Synthesis and comparison of physicochemical, transport, and antithrombic properties of a cyclic prodrug and the parent RGD peptidomimetic

Article abstract of DOI:10.1016/S0040-4020(03)00333-8

Cyclic prodrug 1 was derived from RGD peptidomimetic 2 by linking the amino and carboxylic acid groups via an (acyloxy)alkoxy linker. The formation of a cyclic prodrug can transiently alter the physicochemical properties of the RGD peptidomimetic. Cyclic

Full text of DOI:10.1016/S0040-4020(03)00333-8

Tetrahedron 59 (2003) 2861–2869
Synthesis and comparison of physicochemical, transport,
and antithrombic properties of a cyclic prodrug and the parent
RGD peptidomimetic
*
Christine R. Xu, Henry T. He, Xiaoping Song and Teruna J. Siahaan
Department of Pharmaceutical Chemistry, Simons Laboratories, The University of Kansas, 2095 Constant Ave., Lawrence, KS 66047 USA
Received 6 December 2002; revised 1 March 2003; accepted 1 March 2003
Abstract—Cyclic prodrug 1 was derived from RGD peptidomimetic 2 by linking the amino and carboxylic acid groups via an
(acyloxy)alkoxy linker. The formation of a cyclic prodrug can transiently alter the physicochemical properties of the RGD peptidomimetic.
Cyclic prodrug 1 was synthesized via the key intermediate 8, and the synthesis of this key intermediate was accomplished using two different
routes. Cyclic prodrug 1 has a smaller hydrodynamic radius and higher membrane interaction potential than those of the parent RGD
peptidomimetic 2. The cell membrane permeation of cyclic prodrug 1 is twice that of the parent peptidomimetic 2. The prodrug-to-drug
conversion can be carried out in isolated porcine esterase as well as human plasma. The cyclic prodrug is more stable at pH 4 than pH 7, and is
very unstable at pH 10. The cyclic prodrug has antithrombic activity, suggesting that it can be converted to the RGD peptidomimetic in
platelet-rich plasma (PRP). q 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction
amino and carboxylic acid groups that are necessary for
their biological activity. Unfortunately, these charges make
the RGD peptidomimetics very hydrophilic and prevent
their partitioning into cell membranes of the intestinal
mucosa for oral bioavailability. Therefore, it is proposed
that the membrane permeation of the RGD peptidomimetic
can be improved by modifying its physicochemical
properties.
Arg-Gly-Asp (RGD) peptides and peptidomimetics have
been developed as antithrombic1–3 and antitumor agents.4,5
The biological activities of these molecules are due to their
ability to bind integrin receptors on the cell surface and
inhibit cell adhesion to extracellular matrix (ECM) proteins
such as fibronectin (FN), fibrinogen (FG), and vitronectin
(VN).6–8 Aggrastat is one example of an RGD peptido-
mimetic that has been used clinically to treat thrombosis by
binding to GPIIb/IIIa receptors and inhibiting platelet
aggregation.9,10 Aggrastat is delivered clinically via intra-
venous (IV) injection because it is not orally bioavailable.
Therefore, any method to improve the oral bioavailability of
RGD peptidomimetics will be beneficial for patients.
Aggrastat and other RGD peptidomimetics have charged
The physicochemical properties of RGD peptidomimetics
can be transiently altered with the formation of cyclic
prodrugs (Scheme 1). Cyclic prodrugs can be synthesized by
linking the amino and carboxylic acid groups using different
linkers such as (acyloxy)alkoxy,11,12 phenylpropionic
acid,13,14 modified phenylpropionic acid,15 coumarinic
acid,16–17 and modified coumarinic acid.20 Previously,
cyclic prodrugs have been shown to have better membrane
permeation than their respective parent peptides or peptido-
mimetics.20–23 The improved membrane permeation is
due to the increased hydrophobicity, conformation stability,
and enzymatic stability.24–27 In addition, cyclic prodrugs
have lower hydrodynamic radii and hydrogen-bonding
potentials than their parent compounds.24,28,29
Keywords: RGD peptidomimetic; (acyloxyl)alkoxy linker; cyclic prodrug;
antithrombic activity; membrane permeation.
Abbreviations: ADP, adenosine diphosphate; Boc, t-butyloxycarbonyl;
DIEA, diisopropylethyl amine; DMF, dimethylformamide; DMSO,
dimethylsulfoxide; EDC, 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide
hydrochloride; FAB-MS, fast atom bombardment mass spectrometry;
HBSS, Hanks balanced salt solution; HBTU, O-benzotriazol-1yl-
N,N,N0,N0-tetramethyl uronium hexaflourophosphate; HOBT, 1-hydroxy-
benzotriazole; HPLC, high performance liquid chromatography; HRMS,
high resolution mass spectra; IAM, immobilized artificial membrane;
NMM, N-methylmorpholine; NMR, nuclear magnetic resonance; PBS,
phosphate buffer solution; PRP, platelet-rich plasma; Tce, trichloroethanol;
TFA, trifluoroacetic acid; TMS, tetramethylsilane.
In this work, we developed two methods to synthesize cyclic
prodrug 1 using an (acyloxy)alkoxy linker. The first method
follows the previously described method;11,12 the second
method is a more efficient way to make the (acyloxy)alkoxy
cyclic prodrug. The membrane permeation of cyclic prodrug
1 was compared with that of parent RGD peptidomimetic 2.
The physicochemical properties (i.e. hydrophobicity, size)
*
Corresponding author. Tel.:þ1-785-864-7327; fax:þ1-785-864-5736;
e-mail: siahaan@ku.edu
0040–4020/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4020(03)00333-8

2862
C. R. Xu et al. / Tetrahedron 59 (2003) 2861–2869
Scheme 1.
of both compounds were studied for their correlation to the
transport properties. The antithrombic activity of cyclic
prodrug 1 was compared to that of the parent drug 2. The
rate of prodrug-to-drug conversion was determined using
isolated esterase and human plasma to prove that RGD
peptidomimetic 2 can be released from cyclic prodrug 1.
Finally, the chemical stability of this prodrug was evaluated
at pH 4, 7, and 10.
methylbenzoic acid (3), TFA·H-Asp(OBzl)-OTce (16),
Boc-[(phenylalanyloxyl)methyl]-p-nitrophenyl carbonate
(17) and Boc-PheO2Csþ (18), were prepared according to
published methods.11,12 Due to the side reaction, compound
17 was obtained at 37% yield.
2.2.1.1. 4-(N-Boc-aminomethyl)benzoyl-b-Ala-OBzl
(4). Acid 3 (2.52 g, 10 mmol), EDC (1.92 g, 10 mmol),
HOBT (1.35 g, 10 mmol) and NMM (1.0 g, 10 mmol) were
dissolved in 150 mL of CH2Cl2 in an ice bath. After 1 min,
b-Ala-OBzl (3.57 g, 10 mmol) was added to the mixture
and stirred in the ice bath for 4 h. Following this, it
was stirred for 24 h at ambient temperature. The mixture
was concentrated under reduced pressure. The residue was
dissolved in 150 mL of EtOAc and was successively washed
with 5% KHSO4 (2£50 mL), H2O (2£50 mL), and brine
(2£50 mL). The organic layer was separated, dried over
anhydrous MgSO4, and concentrated to yield a white solid.
The white solid was recrystallized from EtOAc–petroleum
2. Experimental procedures
2.1. General methods
1H NMR spectra were obtained using a Bruker DRX-400
(400 MHz) or Varian XL-300 (300 MHz) instrument.
Chemical shifts are expressed in parts per million (d)
relative to tetramethylsilane (TMS) or residual solvent as an
internal reference. Abbreviations are reported as (s) singlet,
(d) doublet, (t) triplet, (q) quartet, (m) multiplet and (br)
broad. FAB-MS and HRMS were obtained using a VG
Analytical ZAB double-focusing spectrometer; several
matrixes were used, including nitrobenzyl alcohol (NBA)
and thioglycerol/glycerol (TG/G). All starting materials
were purchased from Aldrich Chemical Co., Sigma
Chemical Co., Fluka Chemicals, and Bachem Bioscience
Inc., and were used as received.
1
ether to give a pure compound 4 (3.85 g, 94%). H NMR
(CDCl3, d): 1.48 (9H, s), 2.72 (2H, t, J¼5.6 Hz), 3.75 (2H,
q, J¼5.6 Hz), 4.36 (2H, s), 4.97 (1H, s), 5.17 (2H, s), 6.81
(1H, s), 7.33 (7H, m), 7.68 (2H, d, J¼7.8 Hz). MS (FAB)
m/z: 413 (Mþþ1).
2.2.1.2. 4-(N-Boc-aminomethyl)benzoyl-b-Ala-OH (5).
10% Pd/C catalyst (0.38 g) was added to a solution of 4
(3.85 g, 9.4 mmol) in 50 mL of MeOH, and the reaction
mixture was stirred for 24 h under H2 atmosphere using a
balloon. The reaction mixture was filtered by celite to
remove the catalyst; the filtrate was concentrated under
vacuum to yield a crude product. The crude product was
recrystallized from EtOAc–hexane to give a pure white
HPLC purification and analysis were conducted using a
Rainin HPXL gradient system with a Dynamax UV
detector. The desired product was purified with a semi-
preparative reversed-phase HPLC system using a C-18
1
solid, compound 5 (2.85 g, 95%). H NMR (DMSO-d6, d):
˚
1.39 (9H, s), 3.46 (2H, m), 4.15 (2H, d, J¼6.0 Hz), 7.28
(2H, d, J¼7.9 Hz), 7.46 (1H, t, J¼6.0 Hz), 7.77 (2H, d,
J¼7.9 Hz), 8.48 (1H, t, J¼5.0 Hz). MS (FAB) m/z: 323
(Mþþ1).
column (12 mm, 300 A, 25 cm£21.4 mm i.d., flow rate
10 mL/min) eluting with a gradient of solvents A (0.1%
TFA/H2O:5% MeCN) and B (100% MeCN). The gradient
method used for the semi-preparative HPLC started from 0
to 35% of solvent B in 23 min, followed by an increase to
100% of solvent B in 10 min. Elution at 100% of solvent B
was maintained for 2 min, and ended with gradient elution
to 0% of solvent B in 2 min. The desired peptide was
evaluated by analytical reversed-phase HPLC using a C-18
2.2.1.3. 4-(N-Boc-aminomethyl)benzoyl-b-Ala-Asp-
(OBzl)-OTce (6). Compound 5 (1.61 g, 5.0 mmol) in
100 mL of CH2Cl2 was activated by adding EDC (0.96 g,
5.0 mmol), HOBT (0.68 g, 5.0 mmol), and NMM (2.02 g,
20 mmol). The mixture was stirred at 08C for 30 min before
the addition of Asp(OBz)-OTce (16; 2.34 g; 5.0 mmol).
This reaction mixture was stirred at 08C for 4 h and then
at ambient temperature for 24 h. The solvent was then
evaporated using a rotary evaporator under reduced
pressure. The residue was dissolved in 200 mL of EtOAc
and washed with 10% aqueous citric acid (2£20 mL), H2O
(2£50 mL), saturated aqueous NaHCO3 (2£20 mL), satu-
rated aqueous NaCl (20 mL), and H2O (2£50 mL). The
organic layer was dried over anhydrous Na2SO4, filtered,
and concentrated to give compound 6 as a pale yellow oil in
˚
column (5 mm, 300 A, 25 cm£4.6 mm i.d., flow rate
1 mL/min) eluting with a gradient of solvents A and
B. The total HPLC analysis run time was 18 min. The
gradient started with solvent B from 0 to 50% over 12 min;
then solvent B was increased to 100% in 2 min and kept
constant at 100% of solvent B for 2 min. Finally, the
gradient was changed back to 0% solvent B over 2 min.
2.2. Synthesis
1
91% yield. H NMR (CDCl3, d): 1.47 (9H, s), 2.57 (2H, t,
2.2.1. Synthesis of cyclic prodrug 1. 4-N-Boc-amino-

C. R. Xu et al. / Tetrahedron 59 (2003) 2861–2869
2863
J¼5.2 Hz), 2.29–2.97 (1H, dd, J¼4.3 Hz), 3.15–3.21 (1H,
dd, J¼4.7 Hz), 3.69–3.81 (2H, m), 4.35 (2H, s), 4.72 (2H, q,
J1¼21.1 Hz, J2¼11.9 Hz), 4.91 (1H, s), 5.04–5.14 (3H, m),
6.70 (1H, d, J¼8.4 Hz), 7.11 (1H, s), 7.31–7.40 (7H, d, J¼
8.1 Hz). MS (FAB) m/z: 658 (Mþþ1), 602 (Mþ2C4H9þ1).
2.2.1.4. TFA·(4-Aminomethyl)benzoyl-b-Ala-Asp-
(OBzl)-OTce (7). TFA (10 mL) was added to a stirred
solution of compound 6 (3.0 g, 4.5 mmol) in 10 mL of
CH2Cl2 at 08C followed by stirring for 1 h at room
temperature. The reaction mixture was concentrated under
vacuum to give a crude product, which was triturated and
washed with anhydrous Et2O to give a white solid. The
white solid was isolated by decantation and dried under
7.68 (2H, d, J¼8.0 Hz). MS (FAB) m/z: 634 (Mþþ1), 534
(Mþ2Bocþ1).
2.2.1.8. Boc-Phe-methyleneoxycarbonyl-4-amino-
methyl-benzoyl-b-Ala-OH (15). 10% Pd/C catalyst
(0.20 g) was added to a solution of 14 (1.91 g, 3.01 mmol)
in 50 mL of MeOH, and the reaction mixture was stirred for
24 h under H2 atmosphere using a balloon. The reaction
mixture was filtered by celite to remove the catalyst; the
filtrate was concentrated under vacuum to yield a pure
1
product, 15 (1.40 g, 85%). H NMR (DMSO-d6, d): 1.32
(9H, s), 2.75–3.00 (2H, m), 3.40–3.50 (5H, m), 4.20 (1H,
m), 4.26 (2H, d, J¼6.0 Hz), 5.70 (2H, dd, J1¼5.9 Hz,
J2¼6.1 Hz), 7.20–7.38 (8H, m), 7.77 (2H, d, J¼8.0 Hz),
8.20 (1H, t, J¼ 5.8 Hz), 8.48 (1H, t, J¼5.2 Hz). MS (FAB)
m/z: 544 (Mþþ1), 488 (Mþ2C4H9þ1), 444 (Mþ2Bocþ1).
2.2.1.9. Boc-Phe-methyleneoxycarbonyl-4-amino-
methylbenzoyl-b-Ala-Asp(OBzl)-OTce (8). Method 1. A
solution of compound 17 (2.02 g, 3.0 mmol) and NMM
(1.21 g, 12 mmol) in 10 mL of DMF was added to a mixture
of compound 7 (1.39 g, 3.0 mmol) and HOBT (0.41 g,
3.0 mmol) in 50 mL of DMF at 08C. The mixture was stirred
for 24 h at room temperature. The solvent was removed
under reduced pressure, and the residue was dissolved in
100 mL of CH2Cl2. The organic solution was washed with
saturated aqueous NaHCO3 (2£20 mL), saturated aqueous
NaCl (20 mL), and H2O (2£50 mL). The CH2Cl2 layer was
dried over anhydrous Na2SO4, filtered, and evaporated to
give a crude product. The crude product was purified using
column chromatography (63–200 mm silica gel with
EtOAc–hexane–MeOH (300:100:2)) to give compound 8
(1.88 g, 71%).
1
vacuum to give pure compound 7 (2.93 g, 96%). H NMR
(CD3OD, d): 2.58 (2H, t, J¼6.6 Hz), 2.93–3.07 (2H, m),
3.65 (2H, m), 4.18 (2H, s), 4.80 (2H, dd, J1¼8.2 Hz, J2¼
12.0 Hz), 5.00 (2H, m), 5.13 (2H, s), 7.35 (5H, m), 7.55 (2H,
d, J¼8.1 Hz), 7.90 (2H, d, J¼8.1 Hz). MS (FAB) m/z: 558
(Mþþ1).
2.2.1.5. 4-Aminomethylbenzoyl-b-Ala-OBzl (12).
Compound 4 (2.60 g, 6.31 mmol) was stirred in 25 mL of
50% TFA in CH2Cl2. The mixture was concentrated under
reduced pressure to give a white solid upon addition of cold
ether (Et2O). The solid was filtered and washed with ether.
The residual ether was removed under vacuum to give
1
compound 12 (1.70 g, 87%). H NMR (CD3OD, d): 2.73
(2H, t, J¼6.7 Hz), 3.68 (2H, t, J¼6.7 Hz), 4.20 (2H, s), 5.16
(2H, s), 7.29–7.37 (5H, m), 7.55 (2H, d, J¼8.2 Hz). 7.85
(2H, d, J¼8.2 Hz). MS (FAB) m/z: 313 (Mþþ1).
2.2.1.6. 1-Chloromethyleneoxycarbonyl-4-amino-
methyl-benzoyl-b-Ala-OBzl (13). A solution of 1-chloro-
methyl chloroformate (0.85 g, 6.59 mmol) in dry CH2Cl2
(20 mL) was added dropwise to a mixture of 12 (1.70 g,
5.47 mmol) and TEA (0.66 g, 6.52 mmol) in ice-cold dry
CH2Cl2 (50 mL). After stirring overnight at room tempera-
ture, the mixture was concentrated to give a white residue.
This residue was dissolved in 150 mL of EtOAc and
extracted with 5% KHSO4 (2£50 mL), brine (2£50 mL),
and H2O (2£50 mL); the organic layer was dried over
anhydrous MgSO4. After filtering the MgSO4, the filtrate
was concentrated and the crude product was purified on a
silica gel column using EtOAc–hexane (2:1) as eluent to
give compound 13 (1.70 g, 77%). 1H NMR (CDCl3, d): 2.70
(2H, m), 3.72 (2H, m), 4.43 (2H, m), 5.16 (2H, m), 5.78 (2H,
m), 6.89 (1H, s), 7.24–7.40 (6H, m), 7.64 (2H, d,
J¼8.0 Hz). MS (FAB) m/z: 405 (Mþþ1).
Method 2. EDC (0.37 g, 1.93 mmol), HOBT (0.22 g,
1.63 mmol), and NMM (0.20 g, 1.98 mmol) were added to
a stirred ice-cold solution of 15 (0.72 g, 1.63 mmol) in
30 mL of CH2Cl2. After 1 min stirring, compound 16
(0.80 g, 1.71 mmol) in 10 mL of CH2Cl2 was added and the
mixture was stirred overnight. The mixture was concen-
trated under reduced pressure, and the residue was dissolved
in 100 mL of CH2Cl2. The organic solution was washed
with saturated aqueous NaHCO3 (2£20 mL), saturated
aqueous NaCl (20 mL), and H2O (2£50 mL). The CH2Cl2
layer was dried over anhydrous Na2SO4, filtered, and evapo-
rated to give a crude product. The crude product was
purified using column chromatography (63–200 mm silica
gel with EtOAc–hexane–MeOH (300:100:2)) to give
compound 8 (1.09 g, 94%). 1H NMR (CDCl3, d): 1.42
(9H, s), 2.50–2.58 (2H, m), 2.69 (2H, t, J¼5.6 Hz), 2.90–
3.20 (4H, m), 3.73 (2H, m), 4.43 (2H, m), 4.58–4.77 (5H,
m), 4.07–5.13 (4H, m), 5.41 (1H, br s), 5.80 (2H, dd, J1¼
19.0 Hz, J2¼5.4 Hz), 6.85 (1H, d, J¼8.2 Hz), 7.14 (2H,
J¼6.7 Hz), 7.24–7.40 (10H, m), 7.55–7.76 (4H, m), 7.92
(2H, d, J¼8.1 Hz). MS (FAB) m/z: 881 (Mþþ1), 781
(Mþ2Bocþ1).
2.2.1.7. Boc-Phe-methyleneoxycarbonyl-4-amino-
methyl-benzoyl-b-Ala-OBzl (14). Compound 13 (1.70 g,
4.21 mmol) in 20 mL DMF was added dropwise to a stirred
solution of cesium salt 18 (1.84 g, 4.64 mmol) in ice-cold
DMF (30 mL). The mixture was stirred overnight and the
solvent was removed under reduced pressure. This residue
was dissolved in 150 mL of EtOAc and extracted with 5%
KHSO4 (2£50 mL), brine (2£50 mL), and H2O (2£50 mL).
The organic layer was dried over anhydrous MgSO4. The
mixture was filtered, and the filtrate was concentrated.
The crude product was purified on a silica gel column
using EtOAc–hexane (2:1 to 3:1) as eluent to give
2.2.1.10. Boc-Phe-methyleneoxycarbonyl-4-amino-
methylbenzoyl-b-Ala-Asp(OBzl)-OH (9). Compound 8
(1.09 g, 1.24 mmol) was dissolved in acetic acid (30 mL)
followed by addition of Zn dust (1 g) over a 1 h time period.
The reaction mixture was stirred for 18 h at room
temperature to remove the Tce-protecting group, and
filtered to remove the insoluble material. The filtrate was
concentrated under vacuum to give an oily residue. The
crude product was purified using column chromatography
1
compound 14 (1.91 g, 72%). H NMR (CDCl3, d): 1.42
(9H, s), 2.71 (2H, t, J¼5.8 Hz), 3.06–3.15 (2H, m), 3.74
(2H, q, J¼5.7 Hz), 4.45 (2H, d, J¼5.9 Hz), 4.98 (1H, d, J¼
5.8 Hz), 5.16 (2H, s), 5.51 (1H, s), 5.80 (2H, dd, J¼5.8 Hz),
6.85 (1H, s), 7.15 (2H, d, J¼6.8 Hz), 7.20–7.37 (9H, m),

2864
C. R. Xu et al. / Tetrahedron 59 (2003) 2861–2869
(63–200 mm silica gel with EtOAc–hexane–MeOH
1
(300:100:2)) to give compound 9 (0.75 g, 80%). H NMR
J¼7.5 Hz), 6.98 (1H, d, J¼5.9 Hz), 7.06 (2H, d, J¼3.7 Hz),
7.24–7.36 (12H, m).
(CDCl3, d): 1.42 (9H, s), 2.65–3.13 (6H, m), 3.50–3.70
(3H, m), 4.35 (3H, m), 4.60 (1H, s), 4.80 (1H, m), 5.06 (2H,
s), 5.74–6.00 (3H, m), 7.12–7.34 (13H, m), 7.67 (2H, d, J¼
8.0 Hz). MS (FAB) m/z: 749 (Mþþ1), 649 (Mþ2 Bocþ1).
2.2.1.11. TFA·H-Phe-methyleneoxycarbonyl-4-amino-
methylbenzoyl-b-Ala-Asp(OBzl)-OH (10). Compound 9
was dissolved in 25 mL of 50% TFA in CH2Cl2 and stirred
for 2 h. The mixture was concentrated under reduced
pressure. The reaction mixture was concentrated under
vacuum to give a crude product, which was triturated and
washed with anhydrous Et2O to give a white solid. The
white solid was isolated by decantation and dried under
vacuum to give pure compound 10 in 95% yield. The
peptide was analyzed using analytical reversed-phase HPLC
2.2.2.2. TFA·H-Asp(OBzl)-Phe-OBzl (20). The pro-
cedure to prepare compound 20 was similar to that for
compound 7. Yield: 98%. 1H NMR (CD3OD, d): 2.73–3.03
(3H, m), 3.20 (1H, m), 4.20 (1H, m), 4.78 (1H, t, J¼3.0 Hz),
5.12–5.20 (4H, m), 7.18–7.88 (15H, m).
2.2.2.3. Boc-(4-aminomethylbenzoyl)-b-Ala-Asp(OBzl)-
Phe-OBzl (21). The procedure to prepare compound 21 was
similar to that for compound 4. Yield: 90%. 1H NMR
(CDCl3, d): 1.48 (9H, s), 2.47 (2H, m), 2.70 (1H, dd, J¼
6.6 Hz). 2.93–3.11 (3H, m), 3.70 (2H, m), 4.30 (2H, s), 4.82
(2H, t, J¼6.2 Hz), 4.90 (1H, s), 5.06–5.17 (7H, m), 7.01
(4H, m), 7.30–7.53 (15H, m), 7.73 (2H, d, J¼7.7 Hz), 7.86
(2H, d, J¼7.7 Hz). MS (FAB) m/z: 765 (Mþþ1), 709
(Mþ2C4H9þ1).
˚
with a C-18 column (5 mm, 300 A, 25 cm£4.6 mm i.d., flow
2.2.2.4. Boc-(4-aminomethylbenzoyl)-b-Ala-Asp-Phe-
OH (22). Compound 22 was prepared using the same
1
rate 1 mL/min) to give a retention time of 13.14 min. H
1
NMR (DMSO-d6, d): 2.70 (2H, t, J¼7.0 Hz), 2.76 (2H, m),
3.00 (2H, m), 3.46 (2H, m), 3.87 (1H, q, J¼6.0 Hz), 4.13
(2H, m), 4.91 (1H, m), 5.13 (2H, s), 5.96 (2H, m), 7.28
(12H, m), 7.88 (2H, d, J¼9.0 Hz). MS (FAB) m/z: 649
(Mþþ1), 633 (Mþ2NH2þ1). HRMS: calcd for
C33H36N4O10 649.2509. Found 649.2501.
procedure as that to make compound 5. Yield: 65%. H
NMR (DMSO-d6, d): 1.39 (9H, s), 2.35–2.43 (3H, m), 2.64
(2H, dd, J1¼11.6 Hz, J2¼5.0 Hz), 2.92–3.02 (2H, m),
3.42–3.48 (4H, m), 4.15 (2H, d, J¼5.7 Hz), 4.37 (1H, d,
J¼5.9 Hz), 4.61 (1H, d, J¼6.3 Hz), 7.18–7.45 (7H, m),
7.76 (2H, d, J¼7.7 Hz), 7.98 (1H, d, J¼7.3 Hz), 8.22 (1H, d,
J¼7.7 Hz), 8.41 (1H, s). MS (FAB) m/z: 585 (Mþþ1), 529
(Mþ2C4H9þ1).
2.2.1.12. Cyclic prodrug (1). Compound 10 (0.129 g,
0.2 mmol) and HBTU (0.379 g, 1.0 mmol) were dissolved
in DMF (300 mL) and stirred at room temperature for 0.5 h
under N2. DIEA (0.258 g, 2.0 mmol) in DMF (25 mL) was
added dropwise over 1 h and the reaction mixture was
stirred at room temperature for 8 h under N2. The mixture
was concentrated under vacuum to yield a light brown
residue, which was purified using preparative reversed-
2.2.2.5. TFA·(4-aminomethylbenzoyl)-b-Ala-Asp-Phe-
OH (2). The removal of the Boc-protecting group in
compound 2 was accomplished using a procedure similar to
1
that used for making compound 7. Yield: 98%. H NMR
(CD3OD, d): 2.53 (2H, t, J¼6.6 Hz), 2.62 (1H, dd, J1¼
8.7 Hz, J2¼8.2 Hz), 2.81 (1H, dd, J1¼11.3 Hz, J2¼5.5 Hz),
3.05 (1H, dd, J1¼5.9 Hz, J2¼7.9 Hz), 3.20 (1H, dd, J1¼
8.5 Hz, J2¼5.2 Hz), 3.60–3.68 (2H, m), 4.18 (2H, s), 4.64
(1H, q, J1¼2.6 Hz, J2¼5.3 Hz), 4.77 (1H, q, J1¼2.7 Hz,
J2¼5.5 Hz), 7.18–7.30 (5H, m), 7.54 (2H, d, J¼8.4 Hz),
7.91 (2H, d, J¼8.4 Hz). MS (FAB) m/z: 485 (Mþþ1).
˚
phase HPLC with a C-18 column (12 mm, 300 A,
25 cm£21.4 mm i.d., flow rate 10 mL/min). The desired
fractions were analyzed by analytical HPLC with a C-18
˚
column (5 mm, 300 A, 25 cm£4.6 mm i.d., flow rate
1 mL/min) to give a retention time of 15.47 min. The
desired fractions were combined, concentrated, and lyophi-
lized to give pure compound 11 (0.033 g, 26%). MS (FAB)
m/z: 631 (Mþþ1). HRMS: calcd for C33H34N4O9 631.2404.
Found 631.2379. A mixture of compound 11 (0.033 g,
0.052 mmol) and 10% Pd/C (0.020 g) in MeOH (10 mL)
was stirred for 24 h under a balloon H2 atmosphere. The
reaction mixture was filtered through celite to remove Pd/C,
and the filtrate was concentrated to give a crude product,
which was purified using preparative reversed-phase HPLC
2.3. Platelet aggregation
The platelet aggregation assay was carried out using a
previously described protocol.3,30 The assay was done using
fresh platelet-rich plasma (PRP) in Tyrodes buffer.
2.4. Chemical and enzymatic stabilities
˚
with a C-18 column (12 mm, 300 A, 25 cm£21.4 mm i.d.,
The chemical stability of prodrug 1 was studied in aqueous
phosphate buffer solutions (pH¼4, 7, 10) at 37^0.58C, and
a constant ionic strength of m¼0.15 was maintained using
NaCl.31 The purified prodrug 1 was dissolved in buffer at a
concentration of 0.1 mM. At appropriate time intervals,
aliquots were removed in triplicate, frozen immediately in
dry ice and stored at 2708C until HPLC analysis. Prior to
analysis the appropriate sample was fast-thawed.31
flow rate 10 mL/min) to give pure cyclic prodrug 1 (0.027 g,
96%). Cyclic prodrug 1 was analyzed using analytical
˚
reversed-phase HPLC with a C-18 column (5 mm, 300 A,
25 cm£4.6 mm i.d., flow rate 1 mL/min) to give a retention
1
time of 11.41 min. H NMR (DMSO-d6, d): 2.17 (2H, m),
2.86 (2H, m), 3.04 (2H, m), 3.46 (2H, q, J¼5.8 Hz), 3.97
(1H, m), 4.26 (2H, m), 4.55 (1H, m), 5.46 (2H, m), 7.25 (7H,
m), 7.68 (2H, d, J¼9.0 Hz). MS (FAB) m/z: (Mþþ1).
HRMS: calcd for C26H28N4O9 541.1935. Found 541.1957.
For the enzymatic stability studies, a purified porcine liver
esterase (carboxylic-ester hydrolase; EC 3.1.1.1, Sigma)
was dissolved in phosphate buffer (0.05 M, m¼0.15,
pH¼7.4) to a concentration of 20 unit/mL.22 For stability
studies in plasma, human plasma was obtained from the
Community Blood Center of Greater Kansas City, Kansas
City, MO. Stock solutions of prodrug 1 were diluted with
enzyme solution or human plasma to give the final
2.2.2. Synthesis of parent compound 2
2.2.2.1. Boc-Asp(OBzl)-Phe-OBzl (19). The procedure
to make compound 19 was similar to that for compound 4.
1
Yield: 95%. H NMR (CDCl3, d): 1.43 (9H, s), 2.67–2.73
(1H, m), 3.04 (1H, s), 3.11 (2H, d, J¼5.7 Hz), 4.54 (1H, s),
4.86 (d, J¼6.5 Hz), 5.09–5.17 (4H, m), 5.64 (1H, d,

C. R. Xu et al. / Tetrahedron 59 (2003) 2861–2869
2865
Scheme 2.
concentration of 0.1 mM.22 The milieu was maintained at
37^0.58C. At various time intervals 100 mL samples were
removed and the residual enzyme activity quenched by
adding 100 mL of a freshly prepared 6N guanidinium
hydrochloride solution in Hanks’ balanced salt solution
containing 0.01% (v/v) phosphoric acid. This mixture was
transferred to an Ultrafree-MC 5000 NMWL filter unit
(Milipore, Bedford, MA) and centrifuged at 9000 rpm for
5 min (48C). Aliquots (50 mL) of the filtrate were diluted
with mobile phase and analyzed by HPLC. Recovery for the
prodrug was found to be 96.4^1.1%. The effect of an
esterase on the rate of degradation of cyclic prodrug 1 in
various biological media was determined using an esterase
inhibitor, paraoxon. The biological media was incubated
with paraoxon (final concentration 1 mM) for 15 min at
378C before adding the prodrug.
prodrug 1 was accomplished by cyclization of linear
intermediate 10, which is a deprotection product of
intermediate 8 (Scheme 2). We explored two ways to
make intermediate 8 (Schemes 3 and 4). At first (Scheme 3),
intermediate 8 was formed by conjugation of two key
intermediates, 7 and 17. Unfortunately, our previous
experiment indicated that the synthesis of intermediate 17
produced side product 17a.11,33 It is very difficult to purify
intermediate 17 from the side product intermediate
17a, resulting in the low yield of 17. Therefore, we
investigated a second route to make intermediate 8 that
is more efficient than the first. This route involves
making the key intermediate 13 and reacting this inter-
mediate with Boc-PheO2Csþ (18) to produce the coupling
product 14 (see Scheme 4) without generating any side
product.
2.5. Physicochemical property characterization
The synthetic procedure to make cyclic prodrug 1 using
the first pathway is shown in Scheme 3. The synthesis
was initiated by activating the carboxylic acid group of
compound 3 with EDC and HOBT in the presence of NMM;
the activated acid was then reacted with b-Ala-OBzl to
produce dipeptide 4 in 94% yield. The benzyl ester
protecting group in 4 was removed by hydrogenation with
10% Pd/C to give 5 in 95% yield. In the same way as with 3,
the carboxylic acid group in 5 was activated, followed by
reaction with compound 16 to give compound 6. Treatment
of 6 with TFA in CH2Cl2 produced 7 in quantitative yield; 7
was conjugated with 17 in the presence of HOBT and NMM
to give compound 8 in 71% yield after silica gel purification.
The Tce- and Boc-protecting groups in 8 were removed by
treatment with Zn/AcOH and TFA/CH2Cl2, respectively, to
give linear peptide 10 in an overall 76% yield. The
cyclization was accomplished by adding a dilute solution
of 10 in DMF to a solution of HBTU in the presence of
DIEA. The Bzl-protecting group in 11 was removed in the
The molecular size of the cyclic prodrug and the parent
peptidomimetic was determined by measuring their diffu-
sion coefficients using NMR spectroscopy. The molecular
radius was then calculated from the molecular diffusion
coefficient using the Stokes–Einstein equation.14 Lipo-
philicity of the cyclic prodrugs and the parent compounds
was evaluated by partitioning experiments using immobi-
lized artificial membrane (IAM) column chromatography.32
3. Results and discussion
3.1. Synthesis of cyclic prodrug 1
Cyclic prodrug 1 was formed by linking the N-terminus of
the benzylamine moiety and the C-terminus of the Phe via
the (acyloxy)alkoxy linker (Scheme 1). Synthesis of cyclic

2866
C. R. Xu et al. / Tetrahedron 59 (2003) 2861–2869
Scheme 3. (a) EDC, HOBT, NMM, CH2Cl2, 0–28C; (b) b-Ala-OBzl; (c) H2, 10% Pd/C, MeOH; (d) TFA/CH2Cl2 (1:1); (e) NMM, HOBT, DMF, 0–28C; (f) Zn,
AcOH; (g) HBTU, DIEA, DMF.
Scheme 4. (a) TFA–CH2Cl2 (1:1); (b) 2 M NaOH–CH2Cl2; (c) ClCH2OCOCl, TEA–CH2Cl2, 08C; (d) DMF, 0–258C; (e) H2, 10% Pd/C, MeOH, (f) EDC,
NMM, HOBT, CH2Cl2.
same way as with 4 to give prodrug 1 in an overall 25%
yield (last two steps).
using a solution phase method. Both intermediates 19 and
21 were synthesized in the same way as 4 while the Boc-
protecting group in 19 and 22 and the Bzl-protecting group
in 21 were removed in the same way as with 6 and 4,
respectively, as described in Scheme 3.
For the second pathway (Scheme 4), the TFA salt 12 was
converted to free amine by using a strong base (NaOH) so
that it readily reacted with 1-chloromethyl chloroformate in
the presence of a weak base (TEA) to produce compound 13
in high yield. This method, however, cannot be applied to
compound 7 because its Tce-protecting group could be
cleaved in the presence of a strong base like NaOH.
Metathesis of chloride to iodide atom in compound 13 was
not necessary because the subsequent coupling reaction
between 13 and cesium salt 18 proceeded smoothly to give
the key intermediate 14 in 72% yield after silica gel column
chromatography.33 The Bzl-protecting group in 14 was
removed in the same way as in 4 to give acid 15, which was
activated in the same way as 3 and reacted with 16 to give 8
in 94% yield.
3.2. Biological study
3.2.1. Cell membrane permeation. The membrane per-
meation properties of cyclic prodrug 1 and the parent RGD
peptidomimetic 2 were evaluated using Caco-2 cell
monolayers, an in vitro model of the intestinal mucosa.34,35
The apparent membrane permeability of cyclic prodrug 1
(0.86^0.11£1027 cm/s) was two-fold higher than that of its
parent compound 2 (0.43^0.14£1027 cm/s). This improve-
ment was not very dramatic compared to the previously
studied cyclic prodrugs.21,22,33,36 For example, we have
shown previously that a cyclic prodrug of RGD peptido-
mimetic (1b, Scheme 1) has a 13-fold increase of transport
over the parent RGD peptidomimetic.33 However, unlike
cyclic prodrug 1, this particular cyclic prodrug does not
Finally, the linear RGD peptidomimetic 2 was prepared
according to Scheme 5 by typical peptide bond formation

C. R. Xu et al. / Tetrahedron 59 (2003) 2861–2869
2867
Scheme 5. (a) EDC, NMM, HOBT, CH2Cl2, 0–28C; (b) TFA/CH2Cl2 (1:1); (c) H2, 10% Pd/C, MeOH.
carboxylic acid group of the side chain of the Asp residue.
These hydrophobicity studies support the suggestion that the
two-fold improvement of the cyclic prodrug transport is due
to its size and hydrophobicity compared to those of the
parent compound 2.
Table 1. Summary of physicochemical properties of cyclic prodrug 1 and
the parent drug 2
Compound MW Diffusion coefficient Size
(£1026 cm2/s)
Membrane
interaction
potential
(log kI0AM
˚
(A)
)
3.2.2. Chemical and enzymatic stability. The precondition
of prodrug strategy is that the prodrug can be converted to
the parent drug in order to exert pharmacological effects
(Scheme 1).39 For this purpose, we evaluated the enzymatic
and chemical stability of cyclic prodrug 1 (Table 2). The
prodrug can be converted to the parent compound 2 by
isolated esterase and human plasma. The rate of prodrug-to-
drug conversion in both isolated esterase and human plasma
was impeded by the presence of paraoxon, suggesting that
the prodrug conversion was mediated by esterase. The rate
of conversion in human plasma is faster than in isolated
esterase, suggesting that the human plasma may have
more selective esterase(s) for this conversion. The rate of
enzymatic degradation was faster than the rate of chemical
degradation at pH 7, suggesting that the prodrug can be
converted by enzymatic degradation.
pH¼4
pH¼7.4
2
1
483
539
1.69
1.86
5.93 20.230 20.698
5.38 0.378 20.193
have any charge.33 On the other hand, cyclic prodrug 1 has
one negative charge from the Asp residue side chain, and
this negative charge may prevent the effective partitioning
of the cyclic prodrug into the cell membrane. Therefore, the
transcellular permeation of cyclic prodrug 1 is lower than
that of the non-charged cyclic prodrug. Nonetheless, the
formation of cyclic prodrug improves the transport of
prodrug 1 over linear RGD peptidomimetic 2. This suggests
that there is a partial shift from a paracellular transport for
linear RGD peptidomimetic 2 to an increasing contribution
of transcellular transport for the cyclic prodrug 1. Another
possible explanation is that the low permeation of cyclic
prodrug 1 is due to the recognition of this compound by the
efflux transporter.20,37,38
The chemical stability of cyclic prodrug 1 was also
evaluated at pH 4, 7, and 10. The cyclic prodrug was
more stable at acidic pH than at neutral or basic pH. As
expected, the hydrolysis of the cyclic prodrug is very rapid
at basic pH. The high stability of this prodrug at low pH is
an advantage for oral delivery of this prodrug because it
remains longer in the stomach.
The transport properties of cyclic prodrug 1 and its parent
RGD peptidomimetic 2 are related to their physicochemical
properties such as size, hydrogen bonding potential, and
hydrophobicity. The size of both compounds was deter-
mined using their diffusion coefficients from the NMR
studies. Prodrug 1 is only slightly smaller than its parent
compound 2 (Table 1). The membrane interaction potential
studies (log kI0AM) showed that the cyclic prodrug 1 is more
hydrophobic than the parent drug at both pH 4 and 7. In
addition, the hydrophobicity of cyclic prodrug 1 is higher at
pH 4 than at pH 7. This is due to the protonation of the
3.2.3. Antithrombic activity. Prodrug-to-drug conversion
can also be evaluated using the antithrombic activity of
cyclic prodrug 1 as compared to parent compound 2 and a
standard RGDF peptide. Prodrug 1 itself should not have
antithrombic activity because the critical functional groups
for binding to GP-IIb/IIIa receptors are masked by the
linker. Thus, the biological activity of the prodrug is due to
the conversion of prodrug 1 to the parent drug 2. The
Table 2. The half-lives (t1/2 in min) for cyclic prodrug 1 in biological media and aqueous buffers at various pH values (378C, m¼0.15)
Enzymatic stability
No paraoxon (t1/2 in min)
With paraoxon (t1/2 in min)
Isolated esterase
Human plasma
436^13
179^5
877^61
793^49
Chemical stability
pH¼4 (t1/2 in min)
4310^30
pH¼7 (t1/2 in min)
956^58
pH¼10 (t1/2 in min)
5.69^0.08

2868
C. R. Xu et al. / Tetrahedron 59 (2003) 2861–2869
Table 3. In vitro antithrombic activity of linear peptidomimetic 2 and cyclic
prodrug 1
prodrug compared to the parent compound. The prodrug can
be converted to the parent drug by isolated esterase and
human plasma. The cyclic prodrug is more stable at pH 4.0
than at pH 7.0, and is very unstable at pH 10. Finally, the
prodrug can inhibit platelet aggregation, presumably
because it can be converted to the parent drug in human
platelet-rich plasma.
Compound
IC50^SD (mM)
RGDF
RGD peptidomimetic 2
Cyclic prodrug 1
12^0.01
50^4
250^36
antithrombic activity of these compounds was evaluated by
a platelet aggregation assay using human platelet-rich
plasma (PRP) induced by adenosine diphosphate (ADP).
Table 3 shows that the linear peptidomimetic 2 has an IC50
of 50 mM, which is higher than that of the standard RGDF
peptide (12 mM). Cyclic prodrug 1 has an IC50 of 250 mM,
which is five times higher than the parent peptide. In these
IC50 studies, the induction of platelet aggregation was done
immediately after addition of cyclic prodrug 1. Therefore, it
is possible that the amount of drug 2 in the PRP solution is
not the same as the added prodrug. This is due to the slow
rate of prodrug-to-drug conversion compared to the
aggregation rate of the platelets.
Acknowledgements
Financial support for this research was provided by a grant
from NIH (HL-59931) and the General Research Funds
from The University of Kansas. We would like to
acknowledge technical support from the Mass Spectroscopy
Laboratory at the University of Kansas. We also thank
Nancy Harmony for proofreading this manuscript.
References
1. Nicholson, N. S.; Abood, N. A.; Panzer-Knodle, S. G.;
Frederick, L. G.; Page, J. D.; Salyers, A. K.; Suleymanov,
O. D.; Szalony, J. A.; Taite, B. B.; Anders, R. J. Med. Res. Rev.
2001, 21, 211–226.
To prove this hypothesis, we studied the effect of prodrug
incubation time on platelet aggregation inhibitory activity
(Fig. 1). In this case, the prodrug was incubated in PRP at
different time points prior to addition of ADP to induce the
platelet aggregation. The different incubation times were
used to evaluate the effect of prodrug-to-drug conversion on
activity. These results show that the activity of cyclic
prodrug 1 is dependent on incubation time. The inhibitory
activity of the cyclic prodrug was increased from 20 to 55%
when it was incubated for 10 min in PRP prior to platelet
aggregation compared to no incubation prior to PRP
aggregation. The activity of the prodrug plateaued when it
was incubated at 10 to 20 min prior to PRP aggregation,
suggesting that the prodrug can be converted to a drug that
has antithrombic activity in time dependent manner.
2. Scarborough, R. M.; Gretler, D. D. J. Med. Chem. 2000, 43,
3453–3473.
3. Zablocki, J. A.; Rico, J. G.; Garland, R. B.; Rogers, T. E.;
Williams, K.; Schretzman, L. A.; Rao, S. A.; Bovy, P. R.;
Tjoeng, F. S.; Lindmark, R. J.; Toth, M. V.; Zupec, M. E.;
McMakins, D. E.; Adams, S. P.; Miyano, M.; Markos, C. S.;
Milton, M. N.; Paulson, S.; Herin, M.; Jaqmin, P.; Nicholson,
N. S.; Panzer-Knodle, S. G.; Haas, N. F.; Page, J. D.; Szalony,
J. A.; Taite, B. B.; Salyers, A. K.; King, L. W.; Campion, J. G.;
Feigen, L. P. J. Med. Chem. 1995, 38, 2378–2394.
4. Tsuchiya, Y.; Sawada, S.; Tsukada, K.; Saiki, I. Int. J. Oncol.
2002, 20, 319–324.
5. Buerkle, M. A.; Pahernik, S. A.; Sutter, A.; Jonczyk, A.;
Messmer, K.; Dellian, M. Br. J. Cancer 2002, 86, 788–795.
6. Hynes, R. O. Cell 1992, 69, 11–25.
4. Conclusion
7. Smith, J. W. Integrins Molecular and Biological Responses to
the Extracellular Matrix; Cheresh, D. A., Mecham, R. P., Eds.;
Academic: San Diego, CA, 1994; pp 1–32.
We have been able to synthesize the (acyloxy)alkoxy-
derived cyclic prodrug 1 to improve the synthetic method-
ology. The formation of cyclic prodrug 1 improves the
transport properties of the RGD peptidomimetic 2 due to the
increase in membrane partition potential of the cyclic
8. Ruoslahti, E. In Cell Biology of Extracellular Matrix; 2nd ed.,
Hay, E. D., Ed.; Plenum: New York, NY, 1991; pp 343–363.
9. Batchelor, W. B.; Tolleson, T. R.; Huang, Y.; Larsen, R. L.;
Mantell, R. M.; Dillard, P.; Davidian, M.; Zhang, D.; Cantor,
W. J.; Sketch, M. H., Jr.; Ohman, E. M.; Zidar, J. P.; Gretler,
D.; DiBattiste, P. M.; Tcheng, J. E.; Califf, R. M.; Harrington,
R. A. Circulation 2002, 106, 1470–1476.
10. Bizzarri, F.; Scolletta, S.; Tucci, E.; Lucidi, M.; Davoli, G.;
Toscano, T.; Neri, E.; Muzzi, L.; Frati, G. J. Thorac.
Cardiovasc. Surg. 2001, 122, 1181–1185.
11. Gangwar, S.; Pauletti, G. M.; Siahaan, T. J.; Stella, V. J.;
Borchardt, R. T. J. Org. Chem. 1997, 62, 1356–1362.
12. Bak, A.; Siahaan, T. J.; Gudmundsson, O. S.; Gangwar, S.;
Friis, G. J.; Borchardt, R. T. J. Pept. Res. 1999, 53, 393–402.
13. Wang, B.; Gangwar, S.; Pauletti, G. M.; Siahaan, T. J.;
Borchardt, R. T. J. Org. Chem. 1997, 62, 1363–1367.
14. Wang, B.; Nimkar, K.; Wang, W.; Zhang, H.; Shan, D.;
Gudmundsson, O. S.; Gangwar, S.; Siahaan, T. J.; Borchardt,
R. T. J. Pept. Res. 1999, 53, 370–382.
Figure 1. Time-dependent inhibitory activity of cyclic prodrug 1 in platelet
aggregation.

C. R. Xu et al. / Tetrahedron 59 (2003) 2861–2869
2869
15. Song, X.; He, H. T.; Siahaan, T. J. Org. Lett. 2002, 4,
549–552.
27. Knipp, G. T.; Velde, D. G. V.; Siahaan, T. J.; Borchardt, R. T.
Pharm. Res. 1997, 14, 1332–1340.
16. Wang, B.; Zhang, H.; Wang, W. Biorg. Med. Chem. Lett.
1996, 6, 945–950.
28. Conradi, R. A.; Burton, P. S.; Borchardt, R. T. In Lipophilicity
in Drug Action and Toxicity; Pliska, V., Testa, B., Waterbeend,
H. V., Eds.; VCH: Weinheim, 1996; pp 233–252.
17. Wang, B.; Wang, W.; Zhang, H.; Shan, D.; Smith, D. Biorg.
Med. Chem. Lett. 1996, 6, 2823–2826.
29. Burton, P. S.; Conradi, R. A.; Ho, N. F. H.; Hilgers, A. R.;
Borchardt, R. T. J. Pharm. Sci. 1996, 85, 1336–1340.
30. Jois, S. D. S.; Tambunan, U. S. F.; Chakrabarti, S.; Siahaan,
T. J. J. Biomol. Struct. Dyn. 1996, 14, 1–12.
18. Zheng, A.; Wang, W.; Zhang, H.; Wang, B. Tetrahedron 1999,
55, 4237–4254.
19. Wang, B.; Wang, W.; Camenisch, G. P.; Elmo, J.; Zhang, H.;
Borchardt, R. T. Chem. Pharm. Bull. (Tokyo) 1999, 47, 90–95.
20. Ouyang, H.; Tang, F.; Siahaan, T. J.; Borchardt, R. T. Pharm.
Res. 2002, 19, 794–801.
31. Bogdanowich-Knipp, S. J.; Chakrabarti, S.; Williams, T. D.;
Dillman, R. K.; Siahaan, T. J. J. Pept. Res. 1999, 53, 530–541.
32. Liu, H.; Ong, S.; Glunz, L.; Pidgeon, C. Anal. Chem. 1995, 67,
3550–3557.
21. Bak, A.; Gudmundsson, O. S.; Friis, G. J.; Siahaan, T. J.;
Borchardt, R. T. Pharm. Res. 1999, 16, 24–29.
22. Pauletti, G. M.; Gangwar, S.; Okumu, F. W.; Siahaan, T. J.;
Stella, V. J.; Borchardt, R. T. Pharm. Res. 1996, 13,
1615–1623.
33. Song, X.; Xu, C. R.; He, H. T.; Siahaan, T. J. Bioorg. Chem.
2002, 30, 285–301.
34. Borchardt, R. T.; Hidalgo, I. J.; Hillgren, K. M.; Hu, M. In
Pharmaceutical Applications of Cell and Tissue Culture to
Drug Transport; Wilson, G., Davis, S. S., Illum, L., Eds.;
Plenum: New York, 1991; pp 1–14.
23. Wang, W.; Camenisch, G.; Sane, D. C.; Zhang, H.; Hugger, E.;
Wheeler, G. L.; Borchardt, R. T.; Wang, B. J. Control. Rel.
2000, 65, 245–451.
35. Artursson, P.; Palm, K.; Luthman, K. Adv. Drug Del. Rev.
2001, 46, 27–43.
24. Pauletti, G. M.; Okumu, F. W.; Borchardt, R. T. Pharm. Res.
1997, 14, 164–168.
36. Camenisch, G. P.; Wang, W.; Wang, B.; Borchardt, R. T.
Pharm. Res. 1998, 15, 1174–1181.
25. Pauletti, G. M.; Gangwar, S.; Siahaan, T. J.; Aube, J.;
Borchardt, R. T. Adv. Drug Del. Rev. 1997, 27, 235–256.
26. Okumu, F. W.; Pauletti, G. M.; Velde, D. G. V.; Siahaan, T. J.;
Borchardt, R. T. Pharm. Res. 1997, 14, 169–175.
37. Tang, F.; Borchardt, R. T. Pharm. Res. 2002, 19, 787–793.
38. Tang, F.; Borchardt, R. T. Pharm. Res. 2002, 19, 780–786.
39. Oliyai, R. Adv. Drug Del. Rev. 1996, 19, 275–286.