Percutaneous absorption of cyclizine and its alkyl analogues

Article abstract of DOI:10.1016/j.ejps.2004.11.001

Cyclizine (I) alkyl analogues (II-IV) were synthesized and their skin permeation parameters evaluated in vitro. It was hoped that these compounds would possess physicochemical properties more favourable for percutaneous delivery than (I). The identificati

Full text of DOI:10.1016/j.ejps.2004.11.001

European Journal of Pharmaceutical Sciences 24 (2005) 239–244
Percutaneous absorption of cyclizine and its alkyl analogues
L. Mishack Monenea, Colleen Goosenc, Jaco C. Breytenbachb,
Jonathan Hadgraftd, Jeanetta du Plessisa,∗
a
Department of Pharmaceutics, School of Pharmacy, North-West University, Private Bag X6001, Box16 Potchefstroom Campus,
Potchefstroom 2520, South Africa
b
Department of Pharmaceutical Chemistry, School of Pharmacy, North-West University, Potchefstroom 2520, South Africa
c
University of Michigan, College of Pharmacy, 428 Church street, Ann Arbor, MI 48109-1065, USA
The School of Pharmacy, University of London, 29-39 Brunswick Sq., London WC1N 1AX, UK
d
Received 25 October 2004; received in revised form 25 October 2004; accepted 3 November 2004
Available online 15 December 2004
Abstract
Cyclizine (I) alkyl analogues (II–IV) were synthesized and their skin permeation parameters evaluated in vitro. It was hoped that these
compounds would possess physicochemical properties more favourable for percutaneous delivery than (I). The identification and levels
of purity for the compounds were confirmed by mass spectrometry (MS), nuclear magnetic resonance (NMR) spectrometry, and infrared
spectrometry (IR) while melting points were determined by an electrothermal digital Bv¨chi melting point apparatus. Aqueous solubilities
(25 ◦C) and partition coefficients were determined and in vitro permeation studies were performed in buffer (37 ◦C) at pH 7.4 over a period
of 24 h, using Franz diffusion cells fitted with human epidermal membranes. Generally, the analogues were more lipophilic, but nevertheless
possessed higher aqueous solubilities as compared to (I). (II) and (IV) exhibited two- to three-fold increase in aqueous solubility and their
melting temperatures dropped by more than 55 ◦C. Compound (III) had similar aqueous solubility to (I), but its melting point dropped by about
35 ◦C. Measured steady-state fluxes indicated that (II) is a far better penetrant (J = 6.95 ␮g/cm2/h) of human epidermis than (I). Although
fluxes of (III) and (IV) drop off markedly from that of (II), they remained above the flux of (I), which is (0.132 ␮g/cm2/h). In conclusion,
(II) was the best skin permeant and also exhibited the highest aqueous solubility and lowest level of crystallinity as compared to (I) and other
analogues. (III) and (IV) were more lipophilic. The overall permeation data of this series indicated that the more water-soluble and the lowest
melting point compound was the best skin permeant.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Cyclizine; Alkyl analogues; Physicochemical properties; Percutaneous delivery
1. Introduction
treatment and prevention of motion sickness, post-operative
vomiting, and Menieres disease (Dundee and Jones, 1968).
The synthesis of cyclizine was reported by Baltzly et al.
(1949). Its anti-histaminic action was discovered by Castillo
et al. (1949) and it was reported to be one-fourth as active as
itscongenerchlorcyclizineinblockingthehistamine-induced
spasm of the tracheal chain preparation. The use of cyclizine
hydrochloride for the prevention of seasickness and airsick-
ness has been described by Chinn et al. (1952, 1953). Gutner
et al. (1954), using microcaloric and galvanic stimulation
methods, found that cyclizine notably decreased labyrinthine
sensitivity in human subjects and that cyclizine alleviated
post-operative nausea and vomiting. The same investiga-
Cyclizine is a piperazine derivative, which belongs to the
anti-histamine group of drugs. It is described chemically
as 1-(diphenylmethyl)-4-methylpiperazine. It acts both on
the emetic trigger zones and by damping the labyrinthine
sensitivity. Pharmacologically it has anti-histaminic, anti-
serotonic, local anesthetic, and vagolytic actions (Susan et al.,
1989). Therapeutically it is an anti-emetic agent, the nor-
mal dose being 50 mg for 4–6 h. It is recommended in the
∗
Corresponding author. Tel.: +27 18 299 2274; fax: +27 18 299 2225.
E-mail address: fmsjdp@puk.ac.za (J.d. Plessis).
0928-0987/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejps.2004.11.001

240
L.M. Monene et al. / European Journal of Pharmaceutical Sciences 24 (2005) 239–244
tion also found that the drug partially antagonized vom-
iting induced by the administration of apomorphine to
dogs.
grade double deionized water and n-octanol (BDA labora-
tory suppliers, Poole, England) were used. For permeability
studies, female human abdominal skin was obtained from
Klerksdorp Surgical Clinic (South Africa).
A number of drugs readily passes through the skin and
the rate and extent to which this happens is influenced by
the physicochemical properties of the drug (Beckett, 1982).
Other factors may also have an influence, but if kept constant,
it is possible to determine which physicochemical properties
are most important in determining the rate and extent of ab-
sorption through or into the skin. Recent studies revealed that
alkylation approach to improve the dermal delivery of drugs
offers several advantages, since it changes the physicochem-
ical properties of the drug (i.e. aqueous solubility, lipophilic-
ity, and level of crystallinity). It is important to take into con-
sideration both the relative potency of the analogue and its
ability to permeate the skin. For example, it would be better
to select an active with a 20-fold lower potency but a 100-fold
better flux.
2.1. General synthesis method
R = CH3
1-(diphenylmethyl)-4-methylpiperazine (I)
1-(diphenylmethyl)-4-ethylpiperazine (II)
R = CH2CH3
R = (CH2)2CH3 1-(diphenylmethyl)-4-propylpiperazine (III)
R = (CH2)3CH3 1-(diphenylmethyl)-4-butylpiperazine (IV)
Calpena et al. (1994) highlighted the possibility of deliv-
ering some anti-emetic drugs via the dermal route. Several of
the anti-emetics were found to be likely candidates for for-
mulation into transdermal delivery system. Thus, cyclizine
(I) is one of the anti-emetic drug entities, which could be
considered for possible transdermal delivery. This possibil-
ity, however, may be hindered by its low aqueous solubility
and high melting point. It is, therefore, likely that the deliv-
ery characteristics of (I) can be improved by using its alkyl
analogues, derivatives possessing both a high aqueous sol-
ubility and lipophilicity at physiological pH (pH 7–8) and
also possessing a lower melting point than (I). Therefore,
in the present study, cyclizine alkyl analogues were synthe-
sized, their physicochemical properties determined and their
in vitro skin permeation evaluated.
To0.12 mole(30.28 g)of1-(diphenylmethyl)piperazinein
20 ml dry benzene, 20.6 g anhydrous sodium bicarbonate was
added. The reaction was stirred and heated under reflux. To
the suspension, 0.12 moles of ethyliodide or propyl iodide or
butyl iodide dissolved in 20 ml dry benzene was added drop-
wise over a period of 20 min. The reaction was refluxed until
completion, as followed by TLC. It was filtered, washed with
dry benzene, and the solvent was removed under vacuum. In
each instance, an almost white powder (compounds II, III,
or IV) was purified by column chromatography (ethyl ac-
etate:dichloromethane:methanol, 7:2:1; silica gel PF254) and
recrystallised at room temperature.
2.1.1. 1-(Diphenylmethyl)-4-ethylpiperazine (II)
6.71 g (44.73%) of almost white product were produced
following the analysis by TLC and column chromatogra-
phy (CC) (ethyl acetate:dichloromethane:methanol, 7:2:1).
Rf value: 0.64; mp: 51 ◦C; m/z (EI+, %): M+ 280 (91), 113
(100), 56 (58), 165 (75), 167 (89), 194 (84), 195 (68), 208
(56), and 70 (67); δH (300 MHz, CDCl3): 1.1 (t, 3H, CH3),
2.4 (t, 2H, CH2), 2.5 (bs, 8H, piperazine protons), 4.21 (s, 1H,
CH), 7.3 (m, 10H, aromatic protons); δC (75 MHz, CDCl3):
11.7 (CH3), 51.6 (CH2), 52.2 (CH2), 52.9 (CH2), 76.2 (CH),
126.9 (CH), 127.9 (CH), 128.5 (CH), and 142.8 (C). νmax
(KBr, cm−1): 725, 960, 1160, 1250, 1460, 2800, 2960, and
3460.
2. Materials and methods
Cyclizine analogues (II–IV) were synthesized according
to the methods by Zikolova and Ninov (1972) and Zikolova
et al. (1984). Cyclizine (I), 1-(diphenylmethyl)piperazine,
ethyl iodide, propyl iodide, and butyl iodide were obtained
from Sigma–Aldrich (UK). Identification and levels of pu-
rity were confirmed by electron impact mass spectra (MS)
recorded on a micromass autospec ETOF mass spectrome-
ter, nuclear magnetic resonance (1H NMR and 13C NMR)
spectra recorded on a Varian Gemini 300 spectrometer oper-
2.1.2. 1-(Diphenylmethyl)-4-propylpiperazine (III)
1
ating at frequency of 300 MHz for H and 75 MHz for 13C
8.00 g (53.3%) of almost white product were produced
following the analysis by TLC and column chromatogra-
phy (CC) (ethyl acetate:dichloromethane:methanol, 7:2:1).
Rf value: 0.71; mp: 70 ◦C; m/z (EI+, %): M+ 294 (89), 127
(100), 167 (97), 194 (78), 195 (78) 208 (54), 165 (71), and
56 (55); δH (300 MHz, CDCl3): 0.9 (t, 3H, CH3), 1.5 (st,
2H, CH2), 2.31 (t, 2H, CH2), 2.49 (bs, 8H, piperazine pro-
tons), 4.21 (s, 1H, CH), 7.3 (m, 10H, aromatic protons); δC
(75 MHz, CDCl3): 11.8 (CH3), 19.9 (CH2), 51.8 (CH2), 53.4
(CH2), 60.6 (CH2), 76.2 (CH), 126.9 (CH), 128 (CH), 128.5
using CDCl3. The following abbreviations were used to de-
scribe the splitting pattern of 1H signals: s: singlet, t: triplet,
st: sextet, bs: broad singlet, and m: multiplet. Infrared spectra
(IR) were recorded on a Nicolet 550 series II spectrometer
using KBr pellets. Melting points were determined with an
electrothermal digital Bv¨chi B-540 melting point apparatus.
For solubility studies, HPLC grade acetonitrile (Across Or-
ganic, NJ, USA), potassiumdihydrogenphosphateand ortho-
phosphoric acid (Merck, Johannesburg, South Africa), HPLC

L.M. Monene et al. / European Journal of Pharmaceutical Sciences 24 (2005) 239–244
241
(CH), and 142.9 (C). νmax (KBr, cm−1): 725, 1000, 1160,
1250, 1460, 2800, 2960, and 3460.
2.1.3. 1-(Diphenylmethyl)-4-butylpiperazine (IV)
5.12 g (34.1%) of almost white product were produced
following the analysis by TLC and column chromatogra-
phy (CC) (ethyl acetate:dichloromethane:methanol, 7:2:1).
Rf value: 0.67; mp: 52 ◦C; m/z (EI+, %): M+ 308 (90), 141
(98), 167 (99), 194 (89), 195 (85), 208 (62), 165 (76), and
56 (60); δH (300 MHz, CDCl3): 0.9 (t, 3H, CH3), 1.28 (m,
2H, CH2), 1.42 (m, 2H, CH2), 2.3 (t, 2H, CH2), 2.41 (bs, 8H,
piperazine protons), 4.2 (s, 1H, CH), 7.3 (m, 10H, aromatic
protons); δC (75 MHz, CDCl3): 13.9 (CH3), 20.7 (CH2), 28.9
(CH2) 51.9 (CH2), 53.5 (CH2), 58.5 (CH2), 76.3 (CH), 126.9
(CH), 128 (CH), 128.5 (CH), and 142.9 (C). νmax (KBr,
cm−1): 725, 1000, 1160, 1250, 1460, 2800, 2960, and 3460.
Fig. 1. log D as a function of pH.
Five millilitres of these solutions were transferred to a 10 ml
test tube, each containing equal volumes (5 ml) of phosphate
buffer. Three tubes of each compound were stoppered and ag-
itated for 1 h and another set for 2 h. After centrifugation, the
n-octanol and buffer phases were separated and appropriately
diluted with methanol and mobile phase before injecting onto
the HPLC column. Partition coefficients were calculated as
the ratio of drug concentration in the n-octanol phase to that
in the buffer phase. There was no difference in the values of
Koct when the tubes were agitated for 1 or 2 h.
2.2. Chromatographic procedure
The analytical HPLC system comprised of a Shimadzu
LC-6A delivery system, Shimadzu SPA-6A UV detector, Shi-
madzu SCL-6B system controller with SIL-6B autosampler
and Shimadzu CR6A chromatopac (integrator). A Lichro-
˚
spher (4 mm × 250 mm), 100 A pores – 5 ␮m, RP – 18
¨
endcapped column (Machery–Nagel, Duren, Germany) was
2.5. Melting point
used. Compounds were analysed with a mobile phase com-
prising of acetonitrile: 0.05 M (pH 3) phosphate buffer (6:4)
with UV detection at 200 nm. The pH was adjusted to 3
with 15% ortho-phosphoric acid solution. The flow rate was
1.0 ml/min and column temperature was maintained at 25 ◦C.
Melting points were taken in capillary tubes on an elec-
trothermal digital Bv¨chi B-540 melting point apparatus and
are uncorrected.
2.6. In vitro skin permeation study
2.3. Aqueous solubility
2.6.1. Preparation of donor solutions
The aqueous solubilities of (I) and its alkyl analogues were
determined by equilibrating a large excess of each compound
in HPLC water. The temperature (25 ◦C) was held constant
by means of a water bath. To hasten the attainment of equilib-
rium, magnetic bars continuously stirred each slurry. Equi-
librium was attained within 24 h. Samples were then filtered
through a Millipore filter (0.45 ␮m) and the concentration of
the compounds were determined by HPLC.
Donor solutions of (I)–(IV) were obtained by equilibrat-
ing excess amounts of each compound with phosphate buffer
(pH7.4)intostopperedglasscontainers. Thetemperaturewas
maintained at 37 ◦C by circulating water through the jackets
from a constant temperature water bath. The slurries were
vigorously and continuously mixed for 24 h using magnetic
stirring bars. Saturated state was achieved well within 24 h.
2.4. Partition coefficient
2.6.2. Preparation of the skin
Full thickness female human abdominal skin tissue from
the surgery clinic was sealed in evacuated plastic bags and
frozen at −20 ◦C. The adipose tissue was removed by blunt
dissection and immersion of the skin in 60 ◦C water for 1 min
to separate the epidermis. The epidermis was then peeled
away from the dermis using forceps (Harrison et al., 1984).
The skin sections were cut into squares, wrapped in alu-
minium foil, sealed in plastic bags, and stored in a freezer
at −20 ◦C until utilized (within 1 month of its receipt). The
frozen skin pieces were thawed at room temperature and ex-
amined for defects before mounting them onto the Franz dif-
fusion cells.
Cyclizine has ionizable functional groups and therefore it
isimportanttoknowthepHatwhichpartitioningismeasured.
ACD software was used to predict the effect of pH on the
partitioning and the results of log D plotted as a function of
pHareshownin Fig. 1. Thefigureshowsacomplexbehaviour
but it clearly demonstrates the differences in partitioning that
occur around physiological pH.
Equal volumes of n-octanol and phosphate buffer (pH 7.4)
were saturated with each other for at least 24 h before the ex-
periment. Solutions of (I) and its alkyl analogues were pre-
pared with the pre-saturated n-octanol phase as a solvent.

242
L.M. Monene et al. / European Journal of Pharmaceutical Sciences 24 (2005) 239–244
Table 1
2.6.3. Skin permeation method
Physicochemical properties of (I) and its alkyl analogues
Vertical Franz diffusion cells with a 4 ml capacity receptor
compartment and a 1.07 cm2 diffusion area were used. The
epidermal layer of the skin was mounted carefully onto the
lower half of the diffusion cells with the stratum corneum fac-
ing the donor compartment. The donor compartments were
fastened to the receptor compartments with clamps, the skin
acting as a seal between the half-cells. The receptor com-
partments were filled with phosphate buffer (pH 7.4), taking
care that there were no air bubbles left in the compartments.
The diffusion cells were placed in a water bath at a constant
temperature of 37 ◦C on a submersible magnetic stirring bed,
a small magnetic stirring bar was placed at the bottom of
the lower compartment and the system was allowed to equi-
librate for 1 h before adding the saturated drug containing
solution to the donor compartment. The experiment was ini-
tiated by charging the donor compartment with 700 ␮l of a
freshly prepared saturated solution of the drug and immedi-
ately covering it with Parafilm® to prevent any significant
evaporation from the donor compartment. At pre-determined
intervals (2, 4, 6, 8, 10, 12, and 24 h), the entire contents of
the receptor compartment were withdrawn and replaced with
fresh buffer (37 ◦C). This ensured that sink conditions existed
throughout the experiment. Two hundred microlitres of each
sample was directly assayed by HPLC to determine the drug
concentration in the receptor fluid. The duration of the skin
permeation experiment was 24 h.
Compound
Molecular weight
(g/mol)
ACD log D
pH 7.4
Melting point
(◦C) S.D.
I
II
III
IV
266.4
280.4
294.4
308.5
2.09
2.58
3.11
3.64
107
51
70.1
52.1
lene protons and a triplet at δ 0.9 due to the methyl group of
the propyl moiety. 13C NMR spectrum of (III) shows signals
due to methylene carbon atom at δ 19.9 and at 11.8 from the
methyl group of the propyl moiety. The 1H NMR spectrum
of (IV) shows in addition to signals of (I) a multiplet at δ 1.28
and 1.42 due to the methylene protons and a triplet at δ 0.9
due to the methyl group of the butyl moiety. 13C NMR spec-
trum of (IV) shows signals due to methylene carbon atoms
at δ 20.7 and at 28.9; whereas, the methyl group of the butyl
moiety resonates at δ 13.9.
3.2. Physicochemical properties
Table 1 summarizes the physicochemical properties of (I)
and its alkyl analogues. The melting point of (I) was the same
as reported in literature. The trend in melting point as a func-
tion of alkyl chain length can be seen in Table 1. Although
there is irregularity in the pattern, overall the melting points
decrease as the alkyl chain is extended. Compound (I) had a
melting point of 107 ◦C. Increasing chain length to –CH2CH3
and –CH2CH2CH2CH3 resulted in compounds (II) and (IV)
with melting points of about 55 ◦C lower than (I). Increasing
the chain to –CH2CH2CH3 resulted in a melting point higher
than that of (II) and (IV), however, the melting point was
lower than that of (I). Yalkowsky et al. (1972) studied alkyl
chain length series and found that melting points decreased
overall, but not linearly. This indicates that extra alkyl func-
tionalities disrupt the intracrystalline cohesion of the drug.
It is interesting to compare the predicted log D values,
as determined using ACD software, with the experimentally
measured ones. This is shown in Fig. 2. There is a linear
relationship between the experimental and predicted values
for the four analogues. However, the predicted ones are not
as high as those measured. The precise reasons for this are
unknown but it should be remembered that log D values are
dependent on the ionic nature of the aqueous medium and
the absolute values may depend on different amount of ion
pairs that are present. This will be affected by the nature of
the counter ions and the ionic strength of the solution.
3. Results and discussion
3.1. Synthesis of cyclizine alkyl analogues
The analogues were successfully synthesized according
to the methods by Zikolova and Ninov (1972) and Zikolova
et al. (1984). The molecular mass of each compound was
determined experimentally with mass spectroscopy and by
empirical calculation. The value of molecular mass was the
same in both cases. The structures and purities of the ana-
logues were confirmed by MS, NMR, and IR.
3.1.1. Spectral data of cyclizine analogues
The MS data for the compounds confirmed the presence
of the molecular ions (M+) at m/z 280, 294, and 308, corre-
sponding to the molecular formula C19H24H2 (II), C20H26N2
(III), and C21H28N2 (IV), respectively. M+ 113, 127, and 141
correspond to the N-alkyl piperazine fragment for (II), (III),
and (IV), respectively, and the signal at 167 represents the
aromatic portion without the alkyl piperazine radical. Sig-
nals at 194, 195, and 208 represent the rearrangement and
fragmentation of the alkyl piperazine moiety.
Physicochemical parameters, such as aqueous solubil-
ity and lipophilicity, have shown to influence membrane
flux, therapeutic activity, and pharmacokinetic profiles of
medicines (Goosen et al., 1998). The aqueous solubilities
and lipophilicities of (I) and its alkyl analogues are listed in
Table 2. Further alkylation of the piperazine ring results in
compounds that are more lipophilic but also possess higher
aqueous solubility than (I). Although the first member of the
In addition to the 1H NMR and 13C NMR signals of cy-
clizine (I), compound (II) has a triplet from the methyl group
at δ 1.1 in the 1H NMR spectrum and a signal at δ 11.7 in the
13C NMR spectrum. The 1H NMR spectrum of (III) shows
in addition to signals of (I) a sextet at δ 1.5 due to the methy-

L.M. Monene et al. / European Journal of Pharmaceutical Sciences 24 (2005) 239–244
243
Table 3
Permeation parameters of (I) and its alkyl analogues through human skin
Compound
J
S.D.
TL (h) kp S.D. (cm/h)
D (cm2/h)
× 10−3
(␮g/cm2/h)
32 ◦C
I
II
III
IV
0.132 0.04 0.15
6.952 0.38 0.21
0.250 0.02 0.28
0.686 0.06 0.40
0.003 0.01
0.044 0.18
0.0044 0.09
0.005 1.05
25.0
17.9
13.4
9.38
chemical properties, partitioning and solubility, are both in
the same direction.
3.3. In vitro skin permeation study
The percutaneous permeation of (I) and its alkyl analogues
was studied by an in vitro technique using a Franz diffusion
cell mounted with the epidermal layer of human skin as the
diffusion membrane. The permeation parameters (flux, J; lag
time, TL; permeability coefficient, kp; and diffusion coeffi-
cient, D of (I)) and its alkyl analogues from their saturated
aqueous solutions (pH 7.4) are summarized in Table 3. The
steady-state flux of each compound was determined from the
slope of the linear portion of the cumulative amount versus
time plot and their mean steady-state flux is presented as a bar
chart in Fig. 3. The compounds attained steady state diffusion
in the skin within 0.4 h. The lag times were determined by
extrapolating the linear portion of the curve to its intersection
with the x-axis. There was no significant difference in the lag
times.
The diffusion experiment showed that there is a varying
degree of permeation between (I) and its alkyl analogues.
Generally, the analogues resulted in higher permeation than
(I). Fig. 3 shows the permeation profiles of (I) and its alkyl
analogues. It clearly indicates that, relative to (I), there is
an increase in the steady-state flux of the analogues. The
highest flux is observed with (II) followed by (IV). This
could be attributed to their high aqueous solubility and low
Fig. 2. The relationship between the experimental log D and those predicted
using ACD software.
series (II) was about three times more soluble than (I), com-
pound (III) had aqueous solubility almost equal to that of
(I). Compound (IV) was found to be more than 2.5 times
more soluble than (I). It clearly appears that compounds (II)
and (IV) are more soluble than compound (III) despite the
fact that the aqueous solubility of (IV) is limited by its extra
methylenegroup. Thechangesinaqueoussolubilitiesofthese
analogues mirror the changes in crystallinity seen through the
melting points. The low solubility of (III) can be attributed
to the higher melting point as compared to (II) and (IV).
Comparing the n-octanol/water partition coefficient as a
function of alkyl chain length in Table 2, it can be seen that
longer chain analogues are more lipophilic. Bundgaard and
Falch (1985) indicated that the increase in lipophilicity is
generally accompanied by a decrease in water-solubility, but
inspection of the data in Table 2 reveals that the aqueous
solubility increased relative to (I) despite the higher log Koct
value of the compounds.
Based on the physicochemical properties of the analogues
(Tables 1 and 2), it can be seen that compounds (II) and (IV)
have higher aqueous solubilities and higher lipophilicities
as compared to (I). It can, therefore, be predicted that these
analogues should have the highest skin flux in this series.
Compound (I) and (III) have lower aqueous solubilities and
highermeltingpoints. Thismighthavealessfavourableeffect
on their skin permeability. Compound (III) possesses higher
log Koct, that could result in the compound being more skin
permeable than (I). A prediction of the order of penetration of
compoundsthroughtheskinwouldbe:(II) > (IV) > (III) > (I)
as the result of the fact that the flux determining physico-
Table 2
Aqueous solubility and partition coefficient for (I) and its alkyl analogues
Compound
Aqueous solubility (25 ◦C)
log Koct S.D.
(␮g/ml S.D.)
I
II
III
IV
185.5 0.20
569.6 5.39
189.9 3.30
481.8 11.8
3.11 0.4
3.64 0.4
4.18 0.2
4.71 0.5
Fig. 3. Mean steady-state flux (␮g/cm2/h) S.D. of (I) and its alkyl ana-
logues from saturated aqueous solutions (n = 6).

244
L.M. Monene et al. / European Journal of Pharmaceutical Sciences 24 (2005) 239–244
level of crystallinity, as is evident from their low melting
points. (I) and (III) exhibited lower flux, a reflection of their
higher melting points and low aqueous solubility. Flynn and
Yalkowsky (1972) studied the effects of alkyl chain length
on the flux across a synthetic membrane. They found that a
plot of the logarithm of the steady-state flux from saturated
solution against chain length gave a parabolic curve. The re-
sults obtained in this study, show that the series is capable
of odd–even alternation in aqueous solubility, melting point
and in percutaneous permeation. This kind of behaviour is at-
tributed to the dependence of these physicochemical proper-
ties upon the sum of the energy required to disrupt the crystal
and the intermolecular interactions between like and unlike
species in solution. The linearity of partition coefficient with
alkyl chain length is because the partition coefficient is a
property of the solute and therefore is not dependent on the
crystal structure. It is only dependent upon the interactions
occurring in solution. There is a correlation between aque-
ous solubility, melting point and percutaneous permeation of
the compounds. (II) Showed high percutaneous permeation
hence increased aqueous solubility and low melting point.
Low percutaneous permeation of (I) and (III) is attributed
to their low aqueous solubility and high melting point. Thus,
the more water-soluble member of the series is the most pen-
etrating member through the human epidermis.
The partitioning behaviour of cyclizine and its derivatives
as a function of pH is interesting. In this study the donor
phase was buffered at pH 7.4 at which value partitioning is
quite favourable. In unbuffered conditions the solution at the
skin surface will be at approximately pH 5, the estimated
pH of the skin surface. At this pH, the portioning is much
less favourable and less efficient fluxes may be anticipated.
However, this would need verifying experimentally.
The results of this study suggest that alkylation is a po-
tential useful approach to enhance percutaneous penetration.
The following rank order of penetration of compounds was
established:(II) > (IV) > (III) > (I). Inconclusion, itisclearly
demonstrated that further alkylation of the piperazine ring
in 1-(diphenylmethyl)piperazine molecule beyond a methyl
groupresultsincompoundsthataremorehydrophilicandless
crystalline. These physicochemical properties favour percu-
taneous delivery of drugs. From all the permeability data,
compound (II) showed the best permeation results of the se-
ries and it was clear from experimental data that aqueous
solubility and level of crystallinity played an important role
in the skin permeation process.
References
Baltzly, R., Dubreuil, S., Ide, W.S., Lorz, E., 1949. Unsymmetrically dis-
ubstituted piperazines. II. N-Methyl-Nꢀ-benzhydrylpiperazine as his-
tamine antagonists. J. Org. Chem. 14, 775–782.
Beckett, A.H., 1982. Possibilities and limitations of transdermal absorp-
tion. In: Aulton, M. (Ed.), Pharmaceutics: The Science of Dosage
Form Design. Churchill Livingstone, New York, pp. 154–170.
Bundgaard, H., Falch, E., 1985. Allopurinol prodrugs. II. Synthesis,
hydrolysis kinetics and physicochemical properties of various N-
acyloxymethyl allopurinol derivatives. Int. J. Pharm. 24, 307–325.
Calpena, A.C., Blane, C., Moreno, J., Obach, R., Domenech, J., 1994. A
comparative in vitro study of transdermal absorption of antiemetics.
J. Pharm. Sci. 83, 29–33.
Castillo, J.C., De Beer, E.J., Jaros, S., 1949. J. Pharmacol. Exp. Ther. 96,
388.
Chinn, H.I., Handford, S.W., Cone, T.E., Smith, P.K., 1952. Effectiveness
of various drugs for the prophylaxis of seasickness. Am. J. Med. 12,
433–439.
Chinn, H.I., Gammon, W.R., Frantz, M.E., 1953. J. Appl. Physiol. 5, 599.
Dundee, J.N., Jones, P.O., 1968. The prevention of analgesic-induced
nausea and vomiting by cyclizine. Br. J. Clin. Pract. 22, 379–382.
Flynn, G.L., Yalkowsky, S.H., 1972. Correlation and prediction of mass
transport across membranes. I. Influence of alkyl chain length on
flux-determining properties of barrier and diffusion. J. Pharm. Sci.
61, 838–851.
Goosen, C., Du Plessis, J., Mu¨ller, D.G., Janse Van Rensburg, L.F., 1998.
Correlation between physicochemical characteristics, pharmacokinetic
properties and transdermal absorption of NSAIDs. Int. J. Pharm. 163,
203–209.
Gutner, L.B., Gould, W.J., Cracovaner, A.J., 1954. Effects of
cyclizine–HCl and chlorcyclizine–HCl upon vestibular function. Arch.
Otolaryngol. 59, 503–509.
Harrison, S.M., Barry, B.W., Dugard, P.H., 1984. Effects of freeze drying
on human skin permeability. J. Pharm. Pharmacol. 36, 261–262.
Susan, M.R., Mclean, P.C., Melville, J., 1989. Cyclizine abuse among a
group of opiate dependents receiving methadone. Br. J. Addict. 84,
929–934.
Yalkowsky, S.H., Flynn, G.L., Slunick, T.G., 1972. Importance of chain
length on physicochemical and crystalline properties of organic ho-
mologs. J. Pharm. Sci. 61, 852–857.
Zikolova, S., Ninov, K., 1972. Analogs of N1-benzhydryl-
N4-cinnamylpiperazine
(cinnarizine).
II.
N1-substituted-N4-
benzhydrylpiperazines. Trans. Sci. Chem. Pharm. Inst. 8, 59–67.
Zikolova, S., Slavova, S., Buchvarova, A., 1984. Analogs of
1N-benzhydryl-4N-cinnamylpiperazine (cinnarizine). V. New N1-
benzhydryl-N4-substituted piperazines. Trans. Sci. Chem. Pharm. Inst.
14, 23–28.