Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

AME is potentially valuable as a drug delivery system for large-molecular drugs to the brain. A dynorphin-like analgesic peptide, E-2078, and an ACTH analogue, ebiratide, are transported into the brain capillaries via AME (Terasaki et al., 1991a, 1992; Shimura et al., 1991, 1992). Other large molecules that penetrate the BBB via AME include various polycationic proteins such as -endorphin-cationized albumin complex (Kumagai et al., 1987), histone (Pardridge et al., 1989) and avidin (Pardridge and Boado, 1991). However, the structural specificity of AME at the BBB has not yet been clarified.

In vitro study using BCEC is advantageous to assess the transport mechanisms of the BBB (Tsuji et al., 1992, 1993) because primary cultured BCEC have favorable characteristics as a BBB model morphologically (intercellular tight junctions and no fenestra), biochemically (attenuated pinocytosis) and immunohistochemically (factor VIII) (see review by Borchardt, 1990). They are free from the influence of other cells and tissues that exist in vivo, and they maintain their polarity during cultivation.

The purpose of our study is to clarify the structural specificity of the AME system by synthesizing several peptides that were designed to have various molecular sizes, basicities, hydrophobicities and carboxyl-terminal structures and by using bovine BCEC in primary culture.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals. [³H]PEG 900 (74 to 370 MBq/g) and Na[¹²⁵I] (629 GBq/mg) were purchased from New England Nuclear (Boston, MA). Horse serum was purchased from GIBCO (Grand Island, NY). Rat tail collagen (type I) was purchased from Collaborative Research Inc. (Bedford, MA), and human fibronectin was from Boehringer Mannheim GmbH (Mannheim, Germany). Salmon roe protamine sulfate was purchased from Wako Pure Chemical Industries Ltd. (Osaka, Japan). Poly-L-lysine and bovine serum albumin (fraction V) and dansylcadaverine were purchased from Sigma Chemical Co. (St. Louis, MO) and Fluka Fine Chemical (Tokyo, Japan), respectively. Ebiratide (H-Met(O₂)-Glu-His-Phe-D-Lys-Phe-NH(CH₂)₈NH₂) and E-2078 (H-MeTyr-Gly-Gly-Phe-Leu-Arg-MeArg-D-Leu-NHC₂H₅) were kindly supplied by Hoechst Japan Ltd., (Kawagoe, Japan) and Eisai Co., Ltd. (Tokyo, Japan), respectively. All other chemicals were commercial products of reagent grade.

Synthetic peptides. 001-C8 and its derivative peptides used in our study were synthesized in this laboratory. Briefly, the synthesis of 001-C8 was mainly carried out by use of the conventional DCC or EDC/HOBt method. 1,8-Octanediamine as amide component was converted into monobenzyloxycarbonyl derivative coupled with the D-Leu residue. N-Methylamino acids in 001-C8 were prepared as N-protected derivative by simple N-methylation of N-Boc-O-benzyltyrosine with NaH and CH₃I in THF and based on retro aza Diels-Alder reaction for N-methyltyrosine and N-methylarginine, respectively. In a similar manner, various peptide analogues shown in table 1 were synthesized.

View this table: [in this window] [in a new window]   TABLE 1 The primary structure, molecular weight (MW), isoelectric point (pI) and capacity factor (k) of the synthetic basic peptidesa

Radioiodination of peptides. Each peptide was labeled with ¹²⁵I by the chloramine-T method (Hunter and Greenwood, 1962) as follows: 10 µg of each peptide were mixed with 18.5 MBq of Na[¹²⁵I] in 0.5 M phosphate buffer, pH 7.4 (total volume 60 µl), then 10 µl of 0.25% chloramine-T were added and allowed to react at room temperature for 30 sec. The reaction was stopped by addition of 0.25% sodium metabisulfite (25 µl) and 10 µl of 10% potassium iodide were added as a carrier. The reaction mixture was purified by HPLC using reversed phase analytical column, VYDAC 214TP104 (Separation Group Co., Ltd., Hasperia, CA). Mobile phases were mixture of water and acetonitrile at the ratio 99.9:0.1 for 101-C2 (approximate retention time: 6 min), 99:1 for 101-A (7) and 101-C5 (8), 95:5 for 101-OH (7), 92:8 for 003-C8 (15) and 101-EA (9), 90:10 for 101-C8 (14), 85:15 for 001-C8 (12) and 001-EA (10), 80:20 for 001-OH (9) and 002-C8 (15) and 70:30 for 004-C8 (12), respectively, each containing 0.1% trifluoroacetic acid and their flow rate was 1.0 ml/min. Each [¹²⁵I]peptide obtained had a specific activity of about 11.1 TBq/g and a chemical purity of >95%.

Isolation and culture of BCEC. BCEC were isolated from cerebral gray matter of bovine brains by the reported method (Audus and Borchardt, 1986) with minor modifications (Terasaki et al., 1991b). The isolated BCEC were stored at -100°C in culture medium containing 20% horse serum and 10% dimethylsulfoxide until use for cell culture. Before seeding, dishes (four-well multidish, Nunc, Denmark) were coated with rat tail collagen under UV light and then with human fibronectin. Isolated BCEC were seeded on the dishes and cultured at 37°C with 95% air and 5% CO₂. Transport experiments were performed when the cells reached confluence, after 10 to 12 days. These primary cultured cells have been shown to retain the morphological properties typical of brain capillary endothelial cells in vivo, i.e., tight junctions. Furthermore, the cells were confirmed to be capillary endothelial cells by the immunostaining method using factor VIII-related antigen (Meresse et al., 1989) (data not shown).

Uptake studies using cultured BCEC. Uptake of [¹²⁵I]peptide into cultured monolayers of BCEC was examined by a method reported previously (Hughes and Rantos, 1989) with minor modifications. Briefly, cultured cell monolayers were washed three times with 1 ml of incubation solution (141 mM NaCl, 4 mM KCl, 2.8 mM CaCl₂, 1 mM MgSO₄, 10 mM 2-[4-(2-hydroxyethyl)-1-piperazinyl]ethanesulfonic acid, pH 7.4, 0.1% bovine serum albumin) at 37°C. The uptake experiment was initiated by adding 250 µl of incubation solution containing [¹²⁵I]peptide (1.0 µCi) or [³H]PEG900 (1.0 µCi) to cells. [³H]PEG900 was used as the extracellular space marker. At designated times after incubation, cells were washed three times with 1 ml of the ice-cold incubation solution to terminate uptake. An acid-wash technique (Terasaki et al., 1992) was then used to distinguish the surface-bound and the internalized [¹²⁵I]peptide. The acid treatment removes [¹²⁵I]peptide bound to the cell surface. After the above uptake procedure, cells were incubated for 10 min with 1 ml of ice-cold acetate-barbital buffer (28 mM CH₃COONa, 120 mM NaCl, 20 mM barbital-HCl, pH 3.0) at 4°C. Then, the buffer was removed, and cells were washed four more times with 1 ml of acetate-barbital buffer. The radioactivity in the cells was measured after solubilization with 250 µl of 1 M NaOH for 1 hr at room temperature and represents internalized [¹²⁵I]peptide. Radioactivity was measured with a gamma-counter, ARC-600 (Aloka Co. Ltd., Tokyo, Japan) and a liquid scintillation counter, LSC-700 (Aloka Co. Ltd., Tokyo, Japan) for ¹²⁵I and ³H, respectively. Protein contents of cultured cells were determined as described by Lowry et al. (1951) with bovine serum albumin as a standard. The number of experiments (n) described in table footnote or figure legend represents the number of wells of cultured cells in each experiment.

HPLC analysis. Unchanged [¹²⁵I]001-C8 and its metabolites in the acid-resistant fraction were analyzed by HPLC. Acid-washed cells were solubilized with 1 M NaOH. Each sample was evaporated to dryness under reduced pressure at room temperature and reconstituted in the mobile phase used for HPLC assay. The HPLC analysis conditions were as follows; column, VYDAC 214TP54 (Separations Group Co., Ltd., Hasperia, CA); mobile phase, a mixture of water, acetonitrile and trifluoroacetic acid (15:85:0.1); flow rate, 1.0 ml/min. The eluates were collected with a fraction collector FRAC-100 (Pharmacia, Tokyo, Japan) and the radioactivity in each fraction (0.5 ml) was measured.

Data analysis. Total and acid-resistant bindings were expressed as the cell to medium ratio as follows: acid-resistant binding (µl/mg protein) = ¹²⁵I-R/mg of BCEC protein/(¹²⁵I-M/µl of medium) (equation 1).

Peptide basicity and lipophilicity. Basicity of peptides was represented by the theoretical isoelectric point, which was estimated from the acid-base dissociation constants of ionizable functional groups in each peptide.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Design of peptides for AME. All of peptides were synthesized to have cationic charges at physiological pH with estimated isoelectric points between 10 and 13, which is supposed to be an important factor for AME and were basically designed to have a part of amino acid residues of E-2078 and the carboxyl-terminal structures of ebiratide, both of which have been clarified to cross the BBB by AME (Terasaki et al., 1991a, 1992; Shimura et al., 1991, 1992). Furthermore, to assess the optimal physicochemical and structural features of peptides for AME, a variety of basicities, hydrophobicities, molecular sizes and carboxyl-terminal structures were introduced as shown in table 1.

Time course of the total and acid-resistant bindings of [¹²⁵I]001-C8 to primary cultured bovine BCEC. The total and acid-resistant bindings of a model peptide, [¹²⁵I]001-C8 (H-Me[¹²⁵I]Tyr-Arg-MeArg-D-Leu-NH(CH₂)₈NH₂), to primary cultured BCEC were compared with those of [³H]PEG900, the extracellular space marker. Figure 1 shows the time courses of the total and acid-resistant bindings of both compounds. The results were expressed as the cell to medium ratio, which was calculated from equation (1) as described in "Materials and Methods." Total and acid-resistant bindings of [¹²⁵I]001-C8 increased time-dependently and were significantly higher than those of [³H]PEG900, which showed constant bindings over the 120-min incubation period. The total and acid-resistant bindings of [¹²⁵I]001-C8 both reached a steady-state at 60 min, and the values amounted to over 110 and 25 µl/mg protein, respectively. In the following experiments, the acid-resistant binding at 60 min was used to evaluate the steady-state uptake of peptides.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   The time courses of the total (A) and acid-resistant bindings (B) of [125I]001-C8 and [3H]PEG900 to primary cultured BCEC. BCEC were incubated with [125I]001-C8 (1 µCi) (closed symbols) or [3H]PEG900 (1 µCi) (open symbols) for 5 to 120 min at 37°C. The total and acid-resistant bindings of each substrate were determined as described in "Materials and Methods." Each point represents the mean ± S.E.M. of four experiments.

Temperature-dependency and effects of endocytosis inhibitor and various cationic peptides on acid-resistant binding of [¹²⁵I]001-C8 and [¹²⁵I]101-C8. Table 2 shows the temperature dependency and the effects of an endocytosis inhibitor and several cationic peptides on the acid-resistant binding of [¹²⁵I]001-C8 and [¹²⁵I]101-C8 to BCEC. At 4°C, the acid-resistant binding of [¹²⁵I]001-C8 was reduced to 38% of the control value obtained at 37°C. An endocytosis inhibitor, dansylcadaverine (500 µM) significantly decreased the acid-resistant binding of [¹²⁵I]001-C8. Polycationic peptides, poly-L-lysine (300 µM) and protamine (300 µM) also inhibited the acid-resistant binding of [¹²⁵I]001-C8. It was also reduced by the structurally analogous compounds, E-2078 (200 µM) and ebiratide (200 µM). Similar results were obtained for its dipeptide counterpart [¹²⁵I]101-C8 (H-Me[¹²⁵I]Tyr-Arg-NH(CH₂)₈NH₂).

Structural specificity for synthetic peptides in terms of total and acid-resistant bindings to BCEC. The structural specificity of AME activity of the primary cultured bovine BCEC was assessed by using a series of di- and tetrapeptide derivatives varying in basicity and carboxyl-terminal structure (listed in table 1). Table 3 shows the total and acid-resistant bindings of ¹²⁵I-labeled peptide derivatives and [³H]PEG900 to cultured monolayers of BCEC at 37°C for 60 min. The total binding of peptides with 1,8-octanediamine (003-C8, 001-C8, 004-C8, 002-C8 and 101-C8, in decreasing order), 1,5-pentanediamine (101-C5), 1,2-ethanediamine (101-C2) and ethylamide (001-EA) moieties was significantly higher than that of [³H]PEG900. Peptides containing a free terminal carboxyl group showed similar binding to that of [³H]PEG900. 003-C8 and 001-C8 (which have three and two arginine residues, respectively) showed noticeably higher binding values than peptides with fewer arginine residues. The values of acid-resistant binding were affected by the C-terminal structure and were in the same order as those of the total binding, except for 001-C8 and 003-C8 (the acid-resistant binding of 001-C8 was the highest among the peptides used).

Concentration-dependency of the acid-resistant binding of model peptides. The acid-resistant binding of peptides modified with 1,8-octanediamine and 1,5-pentanediamine to primary cultured BCEC was saturable (fig. 2). The binding parameters for these peptides estimated by nonlinear least-squares analysis (Yamaoka et al., 1981) are listed in table 4. The maximal acid-resistant binding, the K_d and the nonsaturable binding of tetrapeptide derivatives were in the ranges of 1.91 to 91.1 pmol/mg protein, 0.24 to 22.3 µM and 0.31 to 2.94 µl/mg protein, respectively. The K_d values obtained for dipeptide derivatives 101-C5 and 101-C8 were 134 and 0.47 µM, respectively. Saturability was not observed in the acid-resistant bindings of 101-C2 and 101-A.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Concentration dependence of acid-resistant binding of [125I]tetrapeptide (A) and [125I]dipeptide (B) derivatives to primary cultured BCEC 125I-Labeled peptides were incubated with BCEC for 60 min at 37°C with increasing concentrations of unlabeled peptide (100 nM - 1 mM). The acid-resistant bindings were determined as described in the legend to figure 1. (A: open circles, 001-C8; closed circles, 002-C8; open triangles, 003-C8; closed triangles, 004-C8; B: open circles, 101-C8; closed circle, 101-C5). Each point represents the mean ± S.E.M. of three or four experiments.

Relationship between the basicity of peptides and affinity and uptake capacity by AME. The optimal structure for the uptake by AME of the cationic peptides were examined by the relationship between the basicity (represented by isoelectric point) and the values of 1/K_d (representing affinity constant) or maximal acid-resistant binding. As clearly shown in figure 3, there is tendency that the more the basicity increases, the more the affinity increases and the maximal acid-resistant binding value decreases.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Effect of basicity of 1,8-diaminooctane-modified peptides on maximal uptake capacity (Bmax) and half-saturation constant (Kd). The affinity constant (1/Kd) and the maximal uptake capacity (Bmax) are each shown as a function of isoelectric point. Points represent 004-C8, 002-C8, 001-C8 and 003-C8 (from left to right).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

To clarify the structural specificity of AME of the primary cultured bovine BCEC, we used the novel synthetic peptide 001-C8 and its derivatives with different numbers of basic and neutral amino acids and with various carboxyl-terminal structures (table 1). 001-C8 consists of a partial amino acid sequence of E-2078 and the carboxyl-terminal structure of ebiratide, with two arginine residues and an aminooctylamide residue. The estimated isoelectric point, 12.5 (net positively charged at physiological pH), is important for recognition by AME. As reported in the cases of enkephalin (Pardridge and Mietus, 1981) and somatostatin (Pardridge et al., 1985), enzymatic degradation should be considered in in vitro transport studies. The HPLC chromatogram of [¹²⁵I]001-C8 in the acid-resistant binding fraction after incubation at 37°C for 60 min with cultured monolayers of BCEC showed that more than 30% of total radioactivity remained in the intact peptide (data not shown). Because this peptide is possibly metabolized by BCEC after internalization presumably in secondary lysosome containing hydrolytic enzymes (Broadwell and Banks, 1993) and is quite stable in the extracellular medium because of methylations, carboxyl-terminal modification and an introduction of D-isomer of leucine, the initial association of the peptide to negatively charged cell surface components may not be affected by the metabolism observed in the intracellular space. The negligible uptake of 001-OH, 101-OH and 101-A, which are comparable with that of PEG900, may be ascribed to the low affinity to AME and/or metabolism before the internalization.

We first evaluated the transport mechanism of our peptides into BCEC. The apparent uptake of 001-C8 at steady-state, 25 µl/mg protein, was significantly higher than that of PEG900, 0.84 µl/mg protein. The observed uptake of PEG900 was comparable with that of Lucifer Yellow, approximately 0.5 µl/mg protein, which is taken up by fluid-phase endocytosis mechanism by BCEC (Guillot et al., 1990). Because such a significant difference in the uptakes by BCEC between AME and fluid-phase endocytosis has been also demonstrated using horseradish-peroxidase-conjugated lectin (ricin communis agglutinin I) and free horseradish-peroxidase (Raub and Audus, 1990), 001-C8 can be said to be internalized by a specialized mechanism other than fluid-phase endocytosis. Uptake of [¹²⁵I]001-C8 and [¹²⁵I]101-C8 in the steady-state was temperature dependent, and inhibited by an endocytosis inhibitor, dansylcadaverine, that suppresses formation of coated pits by inhibiting transglutaminase at the cell membrane (Haigler et al., 1982), suggesting that an endocytosis mechanism operates in cultured BCEC. It was also inhibited by cationic peptides, including protamine, poly-L-lysine, E-2078 and ebiratide (table 2). E-2078 and ebiratide have been reported to be transported through the BBB via AME based on isolated brain capillary studies (Terasaki et al., 1989; Shimura et al., 1991) or the use of cultured BCEC (Terasaki et al., 1992). The steady-state acid-resistant binding of [¹²⁵I]001-C8 to BCEC (25.0 µl/mg protein at 60 min) was greater than that of ebiratide (1.65 µl/mg protein at 120 min, Terasaki et al., 1992). The K_d value obtained by nonlinear least-squares analysis may offer one criterion for judging AME. Observed K_d values in this study (0.2-22 µM) are comparable with those for substrates reported to be taken up into BCEC via the AME mechanism, including ebiratide (15.9 µM) (Terasaki et al., 1992), E-2078 (4.62 µM) (Terasaki et al., 1989), histone (15.2 µM) (Pardridge et al., 1989) and cationized bovine serum albumin (0.8 µM) (Kumagai et al., 1987). In contrast, the K_d values for receptor-mediated endocytosis reported for atrial natriuretic factor (0.4 nM) (Smith et al., 1988), transferrin (5.6 nM) (Pardridge et al., 1987) and insulin (2.3 nM) (Frank and Pardridge, 1981) are several thousand times smaller than those for AME. All of these data indicate that both 001-C8 and 101-C8 are taken up into BCEC via the AME mechanism.

As shown from the result in figure 3, basicities of peptides significantly affected on the affinity and capacity of AME, with increasing affinity and decreasing capacity by an increase of isoelectric point values. Accordingly, it may be important for the optimal brain delivery of peptides by AME to give an moderate basicity. Based on the C-terminal structure and uptake profile, the peptides were classified into three types. 1) OH-type peptides, which contain a free carboxyl group in the C-terminal. This group was not internalized into cultured BCEC, because total and acid-resistant bindings of 001-OH were as small as that of [³H]PEG900. 2) EA-type peptides, which have an ethylamide group in the C-terminal. The acid-resistant binding of 001-EA was slight but significantly higher than that of PEG900. Peptides of this type may be internalized into BCEC via AME because they include a partial structure of E-2078. 3) Cn-type peptides, which were modified by alkanediamine in C-terminus. Total and acid-resistant bindings of peptides of this type were notably high compared to those of the other two types of peptides. Moreover, total and acid-resistant bindings were well correlated with the length of the C-terminal alkanediamine, as can be seen in the series of 101 peptides. As for the number of amino acid residues of the peptide, the uptake of the dipeptide derivative 101-C8 into the cultured BCEC was comparable to that of C8 modified tetrapeptide, 002-C8 (table 3). The degree of uptake may be affected by the change of certain physicochemical properties, such as charge density in the molecule and/or lipophilicity, rather than by the number of amino acid residues. These results suggest that not the number of constituent amino acids of the peptide, but rather the C-terminal structure and the basicity of the molecule are the most important determinants for uptake by the AME system of cultured BCEC.

In conclusion, in this study, 001-C8 and its derivatives were confirmed to be taken up into BCEC via an AME system in a structure-dependent manner. To achieve the practical delivery of pharmacologically active peptides into the brain interstitial fluid across the BBB, it is necessary to take into consideration the intracellular dynamics of such peptides and also exocytosis to the brain. Because the vectorial movement of micro- and macromolecules from blood-to-brain side has been suggested in adsorptive-mediated transcytosis as well as receptor-mediated transcytosis (Broadwell and Banks, 1993), AME-mediated delivery of peptides into brain is hopeful. We are now addressing these points by means of visual three-dimensional analysis using high-resolution confocal microscopy and/or in vivo studies using capillary depletion and brain microdialysis methods.