinity LAT1 substrate have resulted in poor transport properties [22]. Therefore, it may be useful to create LAT1 substrates with an optimal balance between affinity and transport to achieve efficient brain drug delivery. Thus, the BTDS strategy potentially can be used in the effective brain drug delivery of clinically important neuropharmaceuticals.

Funding for this research was provided by the PHS Grant, CA74377 (PJC) and by the American Foundation for Pharmaceutical Education (DMK). We thank Dr. Jun Oki and Dr. Quentin Smith for the initial data on IBM-BTDS affinity.