Enhanced Permeation of Methotrexate In Vitro by Ion Pair
Formation With L-Arginine
VIJAY D. IVATURI, S. KWONHO KIM
College of Pharmacy and Allied Health Professions, St. John s University, Queens, New York 11439
Received 3 September 2008; revised 29 October 2008; accepted 17 November 2008
Published online 30 December 2008 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21663
ABSTRACT: Ion paired solutions of methotrexate in L-arginine/water/propylene glycol
systems were evaluated for their potential to enhance the permeation of methotrexate
across rabbit nasal mucosa in vitro. The partition coefficient of methotrexate in the
methotrexate: L-arginine ion paired systems was observed to be 24 times greater than
that of the methotrexate system without L-arginine. The ion pair formation between
methotrexate and L-arginine was confirmed by a decrease in the conductivity of the
systems in the presence of propylene glycol, a dielectric constant reducing agent. The
permeation of methotrexate across the rabbit nasal mucosa from the ion paired systems
was observed to be significantly greater ( p < 0.05) as compared to control systems of
methotrexate solution in water and a sodium salt. Furthermore, a threefold increase
in the flux of methotrexate was observed when propylene glycol was added to the ion
paired systems. These results suggest that methotrexate: L-arginine ion paired
systems have potential in improving the permeation of methotrexate across rabbit
nasal mucosa and may form the basis for further development of an intranasal
therapeutic system of methotrexate. ß 2008 Wiley-Liss, Inc. and the American Pharmacists
Association J Pharm Sci 98:3633 3639, 2009
Keywords: methotrexate; ion pair; L-arginine; nasal mucosa; permeation; propylene
glycol
INTRODUCTION gated the possibility of a direct drug transport
of methotrexate from the nasal cavity to the
Methotrexate, an antineoplastic agent, is widely cerebro-spinal fluid after intranasal administra-
used in the treatment of many malignant and tion in rats.
nonmalignant disorders. Presently, malignancies In order to achieve effective systemic absorption
of the central nervous system are usually treated of a drug through the nasal mucosa, the drug
by lumbar puncture or the intrathecal adminis- moiety must have good solubility and lipophilicty
tration of a high dose of methotrexate. These for it to be administered in acceptable volumes of
routes of administration are invasive, require less than 0.3 mL in the nostrils. Unfortunately,
special surgical procedures, and are difficult to methotrexate is a poorly soluble and hydrophilic
use in routine clinical settings. On the other hand, drug, and thus possesses low permeability across
intranasal administration of methotrexate can mucosal membranes. Ion pairing concepts for a
provide for a more convenient and noninvasive hydrophilic drug with an appropriate counter ion,
alternative. Wang et al.1 have previously investi- in order to improve its permeability and hence
bioavailability have been well established by
the research activity of Neubert et al.2,3 Also,
Correspondence to: Vijay D. Ivaturi (Telephone: 347-558-
the permeation of ion paired cationic as well as
3650; Fax: [phone]; E-mail: [e-mail])
anionic hydrophilic drugs has been researched
Journal of Pharmaceutical Sciences, Vol. 98, 3633 3639 (2009)
ß 2008 Wiley-Liss, Inc. and the American Pharmacists Association in vitro and in vivo by numerous investigators.4 15
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 10, OCTOBER 2009 3633
3634 IVATURI AND KIM
The objective of this study was to enhance the nonpolar solvents at room temperature (258C) for
permeation of methotrexate across rabbit s nasal 24 h till equilibrium was achieved.
mucosa by using L-arginine as an ion pairing A solution depletion technique was employed
agent, in order to assess the feasibility for the for the determination of the partition coefficient
development of an intranasal delivery system for based on the following equation:
methotrexate. Furthermore, the effect of addition
C0 C1
of propylene glycol on the stability, partition
Papp ź
C1
coefficient and permeation of methotrexate was
also investigated. where C0 is the initial concentration of drug in the
aqueous phase and C1 is the final concentration of
the drug in the aqueous phase after equilibrium
MATERIALS AND METHODS was established.
Materials
Conductivity Measurements
Methotrexate, ethanol, n-octanol, propylene gly-
The electric conductance of MTX solutions incor-
col, L-arginine, and triethylamine were purchased
porating ARG, and in the presence of different
from Sigma Chemical Co. (St. Louis, MO). Sodium
solvents of varying dielectric constants was
pentobarbital injection was purchased from
determined using a traceable VWR digital con-
Henry Schein, Inc. (Port Washington, NY) and
ductivity meter.
all other chemicals were HPLC or analytical grade
and used as received. Water was deionized and
distilled in the laboratory. Preparation of the Nasal Mucous Membranes
New Zealand white rabbits (2.0 2.5 kg), pur-
chased from Marland Breeding Farms Inc.
Methods (Hewitt, NJ), were euthanized by sodium pento-
barbital injection and two pieces of the nasal
Analytical Methodology
mucous membranes (1.2 cm 2.8 cm) were care-
Methotrexate (MTX) was quantitatively analyzed fully stripped from the rabbit s nasal septum.
by a Hewlett Packard 1050 series HPLC system After excision of the nasal mucous membrane, any
with a symmetrical C18 column (5 mm, 150 mm adhering cartilage was removed and the nasal
3.9 mm ID) (Waters Corporation, Milford, MA). membrane was rinsed in oxygenated Kreb s
The mobile phase consisted of a mixture of Ringer solution. All the permeation experiments
methanol:triethylamine/phosphate buffer (25:75) were conducted using freshly excised nasal
pH 7.3 and the flow rate was set at 0.8 mL/min. mucosal membranes and as per the protocols
MTX was detected at 303 nm with a retention time approved by the Animal Care Committee, St.
of 2.71 min. The limit of detection in the HPLC John s University (Queens, NY).
assay was 0.1 mg/mL, and the limit of quantitation
was 0.5 mg/mL.
In Vitro Permeation Studies
In vitro permeation studies were conducted
Partition Coefficient Studies
using side-by-side permeation cells (Crown Glass,
The apparent partition coefficient of MTX bet- Somerville, NJ) maintained at 37 0.58C by a
ween n-octanol and solutions of L-arginine (ARG) thermostatic circulating water bath (VWR). The
at pH 7.0 was determined using the shake-flask freshly excised nasal mucosa was immediately
method.16 Briefly, ARG solutions of different mounted on the permeation cells with the nasal
concentrations were prepared and 2 mL of each mucosal epithelia facing the donor cell and
solution were taken in screw capped test tubes to exposing a surface area of 0.16 cm2. The receptor
which an equal volume of n-octanol was added. cell was then filled with Kreb s Ringer solution
Each phase had been pre-saturated with the other (3.5 mL, pH 7.4) and the donor cell was filled with
by equilibration overnight before the experi- the test solution (3.5 mL). Samples (200 mL) were
ments. A known quantity of MTX was then added withdrawn from the receptor cell at 0.5, 1, 2, 3, 4,
to the n-octanol/ARG solution mix and MTX 5, 6, 7, and 8 h, and replaced with fresh Kreb s
was allowed to partition between the polar and Ringer solution in order to keep the receptor
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 10, OCTOBER 2009 DOI 10.1002/jps
ENHANCED PERMEATION OF METHOTREXATE BY ION PAIRS 3635
volume constant and to maintain sink conditions. equation:
The withdrawn samples were analyzed by HPLC.
f½MHŠ þ½M Š þ½AþM Šgo
All the test solutions used in the in vitro permea-
Papp ź
tion studies were adjusted to pH 7 using 1 N HCl f½MHŠ þ½M Š þ½AþM Šgw
at which the formulations were considered to be
where [MH] represents the free acidic form
the most stable.
of MTX, [M ] is the ionized form, and [AþM ]
is the ion paired form with ARG [Aþ]. The pH of the
Data Analysis
solution was adjusted to 7.0, so that there would be
0% of [MH], and since the contribution of drug
The steady-state flux of MTX across the rabbit
nasal mucosal membrane was determined from partitioning by [M ], the ionized form would be
the slope of the cumulative amount of MTX relatively small when compared to partitioning
due to the ion paired system, we can attribute the
permeated as a function of time.
increase in partition of the drug into n-octanol as to
being driven by the ion paired form, that is [AþM ].
Statistical Analysis
A 24-fold increase in the partition coefficient of
The results were expressed as mean standard
MTX was observed on increasing the concentra-
deviation of triplicate experiments. Statistical
tion of ARG from 0.01 to 0.10 M or equivalently
comparisons were made using Student s t-test
from 1:3 to 1:59 molar ratio of MTX:ARG.
or ANOVA with the Holm Sidak test as a post
The increase in the concentration of the counter
hoc test. The level of significance was taken as
ion, ARG in the MTX:ARG ion paired systems
p < 0.05.
from 1:3 to 1:59 molar ratio may have increased
the tendency towards the formation of ion pairs
especially in the presence of higher concentrations
RESULTS AND DISCUSSION
of the counter ion. Ion pairs exist in equilibrium
with their two component ions; A þ Bþ $ AB.
Partition Coefficient Studies
Here (MTX þ ARGþ $ MTXARG), an excess of
one ion due to higher concentrations (ARG) may
The partition coefficient of MTX (pKa ź 3.76)
have resulted in a larger fraction of the second
was observed to increase linearly with increase
ion to be in the ion paired state17 (i.e., MTX). Our
in molar concentration of the counter ion, that is,
findings are in agreement with those of Takacs-
ARG (pKa1 ź 2.17, pKa2 ź 9.04, pKa3 ź 12.48), as
Novak and Szasz18 who have investigated the
shown in Figure 1. The partitioning of MTX from
ion pair formation and partition of quaternary
MTX:ARG ion paired solutions, between n-octanol
ammonium compounds, such as propantheline
and water could be represented by the following
bromide, homidium bromide and neostigmine
bromide with organic counter ions of different
lipophilicity such as mesylate, acetate, caprate,
and deoxycholate in the molar ratio of quaternary
ammonium compound: counter ion of 1:1 to
1:50. These authors observed that the ion pair
formation capability increased significantly with
increase in the concentration of the counter
ions and also that the completion of the ion
pair formation and maximum lipophilicity was
obtained only in the presence of high molar excess
of counter ions. Hence, the observed increase in
the partition coefficient of MTX could be mainly
due to the formation of ion pairs between the drug,
MTX and the cationic counter ion, ARG.
Figure 1. Apparent partition coefficient profile ob-
Conductivity Measurements
tained with varying concentrations of L-arginine in
The electric conductivity of MTX:ARG ion pairs
aqueous methotrexate:arginine ion paired solutions at
258C. contained in a 1 mg/mL solution of MTX in 0.1 M
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 10, OCTOBER 2009
3636 IVATURI AND KIM
ARG solution was measured as a function of
propylene glycol (PG) concentration. PG was
selected as the dielectric constant reducing
agent since it resulted in the maximum decrease
in electric conductivity of the ion paired
solutions as compared to other organic solvents
such as ethanol, glycerin and PEG 200 (Fig. 2).
This reduction in solution conductivity is due
to the lowering of the medium s dielectric
constant, which is an important factor that
affects columbic interactions between oppositely
charged ions, and hence the driving force for ion
pairing.19
The concentration of the dielectric constant
reducing agent, that is, PG was observed to have
an effect on the electric conductivity of the ion
Figure 3. The electric conductivity profile of aqueous
paired systems. As shown in Figure 3, there
methotrexate:arginine ion paired solutions as a func-
was an exponential decrease in conductivity
tion of propylene glycol concentration at 258C.
with increase in proportion of PG in the MTX:ARG
ion paired system. It has been previously reported
by Lee et al.19 that ion pair formation is probably In Vitro Permeation Studies
induced by increasing the relative amount of
Effect of ARG Concentration
nonaqueous medium. In addition, with increase
in the concentration of PG in the ion paired The flux of MTX across rabbit nasal mucosa from
systems, the solutions tend to become more solutions with different MTX:ARG molar ratios
viscous. This condition may make the ion pairs is shown in Figure 4A. It can be observed that
less mobile and hence provide a more stable the permeation of MTX increased linearly (r2 ź
environment for MTX. The MTX:ARG ion 0.9976) with increase in MTX:ARG molar ratio
paired solutions containing PG and maintained from 1:0 to 1:3. When compared to a solution of
at pH 7.0 with hydrochloric acid (HCl) were MTX in water maintained at pH 7.0, (i.e., a control
observed to be 100% stable at 458C over a period of study), the flux of MTX across the rabbit nasal
5 weeks (data not shown). mucosa from the ion paired systems was observed
to be significantly greater ( p < 0.05) (Fig. 4A).
This could be because the ion paired systems have
a higher lipophilicity as compared to the pure ionic
species. However, a further increase in the molar
ratios of MTX:ARG from 1:3 to 1:23, 1:45, 1:59,
and 1:118, the flux of MTX appeared to plateau
with no significant difference ( p > 0.05). This is
because, although the apparent partition coeffi-
cient of MTX was observed to increase with
increase in molar ratio of MTX:ARG from 1:0 to
1:45 by 24-fold times, as observed in the partition
coefficient study (Fig. 1), the apparent partition
coefficient value was still in the hydrophilic
region. Furthermore, the solubilizing effect of
high concentrations of ARG on MTX may have
created a low thermodynamic driving force in the
permeation system.
In addition, in order to confirm that ARG was
indeed forming an ion pair and not a salt with
MTX, the permeation of sodium salt of MTX was
Figure 2. Effect of organic solvents with different
dielectric constants on the electric conductivity of aque- compared to that of a 1:3 MTX:ARG solution. As
ous methotrexate:arginine ion paired solutions at 258C. shown in Figure 4B, the flux for sodium salt of
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 10, OCTOBER 2009 DOI 10.1002/jps
ENHANCED PERMEATION OF METHOTREXATE BY ION PAIRS 3637
were reported by Sarveiya et al.20 who observed
that ion-pairs of ibuprofen with counter-ions such
as triethylamine, diethylamine and ethyldiamine
resulted in higher permeation of drug when
compared to the permeability as a sodium salt
of ibuprofen. In other studies, Wang et al. showed
increase in permeability of ion paired scutellarin
with organic amines, whereas Teksin et al.
reported increase in metoprolol permeability
across three-lipid component artificial mem-
branes mediated by ion pair formation.21,22
Another possibility for the enhancement in the
flux of MTX may be due to the facilitated transport
of MTX and MTX:ARG ion pair by the binding of
the positively charged ARGþ to the negatively
charged mucosal membrane.
Effect of PG
The effect of addition of PG in various MTX:ARG
ion paired systems on the permeation of MTX
across rabbit nasal mucosa was studied and is
shown in Figure 5A. It can be observed that at a
fixed concentration of 51.8% (w/v) of PG, the
transmucosal permeation flux of MTX was
observed to significantly increase ( p < 0.05) in a
linear fashion (r2 ź 0.9752) with increase in ARG
concentration from 0.05% to 2.26%.
A comparison of the flux of MTX from ion paired
Figure 4. Comparative flux of methotrexate across
systems with and without PG was also done.
rabbit nasal mucosa from (A) Methotrexate ion paired
As seen in Figure 5B, an approximately twofold
solutions containing arginine in various molar ratios
increase in the flux of MTX was observed in the ion
at 378C. Methotrexate dissolved in water served as a
control. The values in parentheses represent the paired systems that were stabilized by PG, as
percentage of arginine in the system. (B) Methotrexate
compared to ion paired systems without PG. The
sodium salt and methotrexate:arginine ion paired
permeability coefficient of MTX from the formula-
solution.
tion containing the highest ARG concentration
0.13 M stabilized by PG was found to be 1.4 10 5
MTX (i.e., 7.73 1.27 mg/cm2/h) was observed to cm/s, which was approximately 8 times higher
be significantly lower ( p < 0.05) than that of a 1:3 than the permeability coefficient of MTX calcu-
MTX:ARG solution (i.e., 16.25 1.56 mg/cm2/h). lated from the formulation containing MTX in
This result can be explained by the   born energy water. Furthermore, at the highest concentration
of charging.  19 According to this concept, there of ARG, that is, 0.13 M or 2.26%, a threefold
exists a large energy difference between an ion in increase in the permeation flux of MTX was
water and an ion within a nonpolar membrane. observed for the PG stabilized ion paired systems
However, as ion pairs are electrically neutral and (53.60 3.51 mg/cm2/h) as compared to the ion
without an electrostatic energy term, the activa- paired system without PG (17.83 4.65 mg/cm2/h).
tion energy for neutral ion pairs to partition into Hence, this formulation was selected as an
the hydrophobic mucus membrane is much lower optimal formulation for further studies.
as compared to ions. As a result, the permeation of These results can be explained on the basis that
ion pairs through a membrane is greater than the in lower dielectric media, more ion pairs will
permeation of ions when present in a salt form.19 be present, and since this is a diffusible specie
Also, since the salt forms of drugs are usually it would lead to higher flux of the drug.19 In other
polar and less lipophilic, they generally show low words, the ion paired MTX may have been
permeability across membranes. Similar results associated with PG, and during the partitioning
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 10, OCTOBER 2009
3638 IVATURI AND KIM
Effect of MTX Concentration
The effect of varying the MTX concentration in
the donor compartment on the transmucosal
permeation flux of MTX from formulations
containing 2.26% of ARG and 51.8% of PG was
investigated. The MTX concentration levels
investigated included 1.0 (0.1%), 2.5 (0.25%) and
5.0 (0.5%) mg/mL. A 4.92-fold increase in the
MTX flux was observed with increase in the
MTX concentration from 1.0 to 5.0 mg/mL (%). In
addition, a linear relationship (r2 ź 0.9984) was
observed between MTX flux and drug concentra-
tion in the range investigated (Fig. 6). The
permeability coefficient of MTX from this for-
mulation containing 0.5% MTX, 2.26% ARG,
51.8% PG, and 45.44% WT was calculated to be
1.4 10 5 cm/s.
CONCLUSIONS
Based on these results it may be concluded that
methotrexate: L-arginine ion paired systems in
low dielectric media such as propylene glycol show
good chemical stability as well as enhanced
lipophilicity and permeability across rabbit nasal
mucosa in vitro. Thus, this formulation approach
Figure 5. (A) Flux of methotrexate as a function of
may form the basis for further development of
increasing arginine concentrations across rabbit nasal
an effective intranasal therapeutic system for
mucosa from ion paired methotrexate solutions with
methotrexate.
arginine containing dielectric constant reducing agent,
propylene glycol at 378C. (B) Comparative flux of meth-
otrexate across rabbit nasal mucosa from ion paired
solutions of methotrexate:arginine in various molar
ratios, with and without PG at 378C.
of the PG, the partitioning of the ion pairs may
have also taken place. Furthermore, PG may have
exhibited a permeation enhancement effect on
its own via increased passive diffusion of the
drug alone or as an ion pair. The presence of a
significant Y-intercept in Figure 5A may be
interpreted as the inherent flux of MTX in
the presence of PG alone. These results are in
agreement with those of Lee et al.19 who
have reported an increased diffusion of ion
paired sodium salicylate from ethanol and dioxane
(nonaqueous media) as compared to water
Figure 6. Effect of methotrexate concentration on
(aqueous media). The potential use of nonaqueous
the steady-state flux across the rabbit nasal mucosa
solvents in improving the stability and perme-
from the final formulation containing 2.26% arginine,
ability of ion pairs across membranes has been
51.80% propylene glycol, and 45.44% water at 1.0, 2.5,
discussed extensively in the literature.23 26 and 5.0 mg/mL of methotrexate.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 10, OCTOBER 2009 DOI 10.1002/jps
ENHANCED PERMEATION OF METHOTREXATE BY ION PAIRS 3639
REFERENCES 13. Jackson MJ, Airall AA. 1978. Transport of hetero-
cyclic acids across rat small intestine in vitro.
J Membr Biol 38:255 269.
1. Wang F, Jiang X, Lu W. 2003. Profiles of metho-
14. Wilson CG, Tomlinson E, Davis SS, Olejnik O. 1981.
trexate in blood and CSF following intranasal and
Altered ocular absorption and disposition of sodium
intravenous administration to rats. Int J Pharm
cromoglycate upon ion-pair and complex coacervate
263:1 7.
formation with dodecylbenzyldimethylammonium
2. Neubert R. 1989. Ion pair transport across
chloride. J Pharm Pharmacol 33:749 753.
membranes. Pharm Res 6:743 747.
15. Tomlinson E, Dooremalen JAMv, Rooij HHv, Wyne
3. Neubert R, Fisher S. 1991. Influence of lipophilic
HJA. 1982. Ion-pair and complex coacervate effects
counter ions on the transport of ionizable
on large ion flux through polyamide-6 membrane.
hydrophilic drugs. J Pharm Pharmacol 43:204
Int J Pharm 12:87 96.
206.
16. Inagi T, Muramatsu T, Nagai H, Terada H. 1981.
4. Irwin GM, Kostenbauder HB, Dittert LW, Staples
Mechanism of Indomethacin Partition between
R, Misher A, Swintosky JV. 1969. Enhancement of
n-Octanol and Water. Chem Pharm Bull 29:2330
gastrointestinal absorption of a quaternary ammo-
2337.
nium compound by trichloroacetate. J Pharm Sci
17. Jonkman JH, Hunt CA. 1983. Ion pair absorption
58:313 315.
of ionized drugs fact or fiction? Pharmaceutisch
5. Plakogiannis FM, Lien EJ, Harris C, Biles JA. 1970.
Weekblad 5:41 48.
Partition of alkylsulfates of quaternary ammonium
18. Takacs-Novak K, Szasz G. 1999. Ion-pair partition
compounds: Structure dependence and transport
of quaternary ammonium drugs: The influence of
study. J Pharm Sci 59:197 200.
counter ions of different lipophilicity, size, and
6. Gaginella TS, Bass P, Perrin JH, Vallner JJ. 1973.
flexibility. Pharm Res 16:1633 1638.
Effect of bile salts on partitioning behavior and GI
19. Lee SJ, Kurihara-Bergstrom T, Kim SW. 1987. Ion-
absorption of a quaternary ammonium compound,
paired drug diffusion through polymer membranes.
isopropamide iodide. J Pharm Sci 62:1121 1125.
Int J Pharm 47:59 73.
7. Gibaldi M, Grundhofer B. 1973. Enhancement of
20. Sarveiya V, Templeton JF, Benson HA. 2004. Ion-
intestinal absorption of a quaternary ammonium
pairs of ibuprofen: Increased membrane diffusion.
compound by salicylate and trichloroacetate. J Pharm
J Pharm Pharmacol 56:717 724.
Sci 62:343 344.
21. Wang M, Fang L, Ren C, Li T. 2008. Effect of ion-
8. Masaki BW, Lien EJ, Biles JA. 1973. Transport of
pairing and enhancers on scutellarin skin perme-
quaternary ammonium-alkyl sulfate and sodium
ability. J Pharm Pharmacol 60:429 435.
salicylate across rat small intestine. Acta Pharma-
22. Teksin ZS, Hom K, Balakrishnan A, Polli JE. 2006.
ceutica Suecica 10:43 52.
Ion pair-mediated transport of metoprolol across a
9. Newburger J, Kotenbauder HB. 1977. The effect of
three lipid-component PAMPA system. J Control
ion-pair formation on the lethal dose of methanthe-
Release 116:50 57.
line bromide. Life Sci 20:627 630.
23. Matsumoto H, Yamamoto R, Tanioka A. 2005.
10. Bhuta SI, Sugita ET, Niebergall PJ, Schnaare RL.
Membrane potential across low-water-content
1980. Influence of various anions on intestinal
charged membranes: Effect of ion pairing. J Phys
disappearance of hexamethonium chloride and
Chem 109:14130 14136.
pralidoxime chloride in rats. J Pharm Sci 69:923
24. Jones JW, Gibson HW. 2003. Ion pairing and host-
928.
guest complexation in low dielectric constant sol-
11. Ruifrok PG. 1981. Uptake of quaternary ammo-
vents. J Am Chem Soc 125:7001 7004.
nium compounds into rat intestinal brush border
25. Yeh SJ, Yang YM, Chang CH. 2005. Cosolvent
membrane vesicles. Biochem Pharmacol 30:2637
effects on the stability of catanionic vesicles formed
2641.
from ion-pair amphiphiles. Langmuir 21:6179
12. Tomlinson E, Davis SS. 1976. Increased uptake of
6184.
an anionic drug by mucous membrane, upon for-
26. Buchner R, Samani F, May PM, Sturm P, Hefter G.
mation of ion-association species with quaternary
2003. Hydration and ion pairing in aqueous sodium
ammonium salts [proceedings]. J Pharm Pharmacol
oxalate solutions. Chem Phys Chem 4:373 378.
28:75P.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 10, OCTOBER 2009