Human skin permeation of branched-chain 3-O-alkyl ester and carbonate prodrugs of naltrexone

Article abstract of DOI:10.1007/s11095-005-2592-9

Purpose. Physicochemical characterization and in vitro human skin diffusion studies of branched-chain ester and carbonate prodrugs of naltrexone (NTX) were compared and contrasted with straight-chain ester and carbonate NTX prodrugs. Methods. Human skin p

Full text of DOI:10.1007/s11095-005-2592-9

Pharmaceutical Research, Vol. 22, No. 5, May 2005 (© 2005)
DOI: 10.1007/s11095-005-2592-9
Research Paper
Human Skin Permeation of Branched-Chain 3-O-Alkyl Ester and Carbonate
Prodrugs of Naltrexone
1
,2
1
1
1
1
Haranath K. Vaddi, Mohamed O. Hamad, Jianhong Chen, Stan L. Banks, Peter A. Crooks, and
1
,3
Audra L. Stinchcomb
Received November 18, 2004; accepted February 2, 2005
Purpose. Physicochemical characterization and in vitro human skin diffusion studies of branched-chain
ester and carbonate prodrugs of naltrexone (NTX) were compared and contrasted with straight-chain
ester and carbonate NTX prodrugs.
Methods. Human skin permeation rates, thermal parameters, solubilities in mineral oil and buffer, and
stabilities in buffer and plasma were determined. Partition coefficients between stratum corneum and
vehicle were determined for straight- and branched-chain esters with the same number of carbon atoms.
Results. Branched prodrugs had lower melting points, lower buffer solubilities, and higher mineral oil
solubilities than NTX. The transdermal flux values from all of these branched prodrugs were signifi-
cantly lower than flux values from the straight-chain ester and the methyl carbonate prodrugs. Straight-
chain prodrugs had higher partition coefficient values and higher calculated thermodynamic activities
than their branched-chain counterparts. The prodrug hydrolysis to NTX in buffer and plasma was slower
for prodrugs with increased branching.
Conclusions. Branched-chain prodrugs with bulky moieties had smaller stratum corneum–vehicle par-
tition coefficients and lower thermodynamic activities that resulted in smaller transdermal flux values
than straight-chain prodrugs.
KEY WORDS: bioconversion; branched chains; naltrexone; prodrugs; skin permeation.
INTRODUCTION
centration (8,9). Currently, NTX hydrochloride is available as
a 50-mg oral tablet (ReVia) in the United States. Standard
drug regimens for chronic therapy create noncompliance
problems and require high motivation by NTX patients (8,9).
Therefore, sustained-release formulations could be extremely
useful in maintenance therapy and could result in the im-
provement of the therapeutic effectiveness of NTX. Addi-
tionally, a nonoral delivery route for NTX could reduce oral-
NTX associated gastrointestinal side effects, including nau-
sea, abdominal pain, and vomiting (10). NTX bioavailability
would also be increased via a delivery route that bypasses
first-pass metabolism, as the drug undergoes extensive first-
pass metabolism that results in a low oral bioavailability of
Prodrugs are chemically modified therapeutic agents that
will convert to active drug molecules in biological systems (1).
This prodrug strategy can provide various drug delivery ad-
vantages, including improved therapeutic efficacy, increased
bioavailability, decreased toxicity, decreased gastric irritation,
improved organoleptic properties, and others (2–4). Because
the human body has an abundance of esterase enzymes, pro-
drugs with esterase-susceptible linkages can be cleaved by
these enzymes to release the parent molecule in tissues and
plasma (1). Different prodrug moieties can be attached to the
parent drug through various linkages, including ester, carbon-
ate, and carbamate moieties. These linkages are then cleaved
at a variety of rates in the body, depending on the chemistry
of the molecule (1,2,5). An ideal prodrug should have suffi-
cient chemical stability but readily hydrolyze in a biologic
system to release the parent drug for therapeutic effect.
Naltrexone (NTX), an opioid antagonist, is used for al-
cohol dependence and opioid addiction (6,7). In maintenance
therapy for alcohol abstinence, the NTX dose must be given
regularly to maintain a minimum effective plasma drug con-
5
–40% (11).
Drug delivery across the skin to the systemic circulation
offers a number of advantages, including controlled drug re-
lease rates, avoidance of the first-pass effect, and preclusion
of gastrointestinal side effects. Transdermal delivery could be
very useful in the hepato-compromised alcoholic and opioid
addict patient population, as drug doses may be reduced and
the first-pass effect is circumvented. Although transdermal
delivery of NTX appears to be clinically justified, NTX does
not have suitable physicochemical properties for therapeutic
skin absorption rates (12). Penetration enhancers have been
used to enhance the permeation of NTX (12), but these
chemical enhancers manipulate the skin structure and can
1
Department of Pharmaceutical Sciences, University of Kentucky
College of Pharmacy, Lexington, Kentucky, USA.
Current address: Irma L. Rangel College of Pharmacy, Texas A&M
2
University, Kingsville, Texas, USA.
3
To whom correspondence should be addressed. (e-mail: astin2@ cause significant irritation. Prodrugs, with modified physico-
email.uky.edu)
chemical properties from that of the parent drug, provide
0
724-8741/05/0500-0758/0 © 2005 Springer Science+Business Media, Inc.
758

Human Skin Permeation of Branched Naltrexone Prodrugs
759
higher skin permeation rates and simultaneously convert into (CDCl ): ␦ 0.15 (2H, m), 0.56 (2H, m), 0.88 (1H, m), 1.38 (9H,
3
the parent drug. Skin toxicity profiles for topical prodrugs are s), 1.55–1.70 (2H, m), 1.88 (1H, m), 2.15 (1H, m), 2.30 (1H,
similar to that of the parent drug when prompt biotransfor- m), 2.34–2.50 (3H, m), 2.54–2.76 (2H, m), 2.92–3.16 (2H, m),
mation occurs. NTX prodrugs have previously been synthe- 3.21 (1H, d, J ס 5.7 Hz), 4.67 (1H, s), 5.22 (1H, br s), 6.68 (1H,
sized to increase the oral bioavailability of NTX (3) and to d, J ס 8.1 Hz), 6.92 (1H, d, J ס 8.1 Hz) ppm. MS (LC-MS
+
decrease the bitter taste for buccal delivery (4). Prodrugs of electrospray), M ס 426 m/z.
NTX can be synthesized by linking the 3-phenolic hydroxyl
group of the molecule to various moieties. Previously, 3-O-Isovaleryl Naltrexone
straight-chain ester and carbonate prodrugs of NTX were syn-
Naltrexone (708 mg, 2.08 mmol) was added to isovaleryl
thesized and evaluated for human skin permeation, and cor-
chloride (500 mg, 4.15 mmol) and triethylamine (425 mg, 4.2
relations were established for these NTX permeation rates
mmol) in dichloromethane (100 ml). After stirring at ambient
with the extent of prodrug bioconversion and the physico-
temperature overnight, the reaction mixture was diluted with
chemical parameters (13,14). Short-chain ester and carbonate
chloroform (100 ml), washed with brine, dried over potassium
prodrugs of NTX exhibited higher skin permeation. In the
carbonate, filtered, and the solvent removed on a rotary
current study, branched-chain ester and carbonate NTX pro-
evaporator to afford the crude product as a brown oil. Puri-
drugs, pivalyl, isovaleryl, 2Ј-ethylbutyryl, and isobutyryl es-
fication by silica gel column chromatography (chloroform/
ters, and isopropyl and tertiarybutyl carbonates were synthe-
methanol, 10/1) afforded 707 mg (yield 80%) of the title com-
sized, and the in vitro human skin permeation was compared
pound as a white solid. An analytical sample was obtained
with that of the straight-chain prodrugs. Prodrug physico-
1
by recrystallization from hexanes/ethyl alcohol. H NMR
chemical properties and plasma half-lives were determined
(
CDCl ) ␦ 0.10–0.20 (2H, m), 0.50–0.60 (2H, m), 0.85–0.92
3
for potential permeation prediction relationships and as cru-
cial parameters in future prodrug optimization for clinical
application.
(
1H, m), 1.05–1.15(6H, m, two CH ), 1.50–1.65 (2H, m), 1.80–
3
1
2
.95 (1H, m), 2.08–2.18 (1H, m), 2.20–2.33 (2H, m, CH CO ),
2 2
.35–2.42 (1H, m, (CH ) CH), 2.45–2.49 (4H, m), 2.55–2.75
3
2
(
6
4
2H, m), 2.95–3.25 (3H, m), 4.68 (1H, s), 5.22 (1H, brs, OH),
MATERIALS AND METHODS
Materials
.65 (1H, m), 6.82 (1H, m). HRMS: calculated for C H NO
25 31 5
25.2203, found 425.2208.
3
-O-(2Ј-Ethylbutyryl) Naltrexone
NTX base was purchased from Mallinckrodt Inc. (St.
Louis, MO, USA). All prodrugs were synthesized from free
NTX base. Straight-chain ester prodrugs, 3-O-butyryl NTX,
A stirred mixture of naltrexone (2.00 g, 5.88 mmol) and
TEA (1.90 ml, 13.3 mmol) in 20.0 ml of dry methylene chlo-
ride was cooled to 0°C in an ice-bath, and 2-ethylbutyryl chlo-
ride (0.90 ml of 97%, 6.37 mmol) was added. The reaction
mixture was stirred at room temperature for a further 15 h.
The resulting reaction mixture was diluted in methylene chlo-
ride (50 ml) and washed with 10% aqueous sodium carbonate
3
-O-valeryl NTX, 3-O-hexanoyl NTX, and 3-O-propyloxy-
carbonyl NTX, were synthesized by previously published
methods (13,14). Hanks’ balanced salts modified powder, so-
dium bicarbonate, and light mineral oil were purchased from
Sigma Chemical (St. Louis, MO, USA). 4-(2-Hydroxyethyl)-
1
-piperazineethanesulfonic acid (HEPES), gentamicin sul-
(2 × 30 ml), then water (30 ml). Then the organic solution was
fate, trifluroacetic acid (TFA), triethylamine (TEA), metha-
nol, and acetonitrile (ACN) were obtained from Fisher Sci-
entific (Fairlawn, NJ, USA). 1-Octane sulfonic acid sodium
salt was obtained from ChromTech (Apple Valley, MN,
USA).
separated and dried over anhydrous potassium carbonate, fil-
tered, and reduced to a small volume under reduced pressure.
Petroleum ether was added, and the solution stored in the
refrigerator (4°C) overnight. The resulting solid that crystal-
lized out was filtered at the pump, recrystallized from petro-
1
leum ether, and air-dried to afford a white powder. H NMR
Synthetic Procedure for Branched-Chain
Naltrexone Prodrugs
(CDCl ): ␦ 0.15 (2H, m), 0.57 (2H, m), 0.87 (1H, m), 0.9–1.10
3
(6H, m), 1.50–195 (7H, m), 2.15 (m, 1H), 2.28 (1H, m), 2.34–
2
.55 (4H, m), 2.56–2.76 (2H, m), 2.92–3.14 (2H, m), 3.21 (1H,
d, J ס 5.4 Hz), 4.67 (1H, s), 5.20 (1H, br s), 6.66 (1H, d, J ס
.4 Hz), 6.82 (1H, d, J ס 8.4 Hz) ppm. MS (LC-MS electro-
3
-O-Pivalyl Naltrexone
8
A mixture of naltrexone (1.80 g, 5.28 mmol) and TEA
+
spray), M ס 440 m/z.
(1.90 ml, 13.3 mmol) in 20.0 ml of dry methylene chloride was
cooled to 0°C in an ice-bath. To this stirred mixture was
added 0.7ml of the pivaloyl chloride (0.7 ml of 99%, 5.63
mmol). The resulting reaction mixture was stirred at room
3
-O-Isobutyryl Naltrexone
Naltrexone (533 mg, 1.56 mmol) was added to isobutyryl
temperature for 15 h and was diluted in methylene chloride chloride (332 mg, 3.12 mmol) and triethylamine (316 mg, 3.12
50 ml) and washed with 10% aqueous sodium carbonate (2 × mmol) in dichloromethane (100 ml). After stirring at ambient
0 ml) then water (30 ml). The organic solution was sepa- temperature overnight, the reaction mixture was diluted with
(
3
rated, dried over anhydrous potassium carbonate, filtered, chloroform (100 ml), washed with brine, dried over potassium
and reduced to a small volume under reduced pressure. Pe- carbonate, filtered, and the solvent removed on a rotary
troleum ether was added, and the solution stored in the re- evaporator to afford the crude product as a brown oil. Puri-
frigerator (4°C) overnight. The resulting solid that crystal- fication by silica gel column chromatography (chloroform/
lized out was filtered at the pump, washed with chilled pen- methanol, 10/1) afforded 545 mg (yield 85%) of the title com-
1
tane, and air-dried to afford a white powder. H NMR pound as a white solid. An analytical sample was obtained by

7
60
Vaddi et al.
1
recrystallization from hexanes/ethyl alcohol. H NMR × 3.2 mm) was used with the UV detector set at a wavelength
CDCl ) ␦ 0.08–0.18 (2H, m), 0.48–0.58 (2H, m), 0.8–0.92 (1H, of 215 nm. The mobile phase was 70:30 ACN:0.1% TFA with
(
3
m), 1.15–1.40(7H, m, (CH ) CH), 1.50–1.70 (2H, m), 1.80– 0.065% 1-octane sulfonic acid, sodium salt adjusted to pH 3.0
3
2
1
(
5
.95 (1H, m), 2.05–2.20 (1H, m), 2.25–2.45 (4H, m), 2.55–2.75 with triethylamine. The flow rate of the mobile phase was 1.5
2H, m), 2.80–2.90 (1H, m), 2.95–3.25 (2H, m), 4.70 (1H, s), ml/min and 100 ␮l of sample was injected onto the column.
.20 (1H, brs, OH), 6.65 (1H, m), 6.80 (1H, m). HRMS: cal- Diffusion samples were analyzed with a set of standard
culated for C H NO 411.2040, found 411.2041.
samples and exhibited excellent linearity over the entire con-
centration range used in the assays.
2
4
29
5
3
-O-Isopropyloxycarbonyl Naltrexone
The drugs were extracted from the buffer samples by
solid phase extraction (30 mg 1cc Oasis HLB, Waters Corp.).
Before extracting the aqueous drug samples (5 ml), the car-
tridge was conditioned with 1 ml of methanol and 1 ml of
nanopure water. After sample loading, the cartridge was
washed with 1 ml of 5% methanol and the drug was eluted
with ACN and analyzed by HPLC.
A stirred mixture of naltrexone (300 mg, 0.88 mmol) and
TEA (500 ␮l, 3.5 mmol) in 4.0 ml of dry methylene chloride
was cooled to 0°C in an ice-bath, and 1.0 ml of a 1 M chlo-
roformate in toluene (1.0 mmol) was added. The resulting
reaction mixture was stirred at room temperature for 3 h and
was diluted in methylene chloride (30 ml) and washed with
1
0% aqueous sodium carbonate (2 × 30 ml) then water (30
Differential Scanning Calorimetry of NTX and Prodrugs
ml). The organic solution was separated and dried over an-
hydrous potassium carbonate, filtered, and reduced to a small
volume under reduced pressure. The desired product was pre-
cipitated by trituration with excess pentane. The resulting
solid was filtered at the pump, washed with pentane, and
air-dried to afford a white powder. H NMR (CDCl ): ␦ 0.16
Differential scanning calorimetry (DSC) was carried out
for NTX and its prodrugs. Heats of fusion and melting points
were determined with a TA Instruments 2920 DSC (New
Castle, DE, USA). An accurately weighed sample of drug
(2–5 mg) was sealed into the aluminum pans and thermo-
grams were recorded at 10°C/min from ambient to 350°C.
Measurements were repeated once for a total of two scans on
each drug.
1
3
(2H, m), 0.57 (2H, m), 0.88 (1H, m), 1.38 (6H, d, J ס 6.3 Hz),
1.55–1.70 (2H, m), 1.89 (1H, m), 2.14 (1H, m), 2.33 (1H, m),
2.38–2.50 (3H, m), 2.56–2.76 (2H, m), 2.95–3.15 (2H, m),
3.21(1H, d, J ס 6), 4.72 (1H, s), 4.97 1H, (seven line multiplet,
Solubility
J ס 6.3 Hz), 5.22 (1H, br s), 6.68 (1H, d, J ס 8.1Hz), 6.92 (1H,
d, J ס 8.1 Hz) ppm. MS (LC-MS electrospray), M ס 428
m/z.
+
The solubilities of NTX and its prodrugs in free base
form were determined by adding an excess quantity of drug to
mineral oil or pH 7.4 HEPES-buffered Hanks’ balanced salts
solution at 32°C. The drug slurries were equilibrated with
shaking for 8 h. The less chemically stable isobutyryl ester
buffer solubility was determined after equilibrating excess
solid in pre-warmed Hanks’ buffer at 32°C for 10 min. The
solubilities of NTX and 3-O-isopropyloxycarbonyl NTX were
also determined under these instantaneous conditions, and it
was concluded that the instantaneous solubility values were
the same as that of the 8-h equilibration solubilities. After
equilibration, samples were withdrawn in a prewarmed glass
syringe, filtered through a syringe filter discarding the first
3
-O-Tertiarybutyloxycarbonyl Naltrexone
Triphosgene (4.02 g, 13.42 mmol) was added to t-butanol
(500 mg, 6.71 mmol) and triethylamine (5.43 g, 53.68 mmol) in
dichloromethane (100 ml). The reaction mixture was stirred
while cooling with dry ice in acetone. After 2 h, naltrexone
(1.15 g, 3.36 mmol) was added, and the stirred reaction mix-
ture was allowed to warm to ambient temperature overnight.
The reaction mixture was diluted by chloroform (150 ml), the
organic phase was washed with brine, dried over potassium
carbonate, filtered, and the solvent removed on a rotary
evaporator to afford the crude product as a brown oil. Puri-
fication by silica gel column chromatography (chloroform/
methanol, 10/1) gave 770 mg (yield 52%) of the title com-
pound as a white solid. H NMR (CDCl ) ␦ 0.10–0.18 (2H, m),
0
3
0% of the initial filtrate (Mineral oil: Millex FG-13 filter,
Millipore and buffer: nylon filter, Gelman), and measured
with respect to volume and extracted with ACN (mineral oil
samples) or diluted with Hanks’ buffer. The drug from the
buffer samples was immediately extracted by solid phase ex-
traction into ACN. All the samples were analyzed by HPLC.
The sampling procedure was done in triplicate.
1
3
.50–0.60 (2H, m), 0.8–0.95 (1H, m), 1.15–1.35 (6H, m, two
CH ), 1.55–1.70 (2H, m), 1.82–1.92 (1H, m), 2.10–2.20 (1H,
m), 2.25–2.45 (4H, m), 2.55–2.72 (2H, m), 2.95–3.15 (2H, m),
3
3
.15–3.22 (1H, m), 3.30–3.52 (3H, m, CH ), 4.68 (1H, s), 5.15
3
In Vitro Skin Diffusion Studies
(1H, brs, OH), 6.62 (1H, m), 6.82 (1H, m). HRMS: calculated
Human skin harvested during abdominal reduction sur-
gery was used for the diffusion studies. Skin sections were
obtained by using a Padgett dermatome set to 250 ␮m; these
sections were stored at −20°C. Stored skin samples were
for C H NO 441.2151, found 441.2316.
2
5
31
6
Quantitative Analysis
A modified high-pressure liquid chromatography thawed to room temperature at the time of the experiment. A
(HPLC) assay from Hussain et al. (3) was used for the analysis PermeGear flow-through (In-Line, Riegelsville, PA, USA)
of NTX and its prodrugs. The HPLC system consisted of a diffusion cell system was used for the skin permeation studies.
Waters 717 plus autosampler, a Waters 1525 Binary HPLC Diffusion cells were kept at 32°C with a circulating water
pump and a Waters 2487 Dual ␭ Absorbance detector with bath. Data was collected by using human skin from a single
Waters Breeze software (Waters Corp., Milford, MA, USA). donor with three cells of NTX and four cells of the prodrug.
A Brownlee C-18 reversed-phase Spheri-5 ␮m column (220 × Because of normal human skin intersubject permeation vari-
4
.6 mm) with a C-18 reversed phase 7 ␮m guard column (15 ability, the prodrug permeation was compared against NTX

Human Skin Permeation of Branched Naltrexone Prodrugs
761
permeation within each individual skin sample. The receiver Stratum Corneum/Vehicle Partition Coefficient Studies
solution was HEPES-buffered Hanks’ balanced salts with
Human epidermis with the stratum corneum (SC) side
facing up was incubated on filter paper soaked with 0.1%
trypsin in 0.5% sodium bicarbonate solution at 37°C for 3 h
gentamicin at pH 7.4 set to a flow rate of 1.1 ml/h. An excess
quantity (above the saturation solubility) of NTX or prodrug
was added to light mineral oil, sonicated for 10 min, and then
applied to the skin. An excess quantity of the drug was kept
in the donor compartment throughout the diffusion experi-
ment to ensure saturation of the drug in mineral oil in order
to maintain a maximum and constant thermodynamic activity
of the drug. Each cell was charged with 0.25 ml of the respec-
tive drug solution. Samples were collected in six-hour incre-
ments for 48 h. All the samples were stored at 4°C until
processed by solid phase extraction. The human tissue use
was approved by the University of Kentucky Institutional Re-
view Board.
Cumulative quantity of drug collected in the receiver
compartment was plotted as a function of time. The flux value
for a given experiment was obtained from the slope of the
steady state portion of the cumulative amount of drug per-
meated vs. time plot. Apparent permeability coefficient val-
ues were computed from Fick’s First Law of diffusion:
(15). The SC membrane was separated and dried in a vacuum
desiccator. After 24 h, the SC was dipped in acetone for 20 s
to remove sebaceous lipids and dried again.
Approximately 5 mg of SC was equilibrated with sub-
saturated drug solutions in 0.5 g of mineral oil at 32°C for 48
h. An aliquot of the mineral oil solution (10 ␮l) was with-
drawn at the end of the study and was diluted to 1000 ␮l with
acetonitrile. The samples were then analyzed by HPLC. The
amount of the drug partitioned into the SC was measured by
subtracting the amount present in the mineral oil after equili-
bration from the initial drug concentration in the mineral oil.
The partition coefficient value was expressed as the concen-
tration of prodrug in 1 g of SC to the concentration of prodrug
in 1 g of mineral oil.
RESULTS AND DISCUSSION
1
dM
dt
The branched prodrugs were designed, synthesized, and
tested in order to determine physicochemical and permeation
differences from straight-chain prodrugs, differences from
two prodrug linkages (ester vs. carbonate ester), and differ-
ences among this series with varying branching in the prodrug
moiety. All prodrugs studied here, except 2Ј-ethylbutyryl, had
single carbons comprising the branched chains. The pivalyl
ester and tertiary-butyl carbonate ester (3-O-tertiarybutyl-
oxycarbonyl NTX) are structurally identical prodrug moi-
eties, except for the difference in the type of prodrug linkage
ͩ
ͪ
= J
S
= K
p
⌬C
(1)
A
In Eq. (1), J is the steady-state flux, M is the cumulative
amount of drug permeating the skin, A is the area of the skin
0.95 cm ), K is the effective permeability coefficient in cm/h,
and ⌬C is the difference in concentrations of NTX or prodrug
between the donor and receiver compartments. Sink condi-
tions were maintained in the receiver throughout the experi-
ment, so ⌬C was approximated by the drug concentration
s
2
(
p
(
solubility in this case) in the donor compartment.
(
Fig. 1). Likewise, the isobutyryl ester and the isopropyl
carbonate (3-O-isopropyloxycarbonyl NTX) differed only by
the type of prodrug linkage to NTX (Fig. 1). Therefore, these
two sets of prodrugs provided the opportunity for comparison
of just the prodrug linkage effect on physicochemical proper-
ties and transdermal flux. The melting points of all the
prodrugs were lower than that of NTX, suggesting that the
addition of the prodrug moieties to NTX significantly de-
creased the intermolecular cohesion of the resulting prodrug
Chemical Stability of Prodrugs in Buffer
The chemical stability of the prodrugs in Hanks’ buffer
(pH 7.4) was studied by using subsaturated concentrations of
the prodrug. Samples (6 ml) were distributed into sealed vials
and placed at 32°C. One vial was removed immediately after
starting the experiment and at predetermined intervals for 14
days and stored at −70°C until analysis. Drug was extracted
from the buffer using solid phase extraction and the samples
were analyzed by HPLC. Percent drug remaining (on a loga-
rithmic scale) was plotted against time to calculate pseudo-
first order rate constants and half-lives of the prodrugs.
Plasma Hydrolysis of Prodrugs
Human plasma, which was stored at −20°C, was thawed
to room temperature. Known and subsaturated amounts of
the prodrug were added to the human plasma, vortexed, and
incubated at 37°C. Samples were withdrawn immediately af-
ter starting the experiment and at regular intervals for 48 h. A
5
-fold volume of ACN was added to each sample to precipi-
tate protein. Each sample was then vortexed and centrifuged
at 10,000 × g for 15 min. Supernatant was separated and evap-
orated under nitrogen at room temperature and the residue
was reconstituted with ACN for HPLC analysis. A standard
curve was prepared from the drug-spiked plasma samples
processed by the same method. Percent drug remaining (on a
logarithmic scale) was plotted against time, to calculate Fig. 1. Chemical structures, molecular weights, and melting points of
pseudo-first order rate constants and half-lives of the prodrugs.
NTX and its prodrugs.

7
62
Vaddi et al.
Table I. Properties of NTX and its Branched Prodrugs
Light mineral
oil solubility
(mM)
Hanks buffer
solubility
(mM)
Mean permeability
coefficient (Kp × 10 ),
Plasma
half-life
(h)
Half life in
Hanks buffer
(h)
4
Flux from mineral
−2
−1
Drug
oil (nmol cm
h
)
cm/h
NTX
PIV-Ester
Isoval-Ester
Etbut-ester
Isobutyryl-ester
Isoprop-carbonate
Tert-Butyl-carbonate
0.24 ± 0.02
3.22 ± 0.01
1.31 ± 0.07
5.60 ± 0.25
1.17 ± 0.02
5.70 ± 0.60
0.12 ± 0.01
0.14 ± 0.01
0.27 ± 0.02
0.32 ± 0.00
3.06 ± 1.58
1.28 ± 0.49
0.85 ± 0.28
1.97 ± 0.56
2.24 ± 0.80
1.20 ± 0.35
3.73 ± 0.67
127.5
4.0
6.5
3.5
19.1
—
9.37
0.54
44.70
0.41
0.65
ST
—
343
305
1136
54
a
a
a
a
0.87 ± 0.10
1.57 ± 0.04
0.26 ± 0.02
1.64 ± 0.00
14.0
23.8
315
ST
Flux values are mean of three measurements for NTX and four values for prodrugs with standard deviations.
ST, stable, no significant degradation observed over duration of experiment.
a
From Ref. 14.
compared to NTX (Fig. 1). The 2Ј-ethylbutyryl prodrug ex- cantly higher than the NTX control flux (Table I). The 3-O-
hibited the lowest melting point in the series, which corre- tertiarybutyloxycarbonyl NTX prodrug provided an NTX flux
sponded with it having the prodrug moiety with the most most comparable to that of the NTX control (Fig. 4). This
carbons in the series. The lower melting points of the pro- lack of a NTX equivalent flux increase from that of the
drugs are consistent with those observed for straight chain branched prodrugs did not correlate with the straight-chain
ester and carbonate prodrugs of NTX (13,14). Pivalyl and ester and methyl carbonate prodrugs, which exhibited signifi-
isobutyryl esters exhibited higher melting points than their cantly higher fluxes than NTX (13,14). The stratum corneum
corresponding carbonate counterparts, indicating that car- (SC), the uppermost layer of the skin, is the major barrier to
bonate prodrugs have lower intermolecular cohesion than the drug permeation for most compounds. Lipids in the SC are
ester prodrugs. All the prodrugs had higher solubility than arranged in bilayers, and lipophilic drugs are presumed to
NTX in mineral oil, with the 2Ј-ethylbutyryl ester having the permeate through these lipid bilayers. To investigate the
highest mineral oil solubility, which corresponded with its low mechanism of the lower permeabilities of the branched pro-
melting point (Table I). As expected, the solubilities of the drugs in the SC, partition coefficient experiments between
prodrugs were lower than NTX in buffer. The phenolic group human SC and the vehicle (mineral oil) were completed for
of the NTX molecule is important for aqueous solubility some isomeric straight- and branched-chain ester prodrugs.
through hydrogen bonding, and esterification at this group The n-hexyl ester prodrug had a higher partition coefficient
obviously decreases aqueous solubility.
value than the branched-chain 2Ј-ethylbutyryl ester prodrug,
Figure 2 shows a representative permeation profile used with partition coefficient values of 23.14 ± 4.50 and 9.18 ±
to determine the steady-state flux of the prodrugs and NTX 1.63, respectively. Similarly, the valeryl ester had a higher
bioconverted from the prodrugs. The NTX flux from the con- partition coefficient than the pivalyl ester prodrug, with par-
trol treatments and the prodrug treatments in each individual tition coefficient values of 60.56 ± 4.95 and 19.12 ± 3.57, re-
skin sample are shown in Fig. 3. The NTX equivalent flux spectively. These significantly increased straight-chain pro-
obtained from each of the branched prodrugs was not signifi- drug partition coefficients explain the increased human skin
Fig. 2. Representative permeation profiles from saturated solutions for the diffusion
from the NTX control, NTX from isobutyryl ester, and intact isobutyryl ester through
human skin at 32°C. Data represent mean ± SD (n ס 3 for NTX treatment and
n ס 4 for prodrug treatment).

Human Skin Permeation of Branched Naltrexone Prodrugs
763
Fig. 3. In vitro human skin NTX flux from NTX and NTX equivalent (NTX + prodrug)
flux from each prodrug. Data is represented as mean ± SD (n ס 3 for NTX treatment
and n ס 4 for prodrug treatment) using human skin from a single subject.
permeabilities, as compared to branched-chain prodrugs. The also had the lowest activity coefficient. Similarly, the n-hexyl
higher partition and permeability coefficients of the straight- and n-butyl esters had higher thermodynamic activities than
chain prodrugs in SC also correspond with higher calculated their branched counterparts, 2Ј-ethylbutyryl and isobutyryl,
thermodynamic activity coefficients obtained from the DSC respectively. Additionally, the carbonate esters behaved in an
thermogram parameters of melting and heats of fusion (Table analogous fashion, with the 3-O-propyloxycarbonyl NTX hav-
II). Thermodynamic activity of the drug is an important phys- ing a higher thermodynamic activity than the 3-O-isopropyl-
icochemical parameter used in predicting solubility in oil, oxycarbonyl NTX. Pillai et al. (14) observed the highest pro-
which in turn can help predict permeation through the SC drug skin permeability with the methyl carbonate prodrug in
lipids. Stinchcomb et al. (16) demonstrated a direct correla- a series of straight-chain carbonate prodrugs, and this pro-
tion between the thermodynamic activities and hexane solu- drug also had the highest thermodynamic activity value, cal-
bilities of buprenorphine prodrugs. It is possible that higher culated from DSC thermal parameters.
thermodynamic activities result in higher partition coefficient
Although all the branched chain prodrugs had a signifi-
values for drugs in the SC. As can be seen in Table II, the cantly higher solubility than NTX in oil, similar in magnitude
valeryl ester had a higher thermodynamic activity than two to the oil solubilities of the successful straight-chain prodrugs,
isomeric branched alkyl esters, that is, pivalyl and isovaleryl. this did not result in a higher flux. Both aqueous and oil
The bulkier of these two branched prodrugs, the pivalyl ester, solubilities have been identified as important parameters in
Fig. 4. Mean ratio of NTX equivalent flux from prodrug to NTX control flux ± SD (n
ס skin from 2–4 subjects, 4 cells each).

7
64
Vaddi et al.
Table II. Comparison of Thermal Parameters for Straight Chain and Branched Chain Prodrugs with the Same Number of Carbon Atoms
Esters
Hexyl
Carbonates
Thermal parameter
Valeryl
Pivalyl
Isovaleryl
2Ј-Ethylbutyryl
Butyl
Isobutyryl
Propyl
Isopropyl
Heats of fusion (kJ/mol)
Activity coefficient
19.62
0.224
16.84
0.116
14.01
0.202
10.64
0.533
31.67
0.137
20.12
0.151
26.81
0.033
18.94
0.196
24.79
0.049
Calculated from lna ס (−⌬H /RT) (Tf−T/T ); where a is the thermodynamic activity, R is the gas constant, T is room temperature, T is fusion
2
f
f
2
f
temperature, and ⌬H is the heat of fusion (From J. H. Hildebrand and R. L. Scott. The Solubility of Nonelectrolytes. Reinhold, New York,
f
1950, Chap. 12).
the design of prodrugs (18,19). Therefore, a modified Potts skin-stable prodrugs such as tertiary-butyl carbonate esters
and Guy equation that incorporates both aqueous and oil are extremely dependent on their aqueous solubility for trans-
solubility terms has been used to predict the flux of prodrugs port across viable tissue into the capillaries. The low molecu-
from lipophilic vehicles (18). Analysis of the fit of data from lar weight dependence of this data set is most likely due to the
branched ester and carbonate prodrugs of NTX to the fol- narrow molecular weight range of this group of compounds,
2
lowing equation provided an r of 0.92 (MicroMath Scientist). as it is well known that molecular weight influences drug flux
across the skin.
log Jmax = x + y log Soil + ͑1 − y͒log Saq − zMW
(2)
Skin contains esterase activity and the enzymes involved
Soil is the mineral oil solubility, Saq is the Hanks buffer solu- have been reported to be resistant to the stresses of freezing
bility, and MW is the molecular weight of the prodrug. The and storage (19). During skin permeation, it is expected that
parameter estimates were x ס 0.51 (±0.98), y ס 0.46 (±0.09), ester prodrugs are converted to the parent drug by these en-
2
and z ס 0.0003 (±0.0024). The r value indicates a reasonably zymes. In these diffusion experiments, all branched chain al-
good correlation, especially considering the small number of kyl ester prodrugs were bio-transformed to NTX at various
data points (n ס 6) used for the fit. The flux dependency on rates (Fig. 5). The highly branched alkyl ester prodrugs, that
aqueous solubility is considerably higher than the fit for the is, the pivalyl ester and 3-O-tertiarybutyloxycarbonyl NTX,
straight-chain ester compounds (13), but also somewhat less and the moderately branched alkyl ester prodrug, 2Ј-
than the aqueous solubility dependency observed in the ethylbutyryl ester, hydrolyzed to NTX to a lesser extent than
straight-chain carbonate series (14). As described by Pillai et the other prodrugs in these series. The half-lives of the pro-
al. (14), the difference between this group of compounds and drugs in Hanks’ buffer indicated that prodrugs containing a
the straight-chain carbonate series vs. the successful straight- higher degree of branching are generally more stable than the
chain ester compounds was the degree of prodrug bioconver- other prodrugs in this series (Table I); specifically, the greater
sion to NTX that occurred in the skin. In the straight-chain the steric hindrance in the vicinity of the carbonyl linkage the
3
-O-alkyl ester series, the prodrugs almost completely con- greater the stability to hydrolysis by water or esterase en-
verted to NTX, and only trace amounts of prodrug were zymes (20,21). Often, prodrugs may not convert completely to
found in the receiver compartment. Here, the branched alkyl the parent drug during skin permeation, and some intact pro-
esters hydrolyzed to afford NTX in the skin, over a wide drug can appear in the systemic circulation. The plasma hy-
percent conversion range of 3–94% (Fig. 5), and a more sig- drolysis rates should indicate the relative prodrug conversion
nificant dependence on aqueous solubility was observed in rate to the parent drug in the body. Plasma half-lives of the
the Roberts and Sloan equation (2). Therefore, prodrugs that prodrugs were much shorter than Hanks’ buffer half-lives,
undergo rapid bioconversion to NTX should meet less resis- demonstrating that the prodrugs were enzymatically hydro-
tance when traversing the viable aqueous tissue of the skin, lyzed by the plasma esterase enzymes. Plasma half-lives of
and be less dependent on the corresponding aqueous solubil- isovaleryl ester, isobutyryl ester, and 3-O-isopropyloxy-
ity that contributes to the rate of that process. However, very carbonyl NTX prodrugs were less than 1 h, and were also
Fig. 5. Percent NTX converted from prodrug during skin permeation ± SD (n ס skin
from 2–4 subjects, 4 cells each).

Human Skin Permeation of Branched Naltrexone Prodrugs
765
more chemically unstable. Prodrugs containing highly and
moderately branched groups, that is, pivalyl ester, 2Ј-ethyl-
butyryl ester, and 3-O-tertiarybutyloxycarbonyl NTX had
longer plasma half-lives, which again demonstrated that steric
hindrance in the vicinity of the carbonyl linkage decreased the
hydrolysis rates (20,21).
In conclusion, the shape (branching) of the side chain
used in the design of transdermal prodrugs is important to the
success of the compound. These branched-chain prodrugs
contained bulky prodrug moieties, which caused decreased
6. J. R. Volpicelli, A. I. Alterman, and M. Hayashida. Naltrexone in
the treatment of alcohol dependence. Arch. Gen. Psychiatry 49:
8
76–880 (1992).
7
. L. Terenius. Rational treatment of addiction. Curr. Opin. Chem.
Biol. 2:541–547 (1998).
8. J. L. Rothenberg, M. A. Sullivan, S. H. Church, A. Seracini, E.
Collins, H. D. Kleber, and E. V. Nunes. Behavioral naltrexone
therapy: an integrated treatment for opiate dependence. J. Subst.
Abuse Treat. 23:351–360 (2002).
9
. J. R. Volpicelli, K. C. Rhines, and J. S. Rhines. Naltrexone and
alcohol dependence. Role of subject compliance. Arch. Gen. Psy-
chiatry 54:737–743 (1997).
thermodynamic activities, decreased SC/vehicle partition co- 10. H. R. Kranzler, V. M. Lowe, and J. V Kirk. Naltrexone vs nefaz-
odone for treatment of alcohol dependence. A placebo-
efficients, decreased skin permeability coefficients, and re-
sulted in lower transdermal flux rates. Additionally, the pro-
drug skin bioconversion rate is intimately involved in the pre-
controlled trial. Neuropyschopharmacol 22:493–503 (2000).
1. P. D. R. Generics. Medical Economics, 2nd ed., Medical Eco-
1
nomics, Montvale, New Jersey, 1996, pp. 2229–2233.
diction of drug transport rates. A higher degree of branching 12. Y. L. Chen. L. L Chun, and D. J. Enscore. Transdermal thera-
peutic systems for the administration of naloxone, naltrexone,
or bulkier branched groups, although having enhanced
and nalbuphine. Alza U.S. Patent No. 4 573 995. March 4, 1986.
chemical stability, did not convert to the parent molecule
1
3. A. L. Stinchcomb, P. W. Swaan, O. Ekabo, K. E. Harris, J.
Browe, D. C. Hammell, T. A. Cooperman, and M. Pearsall.
Straight-chain naltrexone ester prodrugs: diffusion and concur-
rent biotransformation in human skin. J. Pharm. Sci. 91:2571–
completely during the skin permeation process and did not
enhance the transdermal flux values as previously seen with
the corresponding straight-chain 3-O-alkyl ester prodrugs.
2
578 (2002).
1
4. O. Pillai, M. O. Hamad, P. A. Crooks, and A. L. Stinchcomb.
Physicochemical evaluation, in vitro human skin diffusion, and
concurrent biotransformation of 3-O-alkyl carbonate prodrugs of
naltrexone. Pharm. Res. 21:1146–1152 (2004).
5. A. M. Kilgman and E. Christophers. Preparation of isolated
sheets of human stratum corneum. Arch. Dermatol. 88:702–705
(1963).
ACKNOWLEDGMENTS
The authors would like to thank the National Cancer
Institute Cooperative Human Tissue Network (CHTN) for
supplying the skin. This project was funded by NIH grant
R01DA13425.
1
1
6. A. L. Stinchcomb, R. Dua, A. Paliwal, R. W. Woodard, and G. L.
Flynn. A solubility and related physicochemical property com-
parison of buprenorphine and its 3-alkyl esters. Pharm. Res. 10:
REFERENCES
1
526–1529 (1995).
1
. A. Vigroux, M. Bergon, and C. Zedde. Cyclization-activated pro-
drugs: N-(substituted 2-hydroxyphenyl and 2-hydroxypropyl)car-
bamates based on ring-opened derivatives of active benzoxazo-
lones and oxazolidinones as mutual prodrugs of acetaminophen.
J. Med. Chem. 38:3983–3994 (1995).
1
7. W. J. Roberts and K. B. Sloan. Prediction of transdermal flux of
prodrugs of 5-fluorouracil, theophylline, and 6-mercaptopurine
with a series/parallel model. J. Pharm. Sci. 89:1415–1431 (2000).
18. W. J. Roberts and K. B. Sloan. Correlation of aqueous and lipid
solubilities with flux of prodrugs of 5-fluorouracil, theophylline
and 6-mercaptopurine: A Potts-Guy approach. J. Pharm. Sci. 88:
515–522 (1999).
2. P. Vlieghe, F. Bihel, T. Clerc, C. Pannecouque, M. Witvrouw,
E. D. Clercq, J. P. Salles, J. C. Chermann, and J. L. Kraus. New
3
Ј-azido-3Ј-deoxythymidin-5Ј-yl O-(␻-hydroxyalkyl) carbonate
prodrugs: synthesis and anti-HIV evaluation. J. Med. Chem. 44: 19. P. G. Hewitt, J. Perkins, and S. A. M. Hotchkiss. Metabolism of
7
77–786 (2001).
fluroxypyr, fluroxypyr methyl ester, and the herbicide fluroxypyr
methylheptyl ester. I: during percutaneous absorption through
fresh rat and human skin in vitro. Drug Metab. Dispos. 28:748–
754 (2000).
3
4
5
. M. A. Hussain, C. A. Koval, M. J. Myers, E. G. Shami, and E.
Shefter. Improvement of the oral bioavailability of naltrexone in
dogs. J. Pharm. Sci. 76:356–358 (1987).
. M. A. Hussain, B. J. Aungst, C. A. Koval, and E. Shefter. Im- 20. E. Carvalho, A. P. Francisco, J. Iley, and E. Rosa. Triazene drug
proved buccal delivery of opioid analgesics and antagonists with
bitterless prodrugs. Pharm. Res. 5:615–618 (1988).
. R. B. Greenwald, Y. H. Choe, C. D. Conover, K. Shum, D. Wu,
metabolites. Part 17: Synthesis and plasma hydrolysis of acyloxy-
methyl carbamate derivatives of antitumour Triazenes. Bioorg.
Med. Chem. 8:1719–1725 (2000).
and M. Royzen. Drug delivery systems based on trimethyl lock 21. M. Safadi, R. Oliyai, and V. J. Stella. Phophoryloxymethyl car-
lactonization: poly(ethyleneglycol) prodrugs of amino-containing
compounds. J. Med. Chem. 43:475–487 (2000).
bamates and carbonates-novel water soluble prodrugs for amines
and hindered alcohols. Pharm. Res. 10:1350–1355 (1993).