Introduction

Top Abstract Introduction Methods Results Discussion References

It is very important to develop tissue-specific targeting systems to make drugs safer and more effective. To this end, a number of different methods have been developed and are currently under investigation (Suzuki et al., 1996; Meijer and van der Sluijis, 1989). In such studies, tissue-specific targeting vectors can play an important role (Basu, 1990). So far, some of the vectors investigated include polypeptides (Bron et al., 1994), antibodies (Goren et al., 1996; Muzykantov et al., 1996), viruses (Cristiano et al., 1993; Kasahara et al., 1994), fatty acids (Charbon et al., 1996), and sugars (Gonsho et al., 1994; Nishikawa et al., 1993). As far as the use of sugar moieties is concerned, many studies have been performed using the asialoglycoprotein receptor system in the liver and this has been shown to be very useful for delivering a variety of low-molecular-weight drugs, high-molecular-weight proteins, and genes (Monsigny et al., 1994). However, there have been few reports on other tissues concerning tissue-specific delivery using sugar moieties (Palomino, 1994). During our own investigations, we recently found that the low-molecular-weight model peptide arginine vasopressin (AVP), when modified by linking it to some sugars via an octamethylene group, exhibits renal-selective and efficient uptake (Suzuki et al., 1999). We suggest that the renal uptake of these glycosylated peptides depends on the structure of the sugar; it takes place from blood and involves specific binding to the renal microsomal fraction. After this, these peptides are then distributed in the proximal tubules. In this report, we began by studying the inhibitory effects of glycosides on the binding of glycosylated peptide ([³H]Glc-S-C8-AVP) to a kidney membrane fraction. We found that Glc-S-C7-Me was recognized by the kidney and, therefore, we went on to investigate the possibility that Glc-S-C7-Me might be a useful vector for renal targeting.

	Methods

Top Abstract Introduction Methods Results Discussion References

Materials. AVP and tryptamine were purchased from Peptide Institute (Minoh, Japan). [³H]AVP (vasopressin, 8-L-arginine [phenylalanyl-3,4,5-³H]-, 2890 GBq/mmol), and [³H]tryptamine (758.5 GBq/mmol) were purchased from New England Nuclear (Boston, MA). [³H]Glc-S-C7-Me was synthesized by Amersham (Buckinghamshire, UK). [¹⁴C]Glc-O-C7-Me was purchased from American Radiolabeled Chemicals Inc. (St. Louis, MO) All other chemicals were commercially available reagents.

Synthesis, Purification, and Characterization of Glycosylated Derivatives. The structures of the compounds used in this article, namely, AVP, Glc-S-C5-AVP, Glc-S-C8-AVP, Glc-S-C11-AVP, Glc-S-C8-Tryp, and Glc-S-C8-NBD are shown in Fig. 1. The general procedure for the preparation of these compounds was described in one of our previous articles (Susaki et al., 1994). Preparative high-performance liquid chromatography (HPLC) was performed using a YMC SH 345-5 S5 120A ODS column ( 20 × 300 mm) and the mobile phase was a mixture of 0.05% TFA-containing acetonitrile-water at a flow rate of 10 ml/min. [³H]Glc-S-C5-AVP, [³H]Glc-S-C8-AVP, [³H]S-C11-AVP, and [³H]Glc-S-C8-Tryp were synthesized from [³H]AVP or [³H]tryptamine using the same procedure for nonradiolabeled compounds as described below. The specific radioactivities were as follows: [³H]Glc-S-C8-AVP, 106.9 GBq/mmol; [³H]Glc-S-C5-AVP, 227.9 GBq/mmol; [³H]Glc-S-C11-AVP, 119.1 GBq/mmol; [³H]Glc-S-C8-Tryp, 264.2 GBq/mmol; [³H]Glc-S-C7-Me, 40.9 GBq/mmol; and [¹⁴C]Glc-O-C7-Me, 12.6 GBq/mmol.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Structure of the compounds used: A, AVP; B, Glc-O-C8-AVP; C, Glc-S-Cn-AVP (n = 5, Glc-S-C5-AVP; n = 8, Glc-S-C8-AVP; n = 11, Glc-S-C11-AVP); D, Glc-S-C8-Tryp; and E, Glc-S-C8-NBD.

Synthesis of Glc-S-C8-AVP. A mixture of 8-(methoxycarbonyl)octanol (1.16 g, 6.17 mmol), prepared according to the Kann method (Kann et al., 1990), triphenylphosphine (2.43 g, 9.26 mmol), and tetrabromomethane (3.07 g, 9.26 mmol) in toluene (5 ml) was stirred at room temperature for 3 h and then filtered. The filtrate was concentrated and the residue was chromatographed on a silica gel column with toluene/ethyl acetate (49:1) to give methyl 9-bromononate (1.37 g, 88%) as an oil. A mixture of methyl 9-bromononate (814 mg, 3.24 mmol), 2-(2,3,4,6-tetra-O-acetyl--D-glucopyranosyl)thiopseudourea hydrobromide (1287 mg, 2.70 mmol), prepared by according to method of Saito (Saito and Tsuchiya, 1985), sodium hydrogen sulfite (257 mg, 2.46 mmol), and potassium carbonate (386 mg, 2.80 mmol) in acetone (5 ml) and water (5 ml) was stirred at room temperature for 19 h. The reaction mixture was then diluted with chloroform, washed with water, and the organic phase dried over magnesium sulfate and evaporated. Chromatography of the residue on a silica gel column with toluene/ethylacetate (9:1) gave 8-methoxycarbonyloctyl 2,3,4,6-tetra-O-acetyl-1-thio--D-glucopyranoside (980 mg, 68%) as a colorless solid.

Synthesis of Glc-S-C8-Tryp. Triethylamine (0.025 ml) was added to a solution of tryptamine (32 mg, 0.20 mmol) in methanol (2 ml), and the mixture was evaporated to dryness. To a solution of tryptamine (32 mg, 0.20 mmol) in DMF (2 ml), 9-(1-thio--D-glucopyranosyl)nonanoic acid (88 mg, 0.30 mmol), HOBt (54 mg, 0.40 mmol), and didcyclohexylcarbodiimide (82 mg, 0.40 mmol) were added, and the resulting mixture was stirred for 18 h. After the addition of 10% aqueous TFA to bring the pH to 2.0, the reaction mixture was subjected to preparative HPLC to give Glc-S-C8-Tryp (20 mg, 20%) as a colorless, amorphous solid. []_D 37.6° (c = 1.02, methanol). ¹H NMR (D₂O) : 1.06 to 1.10 (2H, m), 1.16 to 1.22 (2H, m), 1.30 to 1.36 (2H, m), 1.40 to 1.46 (2H, m), 1.60 to 1.64 (2H, m), 2.14 (1H, t, J = 7.3 Hz), 2.69 to 2.78 (2H, m), 3.02 (1H, t, J = 6.6 Hz), 3.31 (1H, dd, J = 9.0, 9.8 Hz), 3.39 to 4.42 (2H, m), 4.49 (1H, d, J = 10.0 Hz), 7.18 (1H, dd, J = 7.1, 7.3 Hz), 7.25 to 7.28 (2H, m), 7.52 (1H, d, J = 8.3 Hz), and 7.71 (1H, d, J = 8.1 Hz). Infared (KBr): 1634, 1546 cm¹. FAB-MS m/z: 495 (MH⁺). Analysis calculated for C₂₅H₃₆O₆N₂S · 3/2 H₂O: C, 57.78; H, 7.56; N, 5.39. Found: C, 57.61; H, 7.21; N, 5.39.

Synthesis of Glc-S-C8-NBD. A mixture of NBD-Cl (900 mg, 4.5 mmol), N-tert-butoxycarbonyl-ethylendiamine (0.178 ml, 1.13 mmol), and sodium hydrogen carbonate (287 mg) in ethanol (30 ml) was stirred at room temperature for 4 h. The reaction mixture was then diluted with chloroform and washed with water. The organic phase was dried over magnesium sulfate and evaporated. Chromatography of the residue on a silica gel column with toluene-ethyl acetate (85:15) gave a red, amorphous solid (220 mg, 67%). To a portion of this red amorphous solid (50 mg, 0.15 mmol), anisole (0.133 ml) and TFA (2 ml) were added and the mixture was stirred at room temperature for 2.5 h. The mixture was then evaporated to dryness and methanol (4 ml) and triethylamine (0.2 ml) were added to the residue, followed by evaporation to dryness. To the residue, DMF (2 ml), 9-(1-thio--D-glucopyranosyl)nonanoic acid (67 mg, 0.21 mmol), HOBt (27 mg, 0.20 mmol), and DCC (47 mg, 0.30 mmol) were added and the mixture was stirred for 16 h. After the addition of 10% aqueous TFA to bring the pH to 2.0, the reaction mixture was subjected to preparative HPLC to give Glc-S-C8-NBD (60 mg). []_D 26.2° (c = 0.21, methanol). ¹H NMR (CD₃OD) : 1.25 to 1.36 (8H, m), 1.53 to 1.60 (4H, m), 2.17 (1H, t, J = 7.6 Hz), 2.64 to 2.74 (2H, m), 3.19 (1H, dd, J = 8.8, 9.5 Hz), 2.64 to 2.74 (2H, m), 3.19 (1H, dd, J = 8.8, 9.5 Hz), 3.26 to 3.36 (2H, m), 3.54 (1H, t, J = 5.9 Hz), 3.63 to 3.67 (1H, m), 3.65 (1H, dd, J = 5.4, 12.0 Hz), 3.85 (1H, dd, J = 1.5, 12.0 Hz), 4.34 (1H, d, J = 9.5 Hz), 6.42 (1H, d, J = 8.8 Hz), 8.53 (1H, d, J = 8.8 Hz). Infared (KBr): 1622, 1588 cm¹. FAB-MS m/z: 558 (MH⁺). Analysis calculated (Anal Calcd) for C₂₃H₃₅O₉N₅S · 9/4 H₂O: C, 45.49; H, 6.72; N, 11.53. Found: C, 45.61; H, 6.72; N, 11.53.

Animals. Male Sprague-Dawley rats were purchased from Sankyo Labo Service Co., Ltd. (Tokyo, Japan). They were used when they were 6 to 7 weeks old and weighed 180 to 230 g. Food and water were available ad libitum. Each group consisted of three to four rats.

Early-Phase Tissue Uptake of Radiolabeled Glycosylated Derivatives. The plasma concentration-time profiles of radioactivity over 5 min and the tissue radioactivity after i.v. administration of [³H]glycosylated AVP derivatives [³H]Glc-S-C7-Me, [¹⁴C]Glc-O-C7-Me, [³H]Glc-S-C-8-Tryp, or [³H]tryptamine were determined as follows. Rats anesthetized with diethyl ether were injected i.v. with 1 nmol/kg radiolabeled carbohydrate conjugates. Blood samples (0.2 ml) were collected from the jugular vein by syringe at 1, 2, 3, and 5 min and plasma was quickly separated by centrifugation. Immediately after the last blood sampling, the rats were sacrificed by bleeding and liver, kidney, heart, lung, small intestine, bone marrow, brain, muscle, and spleen were excised and rinsed with water. The samples of plasma, portions of tissue, or entire tissues were weighed and combusted using an automatic sample combustion system (Aloka ASC-113; Tokyo, Japan) to determine the radioactivity by liquid scintillation counting (Aloka LSC-602). The apparent uptake clearance (CLup,app) for each tissue was calculated from the following equation:

(1)

where X(5 min) is the tissue concentration at 5 min and area under the curve (AUC)_0-5 min is the area under the plasma concentration-time curve from 0 to 5 min.

Early-Phase Tissue Uptake of Glycosylated Derivatives of 4-nitrobenz-2-oxa-1,3-diazole (NBD). The plasma concentration-time profiles over 3 min and the tissue concentration of the derivative after i.v. administration of the glycosylated derivative of NBD (Glc-S-C8-NBD) were determined as follows. Rats anesthetized with diethyl ether were injected i.v. with 10 nmol/kg carbohydrate conjugate. Blood samples (0.2 ml) were collected from the jugular vein by syringe at 20, 60, 100, 140, and 180 s and plasma was quickly separated by centrifugation. Immediately after the last blood sampling, the rats were sacrificed and liver, kidney, heart, lung, and spleen were excised and rinsed with water. Portions of tissues (liver and lung) or entire tissues (kidney, heart, and spleen) were weighed and homogenized in phosphate-buffered saline. To plasma and homogenates (25% w/v), 50 and 200 pmol of internal standard, N-[7-nitrobenz-2-oxa-l,3-diazole-4-yl]aminohexanoic acid (HOCO-C5-NBD), respectively, were added, followed by 67% acetonitrile to denature the proteins immediately and inhibit any further degradation of the NBD derivatives in the homogenate. These solutions were then centrifuged (10,000 rpm, 5 min at 4°C) in a TOMY centrifuge (model MR-150, Tokyo, Japan), and the supernatants were filtered through a 0.45-µm Chromatodisk 13P. Then, 400 µl of 0.05% TFA was added to 200 µl of filtrate and the solutions were analyzed by HPLC as described below. The CLup,app for each tissue was calculated from the following equation:

(2)

where X(3 min) is the tissue concentration at 3 min and AUC_0-3 min is the area under the plasma concentration-time curve from 0 to 3 min. The CLup,app of the reference compound, HOCO-C5-NBD, was determined as described above for Glc-S-C8-NBD. Glc-S-C8-NBD was used as an internal standard in the analysis of HOCO-C5-NBD concentrations by HPLC.

Plasma and Tissue Concentration-Time Profiles of Glc-S-C8-NBD. The plasma and tissue concentration-time profiles over 30 min and the tissue concentration of the derivative after i.v. administration of Glc-S-C8-NBD were determined as follows. Rats anesthetized with diethyl ether were given 10 nmol/kg of the carbohydrate-conjugated derivatives i.v. Blood (0.2 ml) and tissue samples were obtained at 1, 2, 3, 5, 15, and 30 min. The plasma and tissues (liver, kidney, heart, lung, and spleen) were treated as described above.

HPLC Analysis of Glc-S-C8-NBD and HOCO-C5-NBD. All samples (20-100 µl) were analyzed using a JUSCO HPLC system (Tokyo, Japan) fitted with a GL Science Inertsil ODS-2 column (4.6 × 150 mm) at a flow rate of 1 ml/min. For the detection of NBD derivatives, the fluorescence at 540 nm was measured after excitation at 470 nm. The mobile phase was 27% acetonitrile in 0.05% TFA. Typical retention times for Glc-S-C8-AVP and HOCO-C5-NBD were 7.7 and 14.7 min, respectively. Quantification of these compounds was achieved by comparison of peak areas with standard curves whose correlation coefficients were high, r² > 0.99. The limit of detection for Glc-S-C8-AVP and HOCO-C5-NBD was almost 0.8 nM. The recovery of NBD derivatives from the biological samples was over 95%.

Binding of ³H-Labeled Glycosylated Derivatives to Membrane Fractions Prepared from Kidney and Liver. Rat kidney and liver microsomes were prepared by centrifugation (Stassen et al., 1982). The ³H-labeled glycosylated derivative was incubated with the liver or kidney microsomal fractions (1 mg/ml as protein concentration) in phosphate-buffered saline (pH 7.4) on ice. For the analysis of AVP derivatives, phosphate-buffered saline containing 0.1% bovine serum albumin (BSA) was used for incubation. After ultracentrifugation (50,000g for 5 min at 4°C), the supernatant was aspirated and the precipitate was dissolved in 1% Tween 20 to measure the radioactivity using a liquid scintillation counter. [¹⁴C]sucrose was added to the incubation buffer to correct for any supernatant remaining in the precipitate. A Scatchard plot analysis of the ³H-labeled derivatives was performed at different concentrations by measuring the ligand total binding and the nonspecific binding that could not be inhibited by excess cold ligand (100 µM). Then the specific binding was calculated by subtracting the nonspecific from the total binding. The incubation time was determined by a preliminary equilibrium binding experiment as follows: 1 h for Glc-S-C7-Me and Glc-S-C8-AVP and 8 h for Glc-S-C11-AVP, respectively, with the other conditions being the same as described above. The binding parameters were obtained by fitting the relationship between the values of B and F to eq. 3 (Glc-S-C7-Me and Glc-S-C8-AVP) or eq. 4 (Glc-S-C-11-AVP) using the nonlinear least-squares methods (Kim et al., 1991)

(3)

(4)

where B and F represent the binding and free concentration of ligand, respectively.

Inhibition of the Binding of [³H]Glc-O-C8-AVP to the Rat Kidney Membrane Fraction. [³H]Glc-O-C8-AVP (20 pmol/ml), inhibitors, and kidney microsomes (1 mg/ml as protein concentration) were incubated in phosphate-buffered saline (pH 7.4) containing 0.1% BSA on ice for 1 h. The other conditions were as in the binding assay described above. All samples were examined in triplicate and 50% inhibitory concentration (IC₅₀) values were evaluated by log-logit analysis of the mean values from 3 to 4 points involving a ligand concentration at approximately 50% inhibition.

Inhibition of the Binding of [³H]Glc-S-C7-Me to the Rat Kidney Membrane Fraction. [³H]Glc-S-C7-Me (20 pmol/ml), inhibitors, and kidney microsomes (1 mg/ml as protein concentration) were incubated in phosphate-buffered saline (pH 7.4) on ice for 1 h. In the case of analysis of AVP derivatives, phosphate-buffered saline containing 0.1% BSA was used for incubation. The other conditions were as described above for the inhibition of the binding of [³H]Glc-O-C8-AVP to the rat kidney membrane fraction.

	Results

Top Abstract Introduction Methods Results Discussion References

Inhibitory Effects on [³H]Glc-O-C8-AVP Binding to the Kidney Membrane Fraction. The IC₅₀ (µM) values of Glc-O-C7-Me, Glc-S-C7-Me, Gal-S-C7-Me, Tween 20, 3-[3-cholamidopropyl)dimethylamino]-propanesulfonate (CHAPS), and Glc-O-C8-AVP were 2.9, 0.065, 120, 27, 1600, and 0.075, respectively (Fig. 2). The inhibitory effect of Glc-S-C7-Me was 40 times greater than O-glucoside (Glc-O-C7-Me), and more than 1000 times greater than S-galactoside (Gal-S-C7-Me). Also, inhibition by the detergents, Tween 20 and CHAPS, was less than that by Glc-S-C7-Me. The IC₅₀ value of Glc-S-C7-Me (0.065 µM) was almost the same as that of nonlabeled Glc-O-C8-AVP (0.075 µM).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Inhibition of the specific binding of Glc-O-C8-AVP to rat kidney membrane fraction. Inhibition of the binding of [3H]Glc-O-C8-AVP (20 pmol/ml) to membrane (1 mg/ml) was measured in the presence of different inhibitors, i.e., Glc-O-C7-Me (), Glc-S-C7-Me (), Gal-S-C7-Me (), Tween 20 (), CHAPS (), and Glc-O-C8-AVP (). All samples were tested in triplicate and values represent the mean.

Specific Binding of [³H]Glc-S-C7-Me to the Kidney Membrane Fraction. The study of Glc-S-C7-Me binding to the kidney membrane fraction (Fig. 3A) revealed the presence of a specific binding site, because the binding obtained by subtracting the nonspecific binding under excess cold Glc-S-C7-Me from total binding was saturated. Also, from a Scatchard analysis (Fig. 3A), the K_d and B_max were estimated to be 16.4 nM and 24.4 pmol/mg protein, respectively. However, [³H]Glc-S-C7-Me did not bind specifically to the liver membrane fraction (Fig. 3B).

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Change in total binding (), nonspecific binding (), and specific binding () of 3H-labeled Glc-S-C7-Me to rat kidney (A) and liver (B) microsomal fraction. Inset in A, Scatchard plots of the data. Values represent individual data for triplicate assays at each concentration.

In Vivo Tissue Uptake of Glc-S-C7-Me and the S-Glucosylated Derivative of AVP. The AUCs from 0 to 5 min for Glc-O-C7-Me, Glc-S-C7-Me, Glc-S-C5-AVP, Glc-S-C8-AVP, and Glc-S-C11-AVP were 7.1 ± 0.3, 4.3 ± 0.8, 24.3 ± 0.6, 17.5 ± 0.5, and 18.2 ± 1.0 pmol · ml¹ · min, respectively (values represent mean ± S.E. of three rats). There was no clear difference in the CLup,app of Glc-O-C7-Me (1 nmol/kg) in any of the tissues tested, and the CLup,app by kidney (0.39 ml/min/g) was almost the same as that by liver (0.48 ml/min/g) (Fig. 4A). However, in the case of Glc-S-C7-Me, the CLup,app by kidney (5.12 ml/min/g) was about 5 times greater than that by liver (0.93 ml/min/g) (Fig. 4A). The renal CLup,app of Glc-S-C7-Me was about 10 times greater than that of Glc-O-C7-Me (Fig. 4A). There was little difference in the tissue distribution of Glc-S-C5-AVP, Glc-S-C8-AVP, and Glc-S-C11-AVP, S-glucoside derivatives of AVP having 5, 8, and 11 methylenes, respectively, except in the case of the kidney (Fig. 4B). Although the CLup,app by kidney for Glc-S-C5-AVP (0.49 ml/min/g) was the same as that of AVP (0.49 ml/min/g), the corresponding values for Glc-S-C8-AVP and Glc-S-C11-AVP were much greater, 3.4 ml/min/g and 4.5 ml/min/g, respectively (Fig. 4B).

View larger version (12K): [in this window] [in a new window]   Fig. 4.   A, Apparent tissue uptake clearance (CLup,app) of Glc-S-C7-Me (black column) and Glc-O-C7-Me (white column) by various rat tissues. B, CLup,app of Glc-S-C5-AVP (white column), Glc-S-C8-AVP (hatched column), and Glc-S-C11-AVP (black column) by various rat tissues. These derivatives were administered i.v. at a dose of 1 nmol/kg. Blood sampling was performed at 1, 2, 3, and 5 min and tissue sampling was carried out 5 min after administration. CLup,app was calculated as described in Methods. Values represent means ± S.E. of three rats.

Binding of S-Glucoside Derivatives of AVP to the Kidney Membrane Fraction. From an analysis of the binding of Glc-S-C5-AVP, Glc-S-C8-AVP, and Glc-S-C11-AVP to the kidney membrane fraction, it was found that Glc-S-C5-AVP did not exhibit any specific binding (no saturation), whereas Glc-S-C8-AVP and Glc-S-C11-AVP bound with higher affinity than Glc-S-C7-Me. Scatchard analysis showed that Glc-S-C8-AVP had a single binding site on the membrane, the K_d and B_max, of which was 8.63 nM and 45.2 pmol/mg of protein, respectively (Fig. 5A). Glc-S-C11-AVP had both a high-affinity (K_d1 = 1.48 nM and B_max1 = 27.6 pmol/mg protein) and a low-affinity (K_d2 = 1730 nM and B_max2 = 250 pmol/mg protein) binding site (Fig. 5B).

View larger version (11K): [in this window] [in a new window]   Fig. 5.   Scatchard plots for the binding of [3H]Glc-S-C8-AVP (A) and 3H-Glc-S-C11-AVP (B) to rat kidney microsomal fraction. Lines represent the line fitted according to eq. 3 (A) or eq. 4 (B) as described under Methods. Values represent individual data for triplicate assays at each concentration.

In Vivo Tissue Uptake of the S-Glucosylated Derivative of Tryptamine. The AUCs from 0 to 5 min for tryptamine and Glc-S-C8-Tryp were 10.2 ± 0.5 and 5.3 ± 0.5 pmol · ml¹ · min, respectively (values represent mean ± S.E. of three rats). Compared with the CLup,app of tryptamine, that of the S-glucosylated derivative (Glc-S-C8-Tryp) by kidney (5.72 versus 0.63 ml/min/g), liver (1.89 versus 0.12 ml/min/g), and small intestine (1.49 versus 0.056 ml/min/g) was clearly greater (Fig. 6). There was little difference in the other tissues.

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Apparent tissue uptake clearance (CLup,app) of tryptamine (white column) and Glc-S-C8-Tryp (black column) by various rat tissues. These were administered at a dose of 1 nmol/kg. Blood sampling was performed at 1, 2, 3, and 5 min and tissue sampling was carried out 5 min after administration. CLup,app was calculated as described in Methods. Values represent means ± S.E. of three rats.

In Vivo Tissue Uptake of the S-Glucosylated Derivative of NBD. The AUCs from 0 to 3 min for Glc-S-C8-NBD and HOCO-C5-NBD (reference compound) were 113.9 ± 9.3 and 90.9 ± 7.5 pmol · ml¹ · min, respectively (values represent mean ± S.E. of four rats). The CLup,app by kidney of Glc-S-C8-NBD and HOCO-C5-NBD were 3.58 and 0.79 ml/min/g, respectively, and Glc-S-C8-NBD was shown to have renal targeting potential. There was little difference between the CLup,app of Glc-S-C8-NBD (0.29 ml/min/g) and HOCO-C5-NBD (0.19 ml/min/g) by liver. In addition, there was no difference between the CLup,app of these two compounds by spleen, lung, and heart (Fig. 7). Tissue (Fig. 8A) and plasma (Fig. 8B) concentrations were also measured for 30 min after i.v. administration. The compound rapidly disappeared from plasma and the concentration in tissues other than the kidney fell below the limit of detection (ca. 0.1% of the dose/g tissue) by 5 to 15 min. On the other hand, the concentration in kidney remained high, even at 30 min after administration. At 12.1% of the dose/g, this was more than 200 times higher than the concentration in plasma at that time (0.046% of the dose/ml).

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Apparent tissue uptake clearance (CLup,app) of Glc-S-C8-NBD (black column) and HOCO-C5-NBD (white column) by various rat tissues. These were administered at a dose of 10 nmol/kg. Blood sampling was performed at 20, 60, 100, 140, and 180 s and tissue sampling was carried out 180 s after administration. CLup,app was calculated as described in Methods. Values represent means ± S.E. of four rats.

View larger version (12K): [in this window] [in a new window]   Fig. 8.   Tissue concentration-time profile (A) and plasma concentration-time profile (B) of Glc-S-C8-NBD, administered at a dose of 10 nmol/kg. Tissue concentration (, heart; , lung; , spleen; , liver; , kidney) and plasma concentration () were determined by HPLC. Values represent means ± S.E. of four rats.

Inhibitory Effects of Thioglucoside Derivatives on [³H]Glc-S-C7-Me Binding to the Kidney Membrane Fraction. Various thioglucoside derivatives inhibited [³H]Glc-S-C7-Me binding to the kidney membrane fraction, the degree of inhibition depending on the concentration. Among these derivatives, Glc-S-C8-AVP, Glc-S-C11-AVP, and Glc-S-C8-NBD had almost the same inhibitory effect, whereas Glc-S-C8-Tryp was about 50% weaker (Fig. 9). In addition, Glc-S-C5-AVP had a much weaker inhibitory effect, its inhibition being about th that of Glc-S-C8-AVP (Fig. 9).

View larger version (21K): [in this window] [in a new window]   Fig. 9.   Inhibition of the specific binding of [3H]Glc-S-C7-Me to rat kidney membrane fraction. The binding of [3H]Glc-S-C7-Me (20 pmol/ml) to membrane (1 mg/ml) was measured in the presence of different inhibitors, i.e., Glc-S-C5-AVP (), Glc-S-C8-AVP (), Glc-S-C11-AVP (), Glc-S-C8-Tryp (), and Glc-S-C8-NBD (). Values represent mean ± S.D. of triplicate assay.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Our previous analysis of the renal targeting of [³H]Glc-O-C8-AVP suggested that receptors or transporters on the membrane of renal cells are involved in the renal uptake of [³H]Glc-S-C8-AVP, and that the sugar, alkyl, and peptide moieties are important for binding to the kidney membrane fraction (Suzuki et al., 1999). To estimate the structural requirements for renal targeting, we studied the inhibitory effects of various glycosides on [³H]Glc-O-C8-AVP binding to the kidney membrane fraction. Our findings showed that Glc-S-C7-Me inhibited binding to the same degree as Glc-O-C8-AVP (Fig. 2), and this inhibition was much stronger than was the case with Glc-O-C7-Me, suggesting that the S-glucoside has a higher affinity than the O-glucoside. In addition, even for S-glycosides, Gal-S-C7-Me had only th the inhibitory effect of Glc-S-C7-Me (Fig. 2), suggesting that the structure of the sugar moiety is critical in determining the nature of the binding. This corresponded to the result that the galactosylated derivative of AVP exhibited no renal targeting ability, unlike its glucosylated counterparts (Suzuki et al., 1999). Glc-S-C7-Me, Glc-O-C7-Me, and Gal-S-C7-M are widely used as detergents (Tsuchiya and Saito, 1984; Aungst, 1994) and there is very little difference in their physicochemical characteristics. Their detergent effect is exhibited at concentrations greater than millimolar (Ogiso et al., 1994), therefore it is difficult to explain their renal targeting ability simply in terms of detergent properties. This is supported by the finding that Tween 20 and CHAPS, two other types of detergent, had th the inhibitory activity of Glc-S-C7-Me (Fig. 2). The inhibition study showed that Glc-S-C7-Me binds specifically to the kidney membrane fraction with a K_d (17 nM) approximately 3 times higher than that for Glc-O-C8-AVP (Fig. 3). No saturable bindings to the liver membrane fraction were seen, so that binding to the kidney membrane fraction seems to be specific (Fig. 3). In actual fact, Glc-S-C7-Me showed clear renal targeting potential in vivo (Fig. 4), unlike Glc-O-C7-Me, a finding in accordance with the in vitro studies.

To confirm that Glc-S-C8- can act as a renal targeting vector, the in vivo distribution of some derivatives modified with Glc-S-C8- was studied. Initially, the AVP derivative (Glc-S-C8-AVP) was compared with Glc-O-C8-AVP, when the former was shown to have better renal targeting ability in vivo (Fig. 4B) and higher binding to the kidney membrane fraction in vitro (Fig. 5A) than Glc-O-C8-AVP. Consequently, it was concluded that the S-glucoside derivative was more readily recognized by kidney than its O-glucoside counterpart. However, because Glc-S-C8-AVP (K_d = 8.6 nM) had higher affinity for the kidney membrane than Glc-S-C7-Me (K_d = 17 nM), the peptide moiety (AVP) seemed to have a positive effect as far as renal recognition was concerned.

Next, tryptamine and NBD were selected as low-molecular-weight model compounds and modified with Glc-S-C8-. These derivatives (Glc-S-C8-Tryp, Glc-S-C8-NBD) exhibited high renal targeting ability (Figs. 6 and 7) and inhibited the binding of [³H]Glc-S-C7-Me to the kidney membrane fraction (Fig. 9). Therefore, these derivatives appear to be taken up by a mechanism similar to that for Glc-S-C7-Me. However, it should be noted that Glc-S-C8-Tryp was taken up not only by kidney but also by liver and small intestine to some extent (Fig. 6). This distribution pattern was similar to that of oxytocin derivatives in a previous report (Suzuki et al., 1999). On the other hand, AVP derivatives and the NBD derivative distributed specifically to the kidney (Figs. 4B and 7). These results suggest that some unknown structural requirement determines the distribution pattern, i.e., kidney only or also to the liver and small intestine. Further studies are needed to clarify the mechanisms regulating this different distribution pattern. Our study has shown that several compounds modified with Glc-S-C8- have renal targeting ability, leading us to conclude that Glc-S-C7-Me is a suitable vector for renal targeting.

The study comparing the number of methylene groups in the "sugar-alkyl" in the AVP derivatives (Glc-S-Cn-AVP) showed that Glc-S-C5-AVP had no renal targeting ability (Fig. 4B), whereas Glc-S-C8-AVP and Glc-S-C11-AVP (more potent than the C8 derivative) were highly targeted (Fig. 4B). The C5 derivative also had a low affinity for kidney membrane, about th that of other derivatives (Fig. 9). This shows that the length of the alkyl chain in the glycosylated derivatives is important for renal targeting. The sugar-alkyl was introduced into the cyclic portion of the peptide in the AVP derivatives. It is possible that Glc-S-C5-AVP with a short alkyl chain cannot be fitted into the renal recognition sites for the sugar-alkyl because of stereochemical hindrance in the cyclic peptide portion. Further studies are needed to investigate this in more detail.

It has been reported that Glc-O-C7-Me is a competitive inhibitor of the sodium-dependent D-glucose cotransporter (Vincenzini et al., 1987) and, therefore, the "alkylglycoside" structure of the renal targeting vector could be transported by this sugar transporter. Although analysis of Glc-O-C8-AVP showed that glycosylated derivatives are taken up from blood, the sodium-dependent D-glucose cotransporter is distributed on the brush-border membrane of urinary tubules (Hediger and Rhoads, 1994). In addition, 10 mM glucose (a concentration greater than the K_m of glucose for transport by sodium-dependent D-glucose cotransporter) has little inhibitory effect on [³H]Glc-S-C8-AVP-specific binding to the kidney membrane fraction (Suzuki et al., 1999). In light of these findings, the sodium-dependent D-glucose cotransporter may not be involved in the renal uptake of the vector. The involvement of proteinic molecules on the kidney membrane was suggested by our preliminary experiment, showing that there was reduced binding to trypsin-treated membrane (data not shown).

Kidney is a tissue which plays a key role in the homeostasis of the body and it is a target tissue for many drugs. For that reason, some methods have been developed to target the kidney, e.g., using peptide reabsorption after glomerular filtration through the brush-border membrane (Franssen et al., 1992) and prodrug activation by kidney-specific enzymes (Elfarra et al., 1995; Kearney, 1996). However, the renal delivery of drugs from blood via the basolateral membrane has not been reported until now, as distinct from using the asialoglycoprotein receptor system in the liver. Kidney targeting by the alkylglycoside vector described in this report is a new method for delivering drugs from the blood side.

To achieve tissue-specific targeting, it is essential to adopt a pharmacokinetic approach (Suzuki et al., 1996). When we did this, we found that there was a correlation between binding to the kidney membrane fraction (Figs. 2, 5, and 9) and renal uptake in vivo (Figs. 4, 6, and 7). Therefore, the binding assay may be used to design drugs for renal targeting. From the pharmacological point of view, the AVP derivatives described in this report had little biological activity and this will be reported in a future manuscript (Susaki et al., 1998). Here, we focused on the kidney targeting ability of these glycosylated derivatives and did not investigate the biological activity of the other low-molecular-weight derivatives.

As far as applying this renal uptake mechanism is concerned, a prodrug approach would be useful for delivering drugs to the kidney, especially in the case of patients with proximal tubule disorders. In addition, at a basic level, the physiological role of this renal uptake system is interesting and deserves further study. In future studies it will be essential to characterize the agent in the kidney that is involved in the uptake/binding of the glycosylated derivatives and also to identify the endogenous ligand responsible for the recognition system.