Optimization of naltrexone diclofenac codrugs for sustained drug delivery across microneedle-treated skin

Article abstract of DOI:10.1007/s11095-013-1147-8

Purpose: The purpose of this work was to optimize the structure of codrugs for extended delivery across microneedle treated skin. Naltrexone, the model compound was linked with diclofenac, a nonspecific cyclooxygenase inhibitor to enhance the pore lifetim

Full text of DOI:10.1007/s11095-013-1147-8

Pharm Res (2014) 31:148–159
DOI 10.1007/s11095-013-1147-8
RESEARCH PAPER
Optimization of Naltrexone Diclofenac Codrugs for Sustained
Drug Delivery Across Microneedle-Treated Skin
Priyanka Ghosh & DoMin Lee & Kyung Bo Kim & Audra L. Stinchcomb
Received: 1 March 2013 /Accepted: 9 July 2013 /Published online: 14 August 2013
#
Springer Science+Business Media New York 2013
.
.
.
ABSTRACT
KEY WORDS codrugs drug delivery microneedle
stability transdermal
.
Purpose The purpose of this work was to optimize the struc-
ture of codrugs for extended delivery across microneedle treat-
ed skin. Naltrexone, the model compound was linked with
diclofenac, a nonspecific cyclooxygenase inhibitor to enhance
the pore lifetime following microneedle treatment and develop a
7 day transdermal system for naltrexone.
Methods Four different codrugs of naltrexone and diclofenac were
compared in terms of stability and solubility. Transdermal flux, per-
meability and skin concentration of both parent drugs and codrugs
were quantified to form a structure permeability relationship.
Results The results indicated that all codrugs bioconverted in
the skin. The degree of conversion was dependent on the
structure, phenol linked codrugs were less stable compared to
the secondary alcohol linked structures. The flux of naltrexone
across microneedle treated skin and the skin concentration of
diclofenac were higher for the phenol linked codrugs. The
polyethylene glycol link enhanced solubility of the codrugs,
which translated into flux enhancement.
INTRODUCTION
Codrugs or mutual prodrugs are two active pharmaceutical
ingredients (API) linked via a covalent linkage to form a single
chemical moiety (1). They are primarily synthesized to solve
drug delivery issues like stability and solubility of the parent
compounds or to enhance permeation of one or both active
moieties across biological barriers by modification of the
chemical structure (2). Codrugs and prodrugs are pharma-
cologically inactive in their modified form; however they are
bioconverted back to parent drugs either by chemical or
enzymatic hydrolysis in formulation, when crossing biologi-
cal barriers, or in the plasma (1,2).
The overall goal of the current project was to optimize
codrug structures in order to develop a sustained release
transdermal drug delivery system for alcohol addiction. Nal-
trexone (NTX), the model compound (MW-341.4, log P-2.0)
is a μ-opioid receptor antagonist approved for alcohol and
opioid addiction treatment (3). 6β-Naltrexol (NTXol) (MW-
343.4, log P-0.68), the active metabolite of NTX, has also
been shown to have therapeutic efficacy for alcohol addic-
tion (4,5). The current available dosage forms for NTX are
oral and an extended release intramuscular injection. Both
forms have their own drawbacks. The oral dosage form has
compliance issues in an addict population due to daily dosing
and gastrointestinal side effects (5,6). The extended release
intramuscular dosage form has been associated with serious
injection site reactions and difficulty with removal in case of
emergency opiate requirements (7,8).
Conclusion The current studies indicated that formulation sta-
bility of codrugs and the flux of naltrexone can be enhanced via
structure design optimization. The polyethylene glycol linked
naltrexone diclofenac codrug is better suited for a 7 day drug
delivery system both in terms of stability and drug delivery.
Electronic supplementary material The online version of this article
(doi:10.1007/s11095-013-1147-8) contains supplementary material, which is
available to authorized users.
:
:
:
P. Ghosh D. Lee K. B. Kim A. L. Stinchcomb
Department of Pharmaceutical Sciences, College of Pharmacy
University of Kentucky, Lexington, Kentucky 40536-0082, USA
:
P. Ghosh A. L. Stinchcomb (*)
Department of Pharmaceutical Sciences, School of Pharmacy
University of Maryland, 20 N. Pine St.
Baltimore, Maryland 21201, USA
NTX can potentially benefit from an alternative delivery
route like transdermal drug delivery. This noninvasive tech-
nique allows delivery of drugs directly into the systemic
circulation and bypasses gastrointestinal side effects. The
transdermal route is also known to be patient-friendly and
e-mail: astinchc@rx.umaryland.edu
A. L. Stinchcomb
AllTranz, Inc., Lexington, USA

Naltrexone Diclofenac Codrugs
149
comparatively more effective in an addict population with
compliance issues (9). However, NTX doesn’t cross the stra-
tum corneum (SC) barrier at a therapeutically relevant rate
by passive diffusion alone (10). The SC, the topmost layer of
the skin is formed of dead keratinized epidermal cells and a
lipid matrix, and forms the most important barrier to trans-
port of xenobiotics across the skin (11).
solubility, transdermal flux and bioconversion in the skin.
The codrugs were synthesized either using the 3-OH (phenol
position) on NTX or the 6-OH (secondary alcohol) position
on NTXol, and the –COOH group on diclofenac. Ester,
carbamate or amide linkers were used to join the two mol-
ecules either directly, or with the help of a polyethylene
glycol (PEG) linker. Furthermore, hydrochloride salts of
promising codrugs were synthesized and tested to develop a
formulation for in vivo pharmacokinetic studies.
Previous work, in both animal models and humans, has
shown that by using small micron-scale needles or microneedles
(MN), it is possible to deliver NTX at a rate that provides
plasma concentrations in the lower end of the therapeutic range
(10). MNs are a physical enhancement technique used to
permeablize the skin by creating micropores. The effectiveness
of this technique for enhancement of transdermal drug delivery
has been established in the literature over the past decade and a
half (12,13). Using the ‘poke (press) and patch’ approach for
MN treatment drug can be delivered across treated skin for 48–
72 h under occlusion. The drug delivery window can be further
enhanced by local application of a non-steroidal anti-
inflammatory drug, diclofenac sodium (DIC) (MW-296.1, log
P-4.5) (3,14,15). Due to physicochemical incompatibility of the
two molecules in formulation, daily or alternate day application
of DIC was required to keep micropores open for a 7 day
period (16,14). The incompatibility in formulation also leads to
a significant decrease in flux of NTX across micropores in the
presence of DIC, as compared to NTX alone, due to precip-
itation issues upon co-administration (16).
MATERIALS AND METHODS
Chemicals
NTX HCl and NTX base were purchased from Mallinckrodt
(St. Louis, MO), DIC acid from AK Scientific, Inc. (Mountain
view, CA) and DIC sodium salt from TCI America (Portland,
OR). Water was purified using a NANOpure Diamond™,
Barnstead water filtration system. Propylene glycol was pur-
chased from Sigma (St. Louis, MO). Sodium acetate and
acetic acid were obtained from Fisher Scientific (Fairlawn,
NJ). 1-Octanesulfonate, sodium salt was obtained from Regis
Technologies, Inc (Morton Grove, IL). Trifluroacetic acid
(TFA), triethylamine (TEA) and acetonitrile (ACN) were
obtained from EMD chemicals (Gibbstown, NJ). Unless oth-
erwise noted, all other reagents for codrug synthesis were
obtained from Sigma Aldrich (St. Louis, MO).
The codrug approach was thereafter used to combine
NTX and DIC into a single chemical moiety with the goal
of solving the formulation issue. The bioconversion of the
properly designed new chemical entities in the skin or plasma
should not be a problem since it has been previously shown
that prodrugs/codrugs are bioconverted in the skin, both
during passive delivery and MN enhanced delivery (17–20).
The hypothesis was that following bioconversion in the skin,
DIC would act locally and keep the pores open facilitating
systemic delivery of NTX. A proof of concept study was
conducted using a NTX-DIC 3-O-ester linked codrug to
look at stability, solubility, transdermal flux and local skin
concentration of the codrugs and the parent drugs (21).
Although pharmacokinetic profiles were not obtained from
the study, pore visualization techniques were used to verify
the presence of micropores in the skin at the end of 7 days,
using the codrug formulation compared to absence of mi-
cropores for the NTX control. These studies confirmed that
the codrug approach was useful for development of a 7 day
system, however further optimization was required. The
codrug also had a formulation half-life of around 4–5 days,
and the solubility and transdermal flux were an order of
magnitude lower compared to the most optimized formula-
tion for NTX delivery across MN treated skin (22,23).
In the current study, four codrugs of NTX/NTXol and
DIC have been compared to look at structure vs. stability,
Synthesis of the Codrugs
Codrug I and hydrochloride salt of codrug I were synthe-
sized following previously reported procedures (21).
Codrug II: Benzyl bromide (0.522 mL, 4.375 mmol) was
added to a suspension of 6β-NTXol (500 mg, 1.458 mmol)
and potassium carbonate (1 g, 7.3 mmol) in acetone (10 mL).
The reaction mixture was then refluxed for 2 h before it was
concentrated under reduced pressure. The resulting crude
product was further purified by column chromatography on
silica gel using dichloromethane-methanol mixture to obtain
compound 1 as a white solid (500 mg, 79%). A mixture of
compound 1 (500 mg, 1.153 mmol), DIC acid (685 mg,
2.306 mmol), and DMAP (7 mg, 0.06 mmol) were dissolved
in dichloromethane (50 mL) and cooled to −78°C in an
acetone/dry ice bath under N2. To the reaction mixture,
N,N’-dicyclohexylcarbodiimide (685 mg, 2.306 mmol) in
dichloromethane (20 mL) was then added with vigorous
stirring. After 1 h, the reaction mixture was allowed to warm
up to room temperature and concentrated under reduced
pressure. The resulting crude product was purified by col-
umn chromatography on silica gel using hexane-ethyl ace-
tate mixture to yield compound 2 as a white solid (260 mg,
32%). Catalytic hydrogenation of compound 2 (100 mg,

150
Ghosh, Lee, Kim and Stinchcomb
0.140 mmol) over a Pd/C catalyst provided Codrug II as a
solid (80 mg, 92%).
reduced pressure. The crude product was purified by col-
umn chromatography on silica gel using dichloromethane-
methanol mixture to afford Codrug IV as a semi solid
(120 mg, 45%).
Codrug III: Triethylamine (0.123 mL, 0.879 mmol) was
added to a suspension of NTX·HCl (200 mg, 0.586 mmol)
and 4-Nitrophenyl chloroformate (130 mg, 0.644 mmol) in
dichloromethane (5 mL). The resulting solution was stirred
for 4 h at room temperature, and then concentrated under
reduced pressure to yield an oilic crude product. The crude
product was purified by column chromatography on silica
gel using dichloromethane-methanol mixture to provide
compound 3 as a light yellow solid (292 mg, 66%). To a
dichloromethane solution containing compound 3 (61 mg,
0.143 mmol) and DIC-EBE (73 mg, 0.143 mmol) was added
DMAP (35 mg, 0.286 mmol). The reaction mixture was then
stirred for 1 h at room temperature and then concentrated
under reduced pressure. The resulting crude product was
purified by column chromatography on silica gel using
dichloromethane/methanol solvent systems to obtain codrug
III as a light yellow solid (78 mg, 69%).
HPLC Assay
All codrugs and parent drugs were quantified using HPLC.
The HPLC instrument used consisted of a Waters™ e2695
separation module, a 2489 UV/Visible detector and Em-
power™3 software. A Perkin Elmer Spheri5 VL C18 col-
umn (5 μ, 220×4.6 mm) and a C18 guard column
(15×3.2 mm) were used with the UV detector set at a
wavelength of 280 nm/215 nm. The mobile phase consisted
of 70:30(v/v) ACN: (0.1% TFA with 0.065% 1-octane sul-
fonic acid sodium salt, adjusted to pH 3.0 with TEA aqueous
phase). Samples were run at a flow rate of 1.5 ml/min. All
codrugs and parent drugs were quantified using standard
curves in the range of 100–10,000 ng/ml.
Codrug III hydrochloride salt: To a dichloromethane
solution of codrug III (100 mg, 0.126 mmol), 1 M HCl was
slowly added and continuously stirred for 10 min. The
resulting codrug III salt was filtered and dried to yield codrug
III·HCl as a white solid (95 mg, 91%).
Stability Studies
The stabilities of all codrugs were determined in 0.3 M
acetate buffer pH 5.0 at 32°C. For stability studies drug
was presolubilized in ACN. The solution was then used to
spike prewarmed buffer at 32°C (100 μl in 10 ml of
buffer, <1% total ACN concentration). One hundred μl
samples were obtained at regular time intervals and the
reaction was terminated using 900 μl of ACN. All samples
were stored at −80°C until analysis using HPLC. Samples
were stable in −80°C verified using quality control stan-
dards (data not shown). Data for all stability studies were
analyzed using pseudo first-order kinetics. All studies were
run in triplicate.
Codrug IV: To a suspension of 6β-NTXol (300 mg, 0.875-
mmol) and di-tert-butyl dicarbonate (210 g, 0.962 mmol) in
acetone (20 mL) was added potassium carbonate (604 mg,
4.373 mmol) with vigorous stirring. The reaction mixture was
then refluxed for 1 h and concentrated under reduced pressure.
The resulting crude product was further purified by column
chromatography on silica gel using dichloromethane-methanol
solvent systems to provide compound 4 as a white solid
(340 mg, 88%). To a suspension of compound 4 (340 mg,
0.766 mmol) and 4-Nitrophenyl chloroformate (232 mg, 1.149-
mmol) in dichloromethane (10 mL) was added DMAP
(280 mg, 2.298 mmol). The resulting mixture was stirred for
1 h at room temperature and subsequently concentrated under
reduced pressure to yield the crude product, which was further
purified by column chromatography on silica gel using
dichloromethane/methanol solvent systems to obtain com-
pound 5 as a light yellow solid (445 mg, 95%). Next, N, N-
Diisopropylethylamine (0.623 mL, 3.58 mmol) was added to a
suspension of compound 5 (436 mg, 0.716 mmol) and DIC-
EBE (458 mg, 1.074 mmol) in dichloromethane (10 mL) and
stirred for 1 h at room temperature. After the reaction mixture
was concentrated under reduced pressure, the resulting
crude product was further purified by column chromatogra-
phy on silica gel using dichloromethane/methanol solvent
systems to obtain compound 6 as a white solid (455 mg,
71%). Trifluoroacetic acid (1 mL) was added to a solution of
compound 6 (300 mg, 0.334 mmol) in dichloromethane
(10 mL). The resulting solution was stirred at room tempera-
ture for 1 h. The reaction mixture was concentrated under
Donor Solution Preparation and Solubility
Determination
Donor solution was prepared for all codrugs in 0.3 M acetate
buffer pH 5.0 containing 10% propylene glycol (PG). Excess
drug was added to the above mixture, vortexed, sonicated
for 10 min and left overnight at 32°C (The only exception,
codrug I was incubated in donor solution for an hour before
dosing due to stability issues). The choice of the donor was
based on previous studies using codrugs and flux of model
compounds across MN treated skin (20,21). For solubility
determination, codrug solution was removed from the do-
nor, filtered using a 0.2 μm, 500 μl centrifugal filter and
injected onto the HPLC after suitable dilution with ACN.
Saturation solubility was determined for all codrugs except
hydrochloride salt of codrug III due to limited availabil-
ity of codrug. All solubility determinations were carried
out in triplicate.

Naltrexone Diclofenac Codrugs
151
Melting Point Determination of Codrugs
NTX/NTXol from the codrugs using Fick’s first law of diffu-
sion. The cumulative amount permeated was plotted as a
function of time and the slope of the steady state portion of
the curve was used to determine flux. Permeation profiles did
not reach steady state in the 48 h timeframe for DIC and
codrug, so cumulative amount of drug permeated was quan-
tified, where applicable. Receiver samples were concentrated,
if required by evaporating 1 ml of the receiver solution under
nitrogen at 37°C followed by reconstitution in 100 μl of ACN.
For quantification of the drug concentration in the skin samples,
the active permeation area was dissected out and cut into small
pieces. The concentration of parent drugs and codrug in the
skin was determined by incubating the skin in 10 ml of ACN
overnight at 32°C and analyzing the solution by HPLC after
appropriate dilutions. All permeation studies were carried out in
triplicate, at a minimum.
Differential scanning calorimetry (DSC) was used to estimate
melting point of parent drugs and codrugs in the range of 50–
300°C. The apparatus used was a DSC 2920 with a DSC
refrigerated cooling system from Texas Instruments (TA Instru-
ments, New Castle, DE). Two to five mg of drug was weighed
into hermetic aluminum pans and sealed with a lid. Drug was
heated @10°C/min from 50 to 300°C. Data was collected
using Thermal Advantage and analyzed using Universal Anal-
ysis 2000 softwares from TA Instruments. Melting point was
determined from the inflection point of the melting curve.
Diffusion Studies
All diffusion studies were carried out using full thickness
Yucatan miniature pig skin. Studies were approved by the
IACUC protocols at University of Kentucky and University of
Maryland, Baltimore. The skin was obtained from euthanized
animals, dermatomed to a thickness of 1.4–1.8 mm and stored
at −20°C until the day of the experiment. MN’s for this study
were obtained from Dr. Mark Prausnitz’s laboratory at the
Georgia Institute of Technology. A 5 MN in plane array was
used for in vitro studies (Fig. 1). Each MN was 750 μm long,
200 μm wide and 75 μm thick and the interneedle spacing was
1.35 μm. For all experiments skin was cut into small rectan-
gular pieces. They were then placed on a wafer of Sylguard®
184 silicone elastomer to mimic the underlying tissue and
treated with a 5 MN array, 20 times to generate a total of
100 MN pores per diffusion cell (within the 0.95 cm2 active
area). The permeation experiments were carried out using a
PermeGear® flow through system (In-line, Riegelsville, PA).
The receiver solution used was water alkalified to pH 7.4
containing 20% ethanol at 37°C. The flow rate was set at
1.5 ml/h and the temperature of the diffusion cells was
maintained at 32°C to mimic physiological skin surface con-
ditions. Each cell was charged with 250 μl of the relevant
codrug or parent drug for control cells and receiver solution
was collected in 6 h increments for a total of 48 h. At the end of
the study the cells were removed and drug was quantified both
in the receiver solution and skin samples using HPLC. For the
receiver solution, steady state flux values were determined for
Statistical Analysis
Data for all experiments were reported as mean standard
deviation. Statistical analysis of data was carried out with Stu-
dents’ t-test and one way ANOVA with post hoc Tukey’s pairwise
tests, if required, using GraphPad Prism® software, version 5.04
software. P<0.05 was considered to be statistically significant.
RESULTS
Synthesis of Codrugs
Physical properties of parent drugs and codrugs are reported in
Table I. Structures of all codrugs and intermediates are provid-
ed in Figs. 2 and 3. The synthesis of codrug I, base and salt form
has been previously reported (21). Codrugs II, III and IV and
hydrochloride salt form of codrug III were synthesized success-
fully. The structures were verified using 1HNMR and HRMS.
1
Codrug II. H NMR (CDCl3, 500 MHz) δ 7.41−7.24 (m,
10 H),7.09 (m, 1H), 6.96−6.92 (m, 3H), 6.72 (m, 1H), 6.57−6.54
(m, 2H), 5.31 (2H, s), 5.12−5.07 (m, 3H), 4.79−4.71 (m, 2H), 4.04
(m, 2H), 3.86 (s, 2H), 3.49 (m, 2H), 3.08−2.99 (m, 2H), 2.64−2.56
(m, 2H), 2.39−2.18 (m, 4H), 2.09−1.95 (m, 7H), 1.81 (m, 1H),
Fig. 1 Microneedle array with 5 MN array for in vitro studies and 50 MN array for in vivo studies and applicator for 5 MN array(left panel), Yucatan miniature
pig skin before treatment and following treatment with gentian violet for pore visualization (right panel).

152
Ghosh, Lee, Kim and Stinchcomb
Table I Physical Properties of
Parent Drugs and Codrugs
Molecular weight
(MW)
Color/consistency
Retention
time (min)
Melting
point (°C)
Naltrexone
341.4
White/Solid
2.1
2.0
2.9
4.9
178
191.5
184
–
Naltrexol
343.4
Yellowish white/Solid
White/Solid
Diclofenac
296.15
619.53
656.0
Codrug I (hydrochloride salt)
White/Solid
White/Solid
152.5
202
–
Codrug II
621.55
Slightly yellow/Solid
4.4
3.8
Codrug III (hydrochloride salt)
793.73
830.19
Yellow/Solid
Yellow/Solid
Codrug IV
795.75
Cream/Solid
3.7
–
1.71−1.60 (m, 12H), 1.49−1.43 (m, 2H), 1.37−1.27 (m, 6H),
1.17−1.11 (m, 7H), 0.84 (m, 1H), 0.52 (m, 2H), 0.12 (m, 2H)
6.65 (d, 1H, J=8.0 Hz), 6.49 (d, 1H, J=8.0 Hz), 5.28 (m, 1H),
4.51 (m, 2H), 3.76−3.47 (m, 18H), 3.08−2.97 (m, 3H), 2.59−2.57
(m, 2H), 2.35 (m, 2H), 2.18−1.88 (m, 8H), 1.60−1.58 (m, 1H),
1.46−1.25 (m, 9H), 0.82 (m, 1H), 0.51 (m, 2H), 0.11 (m, 2H)
1
Codrug III. H NMR (CDCl3, 500 MHz) δ 7.78 (s, 1H), 7.32
(m, 2H), 7.12−7.07 (m, 2H), 6.97−6.93 (m, 2H), 6.86 (t, 1H,
J=8.5 Hz), 6.71−6.66 (m, 3H), 6.60 (m, 1H), 6.49 (d, 1H,
J=8.5 Hz), 5.98 (m, 1H), 4.69 (m, 2H), 3.64−3.44 (m, 13H),
3.19 (s, 2H), 3.10−3.00 (m, 4H), 2.72−2.56 (m, 4H), 2.44−2.31
(m, 8H), 2.17−2.12 (m, 3H), 1.90−1.26 (m, 8H), 0.86 (m, 2H),
0.56 (m, 4H), 0.15 (m, 4H)
HPLC Assay Development
The first step towards analysis of codrugs and parent drugs
was development of a HPLC method. The same column,
mobile phase composition and wavelength (280 nm) were
used for all compounds. All codrugs and parent drugs had
baseline resolution on the assay. The retention times
( 0.1 min) for all compounds from the HPLC assay are
reported in Table II. Standard curves for both codrugs and
parent drugs were prepared in the range of 100–10,000 ng/ml
for ACN samples and 100–5,000 ng/ml for extracted receiver
samples. All standard curves were linear (r2>0.999).
1
Codrug III HCl Salt. H NMR (DMSO−d6, 500 MHz) δ
8.66−8.58 (m, 2H), 8.08 (s, 1H), 7.96 (s, 1H), 7.47 (s, 1H),
7.08 (m, 2H), 6.77−6.74 (m, 2H), 6.61 (m, 1H), 6.63 (d, 1H,
J=7.0 Hz), 6.41−6.36 (m, 2H), 6.28−6.22 (m, 2H), 5.85 (m,
1H), 4.70 (s, 1H), 4.58 (s, 1H), 3.61−3.56 (m, 2H), 2.82−2.79
(m, 6H), 2.71−2.50 (m, 11H), 2.30−2.23 (m, 3H), 2.09 (m,
12H), 1.74−1.59 (m, 4H), 1.09−1.07 (m, 4H), 0.67−0.64 (m,
6H), 0.27 (m, 2H), 0.20 (m, 2H), 0.10 (m, 2H), 0.00 (m, 2H)
Stability and Solubility Studies
1
Codrug IV. H NMR (CDCl3, 500 MHz) δ 7.62 (s, 1H),
7.33−7.31 (m, 2H), 7.17−7.15 (m, 1H), 7.09−7.08 (m, 1H),
6.96 (t, 1H, J=8.0 Hz), 6.88−6.86 (m, 1H), 6.76−6.69 (m, 2H),
Formulation stability was carried out for all codrugs to deter-
mine the relative rate of conversion due to chemical hydrolysis.
Fig. 2 Structures of NTX/NTXol
and DIC codrugs.

Naltrexone Diclofenac Codrugs
153
Fig. 3 Structures of intermediates
for codrug synthesis. NMR data for
intermediates are provided in
Supplementary Material.
The results indicated that stability was significantly enhanced
by using the 6-O-secondary alcohol linkage over the 3-O-
phenol linkage for both codrugs I and II, and codrugs III and
IV. Stability was also significantly enhanced by using the PEG
covalent link and carbamate/amide linkage compared to the
ester linkage, both for codrug I and III, and codrug II and IV
(Fig. 4).
The solubility studies were carried out to quantify satura-
tion solubility of codrugs in donor formulation and evaluate
its effect on flux. The studies indicated that solubility of the
NTX-DIC codrugs were significantly higher compared to
NTXol-DIC codrugs with similar linkage. The PEG linkage
enhanced solubility for codrug III and IV, as compared to
codrug I and II. Solubilities of codrugs were also enhanced
by using a hydrochloride salt form of the drug compared to a
free base form.
codrugs II and IV (p<0.05). The flux of NTX from codrug
I salt and codrug III, and NTXol from codrug II and codrug
IV were not significantly different from each other (p>0.05)
(Fig. 5). Permeation of DIC and intact codrug was also
observed for codrug III from the 10% PG containing for-
mulation. Cumulative amount of DIC and codrug present in
the receiver solution were 1.09 0.36 nmol and 2.89 1.0-
nmol, respectively, at the end of the 48 h experiment. NTX
flux was further enhanced by 2.5 fold on using the salt form
of codrug III, over free base form (Fig. 6). Significantly
higher permeation of DIC and intact codrug were also
observed using the salt form of codrug III, they were 3.1
0.86 nmol and 24.3 8.8 nmol, respectively.
Skin concentration of codrugs and parent drugs were
quantified at the end of the experiment to evaluate the effect
of codrug structure on bioconversion in the skin. The skin
concentration data indicated that bioconversion was more
complete for NTX-DIC codrugs compared to the NTXol-
DIC codrugs. Intact codrug was present in the skin for all
codrugs, whereas NTX and DIC were only quantifiable for
codrugs I and III. Skin concentrations of codrugs were not
significantly different among the 4 codrugs. Skin concentra-
tion of NTX was not significantly different between codrugs
I and III (p>0.05). Skin concentration of DIC was
Diffusion Studies
The diffusion studies were carried out to quantify the per-
meation of NTX/NTXol across the skin and concentration
of codrugs and parent drugs in the skin. The diffusion studies
indicated that flux of NTX from codrugs I salt and III were
significantly higher compared to flux of NTXol from
Table II Physicochemical Characterization of Codrugs
Solubility (mM)
Pseudo first order rate constant (k) (days −1
)
Stability (approx. half-life) (days)
8.72 1.05
Codrug I (hydrochloride salt)
7.23a
0.0803
83.8 15.4
Codrug II
–
0.0052
0.0238
130.88 4.94
29.14 1.04
Codrug III (hydrochloride salt)
27.16 0.5
87.67 1.5b
Codrug IV
6.8 0.2
0.0035
187.56 8.55
n≥3 for all estimates except melting point determination
a Codrug I solubility was determined in 43% PG, 43% ethanol containing solution (data previously reported in (21))
b Saturation solubility could not be reported due to limited codrug

154
Ghosh, Lee, Kim and Stinchcomb
Fig. 4 Stability of NTX/DIC
codrugs in 0.3 M acetate buffer
pH 5.0. Data analyzed using
pseudo first-order kinetics. n=3
for all codrugs.
significantly different between codrugs I and III (p<0.05).
Skin concentration was also determined for codrug III salt to
evaluate the effect of salt form on skin concentration. No
significant differences were observed for NTX or DIC con-
centration using codrug III base or salt form (p>0.05).
Flux data from codrug I has been previously reported
(21). The salt form of codrug I has been used for all compar-
isons for diffusion experiments in this study since the limited
solubility of the free base form in 10% PG containing for-
mulation prevented direct comparison of results.
codrug examples in the literature where both drugs were
targeted either for systemic delivery or for topical application,
the NTX/NTXol codrugs are designed for systemic delivery
of NTX/NTXol while DIC acts locally in the SC/epidermal
layer of the skin to enhance micropore lifetime. The physico-
chemical properties of NTX allow systemic permeation, and a
higher log P and negative charge facilitates retention of DIC
in the skin (21).
Purity and structure of all codrugs were confirmed using
NMR and HRMS. Quantitative analysis is an important
step in the development of any drug delivery system. HPLC
was used for quantification of codrug and parent drugs in
stability, solubility and permeation studies. Since all com-
pounds were evaluated using reverse phase chromatography
and the same assay parameters, retention time on the HPLC
can be used as a comparative estimate of hydrophobicity or
log P of the molecules (27). NTX and NTXol were compa-
rable in hydrophobicity with NTXol being slightly more
hydrophilic due to the presence of an additional hydroxyl
group at the 6 position compared to the ketone of NTX.
DIC was the most hydrophobic of the parent molecules,
which is expected from the reported log P values from
Scifinder®. Comparing codrugs I and II, codrug I with a
retention time of 4.9 min was more hydrophobic compared
to codrug II at 4.4 min. The difference in hydrophobicity can
be attributed to more significant hydrophilic interactions of a
phenolic hydroxyl as compared to a ketone functional group
substitution. Comparison of codrug III and IV followed the
same pattern, where the retention time shifted from 3.8 min
for codrug III to 3.7 min for codrug IV. Although in both
cases the codrugs with the free phenolic hydroxyl had less
retention on the column, the shift was much less significant
for the PEG linked codrugs. This can be attributed to significant
hydrophilic interactions of the PEG chain itself. Effect of PEG
chain on enhancement of hydrophillicity is also confirmed by the
fact that both codrug III and IV are eluted significantly earlier
DISCUSSION
Prodrugs and codrugs have been explored for drug delivery via
different routes over the last few decades. The approach is
mostly used to solve a drug delivery issue and the major advan-
tage lies in the fact that chemical modification renders the new
chemical entity pharmacologically inactive thus decreasing po-
tential side effects and the need for additional pharmacological
evaluation. Compared to all other routes of drug delivery, the
topical/transdermal field has seen slower progress in terms of
the number of new API’s approved and on the market over the
years (24). The most important reason being the extreme barrier
property of the skin/SC. Chemical modification of the API itself
to form codrugs/prodrugs provides a new approach for flux
enhancement, decreasing some of the need for excess perme-
ation enhancer supplementation. Additionally, two API can be
delivered at the same time in a codrug to have a synergistic
effect. A few examples of topical/transdermal codrugs that have
been evaluated include NTX and bupropion for smoking ces-
sation and alcohol addiction, retinyl ascorbate and ascorbic acid
for oxidative stress, and 5 fluorouracil with triamcinolone
acetonide for actinic keratosis (25,26).
The current study evaluated codrugs for pore lifetime
enhancement following MN treatment. Unlike previous

Naltrexone Diclofenac Codrugs
155
Fig. 5 Flux of NTX/NTXol and
skin concentration of parent drugs
and codrugs. n≥3 for all studies.
Flux of NTX from codrug I and
codrug III are not significantly
different (p>0.05) and they are
higher than NTXol flux from
codrug II and codrug IV (p<0.05).
Skin concentration of NTX and
codrug are not significantly different
for I and III (p>0.05). Skin
concentration of DIC is significantly
different (p<0.05).
than codrug I and II. Since hydrophobicity/hydrophilicity play
an important role in permeation of molecules across intact and
MN treated skin, the comparative analysis provides an estimate
of solubility and transport. For MN treated skin, it has been
shown that lower viscosity aqueous formulations lead to higher
flux. Higher hydrophillicity would translate into enhanced solu-
bility in an aqueous medium; a molecule with lower log P will be
beneficial for the current delivery system in terms of higher
solubility.
Codrugs/prodrugs are primarily designed to solve a drug
delivery problem, but since these molecules are pharmacolog-
ically inactive chemical/enzymatic hydrolysis is imperative for
bioconversion and function of these molecules. In the current
project, codrugs were designed to formulate NTX and DIC in
a single formulation so that precipitation of one drug on
application of the other didn’t lead to reduced NTX flux
across the micropores. Thus, stability in formulation is a very
important criterion for these codrugs. The results from the
stability studies indicated that codrugs I and III (phenol linked
codrugs) were less stable in formulation compared to codrugs
II and IV (secondary alcohol linked codrugs). This was expected
since the resonance stabilization of a phenol makes it a better
leaving group compared to a secondary alcohol. Codrug III was
more stable compared to codrug I, and codrug IV was more
stable compared to codrug II. This can be attributed to the
carbamate/amide linkage in place of the ester linkage. It has
been shown in the literature that ester linked prodrugs are more
susceptible to chemical hydrolysis as compared to carbamate
prodrugs at a formulation pH of 5.0 (20). On the other hand,
carbamate linked prodrugs are less stable at physiological pH 7.4-
at similar buffer concentrations (20). The change in the rate of
chemical degradation within the pH range indicates that the
primary mechanism for hydrolysis of a monosubstituted carba-
mate in an aqueous formulation is probably via proton

156
Ghosh, Lee, Kim and Stinchcomb
Fig. 6 Flux and skin concentration
from codrug III and codrug III salt.
n≥3 for all studies. Flux of NTX
was significantly higher (p<0.05).
No significant difference in skin
concentration except for codrug
(p>0.05).
elimination and not nucleophile attack of hydroxyl groups (28).
Thus employing carbamate codrugs enhances formulation sta-
bility and could be beneficial because they are known to degrade
faster under physiological conditions. Since NTX/NTXol and
DIC codrugs could not be linked directly with a carbamate
linkage, a PEG link was used to serve the dual role of linking
the molecules and enhancing solubility. PEG is one of the most
commonly used topical/transdermal excipients, thus irritation,
toxicity and side effects related to incorporation of PEG in the
formulation should be limited or non-existent. Once inside the
body, PEG is metabolized into a series of phase I and II metab-
olites, with no toxicological concern. Molecular weight of the
PEG subunit does play a role in its metabolism, with PEG’s
>5000 showing little or no metabolism (29). Thus from the
stability studies it can be concluded that codrugs II, III and IV
with more than 85% intact codrug in formulation at 7 days could
be used for the development of a drug delivery system. Since the
stability studies were conducted in aqueous buffers, the codrugs
will be expected to be more stable in a 10% PG containing
formulation since hydrolysis is the primary mechanism for deg-
radation. Once inside the body, the codrugs will be degraded by
enzymatic hydrolysis in the skin, in addition to chemical hydro-
lysis for much faster regeneration of parent drugs. Presence of
esterase, amidase, and protease, in addition to the normal phase I
and II enzymes are well documented in the skin (30,31).
Solubility of drug in the donor is important for transder-
mal drug delivery. The drug concentration in formulation is
the main driving force for transport across the skin. As stated
before, MN-enhanced drug delivery with lower viscosity
aqueous formulations has demonstrated significantly higher
flux compared to viscous PG-rich formulations. Ten percent
PG containing 0.3 M acetate buffer at pH 5.0 was chosen as
the formulation based on previous studies (21,24). The solu-
bility data indicated that the PEG link significantly enhanced

Naltrexone Diclofenac Codrugs
157
solubility for both codrug III and IV, compared to
codrug I and II. Solubility of codrug I and III was
higher compared to II and IV, indicating that phenol
linked codrugs are more soluble compared to secondary
alcohol linked codrugs. This can be further explained in
terms of the melting point (MP) estimates where codrug
I has a lower melting point compared to codrug II.
Lower MP translates into higher aqueous solubility for
the unionized form of the drug (32).
mass spectrometry data (not shown) which indicates that
DIC is not completely bioconverted in the skin from codrug
III. A mixture of DIC (MW: 296) and DIC + PEG linker
conjugate (MW: 452) was found in the receiver and skin
samples by mass scan and ESI-/ESI + ionization. Since an
accurate estimate of local skin concentration of DIC re-
quired for pore lifetime enhancement is not available, it
would be beneficial to test codrug III in vivo to see if it is
efficacious. If not, the linker at the DIC end of the molecule
will have to be optimized further for faster release of DIC.
The hydrochloride salt form of codrug I led to a 13-fold
enhancement in solubility and transdermal flux (21), but
direct comparison of the effect of the salt formulation could
not be evaluated due to changes in donor composition. A
hydrochloride salt form of codrug III was synthesized and
evaluated for further solubility and flux enhancement in vitro.
Codrug III was chosen because formulation stability and
solubility was higher as compared to codrug I. Codrugs II
and IV were comparatively more stable but bioconversion,
transdermal flux of NTXol, and skin concentrations of DIC
were much lower. The permeation studies indicated that flux
of NTX can be further enhanced using a salt form of the
codrug over the free base form. A 2.5 fold enhancement in
NTX flux was obtained using a sub-saturated codrug III salt
as donor compared to free base. (Saturated donor could not
be used due to limited availability of codrug III salt.) The
permeabilities of NTX from codrug and codrug salt were not
significantly different from each other, indicating that the
flux enhancement was due to enhanced solubility. Skin con-
centrations of codrug or parent drugs were not significantly
different between salt and free base. Thus salt form has no
significant effect on skin concentration of the parent drugs or
the codrug, when comparing data from the same donor
condition, irrespective of the physicochemical properties of
the molecule. This will be useful if alternate salt forms are
used to further enhance solubility (23), and additional studies
for determination of local skin concentration of DIC will not
be necessary. The salt form also led to higher cumulative
amounts of DIC and codrug permeation compared to the
base. Intact codrug diffusing through the skin will convert
under physiological conditions to regenerate parent drugs,
thus further contributing to plasma concentrations of NTX.
The amount of DIC permeated through the skin is well
below therapeutic systemic concentrations and thus should
not be of any concern.
Stability and solubility studies provide important infor-
mation about how well drug transport might proceed, how-
ever skin permeation studies are required for estimation of
flux and bioconversion in the skin. Diffusion studies were
carried out using dermatomed Yucatan miniature pig skin.
Yucatan miniature pig was chosen as the appropriate model
based on previously conducted in vitro in vivo correlation
(IVIVC) studies for MN enhanced transdermal delivery.
The studies compared in vitro diffusion data to in vivo phar-
macokinetic studies in order to evaluate the importance of
Yucatan pig skin for MN studies. Modeling of the transder-
mal pharmacokinetic data using in vitro flux as an input
parameter and IV bolus data as the elimination parameter
helped in estimation of the optimum skin thickness required
for IVIVC for the MN geometry used in this study. For
passive transdermal studies, skin thickness of 200–300 μm
is considered optimum, however the MN are 750 μm long,
very thin skin leads to a release rate study across the skin as
opposed to permeation across a dermal barrier. A skin thick-
ness of 1.4–1.8 mm was found to be optimum for MN
enhanced delivery and used for the current study (33,34).
The diffusion studies indicated that the flux of NTX from
codrug I and III was significantly higher compared to
NTXol flux from codrug II and IV. This can be attributed
to the higher formulation stability of the NTXol codrugs;
significant amounts of intact codrug were found in the skin,
and no parent drugs were quantified at the end of the study
suggesting that NTXol was not released efficiently from the
codrug to generate transdermal flux. The difference in phys-
icochemical properties of NTX and NTXol could also be
responsible for reduced flux of NTXol. Since NTXol has an
additional hydroxyl group following conversion, H-bonding
potential of NTXol is higher. It has been shown in the
literature that the higher the number of hydrogen bonds,
the lower the transdermal flux across intact skin (35). The
skin concentration study also showed that DIC was only
present in the skin for the NTX codrugs. Since the presence
of DIC in the skin is imperative for its local effect on pore
lifetime enhancement, the NTXol codrugs would not be able
to keep micropores open beyond 48–72 h in the absence of
DIC. Also, significantly lower flux from these codrugs as
compared to the NTX codrugs is a disadvantage. The skin
concentration of DIC is significantly different between
codrug I and III. This can be explained with the help of
Therefore, from the structure permeability evaluation it can
be concluded that NTX-DIC PEG linked codrug (codrug III) is
the best choice for the current project in terms of stability,
solubility, bioconversion and NTX flux. NTX requires a flux
of 12–90 nmol/cm2/h from a 25 cm2 patch for therapeutic
efficacy, based on current oral dosing and clearance. Codrug
III hydrochloride salt has a flux of 10.8 2.1 nmol/cm2/h from
the currently tested sub-saturated solution, therefore further

158
Ghosh, Lee, Kim and Stinchcomb
flux enhancement is expected using a saturated solution. From
stability data, 85% of codrug III will be intact in formulation
over a 7 day patch application period, rendering it to be a
suitable molecule for in vivo evaluation of the codrug concept via
pharmacokinetic studies.
REFERENCES
1. Strasinger CL, Scheff NN, Stinchcomb AL. Prodrugs and codrugs
as strategies for improving percutaneous absorption. Expert Rev
Dermatol. 2008;3(2):221–33. 04/01; 2012/07.
2. Stella V, Borchardt R, Hageman M, Oliyai R, Maag H, Tilley J.
Prodrugs: challenges and rewards part 1, 1st ed. Springer; 2007.
3. American Chemical Society. https://scifinder.cas.org.
4. McCaul ME, Wand GS, Rohde C, Lee SM. Serum 6-beta-
naltrexol levels are related to alcohol responses in heavy drinkers.
Alcohol Clin Exp Res. 2000;24(9):1385–91.
5. Verebey K, Volavka J, Mule SJ, Resnick RB. Naltrexone: disposi-
tion, metabolism, and effects after acute and chronic dosing. Clin
Pharmacol Ther. 1976;20(3):315–28.
6. Hulse GK, Basso MR. The association between naltrexone compli-
ance and daily supervision. Drug Alcohol Rev. 2000;19(1):41–8.
7. Alkermes. Vivitrol (naltrexone for extended release injectable sus-
pension) http://www.vivitrol.com/.
8. Vivitrol alert, http://www.fda.gov/Safety/MedWatch/Safety
Information/SafetyAlertsforHumanMedicalProducts/ucm210755.
htm.
9. Abelin T, Buehler A, Muller P, Vesanen K, Imhof PR. Controlled
trial of transdermal nicotine patch in tobacco withdrawal. Lancet.
1989;1(8628):7–10.
10. Wermeling DP, Banks SL, Hudson DA, Gill HS, Gupta J,
Prausnitz MR, et al. Microneedles permit transdermal delivery of
a skin-impermeant medication to humans. Proc Natl Acad Sci.
2008;105(6):2058–63.
The goal of the current study was to optimize the codrug
structures based on flux of the parent molecules, bioconversion
in the skin and determination of steady state skin concentration
for codrug and parent molecules. Bioconversion of both
codrugs and prodrugs in Yucatan miniature pig skin have been
previously documented (20,21). However, one limitation of the
study is pore closure or skin resealing, which cannot be evalu-
ated in harvested skin. Pore closure kinetics have been evaluat-
ed in vivo in rats, hairless guinea pigs, Yucatan miniature pigs
and humans (14–16,33,36). Rats and guinea pigs are small
animal models that make pharmacokinetic evaluation and data
quantification easier. Yucatan miniature pigs are more compa-
rable to humans in terms of skin physiology, especially in
dermal thickness, critical to these MN studies. Following opti-
mization in vitro, the codrugs from this study will be evaluated
in vivo either in hairless guinea pig or Yucatan miniature pig
based on avaibility of codrugs and assay sensitivity for detection
of permeated molecules in the plasma.
11. Flynn G. Cutaneous and transdermal delivery–processes and sys-
tems of delivery. Modern pharmaceutics, fourth ed. Informa
Healthcare; 2002.
12. Henry S, McAllister DV, Allen MG, Prausnitz MR. Microfabricated
microneedles: a novel approach to transdermal drug delivery. J Pharm
Sci. 1998;87(8):922–5.
13. Prausnitz MR. Microneedles for transdermal drug delivery. Adv
Drug Deliv Rev. 2004;56(5):581–7.
14. Banks S, Paudel K, Brogden N, Loftin C, Stinchcomb A.
Diclofenac enables prolonged delivery of naltrexone through
microneedle-treated skin. Pharm Res. 2011:1–9.
15. Brogden NK, Milewski M, Ghosh P, Hardi L, Crofford LJ,
Stinchcomb AL. Diclofenac delays micropore closure following
microneedle treatment in healthy human subjects. J Control Re-
lease. 2012;163(2):220–9.
16. Brogden NK. Clinical evalution of novel methods for extending
microneedle pore lifetime. 2012.
17. Stinchcomb AL, Swaan PW, Ekabo O, Harris KK, Browe J,
Hammell DC, et al. Straight-chain naltrexone ester prodrugs: dif-
fusion and concurrent esterase biotransformation in human skin. J
Pharm Sci. 2002;91(12):2571–8.
CONCLUSION
NTX/NTXol and DIC codrugs were designed and evaluated
for development of a 7 day drug delivery system for alcohol
addiction treatment. Stability, solubility, transdermal flux of
NTX and local skin concentration of DIC were determined to
develop a structure permeability relationship for the codrugs.
A secondary alcohol linker enhanced stability of codrug mol-
ecules, however solubility, bioconversion, flux of NTX and
local skin concentration of DIC were higher using phenol
linked codrugs. Transport was further enhanced by utilizing
a hydrochloride salt form of the PEG linked NTX-DIC
codrug making it the most suitable candidate for further
in vivo pharmacokinetic studies.
18. Vaddi HK, Banks SL, Chen J, Hammell DC, Crooks PA,
Stinchcomb AL. Human skin permeation of 3-O-alkyl carbamate
prodrugs of naltrexone. J Pharm Sci. 2009;98(8):2611–25.
19. Kiptoo PK, Paudel KS, Hammell DC, Hamad MO, Crooks PA,
Stinchcomb AL. In vivo evaluation of a transdermal codrug of 6-
beta-naltrexol linked to hydroxybupropion in hairless guinea pigs.
Eur J Pharm Sci Off J Eur Fed Pharm Sci. 2008;33(4–5):371–9.
20. Milewski M, Yerramreddy TR, Ghosh P, Crooks PA, Stinchcomb
AL. In vitro permeation of a pegylated naltrexone prodrug across
microneedle-treated skin. J Control Release. 2010;146(1):37–44.
21. Ghosh P, Pinninti RR, Hammell DC, Paudel KS, Stinchcomb AL.
Development of a codrug approach for sustained delivery across
microneedle treated skin. J Pharm Sci. Accepted January 2013.
22. Milewski M, Stinchcomb A. Vehicle composition influence on the
microneedle-enhanced transdermal flux of naltrexone hydrochlo-
ride. Pharm Res. 2011;28(1):124–34.
ACKNOWLEDGMENTS AND DISCLOSURES
We would like to acknowledge Dr. Mark Prausnitz and Dr.
Vladimir Zarnitsyn at the Georgia Institute of Technology
for their expert advice and assistance with the MN arrays.
This work was funded by NIH grant R01DA13425 and
R42DA32191. Additional funding sources included the Uni-
versity of Kentucky Center for Drug Abuse Research Trans-
lation (CDART.)
Audra L. Stinchcomb is a significant shareholder in and
Chief Scientific Officer of AllTranz Inc., a specialty phar-
maceutical company involved in the development of trans-
dermal formulations for MN delivery.

Naltrexone Diclofenac Codrugs
159
23. Milewski M, Pinninti RR, Stinchcomb AL. Naltrexone salt selection
for enhanced transdermal permeation through microneedle-treated
skin. J Pharm Sci. 2012;101(8):2777–86.
30. Oesch F, Fabian E, Oesch-Bartlomowicz B, Werner C, Landsiedel
R. Drug-metabolizing enzymes in the skin of man, rat, and pig.
Drug Metab Rev. 2007;39(4):659–98.
24. Paudel KS, Milewski M, Swadley CL, Brogden NK, Ghosh P,
Stinchcomb AL. Challenges and opportunities in dermal/transder-
mal delivery. Ther Deliv. 2010;1(1):109–31. 07/01; 2012/03.
25. Cynkowski T, Cynkowska G, Walters KA. Codrugs: potential
therapies for dermatological diseases. Inf Healthc. 2007;255–266.
26. Lau WM, White AW, Gallagher SJ, Donaldson M, McNaughton
G, Heard CM. Scope and limitations of the co-drug approach to
topical drug delivery. Curr Pharm Des. 2008;14(8):794–802.
27. Chin Y, Weber Jr WJ, Voice TC. Determination of partition coefficients
and aqueous solubilities by reverse phase chromatography—II: evalua-
tion of partitioning and solubility models. Water Res. 1986;20(11):1443–
50. 11.
31. Prusakiewicz J, Ackermann C, Voorman R. Comparison of skin ester-
ase activities from different species. Pharm Res. 2006;23(7):1517–24.
32. Ran Y, He Y, Yang G, Johnson JLH, Yalkowsky SH. Estimation of
aqueous solubility of organic compounds by using the general
solubility equation. Chemosphere. 2002;48(5):487–509. 8.
33. Mikolaj Milewski. Microneedle-assisted transdermal delivery of
naltrexone species: in vitro permeation and in vivo pharmacokinet-
ic studies. 2012.
34. Burton SA, Ng CY, Simmers R, Moeckly C, Brandwein D,
Gilbert T, et al. Rapid intradermal delivery of liquid formu-
lations using a hollow microstructured array. Pharm Res.
2011;28(1):31–40.
28. Dittert LW, Higuchi T. Rates of hydrolysis of carbamate and car-
bonate esters in alkaline solution. J Pharm Sci. 1963;52(9):852–7.
29. Webster R, Didier E, Harris P, Siegel N, Stadler J, Tilbury L, et al.
PEGylated proteins: evaluation of their safety in the absence of
definitive metabolism studies. Drug Metab Dispos. 2007;35(1):9–16.
35. Potts RO, Guy RH. A predictive algorithm for skin permeability:
the effects of molecular size and hydrogen bond activity. Pharm
Res. 1995;12(11):1628–33.
36. Kalluri H, Banga A. Formation and closure of microchannels in
skin following microporation. Pharm Res. 2011;28(1):82–94.