Factors that restrict the cell permeation of cyclic prodrugs of an opioid peptide, part 3: Synthesis of analogs designed to have improved stability to oxidative metabolism

Article abstract of DOI:10.1002/jps.23109

Previously, our laboratory reported that cyclic peptide prodrugs of the opioid peptide H-Tyr-d-Ala-Gly-Phe-d-Leu-OH (DADLE) are metabolized by cytochrome P450 (CYP450) enzymes, which limits their systemic exposure after oral dosing to animals. In an attempt to design more metabolically stable cyclic prodrugs of DADLE, we synthesized analogs of DADLE cyclized with a coumarinic acid linker (CA; CA-DADLE), which contained modifications in the amino acid residues known to be susceptible to CYP450 oxidation. Metabolic stability and metabolite identification studies of CA-DADLE and its analogs were then compared using rat liver microsomes (RLM), guinea pig liver microsomes (GPLM), and human liver microsomes (HLM), as well as recombinant human recombinant cytochrome P450 3A4 (hCYP3A4). Similar to the results observed for CA-DADLE, incubation of its analogs with RLM, GPLM, and HLM resulted in monohydroxylation of an amino acid side chain on these cyclic prodrugs. When CA-DADLE was incubated with hCYP3A4, similar oxidative metabolism of the peptide was observed. In contrast, incubation of the CA-DADLE analogs with hCYP3A4 showed that these amino-acid-modified analogs are not substrates for this CYP450 isozyme. These results suggest that the amino-acid-modified analogs of CA-DADLE prepared in this study could be stable to metabolic oxidation by CYP3A4 expressed in human intestinal mucosal cells.

Full text of DOI:10.1002/jps.23109

RESEARCH ARTICLE
Factors that Restrict the Cell Permeation of Cyclic Prodrugs
of an Opioid Peptide, Part 3: Synthesis of Analogs Designed
to Have Improved Stability to Oxidative Metabolism
REBECCA NOFSINGER, TARRA FUCHS-KNOTTS, RONALD T. BORCHARDT
Department of Pharmaceutical Chemistry, The University of Kansas, Lawrence, Kansas 66047
Received 26 March 2011; accepted 17 February 2012
Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23109
ABSTRACT: Previously, our laboratory reported that cyclic peptide prodrugs of the opi-
oid peptide H-Tyr–D-Ala–Gly–Phe–D-Leu–OH (DADLE) are metabolized by cytochrome P450
(CYP450) enzymes, which limits their systemic exposure after oral dosing to animals. In
an attempt to design more metabolically stable cyclic prodrugs of DADLE, we synthesized
analogs of DADLE cyclized with a coumarinic acid linker (CA; CA–DADLE), which con-
tained modifications in the amino acid residues known to be susceptible to CYP450 oxidation.
Metabolic stability and metabolite identification studies of CA–DADLE and its analogs were
then compared using rat liver microsomes (RLM), guinea pig liver microsomes (GPLM), and
human liver microsomes (HLM), as well as recombinant human recombinant cytochrome P450
3A4 (hCYP3A4). Similar to the results observed for CA–DADLE, incubation of its analogs
with RLM, GPLM, and HLM resulted in monohydroxylation of an amino acid side chain
on these cyclic prodrugs. When CA–DADLE was incubated with hCYP3A4, similar oxidative
metabolism of the peptide was observed. In contrast, incubation of the CA–DADLE analogs with
hCYP3A4 showed that these amino-acid-modified analogs are not substrates for this CYP450
isozyme. These results suggest that the amino-acid-modified analogs of CA–DADLE prepared
in this study could be stable to metabolic oxidation by CYP3A4 expressed in human intesti-
nal mucosal cells. © 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
J Pharm Sci
Keywords: cyclic peptide prodrug; DADLE; opioid peptide; peptides; metabolism; cytochrome
P450; oxidation; structure; synthesis
because of their intrinsic receptor specificity and
promising pharmacological properties.1–5 The preclin-
ical and clinical development of these peptides and
peptidomimetics has been limited by their undesir-
able biopharmaceutical properties, including poor in-
testinal mucosa permeation and enzymatic instabil-
ity, which limits the blood levels observed after oral
dosing.6–9 A major challenge in making these chemo-
types orally bioavailable lies in improving the bio-
pharmaceutical properties of the peptides and pep-
tidomimetics while maintaining their high affinity
for their pharmacological receptors. For example, H-
Tyr–D-Ala–Gly–Phe–D-Leu—OH (DADLE) was de-
veloped as a biologically active opioid peptide that is
resistant to peptidase.10,11 However, this opioid pep-
tide exhibits poor permeation across biological bar-
riers such as the blood–brain barrier and intestinal
mucosa.9,12–16
INTRODUCTION
Peptides and peptidomimetics have long been of in-
terest to medicinal chemists and pharmacologists
Abbreviations used: ACN, acetonitrile; Boc, t-butoxy car-
bonyl; CA, coumarinic acid linker; CDCl3, chloroform-D; CH2Cl2,
dichloromethane; DMAP, 4-dimethylamino pyridine; EDC-HCl,
1-(3-dimethylaminopropyl)-3-ethylcarbodiimide
hydrochloride;
GPLM, guinea pig liver microsomes; hCYP3A4, human recombi-
nant cytochrome P450 3A4; HLM, human liver microsomes; HOBt,
1-hydroxybenzotriazole; MeOD, methanol-D4; MgSO4, magnesium
sulfate; NHCO3, sodium bicarbonate; OAll, allyl protecting group;
P-gp, P-glycoprotein; RLM, rat liver microsomes; TEA, triethyl
amine; THF, terahydrofuran.
Correspondence to: Rebecca Nofsinger (Telephone: +215-
652-4864; Fax: +215-616-6082; E-mail: Rebecca Nofsinger@
Merck.com)
Rebecca Nofsinger’s present addresses is Merck & Company,
West Point, Pennsylvania 19486.
Tarra Fuchs-Knotts’s present addresses is Achaogen, South San
Francisco, California 94080.
In an attempt to improve the membrane per-
meability of opioid peptides, we have developed a
Journal of Pharmaceutical Sciences
© 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
JOURNAL OF PHARMACEUTICAL SCIENCES
1

2
NOFSINGER, FUCHS-KNOTTS, AND BORCHARDT
Figure 1. DADLE, CA–DADLE, and modified prodrugs of the cyclic opioid peptide CA–DA-
DLE.
fect modification of Phe4 has on oxidative metabolism
cyclic prodrug [CA–DADLE (1)] of the opioid pep-
tide DADLE in which the N- and C-termini of
the peptide are joined via chemical linkers, specifi-
cally the coumarinic acid linker (CA; Fig. 1).17 Un-
like the zwitterionic DADLE, the resultant cyclic
prodrug is uncharged at physiological pH and it
is lipophilic rather than hydroplilic.17–19 However,
this cyclic DADLE prodrug did not show the ex-
pected increased cell permeation.13,15,20,21 Using cell
culture models, it was shown that the cyclic pro-
drug is a substrate for apically polarized efflux
transporters [e.g., P-glycoprotein (P-gp)] in the in-
testinal mucosa.13,17,22–25 Considering the fact that
many P-gp substrates are also substrates for cy-
tochrome P450 (CYP450) enzymes,26–29 it was not
unexpected to find that CA–DADLE undergoes rapid
oxidative metabolism in the presence of liver micro-
somes and human recombinant cytochrome P450 3A4
(hCYP3A4).21
In this paper, we describe the main sites of ox-
idative metabolism on CA–DADLE (e.g., Tyr1 and
Phe4). Using this knowledge and our general un-
derstanding of the structure–activity relationship of
opioid peptides,30–35 two new CA–DADLE analogs
were designed and synthesized (Fig. 1). Specifically,
chemical modifications of the amino acid sequence
of DADLE were made creating the new cyclic pro-
drugs, CA–[Cha4,D-Leu5]–Enk (2) and CA–[Cha4,D-
Ala5]–Enk (3). Ultimately, synthesis of these new
cyclic prodrug analogs will allow us to probe the ef-
and cell permeation. In the studies described in this
paper, the site of oxidative metabolism on CA–DA-
DLE was elucidated, and the design and synthesis of
two new CA–DADLE analogs are reported, as well as
their metabolism by animal and human microsomes
and a cloned human CYP450.
MATERIALS AND METHODS
The cyclic prodrug CA–DADLE (1) was synthesized
in our laboratory following described procedures.8,12
Pooled male Sprague–Dawley rat liver microsomes
(RLM); pooled, male Hartley albino guinea pig liver
microsomes (GPLM); pooled, mixed-gender human
liver microsomes (HLM); and hCYP3A4 expressed
in bactosomes with reductase +b5 were purchased
from Xenotech, LLC (Lenexa, Kansas). DADLE,
diethyl p-nitrophenyl phosphate (paraoxon), keto-
conazole, $-nicotinamide adenine dinucleotide phos-
phate ($-NADPH; reduced form), sodium phosphate
monobasic (NaH2PO4), and sodium phosphate diba-
sic (Na2HPO4) were obtained from Sigma–Aldrich
(St. Louis, Missouri). Lindlar’s catalysis was pur-
chased from Fluka Chemical Corporation (Milwau-
kee, Wisconsin). Amino acids were purchased from
Novabiochem (La Jolla, California), and the fi-
nal coupling reagent, N-N-bis(2-oxo-3-oxozolidinyl)-
phosphinic chloride was purchased from TCI-America
(Portland, Oregon). Silica gel (Selecto Scientific;
JOURNAL OF PHARMACEUTICAL SCIENCES
DOI 10.1002/jps

FACTORS THAT RESTRICT THE CELL PERMEATION OF CYCLIC PRODRUGS OF AN OPIOID PEPTIDE
3
Suwanee, Georgia)) and solvents [high-performance
liquid chromatography (HPLC) grade] were pur-
chased from Fisher Scientific (Pittsburgh, Pennsyl-
vania). All other chemicals were purchased from
Sigma–Aldrich. Tetrahydrofuran (THF) was distilled
over sodium metal in the presence of benzophenone.
Dichloromethane (CH2Cl2) was distilled over calcium
hydride. All other chemicals and solvents were used
as received. During the synthesis process, nuclear
magnetic resonance (NMR) spectra were taken on
a 400-MHz instrument. The identity of the final
cyclic prodrug products were confirmed with a Quat-
tro Micro triple quadrupole mass spectrometer using
positive-mode electrospray ionization (ESI+)(Waters;
Milford, Massachusetts).
quenched with a stop solution (100 :L of methanol
and 750 :L of ACN). The mixture was vortexed and
then centrifuged at 13,000g for 15 min. The super-
natant (500 :L) was removed and evaporated to dry-
ness (CentriVap concentrator; Labconco). The sample
was reconstituted in 100 :L of 25% ACN for analysis.
Analytical Methods
The metabolic stability and metabolite identifica-
tion samples were assayed using HPLC in tandem
with LC–MS/MS. Separation was preformed on a Wa-
ters 2690 HPLC system and a Quattro Micro triple
quadrupole mass spectrometer was employed using
ESI+ (Waters; Milford, Massachusetts). The com-
pounds were eluted on a Vydac C18 column (50 ×
2
˚
1.0 mm internal diameter, 300 A) (Alltech; Deefield,
In Vitro Microsomal Stability
Illinois) employing a binary gradient [solvent A, wa-
ter with 0.1% formic acid (v/v); solvent B, acetronitrile
with 0.1% formic acid (v/v)]. The total run time was
20 min, with a flow rate of 0.2 mL/min (3 min gra-
dient from 5% to 40% B; 3 min gradient from 40%
to 95% B; 2 min hold at 95%; 2 min return to 5% B;
10 min equilibration at 5% B). For metabolic stability
studies, multiple reaction monitoring (MRM) scans
were used to determine the presence of the cyclic pro-
drug parent and the linear peptide using the following
m/z transitions: parent > 136.1 (Tyr) with parent m/z
set at 698.4 {CA–DADLE (1)}, 704.2 {CA–[Cha4,D-
Leu5]–Enk (2)}, and 662.6 {CA–[Cha4,D-Ala5]–Enk
(3)}. The oxidative metabolites for CA–DADLE were
identified using MRM scans with the following m/z
transitions: parent > 102.6, 136.1, 152.2, and 223.0,
with parent m/z set at 714.4 (CA–DADLE + OH). The
oxidative metabolites of the CA–DADLE analogs were
identified with the following m/z transitions: parent
> 102.6, 142.0, 152.2, and 223.0 with parent m/z set
at 720.3 (CA–[Cha4,D-Leu5]–Enk + OH) and 678.6
(CA–[Cha4,D-Ala5]–Enk + OH).
The in vitro stability studies were preformed as pre-
viously described in our laboratory, with only mi-
nor modifications.36 Briefly, the metabolic stability
studies of the cyclic prodrugs 1–3 were preformed in
500 :L of phosphate buffer solution (100 mM, pH 7.4)
containing 2.5 :M of cyclic peptide prodrug, 1 mM
NADPH, and 300 nM CYP450 enzyme from RLM,
GPLM, or HLM, or 20 nM hCYP3A4. The solution
was incubated for 15 min at 37◦C with and without
inhibitors (100 :M paraoxon and/or 5 :M ketocona-
zole). The mixture was quenched with a stop solu-
tion [100 :L of methanol and 750 :L of acetonitrile
(ACN)] and then vortexed to thoroughly mix, followed
by centrifugation at 13,000g for 15 min. The super-
natant (500 :L) was removed and evaporated to dry-
ness (CentriVap concentrator; Labconco, Kansas City,
Missouri). The sample was reconstituted in 100 :L
of 25% ACN and analyzed using liquid chromatog-
raphy–tandem mass spectrometry (LC –MS/MS).
Side-by-side control experiments were performed to
monitor non-enzyme-mediated degradation. For the
control experiments, methanol (100 :L) was added to
the buffer–NAPDH–prodrug solution before the ad-
dition of hCYP3A4 or microsomes. After the addition
of hCYP3A4 or microsomes, the reaction was quickly
stopped with 750 :L of ACN and processed as stated
above. The amount of substrate remaining was calcu-
lated by dividing the average substrate peak area by
the average substrate peak area in the control sample.
Statistics were performed with SigmaStat 3.5 (Sys-
tat Software, Inc; Chicago, Illinois) using a one-way
ANOVA with a Hohn–Sidik multiple comparison.
Synthesis
Standard Solution-Phase Peptide Synthesis Using
t-Butoxy Carbonyl–Amino Acid Chemistry.
Coupling. The t-butoxy carbonyl (Boc)-protected
amino acid (26 mmol) was dissolved in dry
THF (300 mL) and put on ice. To this solu-
tion, 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide
hydrochloride (EDC-HCl; 7 g, 37 mmol) and 1-
hydroxybenzotriazole (HOBt; 5.7 g, 42 mmol) were
added. The solution was stirred for 5 min, and then to
this solution, triethyl amine (TEA; 5.6 g, 55 mmol), 4-
dimethylamino pyridine (DMAP; 353 mg, 3 mmol),
and C-terminal-protected peptide (30 mmol) were
added. The reaction was stirred at room temperature
for 5 h. The THF was evaporated under vacuum. Pu-
rification was preformed via extraction, where 500 mL
of ethyl acetate was used to dissolve the residue. The
Metabolite Identification
Oxidative metabolites of the cyclic prodrugs 1–3 were
determined in 500 :L of phosphate buffer (100 mM,
pH 7.4) containing 50 :M cyclic peptide prodrug, 5
mM NADPH, and 300 nM P450 enzyme from RLM,
GPLM, or HLM; or 20 nM hCYP3A4 bactosomes. The
solution was incubated for 60 min at 37◦C and then
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES

4
NOFSINGER, FUCHS-KNOTTS, AND BORCHARDT
1
ethyl acetate solution was extracted with an acid so-
lution [10% citric acid (3 × 200 mL)] followed by a
base solution [5% sodium bicarbonate (NHCO3) so-
lution (3 × 200 mL)]. A final wash with saturated
brine (1 × 200 mL) was preformed before magnesium
sulfate (MgSO4) was used to dry the organic layer.
The MgSO4 was filtered off and the collected ethyl
acetate was evaporated, resulting in the desired pep-
tide. Further purification (unless noted) was not pre-
formed until final protected peptide chain had been
synthesized.
Yield: 46%. H NMR (400 MHz, MeOD) δ 1.21 (t, 3H
J1 = 7.0, J2 = 14.1) 1.10 (s, 9H), 2.92 (m, 2H), 3.69
(dd, 2H, J1 = 17.1, J2 = 76.0), 4.21 (s, 1H), 4.31 (m,
1H), 6.72 (d, 2H, J = 7.7), 7.05 (d, 2H, J = 7.7). 13C
NMR (400 MHz, MeOD) δ17.01, 27.37, 36.88, 42.55,
56.95, 79.46, 114.86, 127.44, 130.02, 155.99, 156.42,
173.21, 173.45.
Boc–Tyr–D-Ala–Gly–Cha–OAll (8). Yield: 100%. Rf
= 0.78 (chloroform/methanol, 3:1). 1H NMR (400
MHz, MeOD) 0.91 (m, 2H), 1.25 (m, 6H), 1.41 (s,
9H), 1.71 (m, 7H), 2.90 (m, 2H). 3.90 (m 2H), 4.14
(m, 1H), 4.52 (m, 1H0, 4.61(m, 2H), 5.31 (dd, 2H, J1
= 1.4, J2 = 17.2), 5.91 (m, 1H), 6.72 (d, 2H J = 8.3),
7.03 (d, 2H, J = 8.2). 13C NMR (400 MHz, MeOD)
δ15.72, 25.72, 25.97, 26.17, 27.42, 31.86, 33.35, 33.79,
38.62, 42.02, 49.48, 50.20, 56.91, 65.28, 79.45, 114.83,
117.29, 127.37, 130.02, 131.96, 156.01, 156.54, 170.22,
172.27, 173.27, 173.94.
Boc Deprotection. The Boc-protected peptide (34.3
mmol) was dissolved in ethyl acetate (50 mL) and
cooled to 0◦C. HCl gas was bubbled through the solu-
tion for 15 min. The gas was removed, and the reac-
tion was stirred for an additional 15 min at 0◦C. The
solvent was evaporated under vacuum. Further pu-
rification (unless noted) was not preformed until final
protected peptide chain had been synthesized
H-Tyr–D-Ala–Gly–Cha–OAll (9). Yield: 100%. Rf =
1
0.68 (chloroform/methanol, 3:1). H NMR (400 MHz,
MeOD) δ 0.95 (m, 2H), 1.25 (d, 3H, J = 7.2), 1.40 (m,
1H), 1.70 (m, 7H), 3.07 (m, 2H), 3.91 (s, 2H), 4.10 (t,
1H, J1 = 7.5, J2 = 15.0), 4.25 (d, 1H, J = 7.2), 4.55
(dd, 1H, J1 = 6.5, J2 = 8.4), 4.62 (s, 2H), 5.23 (d,
1H J = 10.5), 5.31 (d, 1H, J = 17.2), 5.92 (m, 1H),
6.80 (d, 2H J = 8.3), 7.11 (d, 2H J = 8.3). 13C NMR
(400 MHz, MeOD) δ16.04, 25.72, 25.93, 26.13, 31.89,
33.29, 33.88, 36.28, 38.74, 41.81, 49.58, 50.10, 54.80,
65.47, 115.38, 117.43, 124.69, 130.22, 131.86, 156.88,
168.73, 170.07, 170.15, 172.67, 173.56.
Peptide Arm
Boc–D-Ala–Gly–OAll (4). Yield: 100%. Rf = 0.51 (chlo-
1
roform/methanol, 3:1). H NMR (400 MHz, CDCl3) δ
1.28 (d, 3H, J = 7.0), 1.34 (s, 9H), 3.96 (m, 2H), 4.21
(s, 1H), 4.53 (d, 2H, J = 5.6), 5.13 (d, 1H, J = 10.4),
5.21 (d, 1H, J = 17.2), 5.54 (s, 1H), 5.79 (m, 1H). 13C
NMR (400 MHz, CDCl3) δ 28.22, 41.08, 49.85, 79.67,
118.63, 124.85, 128.73, 131.48, 143.25, 155.49, 169.4,
173.6.
H–D-Ala–Gly–OAll (5). Yield: 100%. Rf = 0.12
(CH2Cl2/methanol/acetic acid, 20:3:0.1). 1H NMR (400
MHz, MeOD) δ 1.56 (d, 3H, J = 5.8), 3.99 (m, 2H), 4.59
(d, 2H, J = 4.8), 5.18 (d, 1H, J = 10.3), 5.27 (d, 1H, J
= 17.1), 5.89 (m, 1H). 13C NMR (400 MHz, CDCl3) δ
40.74, 49.03, 65.55, 117.58, 131.9, 169.29, 170.39.
Boc–Tyr–D-Ala–Gly–OAll (6). Yield: 100%. Rf =
0.23 (chloroform/ACN, 3:1). 1H NMR (400 MHz,
CDCl3) δ 1.20 (d, 3H, J = 6.5), 1.37 (s, 9H), 2.91
(m, 2H), 4.05 (dd, 2H, J1 = 17.4, J2 = 52.3), 4.16
(m, 1H), 4.33 (d, 1H, J = 7.2), 4.61 (d, 2H, J = 5.6),
5.21 (dd, 1H, J1 = 1.2, J2 = 10.5), 5.31 (dd, 1H, J1 =
1.5, J2 = 17.2), 5.93 (m, 1H), 6.72 (d, 2H, J = 8.5),
7.04 (d, 2H, J = 8.3). 13C NMR (400 MHz, MeOD) δ
17.58, 28.28, 41.27, 48.62, 56.64, 66.05, 80.32, 115.63,
118.94, 127.39, 130.32, 131.43, 155.55, 155.85, 169.67,
172.06, 172.87.
Boc–Tyr–D-Ala–Gly–OH (7). To remove the C-
terminal Allyl group, peptide 6 (8.7 g, 20 mmol) was
dissolved in 50 mL dry THF. To this solution, tetrakis
(triphenylphosphine) Palladium [Pd(PPh3)4] (2.3 g,
2 mmol) and morpholine (17 mL, 200 mmol) were
added. The reaction was then stirred at room tem-
perature for 90 min. The THF was evaporated un-
der vacuum, and the residual oil was purified by
silica gel column chromatography with chloroform/
methanol (3:1). Rf = 0.11 (chloroform/methanol, 3:1).
CA Linker. Phenacyl 3-(2-hydroxyphenyl)-propyno-
ate (10): Compound 10 was synthesized using a
method previously described.17 1H NMR (400 MHz,
CDCl3) δ 5.52 (s,2H), 6.93 (t, 1H, J1 = 7.5, J2 = 16.1),
7.00 (d, 1H, J = 8.4), 7.39 (t, 1H, J1 = 8.3, J2 = 17.4),
7.49 (m, 3H), 7.64 (t, 1H, J1 = 7.4, J2 = 13.7), 7.94
(d, 2H, J = 9.5). 13C NMR (400 MHz, CDCl3) δ 67.18,
83.69, 83.33, 105.84, 115.96, 120.69, 127.87, 128.99,
133.38, 133.75, 133.83, 134.22, 153.14, 159.07, 191.02.
Phenacyl 3-(2ꢀ-Boc-alacylocyl hydroxyphenyl)-
propynoate (11): Phenacyl 3-(2-hydroxylphenyl)-
propynoate (10; 1 g, 3.7 mmol) was dissolved in
40 mL of dry CH2Cl2. In another flask containing
10 mL of dry CH2Cl2, Boc–D-Ala–OH (840 mg,
4.4 mmol), EDC-HCl (707 mg, 3.7 mmol), and DMAP
(452 mg, 3.7 mmol) were dissolved. The mixture
was stirred at 0◦C for 5 min. The solution with
compound 10 was transferred to the flask containing
the activated amino acid via cannula and stirred at
0◦C for 5 h. CH2Cl2 was removed under vacuum.
Initial purification was preformed via extraction,
where 200 mL of ethyl acetate was used to dissolve
the residue. The ethyl acetate solution was extracted
with an acid solution [10% citric acid (3 × 100 mL)]
followed by a base solution [5% NHCO3 solution
(3 × 100 mL)]. A final wash with saturated brine
JOURNAL OF PHARMACEUTICAL SCIENCES
DOI 10.1002/jps

FACTORS THAT RESTRICT THE CELL PERMEATION OF CYCLIC PRODRUGS OF AN OPIOID PEPTIDE
5
28.30, 49.43, 80.25, 121.72, 122.81, 125.78, 128.16,
129.80, 130.47, 137.88.
(1 × 100 mL) was preformed before MgSO4 was used
to dry the organic layer. The MgSO4 was filtered
off and the organic layer was evaporated to foam
(1.05 g, 66%). The foam was further purified by silica
gel column chromatography with chloroform/ethyl
acetate (20:1) to give a yellowish oil (750 mg, 47%).
Compound 14: A mixture of compound 13 (112 mg,
0.31 mmol), EDC-HCl (82 mg, 0.43 mmol), and HOBt
(66 mg, 0.49 mmol) was dissolved in dry THF (6 mL)
and stirred at 0◦C for 5 min. To this solution, DMAP
(4 mg, 0.03 mmol) and TEA (66 mg, 0.64 mmol) were
added. In another flask, H-Tyr–D-Ala–Gly–Cha–OAll
(9; 170 mg, 0.33 mmol) was dissolved in 4 mL dry
THF and transferred via cannula to the reaction flask
containing compound 13. The mixture was stirred at
room temperature for 6 h. THF was removed under
vacuum. Initial purification was preformed via extrac-
tion, where 100 mL of ethyl acetate was used to dis-
solve the residue. The ethyl acetate solution was ex-
tracted with an acid solution [10% citric acid (2 × 50
mL)] followed by a base solution [5% NHCO3 solution
(2 × 50 mL)]. A final wash with saturated brine (1
× 200 mL) was preformed before MgSO4 was used to
dry the organic layer. The MgSO4 was filtered off and
the organic layer was evaporated to a yellowish white
solid (200 mg, 80%). The solid was further purified by
silica gel column chromatography with chloroform/
methanol (20:1) to give a white solid (110 mg, 44%).
1
Rf = 0.37 (chloroform/ ethyl acetate, 20:1). H NMR
(400 MHz, CDCl3) δ 1.46 (s, 9H), 1.68 (d, 3H, J =
7.28), 4.64 (m, 1H), 5.31 (m, 1H), 5.48 (s, 2H), 13C
NMR (400 MHz, CDCl3) δ 17.99, 28.29, 49.56, 67.15,
79.82, 82.52, 84.48, 85.20, 113.51, 122.75, 126.32,
127.77, 128.92, 132.41, 133.77, 134.26, 152.78,
155.36, 171.29, 190.76.
Phenacyl 3-(2ꢀ-Boc-alacylocyl hydroxyphenyl)-
propenoate (12): Phenacyl 3-(2ꢀ-Boc-alacylocyl
hydroxyphenyl)-propynoate (11; 750 mg, 1.72 mmol)
was dissolved in absolute ethanol (60 mL). To this,
Lindlar catalyst [5 wt% of palladium on calcium
carbonate, poisoned with lead; 10% (w/w), 75 mg]
and quinoline [2% (w/w), 15 mg] were added. The
reaction was stirred at 0◦C. Hydrogen gas was
bubbled through the solution for 30 min. The HCl
gas was removed, and the reaction, which was
monitored by 1H NMR, was stirred under a hydrogen
atmosphere until complete. Upon completion, the
catalyst was removed and the solvent was evaporated
under vacuum to give a yellowish foam (680 mg,
82%). The foam was purified by silica gel column
chromatography with hexanes/ethyl acetate (4:1) to
give a light yellow solid (670 mg, 80%). Rf = 0.65
(ethyl acetate/hexanes/methanol, 2:4:1). 1H NMR
(400 MHz, CDCl3) δ 1.46 (s, 9H), 1.56 (d, 3H, J
= 7.3), 4.53 (t, 1H, J1 = 7.2, J2 = 14.5), 5.37 (d,
2H, J = 16.4), 6.22 (d, 1H, J = 12.2), 7.08 (m, 2H),
7.21 (t, 1H, J1 = 7.6, J2 = 15.2), 7.32 (m, 1H), 7.46
(m, 2H), 7.59 (m, 2H), 7.87 (d, 2H, J = 7.2). 13C
NMR (400 MHz, CDCl3) δ 18.35, 28.34, 49.56, 66.09,
79.96, 121.21, 121.74, 125.73, 127.80, 128.83, 129.93,
130.30, 133.92, 134.09, 139.60, 147.86, 155.38,
164.52, 171.62, 192.09.
1
Rf = 0.26 (chloroform/methanol, 20:1). H NMR (400
MHz, CDCl3) δ 0.87 (m, 2H), 1.16 (m, 6H), 1.40 (s,
9H), 1.53 (d, 3H, J = 7.2), 1.63 (m, 7H), 2.55 (m, 1H),
2.74 (m, 1H), 3.90 (m, 2H), 4.35 (m, 2H), 4.59 (m, 3H),
5.24 (d, 2H, J = 10.4), 5.33 (d, 2H, J = 17.3), 5.85 (m,
1H), 6.10 (d, 1H J = 12.1), 6.67 (m, 3H), 6.84 (d, 2H,
J = 7.3), 7.00 (m, 2H), 7.20 (m, 1H), 7.32 (m, 1H). 13
C
NMR (400 MHz, CDCl3) δ 18.11, 25.91, 26.10, 26.36,
28.37, 32.32, 33.40,33.92, 39.40,49.33, 50.35, 65.85,
80.33, 115.65, 118.70, 121.88, 126.52, 127.15, 129.01,
130.28, 131.61, 147.80, 155.51, 166.84, 169.65, 172.59,
173.01.
Compound 15. Removal of Allyl group: Compound
14 (110 mg, 0.13 mmol) was dissolved in 10 mL dry
THF. To this solution, Pd(PPh3)4 (15.5 mg, 1.0 mmol)
and morpholine (117 mg, 1.34 mmol) were added.
The reaction was then stirred at room temperature
for 90 min. The THF was evaporated under vacuum
and the residual oil was purified by silica gel column
chromatography with chloroform/methanol (3:1).
Rf = 0.43 (chloroform/methanol, 3:1).
3-(2ꢀ-Boc-alacylocyl
hydroxyphenyl)-propenoic
(13): A mixture of phenacyl 3-(2ꢀ-Boc-alacylocyl
hydroxyphenyl)-propenoate (12; 100 mg, 0.21 mmol)
and zinc powder (271 mg, 4.14 mmol) was cooled to
10◦C. Acetic acid (22 mL) was slowly added to the
reaction. The mixture was stirred at 10◦C for 30 min.
The supernant was filtered and concentrated under
vacuum. The residual oil was purified by silica gel
column chromatography with ethyl acetate/hexanes/
methanol (2:4:1) to give a colorless oil (90 mg, 75%).
Rf = 0.43 (ethyl acetate/hexanes/methanol (2:4:1). 1H
NMR (400 MHz, MeOD) δ 1.48 (s, 9H), 1.52 (d, 3H, J
= 7.3), 4.38 (m, 1H), 6.10 (d, 1H, J = 12.4), 6.87 (d,
1H, J = 12.4), 7.11 (d, 1H, J = 8.0), 7.22 (t, 1H, J1 =
7.4, J2 = 14.6), 7.35 (t, 1H, J1 = 7.6, J2 = 15.2), 7.64
(d, 1H, J = 7.3). 13C NMR (400 MHz, CDCl3) δ 18.07,
Removal of Boc-group: The OAll deprotected prod-
uct from above was dissolved in ethyl acetate (60 mL)
and cooled to 0◦C. HCl (gas) was bubbled through
the solution for 10 min. The HCl gas was removed,
and the solution was stirred for another 10 min at
0◦C. Vacuum was used to evaporate the solvent, yield-
ing a white solid (15) to which no further purification
was preformed. Total yield for these two deprotection
steps was 49%. Rf = 0.47 (chloroform/methanol, 20:1).
1H NMR (400 MHz, MeOD) δ 0.94 (m, 2H), 1.17 (d, 3H,
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES

6
NOFSINGER, FUCHS-KNOTTS, AND BORCHARDT
J = 7.3), 1.43 (m, 1H), 1.57 (m, 1H), 1.72 (d, 3H, J =
7.1), 1.76 (m, 7H), 2.87 (d, 2H, J = 7.5), 3.66 (dd, 2H, J1
= 16.8, J2 = 50.6), 3.86 (m, 1H), 4.30 (m, 1H), 4.40 (m,
2H), 6.25 (d, 1H, J = 12.1), 6.72 (d, 2H, J = 8.3), 6.83
(d, 1H, J = 12.2), 7.01 (d, 2H, J = 8.1), 7.17 (d, 1H, J =
8.5), 7.23 (t, 1H, J1 = 7.3, J2 = 14.7), 7.38 (d, 2H, J =
7.4). 13C NMR (400 MHz, MeOD) δ 19.38, 25.70, 25.97,
26.18, 31.83, 33.43, 33.81, 36.17, 38.92, 42.06, 48.66,
56.10, 60.15, 114.87, 121.33, 125.53, 126.13, 126.85,
129.05, 129.15, 129.91, 130.13, 133.25, 147.13, 156.12,
166.79, 168.31, 170.40, 172.70.
retention time of about 45 min. Rf = 0.22 (chloroform/
methanol, 20:1) 1H NMR (400 MHz, MeOD) δ 0.99 (m,
4H), 1.20 (d, 3H, J = 7.3), 1.25 (m, 9H), 1.56 (d, 3H, J
= 7.4), 1.74 (m, 7H), 2.85 (m, 2H), 3.88 (m, 2H), 4.03
(m, 1H), 4.28 (m, 1H), 4.46 (m, 1H), 4.65 (m, 1H), 6.20
(d, 1H, J = 12.1), 6.70 (d, 2H, J = 8.5), 6.78 (d, 1H, J
= 12.0), 6.97 (d, 2H, J = 8.3), 7.07 (m, 1H), 7.23 (m,
1H), 7.36 (m, 2H). ESI–MS m/z 662.6 (M + 1).
CA–[Cha4,D-Leu5]–Enk (1). 1H NMR (500 MHz,
MeOD-d4) δ 1.37 (dd, 6H, J1 = 6.0, J2 = 17.3), 1.57 (d,
3H, J = 7.3), 1.65 (m, 3H), 1.82 (m, 2H), 2.11 (m, 9H),
3.10 (dd, 1H, J1 = 7.3, J2 = 13.7), 3.21 (m, 1H), 4.18
(d, 1H, J = 17.1), 4.31 (dd, 1H, J1 = 7.3, J2 = 14.6)
4.68 (t, 1H, J1 = 7.7, J2 = 15.1), 4.86 (dd, 1H, J1 =
5.5, J2 = 10.0), 5.05 (m, 1H), 6.55 (d, 1H, J = 12.1),
7.06 (d, 2H, 8.5), 7.14 (d, 1H, J = 12.1), 7.28 (d, 2H,
J = 8.4), 7.42 (d, 1H, J = 8.0), 7.60 (m, 2H), 7.73 (m,
1H). ESI–MS m/z 704.5 (M + 1).
CA–[Cha4,D-Ala5]–Enk (2). Compound 15 was
initially dissolved in
1
mL anhydrous N,N-
dimethylformamide, and then 50 mL of CH2Cl2 was
slowly added. To this solution, TEA (30 mg, 0.3 mmol)
and bis(2-oxo-3-oxazolidinyl) phosphinic chloride (53
mg, 0.21 mmol) were added. A cloudy mixture re-
sulted, and the reaction was stirred at room temper-
ature for 18 h. The resultant clear yellow solution
was evaporated under vacuum to give a yellow solid.
The solid was dissolved in ethyl acetate (10 mL) and
washed with another 10 mL of water. The ethyl ac-
etate layer was collected and washed with 10% cit-
ric acid (1 × 10 mL) and 5% NaHCO3 solution (1 ×
10 mL). The organic layer was collected, dried with
MgSO4, filtered, and evaporated to a clear oil. The oil
was purified by semi-preparative HPLC with ultravi-
olet detection. The eluate was analyzed by ESI–MS,
and fractions containing peptide prodrug were pooled.
The pooled fractions were lyophilized, resulting in a
white solid (6.7 mg, 34%). Semi-preparative HPLC
was carried out using a C18 Dynamax column (300
RESULTS
Synthesis
The synthesis and characterization of the new CA–
DADLE analogs (Fig. 1) was undertaken using the
methodology developed previously in our laboratory.17
The synthetic design for the new cyclic prodrugs
was based on retrosynthetic analysis leading back
to two fundamental pieces, the peptide arm and a
CA moiety (Fig. 2). The peptide arm was synthesized
utilizing standard solution-phase Boc–amino acid
chemistry, an example of which is described in Figure
3. L-amino acids were used unless otherwise desig-
nated. The general synthesis of the CA base (10) is re-
ported elsewhere.17 Using standard peptide-coupling
chemistry, the CA base was further modified to incor-
porate the fifth amino acid of the peptide sequence
forming the linker moiety (Fig. 4). After protection
and deprotection steps, the linker moiety was cou-
pled with the peptide arm to form the linear precursor
(Fig. 4). The final step in the synthesis of the cyclic
prodrugs was cyclization of the linear precursor. This
step forms the final peptide bond between the fourth
˚
A)(Agilent; Santa Clara, California). The cyclic pep-
tide prodrug was isolated using a gradient from 10%
to 100% B (solvent A, 90% water with 0.01% formic
acid; solvent B, 90% ACN with 0.01% formic acid). The
gradient was applied stepwise for a total run time of
121 min, with a flow rate of 5 mL/min (16 min gra-
dient from 10% to 30% B; 10 min gradient from 30%
to 40% B; 40% B for 10 min; 15 min gradient from
40% to 45% B; 10 min gradient from 45% to 100% B;
15 min at 100% B; 15 min return to 10% B; 30 min
equilibration at 10% B). The peptide prodrug had a
Figure 2. Retrosynthetic analysis leading to the peptide arm and coumarinic acid linker
fragments. aa, amino acid.
JOURNAL OF PHARMACEUTICAL SCIENCES
DOI 10.1002/jps

FACTORS THAT RESTRICT THE CELL PERMEATION OF CYCLIC PRODRUGS OF AN OPIOID PEPTIDE
7
stability of the cyclic prodrugs (i.e., CA–DADLE,
CA–[Cha4,D-Leu5]–Enk, and CA–[Cha4,D- Ala5]–Enk
show only 27%, 60%, and 27% of cyclic prodrug re-
maining, respectively, after a 15 min incubation). The
addition of ketoconazole alone or ketoconazole and
paraoxon to the incubation mixture significantly de-
creased the metabolism of the three cyclic prodrugs.
As shown in Figure 5, more than 90% of the cyclic pro-
drugs 1–3 were detected after a 15 min incubation.
When incubated in the presence of GPLM, more
than 90% of the DADLE was detected after a 15
min incubation with or without paraoxon and/or ke-
toconazole (Fig. 6). However, the cyclic prodrugs 1–3
were very unstable when incubated with GPLM (i.e.,
6% of CA–DADLE, 34% of CA–[Cha4,D-Leu5]–Enk,
and 8% CA–[Cha4,D-Ala5]–Enk were detected after
a 15 min incubation). Inclusion of paraoxon did lit-
tle to enhance the metabolic stability of CA–DADLE,
CA–[Cha4,D-Leu5]–Enk, and CA–[Cha4,D-Ala5]–Enk,
with only 11%, 42%, and 11%, respectively, of cyclic
prodrug remaining after a 15 min incubation. How-
ever, inclusion of ketoconazole in the incubation mix-
ture dramatically increased the stability of the cyclic
prodrugs (i.e., >70% of prodrugs 2 and 3, and 44%
of prodrug 1 were present after incubation with
GPLM). The results from inclusion of ketoconazole
and paraoxon to the incubation mixture were similar
to the results observed with ketoconazole alone (i.e.,
>65% of the cyclic prodrugs 1–3 remained after the
15 min incubation with GPLM).
Figure 3. Synthesis of one of the peptide arms, a funda-
The metabolic stability of cyclic prodrugs 1 and 2
when incubated with HLM were similar to the results
observed with RLM and GPLM. DADLE is stable un-
der all conditions, showing a significant amount of the
opioid peptide remaining after a 15 min incubation
with HLM (Fig. 7). Contrary to DADLE, the cyclic pro-
drugs 1–3 were unstable in the presence of HLM, with
only 15%, 50%, and 20% remaining of CA–DADLE,
CA–[Cha4,D-Leu5]–Enk, and CA–[Cha4,D-Ala5]–Enk,
respectively (Fig. 7). The metabolic stability of the
cyclic prodrugs 1–3 is increased significantly with the
addition of ketoconazole (>80% remaining after a 15
min incubation), but remains low with the addition of
paraoxon (<20% remaining after a 15 min incubation;
Fig. 7). Inclusion of ketoconazole and paraoxon to the
HLM incubation mixture resulted in more than 90%
recovery of the cyclic prodrugs 1–3.
The cyclic prodrugs showed significantly differ-
ent metabolic stability profiles when incubated with
hCYP3A4 (Fig. 8). CA–DADLE did not exhibit high
stability in the presence of hCYP3A4 in the ab-
sence of inhibitors, with only 9% of the parent
drug remaining under this condition. The CA–DA-
DLE analogs 2 and 3, however, displayed relatively
high metabolic stability in the presence of hCYP3A4.
mental building block for the synthesis of cyclic prodrugs.
and fifth amino acids in the peptide sequence and
yields a cyclic final product (Fig. 4). The synthetic
1
intermediates were confirmed by H and 13C NMR.
The structural identities of the cyclic prodrugs were
verified by 1H NMR and ESI–MS.
In Vitro Microsomal Stability
The metabolic characteristics of DADLE and cyclic
prodrugs of DADLE were investigated using RLM,
GPLM, HLM, or hCYP3A4 in the presence or ab-
sence of specific enzyme inhibitors (e.g., paraoxon,
an esterase B inhibitor; or ketoconazole, a CYP450
inhibitor). As shown in Figure 5, DADLE is metabol-
ically stable when incubated with RLM in the ab-
sence or presence of paraoxon and/or ketoconazole.
In contrast, the DADLE cyclic prodrugs 1–3 show
significant metabolic instability when incubated with
RLM (i.e., CA–DADLE, CA–[Cha4,D-Leu5]–Enk, and
CA–[Cha4,D- Ala5]–Enk show only 20%, 48%, and 17%
of cyclic prodrug remaining, respectively, after a 15
min incubation). Inclusion of paraoxon in the incu-
bation mixture did little to improve the metabolic
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES

8
NOFSINGER, FUCHS-KNOTTS, AND BORCHARDT
Figure 4. Final synthetic route to CA–[Cha4,D-Ala5]–Enk (2), including the addition of
the fifth amino acid to the basic linker moiety and formation of the linear precursor
Tyr-D-Ala-Gly-Cha.
The data show 78% of the cyclic parent remain-
ing for CA–[Cha4,D-Leu5]–Enk and 87% remaining
for CA–[Cha4,D-Ala5]–Enk. The addition of ketocona-
zole did improve the metabolic stability of CA–DA-
DLE fivefold, but had relatively little effect on the
metabolic stability of the CA–DADLE analogs 2 and
3. DADLE shows high stability in the presence of
hCYP3A4, with more than 98% of the parent remain-
ing with and without inhibitors.
Metabolite Identification
DADLE, the cyclic prodrugs 1–3, and their metabo-
lites have been analyzed by LC–MS/MS. For all
compounds, similar common immonium product ions
resulted (Fig. 9). The m/z ratio of these common prod-
uct ions correlate with portions of the peptide that
contain various amino acid fragments. The fragmen-
tation shows peaks at m/z of 86.6, 120.1 (or 126.0),
Figure 5. Stability of DADLE and the cyclic prodrugs CA–DADLE (1), CA–[Cha4,D-Leu5]–Enk
(2), and CA–[Cha4,D-Ala5]–Enk (3) in the presence of rat liver microsomes with and without
inhibitors after 15 min incubation (∗p < 0.001). KTZ, ketoconazole.
JOURNAL OF PHARMACEUTICAL SCIENCES
DOI 10.1002/jps

FACTORS THAT RESTRICT THE CELL PERMEATION OF CYCLIC PRODRUGS OF AN OPIOID PEPTIDE
9
Figure 6. Stability of DADLE and the cyclic prodrugs CA–DADLE (1), CA–[Cha4,D-Leu5]–Enk
(2), and CA–[Cha4,D-Ala5]–Enk (3) in the presence of guinea pig liver microsomes with and
without inhibitors after 15 min incubation (∗p < 0.001 and #p < 0.003). KTZ, ketoconazole.
were identified using MRM. The oxidized product ions
being investigated had m/z ratios of 102.6, 136.1 (or
142.0), 152.2, and 223.0 corresponding to the previ-
ously determined product ions with the addition of
a hydroxyl group. It should be noted that with this
method, only the amino acid undergoing oxidation
can be identified. The exact location of the hydroxyl
group on the amino acid was not determined. Oxida-
tion leading to the addition of two or three hydroxyl
groups was not seen under any of the experimental
conditions studied. Also, when no oxidative metabo-
lites were detected, the original unoxidized molecule
was observed, confirming the metabolic stability of
either DADLE or the cyclic prodrugs.
136.2, and 207.0 (Fig. 9). These fragments are re-
lated to the peptide portion of these prodrugs and
have been identified as Leu, Phe (or Cha), Tyr, and
a dipeptide fragment, Tyr–Ala, respectively. In the
cyclic prodrug spectra, there was an additional prod-
uct ion peak identified at m/z 147.0, which corre-
sponds to the coumarin linker of the cyclic prodrugs.
The metabolic stability of the CA in the presence of
animal and human microsomes or hCYP3A4 has been
confirmed [i.e., hydroxylated coumarins have not been
observed, and only m/z 147 has been detected (data
not shown)]. With the common product ions identified
for both the linear and cyclic prodrugs, the MRM tran-
sitions of oxidative metabolites of the cyclic prodrugs
could be predicted.
The metabolite identification studies correlate well
with the metabolic stability studies. DADLE was
shown to be metabolically stable under the four
After incubation of the cyclic prodrugs with RLM,
GPLM, HLM, or hCYP3A4, the oxidative product ions
Figure 7. Stability of DADLE and the cyclic prodrugs CA–DADLE (1), CA–[Cha4,D-Leu5]–Enk
(2), and CA–[Cha4,D-Ala5]–Enk (3) in the presence of human liver microsomes (HLM) with and
without inhibitors after 15 min incubation (∗p = 0.022 and #p < 0.001). KTZ, ketoconazole.
Figure 8. Stability of DADLE and the cyclic prodrugs CA–DADLE (1), CA–[Cha4,D-Leu5]–Enk
(2), and CA–[Cha4,D-Ala5]–Enk (3) in the presence of hCYP3A4 with and without ketoconazole
after 15 min incubation (∗p < 0.001).
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES

10
NOFSINGER, FUCHS-KNOTTS, AND BORCHARDT
3 showed no significant metabolite formation under
the same experimental conditions.
DISCUSSION
Synthetic Strategy
Previously, our laboratory has reported that in spite
of optimal physicochemical properties, cyclic prodrugs
of DADLE, formed using different chemical link-
ers, do not exhibit increased cell membrane perme-
ation when compared with the linear opioid pep-
tide permeation.12,13,20–22,25 The poor cell permeation
characteristics of these cyclic prodrugs appear to arise
from their substrate activity for efflux transporters
and/or their potential to be metabolized by CYP450
enzymes.
In an attempt to improve the permeation prop-
erties of one of these cyclic DADLE prodrugs, two
structural analogs of CA–DADLE were designed and
synthesized. The main goal in designing and syn-
thesizing these CA–DADLE analogs was to stabilize
the cyclic prodrug toward oxidative metabolism while
still maintaining high affinity for the opioid receptor.
Metabolite identification studies described in this pa-
per showed that CA–DADLE was oxidized at the Tyr1
and/or Phe4 residues when incubated with animal or
human microsomes or a recombinant CYP450 (e.g.,
hCYP3A4). Because it has been well established that
modification of the Tyr1 residue on opioid peptides
dramatically decreases their affinities for the opioid
receptor,31,37,34,35 structural changes to this critical
amino acid residue were not attempted in our studies.
Instead, a modification at the Phe4 residue of CA–DA-
DLE was made, which should have had minimal effect
on the peptides’ affinity for the opioid receptors.30,33
One prodrug analog was designed to differ from
CA–DADLE by modification of the fourth amino acid
residue from Phe4 to Cha4, creating CA–[Cha4,D-
Leu5]–Enk (2; Fig. 1). Enkephalin analogs having this
type of structural change (e.g., Phe4 to Cha4) have
been shown to retain opioid-like activity in vivo,30
with only slight modification of the potency and bind-
ing affinity for the opioid receptor.33 Small opioid
peptides have also been shown to maintain modest
binding affinity to the opioid receptors when Phe4
to Cha4 modifications are made in the amino acid
sequence.38–40
Figure 9. Structures of the common ion products of DA-
DLE and CA–DADLE.
oxidative conditions studied (e.g., incubation with
RLM, GPLM, HLM, or hCYP3A4). The metabolism
of the cyclic prodrugs 1–3 was studied under the
same experimental conditions, but yielded very dif-
ferent results. The metabolites arising from the ox-
idation of CA–DADLE, CA–[Cha4,D-Leu5]–Enk, and
CA–[Cha4,D-Ala5]–Enk are shown in Table 1. These
data showed that CA–DADLE is hydroxylated on
Tyr1 or Phe4 in the presence RLM, but CA–DA-
DLE analogs 2 and 3 showed oxidation on only Tyr1.
Incubations with GPLM yielded similar results for
CA–DADLE (e.g., hydroxylation at Tyr1 or Phe4).
CA–[Cha4,D-Leu5]–Enk when incubated with GPLM
showed an identical metabolic profile (e.g., hydroxy-
lation at Tyr1 or Cha4). However, when CA–[Cha4,D-
Ala5]–Enk was incubated with GPLM, only the Tyr1
metabolite was formed. Incubation with HLM yielded
the same metabolites as incubation with GPLM for
the cyclic prodrug analogs 2 and 3 (e.g., hydroxylation
at Tyr1 or Cha4 for CA–[Cha4,D-Leu5]–Enk (2), and
Tyr1 for CA–[Cha4,D-Ala5]–Enk). In contrast, CA—
DADLE when incubated with HLM generated only
one metabolite (e.g., hydroxylation at Tyr1), not the
metabolite arising from hydroxylation of Phe4. Inter-
estingly, in the presence of hCYP3A4, CA–DADLE
yields the metabolite arising from the oxidation at
Phe4 position, whereas cyclic prodrug analogs 2 and
The design of the second analog of CA–DADLE in-
cludes the aforementioned Phe4-to-Cha4 structural
Table 1. Metabolites Arising from Phase I Oxidation of CA–DADLE, CA–[Cha4,D-Leu5]–Enk, and CA–[Cha4,D-Ala5]–Enk
Compound
RLM Oxidation
GPLM Oxidation
HLM Oxidation
hCYP3A4 Oxidation
CA–DADLE
Tyr1, Phe4
Tyr1
Tyr1, Phe4
Tyr1, Cha4
Tyr1
Tyr1
Tyr1, Cha4
Tyr1
Phe4
CA–[Cha4,D-Leu5]–Enk
CA–[Cha4,D-Ala5]–Enk
No significant metabolites
No significant metabolites
Tyr1
JOURNAL OF PHARMACEUTICAL SCIENCES
DOI 10.1002/jps

FACTORS THAT RESTRICT THE CELL PERMEATION OF CYCLIC PRODRUGS OF AN OPIOID PEPTIDE
11
Figure 10. Scheme for the bioconversion of the ester-based cyclic prodrugs.
change as well as an additional modification at the
fifth amino acid residue (e.g., Leu5 to Ala5), creating
CA–[Cha4,D-Ala5]–Enk (3; Fig. 1). The rationale for
this change in the fifth amino acid from D-Leu5 to D-
Ala5 was to hopefully increase the rate of bioconver-
sion of the prodrug by esterases (Fig. 10). Cleavage
of the ester bond formed between the carboxylic acid
of the fifth amino acid and the phenolic group of the
coumarin linker results in the formation of a linear
intermediate (Fig. 10). This intermediate then under-
goes intramolecular lactonization, resulting in the for-
mation of coumarin and the linear peptide. Because
hydrolysis of the ester bond is the rate-determining
step in the bioconversion, a rapid rate of hydrolysis
of this bond would be preferred. The presences of D-
Ala as the fifth amino acid in CA–[Cha4,D-Ala5]–Enk
provides reduced steric bulk around the ester bond,
leaving it more exposed to esterases and potentially
aiding in the bioconversion of the cyclic prodrug. In
support of this idea, it has been shown that a de-
crease in steric bulk around the ester bond increases
the lactonization rate in similar cyclic prodrugs.41 The
D-enantiomer of Ala instead of the L-enantiomer was
used in this analog because in the enkephalin class
of opioid peptides, the chirality of this amino acid has
been shown to greatly affect the opioid peptide’s sta-
bility toward peptidases.10,11
The strategy for the synthesis of the two CA–DA-
DLE analogs was taken from a retrosynthetic anal-
ysis of the structures of the desired molecules. Ret-
rosynthetically, the bond between the fourth and fifth
amino acid was broken, leading to a linear precur-
sor (Fig. 2). The molecule was then further broken
down into two parts: (i) a tetrapeptide arm and (ii) a
linker moiety containing the fifth amino acid. These
two parts were the building blocks for the conver-
gent synthesis of the cyclic peptide prodrug analogs.
The advantage of using this method is that all the
cyclic prodrugs can be constructed from these build-
ing blocks. Thus, this strategy provides a reduction in
the amount of synthetic work required to make the
desired cyclic prodrugs.
In Vitro Microsomal Stability
The metabolic stability and metabolite identification
studies of CA–DADLE, CA–[Cha4,D-Leu5]–Enk, and
CA–[Cha4,D-Ala5]–Enk were performed in the pres-
ence of animal and HLM (e.g., RLM, GPLM, and
HLM). All of the cyclic prodrugs were found to be
very unstable in the presence of animal and hu-
man microsomes (Figs. 5–7). In contrast, DADLE was
found to be metabolically stable in these microsomal
preparations.
The instabilities of cyclic prodrugs 1–3 in these
microsomal preparations were shown to arise from
oxidative metabolism mediated by CYP450 en-
zymes. For example, the stability of CA–DADLE
and CA–[Cha4,D-Ala5]–Enk could be increased four-
fold when ketoconazole, a CYP450 inhibitor, was in-
cluded in the incubation media. For CA–[Cha4,D-
Leu5]–Enk, the results are less dramatic, showing
only a twofold increase. Overall, these results suggest
that CA–[Cha4,D-Leu5]–Enk is slightly more stable
toward CYP450-mediated oxidative metabolism than
cyclic prodrugs 1 and 3. Also, although the addition
of ketoconazole results in a significant increase in the
amount of cyclic prodrugs recovered, the addition of
paraoxon, an esterase B inhibitor, only marginally in-
creased the amount of cyclic prodrugs recovered from
the incubation media (Figs. 5–7). This suggests that
the metabolic instability of cyclic prodrugs 1–3 does
not arise from premature bioconversion of the ester
prodrugs.
Interestingly, when metabolic stability assays are
performed in the presence of hCYP3A4, very dif-
ferent results were observed. Although CA–DADLE
was found to be metabolically unstable in the pres-
ence of hCYP3A4, the CA–DADLE analogs 2 and 3
were found to be quite stable under the same exper-
imental conditions (Fig. 8). CA–[Cha4,D-Leu5]–Enk
and CA–[Cha4,D-Ala5]–Enk both showed more than
75% cyclic prodrug recovered in the presence of
hCYP3A4 without inhibitors as compared with 9% re-
covered for CA–DADLE (Fig. 8). With the addition of
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES

12
NOFSINGER, FUCHS-KNOTTS, AND BORCHARDT
ketoconazole, the recovery of cyclic prodrugs 2 and
3 increased to more than 95%. These results indi-
cate that CA–[Cha4,D-Leu5]–Enk and CA–[Cha4,D-
Ala5]–Enk are not metabolized by hCYP3A4 to the
same extent as CA–DADLE. However, when these re-
sults using recombinant hCYP3A4 are compared with
the animal and human microsomal stability data de-
scribed above, one must conclude that CA–[Cha4,D-
Leu5]–Enk and CA–[Cha4,D-Ala5]–Enk are good sub-
strates for other CYP450 isozymes.
The stability of the CA moiety to oxidative
metabolism reveals that the addition of this chemical
moiety is not directly responsible for the metabolic
instability of the cyclic prodrugs 1–3. Also having
shown separately that DADLE is stable to oxidative
metabolism leads to the conclusion that the physio-
chemical properties (e.g., increased lipophilicity) and/
or the increased rigidity of the cyclic produgs is re-
sponsible for the metabolic instability of the cyclic
prodrugs 1–3. By restricting the solution conforma-
tion of the peptide portion of these prodrugs, amino
acid side chains (e.g., Phe4) must assume conforma-
tions that facilitate their binding to CYP450 isozymes
and their subsequent conversion to oxidative metabo-
lites. In addition, it is not unexpected to observe what
appear to be differences in substrate activity of these
cyclic prodrugs for different CYP450 isozymes. It is
interesting to note that opioid peptides themselves
have been shown to have a diminished binding affin-
ity to their receptor upon cyclization, which is hy-
pothesized to arise from the decreased flexibility of
the cyclic compounds.26–28
liver microsomal assays. After a 60 min incuba-
tion with hCYP3A4, CA–DADLE showed hydroxy-
lation at Tyr1, whereas the cyclic prodrugs 2 and 3
showed no significant metabolite formation (Table 1).
This is consistent with the metabolic stability re-
sults, which showed cyclic prodrugs 2 and 3 to have
greatly enhanced metabolic stability in the presence
of hCYP3A4 (Fig. 8). Because CYP3A4 is an impor-
tant barrier to intestinal absorption, proving that
CA–[Cha4,D-Leu5]–Enk and CA–[Cha4,D-Ala5]–Enk
are metabolically stable, showing no metabolite for-
mation in the presence of hCYP3A4, is a significant
result. The increased metabolic stability of the chem-
ically modified prodrugs 2 and 3 in the presence of
hCYP3A4 makes them good candidates for evaluation
of their absorption after oral administration.
The potential oxidative metabolism of the CA in
cyclic prodrugs 1–3 was carefully monitored in our
metabolic stability studies. Oxidation at the site of
the CA is considered a high probability because of
the known metabolic susceptibility of coumarin.42–44
Metabolite identification studies of the cyclic pro-
drugs 1–3, however, provided no evidence for the for-
mation of hydroxy-coumarin. In fact, the detection
of unhydroxylated coumarin from the cyclic prodrugs
1–3 after incubation with animal and human micro-
somes confirmed the metabolic stability of the CA moi-
ety in these molecules.
It should be noted that in-depth investigations
into the exact structures of the oxidative metabolites
formed from cyclic prodrugs 1–3 were not undertaken
in these studies. Instead, the focus of this research
was to identify the amino acids that are the sites of
oxidative metabolism in these cyclic prodrugs and to
use this information to design cyclic prodrug analogs
that are more metabolically stable to CYP450.
Metabolite Identification
Metabolite identification studies were performed
in order to identify the specific sites of oxidative
metabolism on the cyclic prodrugs 1–3. The primary
sites of oxidation were at Phe/Cha4 and Tyr1. Because
the opioid activity of peptides and peptidomimetics
has been shown to be highly dependent on an un-
modified Tyr1 on the peptide chain,31,37,34,35 it was
decided not to modify this amino acid in an attempt
to improve its metabolic stability. This decision was
made in spite of the fact that the oxidation on Tyr1
gives rise to the majority of the metabolites for all
three cyclic prodrugs (Table 1). In the presence of
microsomes, cyclic prodrug 3 shows fewer oxidative
metabolites than CA–DADLE. However, analog 3 did
not show better metabolic stability in the microsomal
assays as compared with CA–DADLE (Figs. 5–7). The
reason for the similarities in the extent of oxidative
metabolism of cyclic prodrugs 1 and 3 could arise from
the change of Phe4 to Cha4 in 3, making Tyr1 more
susceptible to hydroxylation in the presence of liver
microsomes.
CONCLUSION
The primary objective of the research described in
this paper was to improve the biopharmaceutical
properties, particularly metabolic stability, of a cyclic
prodrug (CA–DADLE) of the opioid peptide DA-
DLE, which had been previously synthesized in our
laboratory.17 To this end, we have designed and syn-
thesized two new analogs of CA–DADLE and char-
acterized their stability to oxidative metabolism me-
diated by CYP450 enzymes. The synthesis of these
analogs involved a convergent method where a linker
moiety was combined with a peptide arm and then
cyclized in a final step, yielding the cyclic prodrugs
2 and 3. The design of these analogs of CA–DADLE
was based on literature knowledge about the struc-
tural requirements for the binding of opioid peptides
to their pharmacological receptor and knowledge gen-
erated in studies about the oxidative metabolism of
CA–DADLE. On the basis of the knowledge that Phe4
The metabolite identification results are very dif-
ferent in the presence of hCYP3A4 than in the
JOURNAL OF PHARMACEUTICAL SCIENCES
DOI 10.1002/jps

FACTORS THAT RESTRICT THE CELL PERMEATION OF CYCLIC PRODRUGS OF AN OPIOID PEPTIDE
13
on CA–DADLE was one of two major sites of ox-
idative metabolism, CA–[Cha4,D-Leu5]–Enk (2) and
CA–[Cha4,D-Ala5]–Enk (3) were designed and synthe-
sized. When incubated with animal and human micro-
somes, cyclic prodrug analogs 2 and 3 did not show in-
creases in metabolic stability as compared with CA—
DADLE. This finding was not totally unexpected be-
cause the second major site of oxidative metabolism
(Tyr1) on CA–DADLE was still present in analogs 2
and 3. In the design of prodrug analogs 2 and 3, we de-
cided not to alter the structure of Tyr1 because of its
critical nature for pharmacological activity.31,37,34,35
However, surprisingly, it was found that cyclic pro-
drug analogs 2 and 3 were significantly more sta-
ble than CA–DADLE when incubated with hCYP3A4.
These results suggest that other CYP450 isozymes
may be responsible for the oxidative metabolism of
cyclic prodrugs analogs 2 and 3, which was observed
when the compounds were incubated with animal and
HLM. The improved metabolic stability of analogs 2
and 3 toward hCYP3A4 may, however, be very im-
portant from an oral absorption perspective because
this CYP450 isozyme is highly expressed in intestinal
mucosal cells and is known to limit the oral bioavail-
ability of some drugs.45–50
Tyr–D-Ala–Gly–Phe–D-Leu–OH (DADLE), and its cyclic pro-
drugs. J Pharmacol Exp Ther 303(2):840–848.
10. Beddell CR, Clark RB, Lowe LA, Wilkinson S, Chang KJ,
Cuatrecasas P, Miller R. 1977. A conformational analysis for
leucine–enkephaline using activity and binding data of syn-
thetic analogues. Br J Pharmacol 61:351–356.
11. Belluzzi JD, Stein L, Dvonch W, Dheer S, Gluckman MI,
McGregor WH. 1978. Enhanced analgesic activity of D-Ala2
enkephalinamides following D-isomer substitutions at posi-
tion five. Life Sci 23(2):99–104.
12. Ouyang H, Tang F, Siahaan TJ, Borchardt RT. 2002. A modi-
fied coumarinic acid-based cyclic prodrug of an opioid peptide:
Its enzymatic and chemical stability and cell permeation char-
acteristics. Pharm Res 19(6):794–801.
13. Tang F, Borchardt RT. 2002. Characterization of the efflux
transporter(s) responsible for restricting intestinal mucosa
permeation of the coumarinic acid-based cyclic prodrug of the
opioid peptide DADLE. Pharm Res 19(6):787–793.
14. Chen W, Yang JZ, Andersen R, Nielsen LH, Borchardt RT.
2002. Evaluation of the permeation characteristics of a model
opioid peptide, H-Tyr–D-Ala–Gly–Phe–D-Leu–OH (DADLE),
and its cyclic prodrugs across the blood–brain barrier using
an in situ perfused rat brain model. J Pharmacol Exp Ther
303:849–857.
15. Ouyang H, Anderson TE, Chen W, Nofsinger R, Steffansen
B, Borchardt RT. 2009. A comparison of the effects of P-
glycoprotein inhibitors on the blood–brain barrier permeation
of cyclic prodrugs of an opioid peptide (DADLE). J Pharm Sci
98:2227–2236.
16. Uchiyama T, Kotani A, Tatsumi H, Kishida T, Okamoto
A, Okada N, Murakami M, Fujita T, Fujiwara Y, Kiso Y,
Muranishi S, Yamamoto A. 2000. Development of novel
lipophilic derivatives of DADLE (leucine enkephalin analog):
Intestinal permeability characteristics of DADLE derivatives
in rats. Pharm Res 17:1461–1467.
17. Ouyang H, Borchardt RT, Siahaan TJ, Vander Velde DG.
2002. Synthesis and conformational analysis of a coumarinic
acid-based cyclic prodrug of an opioid peptide with modified
sensitivity to esterase-catalyzed bioconversion. J Pept Res
59(4):183–195.
ACKNOWLEDGMENTS
The authors would like to thank Dr. Teruna Siahaan
for his helpful discussion and insight into the syn-
thetic strategy. This research was made possible by
the financial support of National Institute of General
Medical Sciences Predoctoral Biotechnology Training
Grant and a research grant from National Institute
on Drug Abuse (DA-03315).
18. Gudmundsson OS, Jois SDS, Velde DGV, Siahaan TJ,
Borchardt RT, Wang B. 1999. The effect of conformation on
the membrane permeation of coumarinic acid and phenylpro-
pionic acid based cyclic prodrugs of opioid peptides. J Pept Res
53:383–392.
REFERENCES
19. Wang B, Nimkar K, Wang W, Zhang H, Shan D, Gudmundsson
O, Gangwar S, Siahaan T, Borchardt RT. 1999. Synthesis
and evaluation of the physicochemical properties of esterase
sensitive cyclic prodrugs of opioid peptides using coumarinic
acid and phenylpropionic acid linkers. J Pept Res 53(4):370–
382.
20. Ouyang H, Chen W, Anderson TE, Steffansen B, Borchardt
RT. 2009. Factors that restrict intestinal cell permeation of
cyclic prodrugs of an opioid peptide (DADLE): Part I. Role
of efflux transporters in the intestinal mucosa. J Pharm Sci
98:337–348.
1. Gentilucci L. 2004. New trends in the development of opioid
peptide analogs as advanced remedies for pain relief. Curr Top
Med Chem 4:19–38.
2. Knipp GT, Vander Velde DG, Siahaan TJ, Borchardt RT. 1997.
The effect of $-turn structure on the passive diffusion of pep-
tides across Caco-2 cell monolayers. Pharm Res 14:1332–1340.
3. Schiller PW. 1993. Development of receptor-selective opioid
peptide analogs as pharmacological tools and as potential
drugs. Handb Exp Pharmacol 681–710.
4. Schiller PW. 2005. Opioid peptide-derived analgesics. AAPS J
7(3):E560–E565.
21. Ouyang H, Chen W, Anderson TE, Steffansen B, Borchardt
RT. 2009. Factors that restrict intestinal cell permeation of
cyclic prodrugs of an opioid peptide (DADLE): Part II. Role
of metabolic enzymes in the intestinal mucosa. J Pharm Sci
98:349–361.
22. Bak A, Gudmundsson OS, Friis GJ, Siahaan TJ, Borchardt RT.
1999. (Acyloxy)alkoxy-based cyclic prodrugs of opioid peptides:
Evaluation of the chemical and enzymatic stability as well
as their transport properties across Caco-2 cell monolayers.
Pharm Res 16:24–29.
5. Witt KA, Davis TP. 2006. CNS drug delivery; opioid peptides
and the blood–brain barrier. AAPS J 8(1):E76–E88.
6. Borchardt RT. 1999. Optimizing oral absorption of peptides
using prodrug strategies. J Control Release 62:231–238.
7. Han C, Liu S, Wang B. 2002. Major factors influencing peptide
drug absorption. Front Biotechnol Pharm 3:270–290.
8. Wang W, Camenisch G, Sane DC, Zhang H, Hugger E, Wheeler
GL, Borchardt RT, Wang B. 2000. A courmarin-based prodrug
strategy to improve the oral absorption of RGD peptidomimet-
ics. J Control Release 65:245–251.
23. Gudmundsson OS, Nimkar K, Gangwar S, Siahaan T,
Borchardt RT. 1999. Phenylpropionic acid-based cyclic
9. Yang JZ, Chen W, Borchardt RT. 2002. in vitro stability and
in vivo pharmacokinetic studies of a model opioid peptide, H-
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES

14
NOFSINGER, FUCHS-KNOTTS, AND BORCHARDT
prodrugs of opioid peptides that exhibit metabolic stability
to peptidases and excellent cellular permeation. Pharm Res
16:16–23.
peptides in rats and guinea pigs following IV administration.
J Pharm Sci 94:2676–2687.
37. Lazarus L, Salvador S, Balvoni G, Tomatis R, Wilson W.
1992. Stereospecificity of amino acid side chains in deltor-
phin defined binding to opioid receptors. J Med Chem 35:1222–
1227.
38. Becker JA, Wallace A, Garzon A, Ingallinella P, Bianchi E,
Cortese R, Simonin F, Kieffer BL, Pessi A. 1999. Ligands for
K-opioid and ORL1 receptors identified from a conformation-
ally constrained peptide combinatorial library. J Biol Chem
274:27513–27522.
39. Weltrowska G, Lemieux C, Chung NN, Schiller PW. 2004. A
chimeric opioid peptide with mixed mu agonist/gamma antag-
onist properties. J Pept Res 63:63–68.
24. Gudmundsson OS, Pauletti GM, Wang W, Shan D, Zhang H,
Wang B, Borchardt RT. 1999. Coumarinic acid-based cyclic
prodrugs of opioid peptides that exhibit metabolic stability
to peptidases and excellent cellular permeability. Pharm Res
16:7–15.
25. Tang F, Borchardt RT. 2002. Characterization of the efflux
transporter(s) responsible for restricting intestinal mucosa
permeation of an acyloxyalkoxy-based cyclic prodrug of the
opioid peptide DADLE. Pharm Res 19(6):780–786.
26. Patel J, Mitra AK. 2001. Strategies to overcome simulta-
neous P-glycoprotein-mediated efflux and CYP3A4-mediated
metabolism of drugs. Pharmacogenomics 2:401–415.
27. Wacher VJ, Silverman JA, Zhang Y, Benet LZ. 1998. Role
of P-glycoprotein and cytochrome P450-3A in limiting oral
absorption of peptides and peptidomimetics. J Pharm Sci
87:1322–1330.
40. Toth G, Ioja E, Tomboly C, Ballet S, Tourwe D, Peter A,
Martinek T, Chung NN, Schiller PW, Benyhe S, Borsodi
A. 2007. Beta-methyl substitution of cyclohexylalanine in
Dmt–Tic–Cha–Phe peptides results in highly potent gamma
opioid antagonist. J Med Chem 50:328–333.
28. Wang Ej, Lew K, Barecki M, Casciano CN, Clement RP, John-
son WW. 2001. Quantitative distinctions of active molecular
recognition by P-glycoprotein and cytochrome P450 3A4. Chem
Res Toxicol 14:1596–1603.
41. Zaych LA, Yang W, Liao Y, Griffin KR, Wang B. 2004. The
effect of substitution patterns on the release rates of opioid
peptides DADLE and [Leu5]–enkephalin from coumarin pro-
drug moieties. Bioorg Chem 32:109–123.
29. Zhang Y, Guo Z, Lin ET, Benet LZ 1998. Overlapping sub-
strate specificties of cytochrome P450-3A and P-glycoprotein
for a novel cysteine protease inhibitor. Drug Metab Dispos
26:360–366.
30. Filippi B, Giusti P, Cima L, Borin G, Ricchelli F, Marchiori
F. 1979. Addicting properties and conformational studies in
solution. Int J Pept Protein Res 14:34–40.
31. Hruby VJ, Bartosz-Bechowski H, Davis D, Slaninova J,
Zalewska T, Stropova D, Porreca F, Yamamura HI. 1997.
Cyclic enkephalin analogues with exceptional potency and se-
lectivity for *-opioid receptors. J Med Chem 40:3957–3962.
32. Lazarus LH, Salvador S, Balvoni G, Tomatis R, Wilson WE.
1992. Stereospecificity of amino acid side chains in deltorphin
defines binding to opioid receptors. J Med Chem 35:1222–1227.
33. McFadyen IJ, Sobczyk-Kojiro K, Schaefer MJ, Ho JC, Omnaas
JR, Mosberg HI, Traynor JR. 2000. Tetrapeptide derivatives of
[D-Pen2, D-Pen5]–enkephalin (DPDPE) lacking an N-terminal
tyrosine residue are agonists at the opioid receptor. J Pharma-
col Exp Ther 295:960–966.
42. Lewis DFV, Ito Y, Lake BG. 2006. Metabolism of coumarin by
human P450’s: A molecular modeling study. Toxicol In Vitro
20:256–264.
43. Mead JA, Smith JN, Williams RT. 1958. Studies in detoxica-
tion.72. The metabolism of coumarin and of o-coumarinc acid.
Biochem J 68:67–74.
44. Mead JA, Smith JN, Williams RT. 1958. Studies in detoxi-
cation.71. The metabolism of hydroxycoumarins. Biochem J
68:61–67.
45. Bailey DG, Malcolm J, Arnold O, Speuce J. 1998. Grape-
fruit juice drug interactions. Br J Clin Pharmacol 46:101–
110.
46. Hebert MF. 1997. Contributions of hepatic and intestinal
metabolism and P-glycoprotein to cyclosporine and tacrolimus
oral drug delivery. Adv Drug Deliv Rev 27:201–214.
47. Kolars JC, Schmiedlin-Ren P, Scheutz JD, Fang C, Watkins
PB. 1992. Identification of rifampin-inducible P450IIIA4
(CYP3A4) in human small bowl enterocytes. J Clin Invest
90:1871–1878.
34. Morgan BA, Smith CFC, Waterfield AA, Huges J, Kosterlitz
HW. 1976. Structure–activity relationships of methion-
ine–enkephalin. J Pharm Pharmacol 28:660–661.
35. Schmidt R, Menard D, Mrestani-Klaus C, Chung NN, Lemieux
C, Schiller PW. 1997. Structural modifications of the N-
terminal tetrapeptide segments of [D-Ala2]-deltrophin I: Ef-
fects on opioid receptor affinities and activities in vitro and on
antinociceptive potency. Peptides 18:1615–1621.
48. Obach SR, Zhang QY, Dunbar D, Kaminsky LS. 2001.
Metabolic characterization of the major human small in-
testinal cytochrome P450s. Drug Metab Dispos 29:347–
352.
49. Watkins PB. 1997. The barrier function of CYP3A4 and
P-glycoprotein in the small bowel. Adv Drug Deliv Rev
27:161–170.
50. Zhou SF. 2008. Drugs that behave as substrates, inhibitors
and inducers of human cytochrome P4503A4. Curr Drug
Metab 9:310–322.
36. Liederer BM, Phan KT, Ouyang H, Borchardt RT. 2005. Signif-
icant differences in the disposition of cyclic prodrugs of opioid
JOURNAL OF PHARMACEUTICAL SCIENCES
DOI 10.1002/jps