Topical delivery of a model phenolic drug: Alkyloxycarbonyl prodrugs of acetaminophen

Article abstract of DOI:10.1023/B:PHAM.0000029281.12753.25

Purpose. To determine whether the delivery of a phenolic parent drug by its alkyloxycarbonyl (AOC) prodrugs through hairless mouse skin would show similar dependencies on water and lipid solubilities that similar prodrugs of more polar heterocyclic amide and imide parent drugs have shown. Methods. Flux through hairless mouse skin from suspensions in isopropyl myristate (J_MIPM), solubilities in IPM (S_IPM) and water (S_AQ), and partition coefficients between isopropyl myristate (IPM) and pH 4.0 buffer (K _IPM:4.0) were measured for two series of AOC derivatives of acetaminophen (APAP); their solubilities in pH 4.0 buffer (S_4.0) were estimated from S_IPM/K_{IPM: 4.0}. Log J_MIPM values were calculated from the n = 43 coefficients for the parameters in the transformed Potts-Guy (Roberts-Sloan) equation, and the average error of prediction (Δ log J′_IPM) was calculated. The J _MIPM, S_IPM, S_4.0, and molecular weight (MW) data for this series and two other series were combined with the n = 43 database to give a n = 61 database, and new best fit coefficients were determined for the Roberts-Sloan equation: log J_MIPM = x + y log S_IPM (1 - y) log S_4.0 - z MW. Results. All of the 4-AOC-APAP derivatives underperformed based on their predicted log J_MIPM (Δ log J′_MIPM = 0.275 ± 0.147 log units) and, although the two more water soluble members of this more lipid soluble series were more effective than APAP, they were only marginally so: <2 times. Addition of three new series to the n = 43 database for the Roberts-Sloan equation did not substantially change the coefficients to the parameters: x, y, z, and r ² = -0.322, 0.530, 0.00337 and 0.92, respectively. Conclusions. The topical delivery of a model phenolic drug by its AOC prodrugs through hairless mouse skin from IPM shows the same dependence on S_IPM, S _4.0, and MW as the delivery of polar heterocycles by their similar prodrugs.

Full text of DOI:10.1023/B:PHAM.0000029281.12753.25

Pharmaceutical Research, Vol. 21, No. 6, June 2004 (© 2004)
Research Paper
have been successful in enhancing the delivery of the parent
drug from a variety of vehicles including mineral oil and wa-
ter. However, the stability of the AC type derivative of phe-
nolic groups in water will be less than that of other types of
acyl derivatives due to the facts (a) that most acyl derivatives
are hydrolyzed by an addition-elimination type mechanism
and (b) that the carbonyl carbon of an ester functional group
will be more electrophilic than the carbonyl carbon of a car-
bonate or carbamate acyl derivative. The result is that, if
hydrolysis takes place by an addition-elimination reaction,
AC type derivatives will be more reactive in the first step.
As part of the development of broader databases for
topical delivery of drugs through hairless mouse skin from
isopropyl myristate (IPM) and water (AQ) (5), it was decided
to use carbonate derivatives (alkyloxycarbonyl; AOC) of phe-
nols instead of AC derivatives. Even though AC derivatives
would be stable in IPM, the AOC derivatives would be more
stable toward an aqueous environment, and we needed pro-
drugs that would be stable in both vehicles for the two data-
bases. Acetaminophen (APAP) was chosen as the model
phenolic drug because there was already some data available
on the stability and solubilities of some of the targeted pro-
drugs (6).
In this paper, we report the synthesis, characterization,
and evaluation in diffusion cell experiments, using IPM as the
vehicle and hairless mouse skin, of two series of AOC deriva-
tives of APAP. The first series is one in which the alkyl part
of the promoiety is a straight alkyl chain group, and the sec-
ond is one in which an oxygen is inserted in the alkyl chain. In
addition, the performance of the two series in diffusion cell
experiments will be compared with that predicted by the
transformed Potts-Guy (Roberts-Sloan) equation (7). Finally,
the results from the series, as well as results from other series
of prodrugs published since the initial publication of the Rob-
erts-Sloan equation, will be added to the Roberts-Sloan da-
tabase to generate a new set of coefficients to the parameters
in the equation and give a more robust Roberts-Sloan
equation.
Topical Delivery of a Model Phenolic
Drug: Alkyloxycarbonyl Prodrugs of
Acetaminophen
Scott C. Wasdo1 and Kenneth B. Sloan1,2
Received January 6, 2004; accepted March 6, 2004
Purpose. To determine whether the delivery of a phenolic parent
drug by its alkyloxycarbonyl (AOC) prodrugs through hairless mouse
skin would show similar dependencies on water and lipid solubilities
that similar prodrugs of more polar heterocyclic amide and imide
parent drugs have shown.
Methods. Flux through hairless mouse skin from suspensions in iso-
propyl myristate (JMIPM), solubilities in IPM (SIPM) and water (SAQ),
and partition coefficients between isopropyl myristate (IPM) and pH
4.0 buffer (KIPM:4.0) were measured for two series of AOC derivatives
of acetaminophen (APAP); their solubilities in pH 4.0 buffer (S4.0
)
were estimated from SIPM/KIPM: 4.0. Log JMIPM values were calculated
from the n ס 43 coefficients for the parameters in the transformed
Potts-Guy (Roberts-Sloan) equation, and the average error of pre-
diction (⌬ log JЈIPM) was calculated. The JMIPM, SIPM, S4.0, and mo-
lecular weight (MW) data for this series and two other series were
combined with the n ס 43 database to give a n ס 61 database, and
new best fit coefficients were determined for the Roberts-Sloan equa-
tion: log JMIPM ס x + y log SIPM + (1 − y) log S4.0 − z MW.
Results. All of the 4-AOC-APAP derivatives underperformed based
on their predicted log JMIPM (⌬ log JЈMIPM ס 0.275 0.147 log units)
and, although the two more water soluble members of this more lipid
soluble series were more effective than APAP, they were only mar-
ginally so: <2 times. Addition of three new series to the n ס 43
database for the Roberts-Sloan equation did not substantially change
the coefficients to the parameters: x, y, z, and r2 ס −0.322, 0.530,
0.00337 and 0.92, respectively.
Conclusions. The topical delivery of a model phenolic drug by its
AOC prodrugs through hairless mouse skin from IPM shows the
same dependence on SIPM, S4.0, and MW as the delivery of polar
heterocycles by their similar prodrugs.
MATERIALS AND METHODS
KEY WORDS: diffusion cell experiments; flux; lipid solubility; pro-
drugs; Roberts-Sloan equation; transformed Potts-Guy equation; wa-
ter solubility.
Melting points were determined with a Meltemp capil-
lary melting point apparatus and are uncorrected. H-NMR
1
spectra were obtained at 90 MHz on a Varian EM-390 spec-
trometer (Palo Alto, CA, USA). Ultraviolet (UV) spectra
were obtained on a Shimadzu UV-265 spectrophotometer
(Kyoto, Japan). The vertical, Franz type diffusion cells were
from Crown Glass (Somerville, NJ, USA) (surface area 4.9
cm2, 20 ml receptor phase volume, 15 ml donor phase vol-
ume). The diffusion cells were maintained at 32°C with a
Fisher (Pittsburgh, PA, USA) circulating water bath model
25. Thin layer chromatography (TLC) analyses were run on
Brinkman Polygram Sil G/UV 254 plates. Isopropyl myristate
(IPM) was obtained from Givaudan (Clifton, NJ, USA). The-
ophylline (Th) was purchased from Sigma Chemical Co. (St.
Louis, MO, USA); all other reagent chemicals were from
Aldrich Chemical Co. (Milwaukee, WI, USA), and all other
solvents were from Fisher. The female hairless mice (SKH-
hr-1) were from Charles River (Boston, MA, USA). The ani-
mal research adhered to the NIH “Principles of Laboratory
Animal Care.”
INTRODUCTION
Although the phenolic functional group is one of the
most common functional groups found in drug molecules,
only one type of promoiety has been extensively synthesized,
characterized, and evaluated in diffusion cell experiments for
its effect on enhancing the topical delivery of drugs containing
phenolic groups: alkylcarbonyl (AC) or ester derivatives (1–
4). These derivatives have mainly been of various narcotic
analgetics such as morphine, buprenorphine, nalbuphine, and
naltrexone because narcotic analgetics are attractive targets
for topical delivery. In most instances, the ester derivatives
1 Department of Medicinal Chemistry, University of Florida, Gaines-
ville, Florida 32610, USA.
2 To whom correspondence should be addressed. (e-mail: sloan@cop.
ufl.edu)
0724-8741/04/0600-0940/0 © 2004 Plenum Publishing Corporation
940

Alkyloxycarbonyl Prodrugs of Acetaminophen
Syntheses
941
stirred in water before application in diffusion cell experi-
ments to be reported elsewhere. Partition coefficients be-
tween IPM and pH 4.0 buffer (KIPM:4.0) were determined as
previously described (8) where the volume ratios between pH
4.0 buffer and IPM (4.0/IPM) were 2, 4, 4, 4, 20, 4, and 4 for
2, 3, 4, 5, 6, 7 and 8, respectively. The solubilities in pH 4.0
buffer (S4.0) were estimated from S4.0 ס SIPM/KIPM:4.0. All
solubilities and partition coefficients were measured in tripli-
cate. The variations in measurements are reported below.
Intact prodrugs and APAP in the diffusion cell receptor
phases were quantitated by UV spectrophotometry at 240 and
280 nm using the ␧ values in Table II and solving the following
To a well-stirred suspension of 4-hydroxyacetanilide
(APAP, 1, 0.01 mol) in 30 ml of CH2Cl2 containing pyridine
(0.01 mol) was added drop-wise a 10 ml CH2Cl2 solution of
the desired chloroformate (0.01 mol). The reaction was al-
lowed to proceed for 2 h, then the solution was diluted to 200
ml with CH2Cl2, extracted with 10 ml 0.6 N HCl and 10 ml
water, dried over Na2SO4 for 2 h, and concentrated using a
rotary evaporator to give a residue. The residue was purified
by recrystallization (or column chromatography in one case
(6)) until a sharp melting point (mp) was observed, only one
1
component was observed on TLC and the H NMR and UV
equations simultaneously: ␧
CA + ␧
CP ס A240 and
A240
P240
spectra were clean of any unreacted APAP. For 6, 7, and 8,
the chloroformate was generated in situ from the reaction of
triphosgene (0.0033 mol) in 20 ml of CH2Cl2 with a 10 ml
CH2Cl2 solution of pyridine (0.01 mol) and the corresponding
alcohol (0.01 mol) added drop-wise over 10 min. In this way,
the following 4-alkyloxycarbonyl (AOC) derivatives of
APAP were synthesized: 4-methyloxycarbonyloxyacet-
anilide (2), 49% from diethyl ether/hexane; 4-ethyloxycar-
bonyloxyacetanilide (3), 82% from ethyl acetate/hexane;
4-propyloxycarbonyloxyacetanilide (4), 59% from ethyl ac-
etate/hexane; 4-butyloxycarbonyloxyacetanilide (5), 63%
from ethyl acetate/hexane; 4-hexyloxycarbonyloxyacetanilide
(6), 51% from ethyl acetate/hexane; 4-(2Ј-methoxyethyloxy-
carbonyloxy) acetanilide (7), 44% from ethyl acetate/hexane;
and 4-(1Ј-methyl-2Ј-methoxyethyloxycarbonyloxy)acetanilide
(8), 29% from diethyl ether/hexane. Melting points (Table I)
and 1H NMR and UV spectra and elemental analyses (Table
II) are given elsewhere.
␧
CA + ␧
CP ס A280, where CA and CP are the con-
A280
P280
centrations of APAP and prodrug, A is the absorbance at
either 240 (A240) or 280 nm (A280), and ␧ is the molar ab-
sorptivity of either APAP (␧A) or prodrug (␧P) at the desig-
nated wavelengths. For the concentration of APAP, the so-
lution is CA ס (␧
␧
A240 − ␧
A280)/(␧
␧
− ␧
P280
P240
A240 P280 P240
A280) and for prodrug it is CP ס (␧
A280 − ␧
A240)/
A240
A280
(␧
␧P280 − ␧
␧A280). The concentrations of both species
A240
P240
were combined to give concentrations of total APAP species.
Theophylline in the receptor phase samples was quanti-
tated by UV spectrophotometry from its absorbance at 270
nm (␧ ס 10,200 L mol−1) in pH 7.1 phosphate buffer contain-
ing 0.11% formaldehyde.
The donor phase compositions at the end of the first
application period were determined by 1H NMR spectros-
copy. The donor phases were removed, filtered, and the resi-
1
dues analyzed by H NMR. In DMSO-d6, the aromatic C–H
absorptions were centered at ␦ 7.10 and 7.57 for the prodrugs,
whereas those for APAP were centered at ␦ 6.63 and 7.30.
Physicochemical Properties and Analysis
Solubilities in IPM (SIPM) were determined as previously
described (8) using 72 h stirring of suspensions before filtra-
tion, dilution of the filtrate with CH3CN, and quantitation by
Diffusion Cell Experiments
The diffusion cell experiments were run in essentially the
UV spectrophotometry from the absorbances of 1 to 8 at 240 same way as previously described (10). Briefly, female hair-
nm using their molar absorptivities (␧) in Table II. Solubilities less mice were sacrificed by cervical dislocation. Their skins
in water (SAQ) were determined as above except using only 1 were removed by blunt dissection and placed epidermal side
h stirring of suspensions before filtration (9). A 1 h stirring up in contact with pH 7.1 phosphate buffer (0.05 M, I ס 0.11
time was used to match the time samples of the prodrugs were M, 32°C) containing 0.11% formaldehyde (2.7 ml of 36%
Table I. Molecular Weights (MW), Melting Points (mp°C), Log Solubilities in Isopropyl Myristate (IPM) (log
SIPM), Log Solubilities in Water (log SAQ), and Log Estimated Solubilities in pH 4.0 Buffer from Log SIPM − log
KIPM:4.0 for the 4-Alkyloxycarbonyl Prodrugs of 4-Hydroxyacetanilide (APAP)
e
e,f
g
Compounda
MW
mp°Cd
log SIPM
log SAQ
0.45
log S4.0
1, APAP
2, C1
3, C2
4, C3
5, C4
151
209
223
237
251
279
253
267
167–170
0.28
1.08
0.97
1.37
1.14
1.22
1.01
0.53
112–115 (115.5–116.5)
120–122 (121–122)
104–106 (105–108)
118–120 (119–121)
108–110 (112.5–113.5)
78–81
1.31 (1.46)
0.58 (0.69)
0.43
−0.37 (−0.20)
−1.32 (−0.88)
1.54
1.23
0.65
0.48
−0.35
−1.49
1.31
6, C6
7,b MOC2
8,c MOC3i
120–123
0.52
0.39
a C1, C2 . . . refer to the chain length of the alkyl group.
b 2-Methoxyethyloxycarbonyl, CH3OCH2CH2OCסO.
c 2-Methoxy-1-methylethyloxycarbonyl, CH3OCH2CH(CH3)OCסO.
d All mp are uncorrected. Literature references are in parentheses: 2, 3, 5, and 6 are from Ref. 6; 4 from Ref. 14.
e Units of mM measured at 23 1°C
f Literature references are in parentheses from Ref. 6, measured at 37°C.
g Log of the solubility in pH 4.0 buffer estimated from log SIPM − log KIPM:4.0
.

942
Wasdo and Sloan
Table II. Molar Absorptivities (⑀), Chemical Shifts (␦), and Elemental Analyses for 1–8
Buffera,b
1H NMRc
CH3CNa
c
Compound
␧
␧
␧
280
CH3CסOd
C6H4
OסCOCH2Xf
240
240
1, APAP
2
3
4
1.36
1.67
1.64
1.63
1.75
1.79
1.74
1.60
0.98
1.25
1.24
1.28
1.24
1.12
1.26
1.28
0.191
0.0560
0.0483
0.0560
0.0376
0.0623
0.0623
0.0520
2.15
2.12
2.08
2.14
2.14
2.14
2.08
7.06, 7.45
7.03, 7.42
7.04, 7.42
7.08, 7.45
7.04, 7.42
7.06, 7.46
7.04, 7.42
s, 3.95
q, 4.28
t, 4.20
t, 4.28
t, 4.20
t, 4.38
m, 5.00
5
6
7g
8h
a Units of 1 × 104 L/mole.
b Buffer: pH 7.1 phosphate buffer with 0.11% formaldehyde.
c Spectra run in CDCl3. APAP, 1, was not sufficiently soluble in CDCl3 to obtain a spectrum in CDCl3.
d Singlet (s).
e Two doublets (d) each with J ס 6 Hz centered at designated positions.
f Four absorptions (quartet, q) with J ס 5 Hz; three absorptions (triplet, t) with J ס 5 Hz; a doublet of quartets or multiplet (m).
g Anal calcd for C12H15NO5: C, 56.91; H, 5.97; N, 5.53. Found: C, 56.84; H, 5.98; N, 5.53.
h Anal calcd for C13H17NO5: C, 58.42; H, 6.41; N, 5.24. Found: C, 58.43; H, 6.43; N, 5.27.
aqueous formaldehyde/liter) to prevent microbial growth and mined by plotting the cumulative amount (␮mol) of total
to insure the integrity of the mouse skins during the course of APAP species or Th measured in the receptor phase against
the experiment (11). The flux of theophylline from propylene time and dividing the steady-state portions of those plots by
glycol (Th/PG) through hairless mouse skin after 4, 24, 48, the surface area of the diffusion cells.
and 120 h of contact with the receptor phase did not change
Solubility Parameters
significantly: 0.0061 0.0006, 0.0083 0.0019, 0.0094 0.0012,
and 0.010
0.0012 ␮mol cm–2 h–1, respectively. The skins
The solubility parameters were obtained using the
method of Fedors (12) as illustrated by Martin et al. (13) and
Sloan et al. (10).
were kept in contact with the buffer for at least 48 h to con-
dition the skins and to allow UV absorbing materials to leach
out of the skins. The receptor phases were changed at least
three times during this time to facilitate the leaching process,
and the UV spectra of the receptor phase at the of this period
was clear of UV absorptions in the 220–350 nm range.
Aliquots (0.5 ml, 0.2–0.3 M) of suspensions of each pro-
drug in IPM were applied to the donor side of each of three
diffusion cells in the first application. After the donor phase
suspensions were applied, 3 ml samples of the receptor phases
were removed, generally at 8, 19, 22, 25, 28, 31, 34, 44, and 48
h. The entire receptor phase was replaced with fresh receptor
fluid each time a sample was removed. The amount of APAP
and prodrug in each sample was determined immediately by
Statistical Analyses and Regression Analyses
Statistical analysis was accomplished using Student t test.
Unless otherwise indicated, statistical significance is for p <
0.05. Multiple linear regression analysis was accomplished us-
ing the SPSS 7.5 statistical software package.
RESULTS AND DISCUSSION
Syntheses
The AOC derivatives of APAP were synthesized from
UV spectrophotometry using the equations described above. the corresponding commercially available chloroformates us-
After the initial application period of 48 h (96 total h) ing an organic base, usually pyridine, as an acid scavenger.
and the donor phases had been removed, the donor surfaces For the hexyl-, 6, 2-methoxyethyl-, 7, and 2-methoxy-1-
were quickly washed with 3 × 2 ml portions of methanol to methylethyloxycarbonyl, 8, derivatives, the chloroformates
removed any residual prodrug or APAP. After the methanol were generated in situ from triphosgene, pyridine, and the
wash, the skins were kept in contact with fresh receptor fluid corresponding commercially available alcohols, then allowed
for 23–24 h to allow any APAP or prodrugs to leach out. to react with APAP as above. The yields of analytically pure
Samples of the receptor phases were removed and analyzed derivatives, which were used in the diffusion cell experiments
for total APAP species as above. The receptor phases were and in the solubility determinations, were in the range of
replaced with fresh receptor fluid, and 0.5 ml aliquots of a 42–82% except for 8, which was only 29%. Five of the de-
standard drug vehicle (Th/propylene glycol, PG) were applied rivatives had been previously reported, 2 to 6 (6,14), and the
(120 total h). Samples of the receptor phases (3 ml) from this uncorrected melting points reported here were essentially
second application were removed at 1, 2, 3, and 4 h, and the identical with those previously reported (Table I). For the
amounts of theophylline in the receptor phases were quanti- two new derivatives, 7 and 8, elemental analyses were 0.40%
tated by UV spectrophotometry as above. Each time a sample (Table II). The UV and 1H NMR spectra were consistent with
was removed, the entire receptor phase was replaced with 4-AOC-APAP structures. When 1H NMR spectra were run in
fresh receptor fluid.
DMSO-d6, the upfield doublet in the aromatic region was
In all cases, the rates of delivery of total APAP species shifted farther downfield than was the downfield doublet for
(JMIPM) or theophylline (JJIPM) through skin were deter- the derivatives compared to APAP, as expected if the 4-OH

Alkyloxycarbonyl Prodrugs of Acetaminophen
943
Table III. Log Solubility Ratios Between Isopropyl Myristate (IPM)
and Water (log SRIPM:AQ), the Differences Between Log SRIPM:AQ
(␲SR), the Log of Partition Coefficients Between IPM and pH 4.0
Buffer (log KIPM:4.0), and the Differences Between Log KIPM:4.0 (␲K)
and Calculated Solubility Parameters (␦i)
group in APAP was acylated. The molar absorptivities of the
derivatives (␧ values in Table II) were also higher. The 1H
NMR spectra of the alkyl portions of the AOC derivatives
were also consistent with the proposed structures where the
O ס C-OCH2X absorption was shifted downfield compared to
the HOCH2X absorption in the corresponding alcohol by the
electron withdrawing effect of the carbonyl group in the
AOC. The spectra were also internally consistent; the values
for ␧ and the chemical shifts were reasonably constant from
derivative to derivative.
b
SR
d
K
e
loga SRIPM:AQ
␲
logc KIPM:4.0
␲
␦
i
2
3
4
5
6
7
8
−0.24
0.39
0.94
1.52
2.54
−0.52
0.01
−0.16
0.32
0.90
1.50
2.71
−0.30
0.13
12.09
11.80
11.55
11.35
10.99
11.62
11.31
0.63
0.55
0.58
0.51
0.48
0.58
0.60
0.60
Solubilities
0.53
0.43
The solubilities in isopropyl myristate (SIPM) and water
(SAQ) in Table I were all determined at room temperature
using UV spectroscopy and ␧ values in acetonitrile (Table II).
The relative SD of the SAQ values were all less than 5%
except for 4 ( 7.1%) and 6 ( 9.4%), and those values of the
previously (6) characterized derivatives (2, 3, 5, and 6) were
reasonably consistent with the previous SAQ values (Table I),
considering that the previous values were determined at 37°C
and so are all higher: 40 10% for 2, 3, and 5, and 280% for
6. There are not great differences among the melting points of
the derivatives (Table I), and there are not great differences
among the SIPM values (Table I), which were all < 5%. How-
ever, the mp for 3 (C2) and 5 (C4) are each higher than the
derivatives on either side, and 3 and 5 are each less soluble in
IPM than 2 and 4 or 4 and 6, respectively. The derivative 4
exhibiting the lowest mp for the straight chain derivatives 2 to
6 also exhibited the highest SIPM. Similar results can be seen
for the short series where an oxygen has been inserted in the
alkyl chain, 7 and 8; 7 exhibited the lowest mp and the highest
a Log of the ratio of the solubilities in IPM (SIPM) and water (SAQ).
␲ ס (log SRn + m − log SRn)/m, where n is the number of methylene
b
units in the promoiety of one prodrug and m is the number of
additional methylene units in the promoiety with which it is com-
pared.
c Log of the partition coefficient between IPM and pH 4.0 buffer.
d The same definition as b above except that K is substituted for SR.
e Calculated according to Ref. 12. Units of (cal cm−3 1/2
) .
Analyses and Stability
The 4-AOC-APAP derivatives were stable in the IPM
donor phases. When the residues from each of the donor
1
phases was filtered from the IPM and analyzed by H NMR
spectroscopy, only intact prodrug was observed based on the
absence of the aromatic C–H absorptions due to APAP and the
correct ratios between CH3(CסO)NH and O(CסO)OCH2X
absorptions. The 4-AOC-APAP derivatives should also be
stable in the receptor phase. Previously determined t1/2 values
at pH 7.4 and 37°C were 150, 200, and 300 h for 2, 3, and 5,
respectively (6). On the other hand, the t1/2 values for the
same derivatives in pH 7.4 buffer 2% human plasma (V/V)
were 180, 45, and 15 min (6), respectively, so the derivatives
should hydrolyze during permeation of the skin to some degree.
S
IPM. The values for SAQ for both series decreased as the
number of carbons in the alkyl chain increased, although the
AQ value for 3 appeared to decrease disproportionately fast
S
compared to 2 and 4.
Regardless of the irregular behavior of the absolute solu-
bilities, the ratios of the solubilities in IPM and AQ (SRIPM:AQ
were reasonably well behaved. The average methylene ␲SR in
Table III was 0.57 0.06 for the first series and the only
methylene ␲SR for the second series was 0.54, which was well
)
Diffusion Cell Experiments
The values for the delivery of total APAP containing
within the SD of the average for the first series. The SD for species from IPM suspensions of the 4-AOC-APAP deriva-
experimentally determined partition coefficients between tives (JMIPM) are given in Table IV. All of the JMIPM values
IPM and pH 4.0 buffer (KIPM:4.0, Table III) were all less then were well within the expected 30% variation in J values seen
10%. The average methylene ␲ values from KIPM:4.0 were for in vitro hairless mouse skin diffusion cell experiments
K
also reasonably well behaved for the first series (␲K ס 0.55
(16.6 6.7%) except for APAP itself, which was 37%. Only
0.06) and are consistent with methylene ␲K values determined two of the derivatives performed better than APAP itself and
for other series of prodrugs (9). However the only ␲ value then only marginally: 2 (1.7×) in the first series and 7 (1.4×) in
K
for the second series was not consistent with that for the first the second series. Although all of the derivatives were more
series (␲ ס 0.44). All K values were determined by parti- soluble in IPM than APAP, as seen for previous series of
K
tioning an IPM solution against pH 4.0 buffer to be consistent derivatives the derivative that gave the highest rate of deliv-
with previous determinations of K and to depress ionization ery was not the most lipid-soluble member. Instead it was that
in the aqueous phase.
member of each more lipid soluble series that also exhibited
In order to be consistent with data from the transformed the higher water solubility that gave the highest rate of de-
Potts-Guy database, which consisted primarily of log SAQ val- livery (15–17). Compared to increases in SIPM seen for the
ues estimated from log SIPM − log KIPM:4.0, estimated log SAQ best derivatives of drugs containing polar heterocyclic NH
values were also calculated from the data in Tables I and III, groups (>100×), the SIPM values measured for the 4-AOC
respectively, and are given as log S4.0 values in Table III. The derivatives were only 8× and 5× for the best performing mem-
estimated S4.0 values were somewhat higher than those for the ber of each series (2 and 7). On the other hand, the increases
directly measured values for 3, 4, and 5 (12 7%) and some- in SAQ values seen for the 4-AOC derivatives were greater
what lower for 2, 6, 7, and 8 (29 9%). For both series (2 to than those measured for most of the best derivatives of drugs
8), the variation was 21 12%.
containing heterocyclic NH groups previously studied (<2×):

944
Wasdo and Sloan
Table IV. Log Values for Experimental Fluxes of Total 4-Hydroxyacetanilide (APAP) Species from Suspensions of 1 to 8
in Isopropyl Myristate (IPM) Through Hairless Mouse Skin (log JMIPM), Error in Predicting Log JMIPM from Transformed
Potts-Guy Equation (n ס 43) (⌬ log JЈMIPM), Error in Calculating Log JMIPM from transformed Potts-Guy Equation (n ס
61) (⌬ log JMIPM), Values for the Fluxes of Theophylline from PG as a Second Application (JJIPM), and Log Permeability
Coefficients (log PMIPM
)
a
b
c
a
d
Compound
log JMIPM
⌬ log JЈMIPM
⌬ log JMIPM
JJIPM
log JMIPM
1, APAP
2
3
4
−0.29
−0.00
−0.76
−0.45
−1.01
−1.49
−0.11
−1.06
0.11
0.18
0.56
0.33
0.33
0.22
0.13
0.34
0.18
0.12
0.51
0.28
0.28
0.18
0.09
0.30
0.74
1.12
0.64
1.14
0.85
0.76
0.98
0.94
1.02
−0.57
−1.08
−1.73
−1.82
−2.15
−2.71
−1.12
−1.59
5
6
7
8
Controle
Controlf
Controlg
0.013
0.015
a Units of ␮mole cm−2 h−1
b Predicted log JMIPM ס −0.216 + 0.534 log SIPM + 0.466 log S4.0 − 0.00361 MW from Ref. 7 (n ס 42, + C5 bis-6,9-ACOM-
.
6-MP). Error in prediction was from experimental log JMIPM − predicted log JMIPM
.
c Calculated log JMIPM ס −0.322 + 0.53 log SIPM + 0.47 log S4.0 − 0.00337 MW from Ref. 7 (n ס 42, + C5 bis-6,9-ACOM-
6-MP), Ref. 9 (Table V, n ס 4), Ref. 19 (Table V, n ס 6), and current data (n ס 8) to give n ס 61. Error in calculation
was from experimental log JMIPM − calculated log JMIPM
.
d Values calculated from log JMIPM − log SIPM ס log PMIPM, units of cm h−1
.
e 48 hours conditioning, 48 hours IPM treatment, methanol wash, 24 hours leaching (Ref. 10).
f 48 hours conditioning, 48 hours no treatment, methanol wash, 24 hours leaching (Ref. 21).
g 48 hours conditioning, 48 hours Th/PG treatment, methanol wash, 24 hours leaching (ref. 21).
7× for 2 and 15× for 7. Thus, the limited increases in delivery the permeability of the skins. The fluxes of suspensions of
of total APAP containing species seems to have been caused theophylline from propylene glycol (Th/PG) applied subse-
by the lower than expected SIPM values. As mentioned above, quent to the application of the prodrugs and APAP in IPM
3 and 5 exhibited higher mp and lower SIPM values than 2 and (JJIPM, Table IV) were not significantly different from the
4 or 4 and 6, respectively. Derivative 3 also exhibited a rela- flux of Th/PG applied after IPM itself: JJIPM ס 0.89 0.18 vs.
tively lower SAQ value than expected. Thus it is not surpris- 1.02 0.13, respectively. Thus, using the second application of
ing, based on analogy to previously studied series (15–17), a standard solute/vehicle (Th/PG) as a measure of changes in
that 3 and to some extent 5 performed much worse (lower the resistance of the skin to permeation (10), there was no
J
MIPM values) than the contiguous members of the series 2, 4, difference in the resistance of the skins to permeation caused
and 6. by the application of 1 to 8 in IPM compared to application of
The percent of intact prodrug found in the receptor IPM alone. Also, normalization of the JMIPM values by the
phase was in direct proportion to the flux values obtained respective JJIPM values did not change the rank order of the
from the steady-state region: 25–34 h. The prodrug giving the performances of 1 to 8 (data not shown).
highest rate of delivery of total APAP-containing species de-
livered the greatest percentage of intact prodrug in each se- dividing by their respective SIPM values (log JMIPM –log SIPM
ries using the data for the 31 h sample, which was represen- to give their corresponding permeability coefficients (log
tative of values obtained at the other times when samples MIPM, Table IV), the PMIPM values decreased as their re-
When the JMIPM values for 1 to 8 were normalized by
)
P
were analyzed. For the first series, the percentages of intact spective KIPM:4.0 values increased (Table III, Fig. 1) and their
C1 to C6 in the receptor phase were 64, 14, 25, 0, and 0%, respective calculated solubility parameter values decreased
respectively. The fact that the C2 derivative (3) gave a lower (␦i, Table III, Fig. 2) for the two series. A plot of log PMIPM
JMIPM value than the C1 or C3 (2 or 4) derivatives was re- vs. ␦i values for the first series (2 and 6) gave a positive
flected in the lower amounts of intact prodrug measured in relationship (slope ס 1.40, r ס 0.98), which is characteristic
the receptor phase. Similarly, for the second series the per- of relationships between PMIPM and ␦i previously seen for
centages of intact prodrug were 51% and 0% for 7 and 8, other series of more lipophilic prodrugs (19). However, in
respectively. This trend in both series is probably the result of those previous examples, the more lipophilic prodrugs were
the flux being too great for the hydrolytic enzyme system to of polar heterocyclic parent drugs derivatized on their amide-
hydrolyze the prodrugs completely. A similar result was ob- or imide-like NH functional groups and not of a more lipo-
served for the alkyloxycarbonyl derivatives of 5-fluorouracil philic phenolic parent drug. Thus, although there is no readily
where up to 90% of intact prodrug was obtained in the re- apparent trend in the decreasing JMIPM values (and conse-
ceptor phases at steady-state for the most effective prodrug quently decreasing PMIPM values) with increasing SIPM values
(18).
because the SIPM values are fairly flat, other measures for
The differences in the JMIPM values for 1 to 8 were not increasing lipophilicity, such as steadily increasing KIPM:4.0
dependent on different effects of the prodrugs and APAP on and decreasing ␦i with increasingly longer alkyl chain length,

Alkyloxycarbonyl Prodrugs of Acetaminophen
945
Table V. Molecular Weights (MW), Log Solubilities in Isopropyl
Myristate (IPM) (log SIPM), Estimated Log Solubilities in pH 4.0
Buffer (log S4.0), from Log SIPM − Log Partition Coefficients Between
IPM and pH 4.0 Buffer (log KIPM:4.0), and Log Fluxes from Suspen-
sion in IPM Through Hairless Mouse Skin (log JMIPM) for 3-Alkyl-
carbonyl-5-FU (9–12) (Ref. 9) and 3-Alkylcarbonyloxymethyl-5-FU
(13–18) (Ref. 19)
b
b,c
d
Compounda
MW
log SIPM
log S4.0
log JMIPM
9, C1
172
186
200
214
202
216
230
244
258
280
0.63
1.15
1.34
0.96
0.09
1.20
1.42
1.47
1.63
1.60
2.02
2.13
1.36
0.74
1.71
2.23
1.87
1.34
0.89
−0.29
0.64
0.72
0.34
−0.26
−0.22
0.34
0.46
0.12
0.00
−0.77
10, C2
11, C3
12, C4
13, C1
14, C2
15, C3
16, C4
17, C5
18, C7
a C1, C2 . . . refer to the chain length of the alkyl group.
b Units of mM.
Fig. 1. Plot of log KIPM:4.0 vs. log PMIPM for the 2–6 (diamonds) and
7, 8 (circles) prodrugs of APAP.
c Estimated from log SIPM − log KIPM:4.0
.
d Units of ␮mole cm−2 h−1
.
identify the existence of such a trend. Increasing lipophilicity
theophylline, 5-fluorouracil, and 6-mercaptopurine to give
n ס 43 and slightly different coefficients for x, y, and z: −0.216,
0.534, and 0.00361, respectively, for the n ס 43 coefficients.
Using the latter coefficients and the values for log SIPM, log
leads to lower JMIPM and PMIPM
.
In order to quantify the relationship between JMIPM
,
solubilities (S4.0 and SIPM) and molecular weight (MW), Rob-
erts and Sloan (7) recently developed a transformation of the
Potts-Guy equation that accommodated flux values from ve-
hicles other than water and explicitly contained both lipid
(SIPM) and aqueous (S4.0) parameters: log JMIPM ס x + y log
SIPM + (1 – y) log S4.0 – z MW. The coefficients for x, y, and
z had values of –0.211, 0.535, and 0.00364, respectively, for the
fit of the original database minus the data for 6-mercaptopu-
rine (n ס 42) to the Roberts-Sloan equation. Data for one
more member from one of the original series (bis-6,9-
hexanoyloxymethyl-6-mercaptopurine) was subsequently
added to the original database of seven series of prodrugs of
S
4.0, and MW (Tables I and III), log JMIPM values for the
current series were predicted (data not shown), and the val-
ues for the error of prediction (experimental log JMIPM
−
predicted log JMIPM ס ⌬ log JЈMIPM) were calculated (Table
IV). The average ⌬ log JЈMIPM for the 4-AOC-APAP series
was 0.27 0.15 log units, which is quite a bit larger than the
average ⌬ log JMIPM value (see below) for any of the other
series in the original database. In addition, all the derivatives
performed worse than predicted using the n ס 43 coefficients.
The current 4-AOC-APAP series was not the only series
that uniformly performed worse than predicted using the n ס
Fig. 2. Plot of solubility parameter vs. log PMIPM for the 2–6 (dia- Fig. 3. Calculated vs. experimental flux through hairless mouse skin
monds) and 7, 8 (circles) prodrugs of APAP. for the 61 compound date set.

946
Wasdo and Sloan
2. A. L. Stinchcomb, A. Paliwal, R. Dua, H. Imoto, R. W. Woodard,
and G. L. Flynn. Permeation of buprenorphine and its 3-alkyl-
ester prodrugs through human skin. Pharm. Res. 13:1519–1523
(1996).
3. K. C. Sung, R.-Y. Han, O. Y. P. Hu, and L.-R. Hsu. Controlled
release of nalbuphine prodrugs from biodegradable polymeric
matrices: influences of prodrug hydrophilicity and polymer com-
position. Int. J. Pharm. 172:17–25 (1998).
4. A. L. Stinchcomb, P. Swaan, O. Ekabo, K. K. Harris, J. Browe, P.
D. Hammell, T. A. Cooperman, and M. Pearsall. Straight-chain
naltrexone ester prodrugs: diffusion and concurrent esterase bio-
transformation in human skin. J. Pharm. Sci. 91:2571–2578
(2002).
5. K. B. Sloan, S. Wasdo, U. Ezike-Mkparu, T. Murray, D. Nickels,
S. Singh, T. Shanks, J. Tovar, K. Ulmer, and R. Waranis. Topical
delivery of 5-fluorouracil and 6-mercaptopurine by their alkyl-
carbonyloxymethyl prodrugs from water: vehicle effects on de-
sign of prodrugs. Pharm. Res. 20:639–645 (2003).
6. L. W. Dittert, H. C. Caldwell, H. J. Adams, G. M. Irwin, and J. V.
Swintosky. Acetaminophen prodrugs I. Synthesis, physicochem-
ical properties and analgesic activity. J. Pharm. Sci. 57:774–780
(1963).
7. W. J. Roberts and K. B. Sloan. Correlation of aqueous and lipid
solubilities with flux for prodrugs of 5-fluorouracil, theophylline
and 6-mercaptopurine: a Potts-Guy approach. J. Pharm. Sci. 88:
515–532 (1999).
8. H. D. Beall, J. J. Getz, and K. B. Sloan. The estimation of relative
water solubility for prodrugs that are unstable in water. Int. J.
Pharm. 93:37–47 (1993).
9. H. D. Beall and K. B. Sloan. Topical delivery of 5-fluorouracil
(5-FU) by 3-alkylcarbonyl-5-FU prodrugs. Int. J. Pharm. 217:127–
137 (2001).
43 coefficients; the 3-alkylcarbonyloxymethyl-5-fluorouracil
(3-ACOM-5-FU) series did as well (19), where the average
error of predicting log JMIPM was ⌬ log JЈMIPM ס 0.19 0.09
log units. In order to update the original database for the
Roberts-Sloan equation and also to see what effect the inclu-
sion of two new series, both of which underperformed, in the
database would have on the calculated performance of the
other series, the members of the 4-AOC-APAP series, the
3-ACOM-5-FU series, (19) and the 3-alkylcarbonyl-5-FU (9)
series (Table V) were added to the original database to give
n ס 61, and new coefficients for the Roberts-Sloan equation
were calculated. The new values for the x, y, and z coefficients
were –0.322, 0.530, and 0.00337, respectively, for the n ס 61
database (Fig. 3). The average error in calculating log JMIPM
(experimental log JMIPM – calculated log JMIPM ס ⌬ log
J
MIPM) was 0.15 0.11 log units for n ס 61, which was some-
what larger than the average ⌬ log JMIPM for the n ס 43
database: 0.13 0.09 log units using the n ס 43 coefficients.
Although the average error in calculating log JMIPM did not
change more than 0.01 log units for most of the series, there
were three series that did change substantially. The average ⌬
log JMIPM for the 1-alkylcarbonyloxymethyl-5-FU series (1-
ACOM-5-FU) worsened from 0.12 to 0.17 log units, whereas
the average ⌬ log JMIPM for the 3-ACOM-5-FU and the
4-AOC-APAP series improved from 0.19 to 0.15 and from
0.27 to 0.24 log units, respectively.
10. K. B. Sloan, S. A. M. Koch, K. G. Siver, and F. P. Flowers. The
use of solubility parameters of drug and vehicle to predict flux. J.
Invest. Dermatol. 87:244–252 (1986).
11. K. B. Sloan, H. D. Beall, W. R. Weimar, and R. Villaneuva. The
effect of receptor phase composition on the permeability of hair-
less mouse skin in diffusion cell experiments. Int. J. Pharm. 73:
97–104 (1991).
12. R. F. Fedors. A method for estimating both the solubility param-
eters and molar volumes of liquids. Polym. Eng. Sci. 14:147–154
(1974).
13. A. Martin, P. L. Wu, and T. Velasquez. Extended Hildebrand
solubility approach: sulfonamides in binary and ternary solvents.
J. Pharm. Sci. 74:277–282 (1985).
14. E. Merck. Zur kenntnis der einwirkung von phosgen bxw
chlorkohlensaure ester auf p-acetylaminophenole und p-
oxyphenylurethane. Chem. Zentralbl. I:468–469 (1897).
15. K. B. Sloan. Prodrugs for dermal delivery. Adv. Drug Del. Rev.
3:67–101 (1989).
16. K. B. Sloan. Functional group considerations in the development
of prodrug approaches to solving topical delivery problems. In K.
B. Sloan (ed.), Prodrugs. Topical and Ocular Drug Delivery, Mar-
cel Dekker, New York, 1992, pp. 17–116.
17. K. B. Sloan and S. Wasdo. Designing for topical delivery: pro-
drugs can make the difference. Med. Res. Rev. 23:763–793 (2003).
18. H. D. Beall, R. J. Prankerd, and K. B. Sloan. Transdermal deliv-
ery of 5-fluorouracil (5-FU) through hairless mouse skin by 1-al-
kyloxycarbonyl-5-FU prodrugs: physicochemical characterization
of prodrugs and correlations with transdermal delivery. Int. J.
Pharm. 111:223–233 (1994).
19. W. J. Roberts and K. B. Sloan. Topical delivery of 5-fluorouracil
(5-FU) by 3-alkylcarbonyloxymethyl-5-FU prodrugs. J. Pharm.
Sci. 92:1028–1036 (2003).
CONCLUSIONS
The delivery of the total acetaminophen (APAP) con-
taining species (JMIPM) by its alkyloxycarbonyl (AOC) de-
rivatives from IPM, although reasonably predicted by the
Roberts-Sloan equation, did not give substantially higher
JMIPM values than by APAP. Again, the more water soluble
members of this more lipophilic series were the most effective
at enhancing delivery of total APAP species through mouse
skin from IPM (15–17). The addition of the 4-AOC-APAP
series data and that of two other series to the database for the
Roberts-Sloan equation and the subsequent fitting of all the
data (n ס 61) to the Roberts-Sloan equation resulted in the
determination of new coefficients for the parameters that are
not substantially different from the previous coefficients, but
do account for the two underperforming series (3-ACOM-5-
FU and 4-AOC-APAP) better. Although these results, which
reinforce the importance of SAQ on predicting flux from a
lipid vehicle, are based on in vitro hairless mouse data, re-
gression analysis of in vivo data for the delivery of nonsteroi-
dal anti-inflammatory drugs from mineral oil through human
skin also showed a significant and substantial dependence on
S
AQ (0.28 vs. 0.47 for hairless mouse skin data) (20). Regard-
less of the difference in degree of hydration between in vitro
and in vivo, and between hairless mouse skin and human skin,
water solubility is an important parameter in predicting flux
from lipid vehicles such as IPM and mineral oil.
20. W. J. Roberts and K. B. Sloan. Application of the transformed
Potts-Guy equation to in vivo human skin data. J. Pharm. Sci.
90:1318–1323 (2001).
REFERENCES
1. J. Drustrup, A. Fullerton, L. Christrup, and H. Bundgaard. Uti- 21. K. B. Sloan, H. B. Beall, H. E. Taylor, J. J. Getz, R. Villaneuva,
lization of prodrugs to enhance the transdermal absorption of
morphine. Int. J. Pharm. 71:105–116 (1991).
R. Nipper, and K. Smith. Transdermal delivery of theophylline
from alcohol vehicles. Int. J. Pharm. 171:185–193 (1998).