Hydrophobic ion pairing of isoniazid using a prodrug approach

Article abstract of DOI:10.1002/jps.10116

Inhalation therapy for infectious lung diseases, such as tuberculosis, is currently being explored, with microspheres being used to target alveolar macrophages. One method of drug encapsulation into polymeric microspheres to form hydrophobic ion-paired (HIP) complexes, and then coprecipitate the complex and polymer using supercritical fluid methodology. For the potent antituberculosis drug, isoniazid (isonicotinic acid hydrazide, INH), to be used in this fashion, it was modified into an ionizable form suitable for HIP. The charged prodrug, sodium isoniazid methanesulfonate (Na-INHMS), was then ion paired with hydrophobic cations, such as alkyltri-methylammonium or tetraalkylammonium. The logarithms of the apparent partition coefficients (log P′) of various HIP complexes of INHMS display a roughly linear relationship with the numbers of carbon atoms in the organic counterions. The water solubility of the tetraheptylammonium-INHMS complex is about 220-fold lower than that of Na-INHMS, while the solubility in dichloromethane exceeds 10 mg/mL, which is sufficient for microencapsulation of the drug into poly(lactide) microspheres. The actual logarithm of the dichloromethane/water partition coefficient (log P) for tetraheptylammonium-INHMS is 1.55, compared to a value of - 1.8 for the sodium salt of INHMS. The dissolution kinetics of the tetraheptylammonium-INHMS complex in 0.9% aqueous solutions of NaCl was also investigated. Dissolution of tetraheptylammonium-INHMS exhibited a first-order time constant of about 0.28 min^-1, followed by a slower reverse ion exchange process to form Na-INHMS. The half-life of this HIP complex is on the order of 30 min, making the enhanced transport of the drug across biological barriers possible. This work represents the first use of a prodrug approach to introduce functionality that would allow HIP complex formation for a neutral molecule.

Full text of DOI:10.1002/jps.10116

Hydrophobic Ion Pairing of Isoniazid Using a
Prodrug Approach
HUIYU ZHOU,1 CORINNE LENGSFELD,2 DAVID J. CLAFFEY,1 JAMES A. RUTH,1 BROOKS HYBERTSON,4
THEODORE W. RANDOLPH,3 KA-YUN NG,1 MARK C. MANNING1
1Department of Pharmaceutical Sciences, University of Colorado Health Sciences Center, Campus Box C238,
4200 E. Ninth St., Denver, Colorado
2Department of Engineering, University of Denver, Denver, Colorado
3Department of Chemical Engineering, University of Colorado at Boulder, Boulder, Colorado
4Webb-Waring Institute for Cancer, Aging, and Antioxidant Research and School of Medicine,
University of Colorado Health Sciences Center, Denver, Colorado
Received 5 September 2001; revised 1 January 2002; accepted 11 January 2002
ABSTRACT: Inhalation therapy for infectious lung diseases, such as tuberculosis, is
currently being explored, with microspheres being used to target alveolar macrophages.
One method of drug encapsulation into polymeric microspheres to form hydrophobic
ion-paired (HIP) complexes, and then coprecipitate the complex and polymer using
supercritical fluid methodology. For the potent antituberculosis drug, isoniazid
(isonicotinic acid hydrazide, INH), to be used in this fashion, it was modified into an
ionizable form suitable for HIP. The charged prodrug, sodium isoniazid methanesulfo-
nate (Na–INHMS), was then ion paired with hydrophobic cations, such as alkyltri-
methylammonium or tetraalkylammonium. The logarithms of the apparent partition
coefficients (log P0) of various HIP complexes of INHMS display a roughly linear
relationship with the numbers of carbon atoms in the organic counterions. The water
solubility of the tetraheptylammonium–INHMS complex is about 220-fold lower than
that of Na–INHMS, while the solubility in dichloromethane exceeds 10 mg/mL, which is
sufficient for microencapsulation of the drug into poly(lactide) microspheres. The actual
logarithm of the dichloromethane/water partition coefficient (log P) for tetraheptylam-
monium–INHMS is 1.55, compared to a value of ꢀ 1.8 for the sodium salt of INHMS.
The dissolution kinetics of the tetraheptylammonium–INHMS complex in 0.9%
aqueous solutions of NaCl was also investigated. Dissolution of tetraheptylammo-
nium–INHMS exhibited a first-order time constant of about 0.28 minꢀ1, followed by a
slower reverse ion exchange process to form Na–INHMS. The half-life of this HIP
complex is on the order of 30 min, making the enhanced transport of the drug across
biological barriers possible. This work represents the first use of a prodrug approach to
introduce functionality that would allow HIP complex formation for a neutral molecule.
ß 2002 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 91:1502–1511,
2002
Keywords: hydrophobic ion pairing; prodrugs; isoniazid
INTRODUCTION
Correspondence to: Mark C. Manning (Telephone: 303-315-
Inhalation therapy for infectious lung diseases
6162; Fax: 303-315-6281; E-mail: Mark.Manning@uchsc.edu)
is currently being explored,1 with microspheres
Journal of Pharmaceutical Sciences, Vol. 91, 1502–1511 (2002)
ß 2002 Wiley-Liss, Inc. and the American Pharmaceutical Association
being used to target alveolar macrophages.2 One
1502
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

HYDROPHOBIC ION PAIRING OF ISONIAZID
1503
method we developed for drug encapsulation into
polymeric microspheres is to form hydrophobic
ion-paired complexes, and then coprecipitate the
complex and polymer using supercritical fluid
methodology termed precipitation with a com-
pressed antisolvent (PCA).3 Hydrophobic ion
pairing (HIP) is a technique that increases the
hydrophobicity of molecules containing ionizable
groups by stoichiometrically replacing polar coun-
terions with more hydrophobic ones. This pro-
cess has been used to solubilize ionic molecules
in nonpolar solvents,4–7 enhance the transport of
proteins and DNA,8,9 and increase the bioavail-
ability of ionic drugs.10–14
INH is an important drug used for the treat-
ment of tuberculosis (TB).15 Due to noncompli-
ance with current TB therapies,16–23 there has
been an effort to develop alternative drug delivery
systems to provide sustained release of anti-TB
drugs at the site of infection (e.g., the lungs).1,24–30
Therefore, incorporation of anti-TB drugs, like
INH, into microparticles that would be small
enough to provide efficient pulmonary delivery, is
desirable. However, because INH is an uncharged,
hydrophilic molecule, it is not amenable to ion
pairing and its solubility in organic solvents is
insufficient to use it directly in the PCA process.
Therefore, to increase the solubility of INH in
nonpolar solvents and to allow the use of the HIP
and PCA processes, an ionizable prodrug of INH,
Na–INHMS, was synthesized. Although HIP has
been described in many systems,4,5,10,11 this repre-
sents the first example of using a prodrug
approach to introduce functionality that would
allow ion pairing to take place, with the aim of
subsequent encapsulation into microspheres for
inhalation therapy.
WI). Ethyl alcohol (denatured) was from Aldrich
(Milwaukee, WI). 2-Propanol was from Fisher
Scientific (Fair Lawn, NJ). Silica gel (‘‘Flash,’’
˚
32–63 mm, 60 A) was obtained from Scientific
Adsorbent Inc. (Atlanta, GA). Medical grade
compressed nitrogen was from General Air
Service & Supply (Denver, CO).
Synthesis and Characterization of Na–INHMS
Briefly, Na–INHMS was synthesized using the
procedure previously described by Logemann.31
An aqueous (100 mL) solution of sodium bisulfite
(52 g, 0.5 mol) was added to 0.5 mol of isoniazid
(68.5 g) in 38 mL of 39% formaldehyde. The
mixture was then heated at 1008C for 8 h and
subsequently evaporated under reduced pressure
to give a yellow solid. The yellow solid was
dissolved in ethanol/water mixture. Recrystalli-
zation from the mixture gave Na–INHMS as a
white solid. The purity of the derivative was
characterized by both 1H- and 13C-NMR. 1H-NMR
(D2O) spectrum of Na–INHMS: d 3.9 (s, 2H),
7.5 (d, 2H, J ¼ 5.6 Hz), 8.5 (d, 2H, J ¼ 5.56 Hz).
13C-NMR (D2O) spectrum of Na–INHMS: d 66.3,
121.8, 121.8, 140.6, 140.6, 149.6, 149.7, 166.9.
Determination of Extinction Coefficient
of INHMS in Water
Approximately 78 mg of Na–INHMS powder was
accurately weighed into a 50-mL beaker and
dissolved in 20 mL of double distilled deionized
water (DDW). The solution was transferred into a
100-mL volumetric flask. The beaker was rinsed
with approximately 10 mL of DDW for three
times, and all the liquid was transferred into the
flask and diluted to volume with additional DDW.
A series of aliquots (0.1, 0.2, 0.4, 0.6, 0.8, and
1 mL) of this stock solution were diluted to
10.0 mL with DDW and analyzed by UV spectro-
photometry. The absorbance was measured at
262 ꢁ 1 nm. A plot of absorbance versus drug
concentration was linear within the range of
7.82–78.2 mg/mL (r2 ¼ 0.99). The extinction coeffi-
MATERIALS AND METHODS
Materials
INH, dodecyltrimethylammonium bromide,
tetradecyltrimethylammonium bromide, hexade-
cyltrimethylammonium bromide, tetraethylam-
monium bromide, tetraoctylammonium bromide
were obtained from Sigma Chemical Co. (St.
Louis, MO). Tetrabutylammonium bromide, tetra-
pentylammonium bromide, tetraheptylammo-
nium bromide were purchased from Fluka
(Switzerland). Dichloromethane was obtained
from Sigma-Aldrich (Sigma Chemical Co., St
Louis, MO and Aldrich Chemical Co., Milwaukee,
cient was determined to be 3.16 ꢂ 103 Mꢀ1 ꢃ cmꢀ1
.
Determination of the Extinction Coefficient
of INHMS in Dichloromethane
Approximately 175 mg of tetraheptylammonium–
INHMS complexes were accurately weighed and
dissolved into 10 mL of dichloromethane. A
0.01-mL volume of this solution was then diluted
to 1 mL with dichloromethane and analyzed by
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

1504
ZHOU ET AL.
UV spectrophotometry. The absorbance was mea-
chromatography utilizing the Aldrich1 flash-
chromatography assembly with frit & System 451
connections (Sigma-Aldrich, Inc.) was used.
Briefly, the ‘‘flash’’ silica gel was packed into the
column to a height of 15–22 cm. The column was
then flushed with a solvent consisting of 60%
2-propanol and 40% ethanol. After this initial
equilibration process, samples were applied to
the column as a 20–30% solution. Medical grade
compressed nitrogen was applied to the column
to maintain a flow rate of 1.5–2.5 mL/min. The
eluents were collected as 5-mL aliquots and were
analyzed for drug content and purity using UV
spectrophotometry and thin layer chromato-
graphy. Aliquots containing the purified ion-
paired complex was then dried using a SpeedVac
dryer and saved for subsequent experimental use.
sured at 262 ꢁ 1 nm. The extinction coefficient was
calculated to be (2.43 ꢁ 0.02) ꢂ 103 Mꢀ1 ꢃ cmꢀ1
.
Formation and Partitioning of HIP–INHMS
Various HIP–INHMS were formed and extracted
into organic solvents using the conventional flask-
shake partitioning procedure.32 Na–INHMS was
first dissolved in DDW to make a 1 mM solution.
To 2 mL of this solution, 2 mL of dichloromethane
containing an organic ammonium salt at a con-
centration of 1 mM was added. The mixture was
mixed by vortexing for 10 min and then was
placed on a laboratory rocker (Enprotech, Inte-
grated Separation Systems) for 2.5 h at approxi-
mately 18 cycles/min to ensure full interaction
between the drug and the organic cations. The
aqueous and the organic phases were then sepa-
rated by centrifugation at 2000 rpm for 10 min.
The drug concentrations in the aqueous and the
organic layer were measured using UV spectro-
photometry at 262 ꢁ 1 nm. For the alkyltrimethy-
lammonium complexes, the mass balance was
quantitative. In the cases of the tetraalkylammo-
nium salts, the recoveries ranged from 95–107%.
Because this measurement did not distinguish
between the different forms of INHMS (Na–
INHMS, HIP–INHMS), the apparent partition
coefficient (P0) is reported as the value observed
by direct determination of the total INHMS
concentration in each phase (eq. 1) and cannot
be considered a true partition coefficient.
Measurement of the (True) Dichloromethane/Water
Partition Coefficient of the Tetraheptylammonium–
INHMS Complex
Two milliliters of dichloromethane containing
5 mM of the tetraheptylammonium–INHMS com-
plex was mixed with 2 mL of DDW by vortexing
for 5 min. The mixture was then centrifuged at
2000 rpm for 10 min. As no salt is present, there is
no need to wait for equilibration between charged
species, and this time has been found to be
sufficient to allow partitioning of HIP complexes
into two immiscible solvents.2 At the end of the
centrifugation, the organic layer was collected
and analyzed for drug content by UV spectro-
photometry. The drug content in the aqueous
phase was determined by subtracting the drug
content in the organic phase from the total drug
content. The dichloromethane/water partition co-
efficient (P) was calculated according to eq. 2.
½Sþꢃ Dꢀꢄorgþ ½NaþꢃDꢀꢄorg
P0 ¼
ð1Þ
½SþꢃDꢀꢄaqþ ½NaþꢃDꢀꢄaq
Here, Dꢀ is the drug anion, which interacts
with an organic counterion (Sþ) to form the ion
pair (Sþ ꢃ Dꢀ).
½SþꢃDꢀꢄorg
P ¼
ð2Þ
½SþꢃDꢀꢄaq
Isolation and Purification of the
Tetraheptylammonium–INHMS Complex
In this case, INHMS exists only in the ion-
paired form, so P represents the true partition
coefficient for the complex.
To generate tetraheptylammonium–INHMS com-
plexes for further characterization, the partition-
ing experiment described above was scaled up.
The concentration of Na–INHMS and tetrahep-
tylammonium bromide solutions was raised to
50 mM, and the volumes of both phases were
increased to 10 mL. After phase separation, the
organic layer was collected, dried under nitrogen
and redissolved in a solvent consisting of 60%
2-propanol and 40% ethanol. To purify the
tetraheptylammonium–INHMS complex, flash
Measurement of the Aqueous Solubility
of Na–INHMS and the Tetraheptylammonium–
INHMS Complex
Approximately 1000 mg of Na–INHMS or 250 mg
of tetraheptylammonium–INHMS complexes
were placed in a 20-mL glass vial (27 ꢂ 57 mm).
Five milliliters of DDW were then added to the
vial and the mixture was stirred at a speed of
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

HYDROPHOBIC ION PAIRING OF ISONIAZID
1505
approximately 300 rpm for 24 h at room tempera-
ture. At the end of the stirring, the supernatant
was collected for determination of the drug con-
centration using UV spectrophotometry. To re-
move any particulates of the undissolved complex,
supernatant was subjected to filtration using
0.2 mm pore size syringe filters.
and dichloromethane. The partitioning of HIP–
INHMS (Sþ ꢃ Dꢀ) between the aqueous and
organic phases can be characterized by the (true)
partition coefficient (P) (eq. 2). However, in this
study no attempt was made to distinguish be-
tween Naþ ꢃ Dꢀ and Sþ ꢃ Dꢀ in solutions. There-
fore, the partitioning behavior of the drug is
reported as the apparent dichloromethane/water
partition coefficient of the ion pair (P0) (eq. 1).
These are termed apparent partition coefficients,
as no distinction is made to the nature of the
counterion for the drug molecule in each phase.
In each phase, the drug could be present either as
the HIP complex or as the sodium salt.
Hydrophobic counterions except dodecyltri-
methylammonium increased the log P0 of INHMS
(Figure 1). In contrast, no surfactant-facilitated
phase transfer of isoniazid was observed (data not
shown), demonstrating the importance of the
added negative charge on the isoniazid derivative.
When log P0 is very low, the errors in the mea-
surement may be large, and may explain the
slightly decreased log P0 by dodecyltrimethylam-
monium compared to Na–INHMS alone. Three
complexes—tetrapentylammonium–INHMS, te-
traheptylammonium–INHMS, and tetraoctylam-
monium–INHMS—exhibited greater log P0 than
INH, suggesting these complexes are more hydro-
phobic than the parent drug. Note that the
partition coefficients were measured in a di-
chloromethane/water, rather than the traditional
octanol–water system. First, this is the most
relevant solvent, as these HIP complexes must be
solubilized in chlorocarbons to use PCA with
poly(lactides) and poly(glycolides). Second, the
amount of water present in 1-octanol after par-
titioning is significant (ꢅ1.7 M), and this might
give an artificially low log P0 value for such
hydrophobic complexes. Finally, octanol–water
measurements are most relevant when examin-
ing permeability across biological barriers, and
this was not the goal of this study.
Study of the Dissolution Kinetics of the
Tetraheptylammonium–INHMS Complex
The purified tetraheptylammonium–INHMS
complex is an extremely viscous semifluid. To
maintain the amount as well as the surface area of
the samples consistent, 2 mL of tetraheptyl-
ammonium–INHMS solution (200 mM) in dichlo-
romethane was placed in a 20-mL glass vial
(27 ꢂ 57 mm) and then allowed to evaporate to
dryness in the hood. A film of tetraheptylam-
monium–INHMS would form on the bottom of
the vial. Five milliliters of DDW or sodium chlo-
ride solution (0.9%) were then added to the vial.
The solution was stirred at a speed of approxi-
mately 300 rpm. At designated time points,
0.1-mL aliquots were removed, filtered using a
0.2 mm pore size syringe filter, and assayed for
drug content using UV spectrophotometry.
RESULTS AND DISCUSSION
The synthesis of INHMS proceeded using a
modification of Logemann.31 This allowed the
incorporation of a readily ionizable group onto the
INH pharmacophore. However, unless the group
can be readily removed in vivo, it will not function
as a true prodrug. Previously, INHMS and INH
had been shown to have equal anti-TB activity
both in vivo and in vitro.33 Furthermore, Hsu and
Ho demonstrated that INHMS was rapidly con-
verted into INH and acetyl-INH upon adminis-
tration to rabbits, suggesting the anti-TB activity
of INHMS comes from its active metabolite,
INH.34 Thus, INHMS does appear to act as a
prodrug of INH. Because INHMS is fully ionized
under physiological conditions, it can be readily
ion paired with hydrophobic counterions, with the
ultimate goal of encapsulation into polymeric
microspheres using PCA.
The partitioning of a HIP complex between an
aqueous and an organic phase depends on its in-
trinsic hydrophobicity as well as the equilibrium
constant for the ion-pair formation and the equi-
librium constants for dissociation of all species in
both phases.4
Two groups of cationic surfactants (tetra-
alkylammonium and trimethylalkylammonium)
were examined for their capability of transfer-
ring INHMS to dichloromethane after the HIP
complex formation. In general, within a parti-
cular homologous series, counterions with longer
Formation and Partitioning of HIP–INHMS
In the partitioning experiments, HIP–INHMS
were formed and then distributed between water
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

1506
ZHOU ET AL.
Figure 1. Extraction of HIP–INHMS by dichloromethane. INH, isoniazid;
INHMS, isoniazid methanesulfonate; DTMA, dodecyltrimethylammonium; TTMA,
tetradecyltrimethylammonium; HTMA, hexadecyltrimethylammonium; TEA, teraethyl-
ammonium; TBA, tetrabutylammonium; TPA, tetrapentylammonium; THA, tetrahep-
tylammonium; TOA, tetraoctylammonium. The values are the averageꢁ standard
deviation; n ¼ 3.
hydrocarbon chains results in a larger P0 for the
HIP complex. Because the length of alkyl chains
influences the size as well as the hydrophobicity of
the organic cation, the relationship between the
apparent partition coefficients of HIP–INHMS
and the size/hydrophobicity of the organic ca-
tions was examined (Figure 2). The number of
carbon atoms in the cation was used as an estima-
tion of the relative size of the cation. Overall,
the results indicate that log P0 of HIP–INHMS
follows a roughly linear relationship with respect
to the number of carbon atoms in the organic
counterions.
This type of behavior has been observed pre-
viously for HIP complexes. For example, Adjei
et al. reported in a study on alkylsulfonate com-
plexes with prolide that, for alkylsulfonates with
short chains (C6–C10), there was a linear depen-
dence of log P on the chain length.35 Takacs-
´
´
´
Novak and Szasz also reported a linear correla-
tion between log P0 of the ion pair and the size
of the counterion, which was expressed as the
solvent (water)-accessible surface area (SASA).36
Using either the true partition coefficient (log P)
of the pure ion pair35 or the apparent partition
coefficient (log P0) values at 1:50 molar ratio
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

HYDROPHOBIC ION PAIRING OF ISONIAZID
1507
Figure 2. Relationship between the apparent partition coefficient (P0) of HIP–
INHMS and the number of carbon atoms in the organic counterions.
between the drug and the organic counter (dimi-
nishing the ion-pair dissociation by the excess
organic counterions), the results are similar.36
In our study, log P0 of HIP complexes was mea-
sured at a 1:1 molar ratio between the drug and
the organic counterion. Therefore, it appears that
the presence of excess counterion does not greatly
affect the relationship between the hydropho-
bicity of the species and the apparent partition
coefficient. Together, these results indicate that
increasing the size and overall hydrophobicity of
the organic counterion produces a HIP complex
that is more able to partition into a lipophilic
phase than the parent compound, and that the
relationship between counterion properties and
the partitioning behavior is relatively well defined.
rithm of the (true) water/dichloromethane parti-
tion coefficient (log P) from ꢀ 1.8 to 1.55. The
approximate 1700-fold increase in the actual
partition coefficient is comparable to the increases
reported for other HIP complexes.3,6,7 The con-
centration of the tetraheptylammonium–INHMS
complex in dichloromethane after direct partition-
ing approached 11 mg/mL, which is more than
sufficient to allow subsequent incorporation in
poly(L-lactide) microspheres using an antisolvent
process.12 Similarly, compared to that of Na–
INHMS, the aqueous solubility of the tetraheptyl-
ammonium–INHMS complex decreased by about
220-fold (Figure 3B).
Dissolution Kinetics of the INHMS–
Tetraheptylammonium Complex
Hydrophobicity of the Tetraheptylammonium–
INHMS Complex
An in vitro dissolution study was performed
on the tetraheptylammonium–INHMS complex.
One would expect that HIP complexes display
biphasic kinetics when dissolved in an aqueous
solution containing salts, with dissolution fol-
lowed by dissociation of the HIP complexes.37
However, little is known about the relative rates
of these two processes for HIP complexes. After
dissolution, dissociation of the complex is driven
by ion exchange, and the drug concentration
should increase, as the nonion paired drug is
much more soluble in water than the correspond-
ing HIP complex. The dissolution and dissociation
behavior of HIP complexes can be determined
by kinetic modeling. The first rate constant, k1,
Because the tetraheptylammonium–INHMS
complex is readily extracted by dichloromethane,
it can be easily produced on a larger scale, allow-
ing for a more accurate measurement of the
partition coefficient of a HIP complex of INHMS.
Because the starting material is the intact,
purified HIP complex and no salt is present,
this value should represent the true dichloro-
methane–water partition coefficient.
The large P value suggests that appreciable
solubility in dichloromethane should be obtained
for this complex (Figure 3A). Ion pairing with
tetraheptylammonium also increased the loga-
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

1508
ZHOU ET AL.
Figure 4. Plots of the ln(1-fraction released) versus
time for the tetraheptylammonium–INHMS complex
in DDW (A) and 0.9% sodium chloride solution (B).
(A) Concentrations of the dissolved tetraheptylam-
monium–INHMS complex ([DS]), normalized by its
equilibrium solubility in DDW ([DS]eq), are plotted on a
log scale versus time. The resulting slope, k1, is the
observed first-order rate constant for the HIP complex
dissolution in DDW. (B) Concentrations of the dissolved
INHMS ([D]), normalized by its equilibrium solubility
in 0.9% sodium chloride solution ([D]eq), are plotted on
a log scale versus time. The dashed line indicates a
simple model with a single dissolution rate; the solid
line indicates a full model including reverse ion pairing.
Figure 3. Effect of ion pairing with tetraheptylam-
monium on the water/dichloromethane partition coeffi-
cient (P) (A) and the aqueous solubility of INHMS (B).
The values are the average ꢁ standard deviation; n ¼ 3.
describes the dissolution of the solid complex into
liquid media. The other two rate constants, k2f
and k2r, describes the rate of reverse ion pairing
and ion pairing, respectively. It is important to
note that the dissolution process (eq. 3), has been
modeled as a kinetic phenomenon. As it will be
dependent on stirring rates and other factors, it
should be considered a relative value. Conversely,
the ion exchange process has both a forward and
reverse component, making it a true thermody-
namic equilibrium. These factors were included in
derivation of the model.37 Figures 4 and 5 show
dissolution/dissociation profiles of the tetrahep-
tylammonium–INHMS complex in DDW and
sodium chloride solutions. In DDW, only dissolu-
tion is possible, and the drug concentration
plateaus at the solubility limit of the HIP complex
(Figure 5).
k1
ðSþꢃDꢀÞsolid ! ðSþꢃDꢀÞaq
ð3Þ
ð4Þ
k2f
ðSþꢃDꢀÞaq þ NaCl aq
ðNaþꢃDꢀÞaq þ ðSþꢃClꢀÞaq
,
k2r
The dissolution behavior is well modeled by a
single first order rate constant k1, which was
found to be 0.28 minꢀ1 (Figure 4A and Table 1).
This corresponds to a half-life (t1/2) for dissolution
of 2.5 min. This observed rate constant can be
influenced by a number of factors, for example,
the amount of samples, the sample–water surface
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

HYDROPHOBIC ION PAIRING OF ISONIAZID
1509
approach saturation under these conditions. The
predicted values did not match data points at
early time points very well. Because the drug
concentrations at the initial stage were very low,
the poor agreement is probably due to larger
errors in making small measurements.
Detergent-enabled transport of proteins and
nucleic acids through hydrophobic solvents by ion
pairing has been shown in previous work.8,9 In
addition, formation of HIP complexes appears to
increase bioavailability of ionic drugs.10–14 All of
this previous work suggests that the dissociation
rates of HIP complexes are relatively slow and
affected by the nature of the organic counterions
used,37 but none of these studies addressed this
issue directly. This study provides a detailed des-
cription of not only the partitioning behavior of
HIP complexes of INHMS, but also their propen-
sity to dissociate in isotonic solutions.
Figure 5. Dissolution profiles of the tetraheptylam-
monium–INHMS complex in DDW (&) and 0.9%
sodium chloride solution (*) with profiles predicted
from the model. For the measured concentrations,
values are the average ꢁ standard deviation; n ¼ 3.
For the model prediction, the estimated standard error
is 2.12 mM.
CONCLUSIONS
area, and the mixing conditions, some of which
were not precisely controlled or measured. There-
fore, k1 cannot be considered an intrinsic property
of the complex, in contrast to k2f and k2r.
A prodrug approach was employed to introduce a
charged moiety into a neutral molecule, thereby
allowing a hydrophobic ion-paired complex to be
formed. As a result, the hydrophobicity of the
prodrug is significantly enhanced, resulting in a
220-fold decrease in the water solubility and a
1700-fold increase in the dichloromethane/water
partition coefficient. The dissolution kinetics of
the HIP complex in an aqueous electrolyte solu-
tion exhibits a rapid dissolution phase followed by
a slower reverse ion-pairing process. Although the
half-life is on the order of 30 min, it is possible
that the reverse ion pairing is still sufficiently
slow to allow enhanced penetration of biological
barriers by HIP complexes. Together, this study
demonstrates that a prodrug approach can be
taken to introduce a changed moiety into a neutral
molecule, and that this compound can form stable
HIP complexes. This now allows alternative meth-
ods of encapsulation into polymeric microspheres.
In normal saline, there are two distinct types
of decays observed: one similar to the dissolution
curve observed in pure water, and the other some-
what slower (Figures 4B and 5). Presumably, the
first is due to dissolution, again with a half-life
near 2.5 min. The second phase is slower, and
must be due to reverse ion pairing with the
sodium ions in solution. Table 1 lists the kinetic
rate constants for the decay behaviors as well as
the saline solubility of the HIP complex and total
drug in 0.9% sodium chloride solution. They are
derived based on the model built by Randolph and
coworkers.37 Figure 5 shows the dissolution
profiles predicted by this model. The half-life for
the second decay process appears to be about
25 min, with about 200 min being required to
Table 1. Solubilities and Kinetic Rate Constants for Dissolution and Reverse Ion Exchange of the
Tetraheptylammonium–INHMS Complex
Kinetic Rate Constants
Solubility (mM)
Measured
Measured
Computed
k1
k2f
k2r
Correlation Tetraheptylammonium– INHMS in Tetraheptylammonium–
(l/min) (L/mol min) (L/mol min) Coefficient r2
INHMS in DDW
0.9% NaCl
INHMS in 0.9% NaCl
0.28 0.55 0.56 0.984
3.7
38.7
7.8
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

1510
ZHOU ET AL.
REFERENCES
16. Friedman L, Berg J, Sipan C, Hovell M,
Catanzaro A, Moser K, Kelley N. 2000. Past medi-
cation compliance and adherence to current tuber-
culosis treatment Hispanic teenagers. J Gen Intern
Med 15:115.
17. Jerant AF, Bannon M, Rittenhouse S. 2000. Identi-
fication and management of tuberculosis. Am Fam
Phys 61:2667–2678.
18. Johnson JL, Okwera A, Nsubuga P, Nakibali JG,
Whalen CC, Hom D, Cave MD, Yang ZH, Mugerwa
RD, Ellner JJ. 2000. Efficacy of an unsupervised
8-month rifampicin-containing regimen for the
treatment of pulmonary tuberculosis in HIV-
infected adults. Int J Tuberc Lung Dis 4:1032–1040.
19. Lazarus A, Sanders J. 2000. Management of
tuberculosis—Choosing an effective regimen and
ensuring compliance. Postgrad Med 108:71–78.
20. Malotte CK, Hollingshead JR, Larro M. 2001.
Incentives vs outreach workers for latent tuber-
culosis treatment in drug users. Am J Prev Med
20:103–107.
1. Suarex S, O’Hara P, Kazantseva M, Newcomer CE,
Hopfer R, McMurray DN, Hickey AJ. 2001. Respir-
able PLGA microspheres containing rifampicin for
the treatment of tuberculosis: Screening in an in-
fectious disease model. Pharm Res 18:1315–1319.
2. Sharma R, Saxena D, Dwivedi AK, Misra A. 2001.
Inhalable microparticles containing drug combina-
tions to target alveolar macrophages for treatment
of pulmonary tuberculosis. Pharm Res 18:1405–
1410.
3. Falk R, Randolph TW, Meyer JD, Kelly RM,
Manning MC. 1997. Controlled release of ionic
compounds from poly (L-lactide) microspheres pro-
duced by precipitation with a compressed antisol-
vent. J Control Rel 44:77–85.
4. Quintanar-Guerrero D, Allemann E, Fessi H,
Doelker E. 1997. Applications of the ion-pair concept
to hydrophilic substances with special emphasis on
peptides. Pharm Res 14:119–127.
21. Archbald LR, Neuzil KM, Griffin MR, Schaffner W,
Arons MM, Chapdelaine PA, Harter TV. 1999. Risk
factors for non-compliance with tuberculosis pre-
ventive therapy among subjects treated by a metro-
politan health department. Am J Resp Crit Care
159:A224–A228.
22. Coker R. 2000. Tuberculosis, culture and coercion.
Eur J Public Health 10:223–227.
23. Coker R. 2000. Tuberculosis, non-compliance and
detention for the public health. J Med Ethics 26:
157–159.
24. Gangadharam PRJ, Kailasam S, Srinivasan S,
Wise DL. 1994. Experimental chemotherapy of
tuberculosis using single-dose treatment with iso-
niazid in biodegradable polymers. J Antimicrob
Chemother 33:265–271.
5. Meyer JD, Manning MC. 1998. Hydrophobic ion
pairing: Altering the solubility properties of bio-
molecules. Pharm Res 15:188–193.
6. Powers ME, Matsuura J, Brassell J, Manning MC,
Shefter E. 1993. Enhanced solubility of proteins
and peptides in nonpolar solvents through hydro-
phobic ion pairing. Biopolymers 33:927–932.
7. Meyer JD, Kendrick BS, Matsuura JE, Ruth JA,
Bryan PN, Manning MC. 1996. Generation of
soluble and active subtilisin and alpha-chymotryp-
sin in organic solvents via hydrophobic ion pairing.
Int J Pept Protein Res 47:177–181.
8. Bromberg LE, Klibanov AM. 1994. Detergent-
enabled transport of proteins and nucleic acids
through hydrophobic solvents. Proc Natl Acad Sci
USA 91:143–147.
25. Hsu YY, Gresser JD, Trantolo DJ, Lyons CM,
Gangadharam PRJ, Wise DL. 1994. In-vitro con-
trolled-release of isoniazid from poly(lactide-co-
glycolide) matrices. J Controlled Rel 31:223–228.
26. Batyrbekov EO, Iskakov R, Zhubanov BA. 1998.
Synthetic and natural polymers as drug carriers for
tuberculosis treatment. Macromol Symp 127:251–
256.
27. Gangadharam PRJ, Geeta N, Hsu YY, Wise DL.
1999. Chemotherapy of tuberculosis in mice using
single implants of isoniazid and pyrazinamide. Int
J Tuberc Lung Dis 3:515–520.
28. O’Hara P, Hickey AJ. 2000. Respirable PLGA
microspheres containing rifampicin for the treat-
ment of tuberculosis: Manufacture and character-
ization. Pharm Res 17:955–961.
29. Dutt M, Khuller GK. 2001. Therapeutic efficacy of
poly(DL-lactide-co-glycolide)- encapsulated antitu-
bercular drugs against Mycobacterium tuberculo-
sis infection induced in mice. Antimicrob Agents
Chemother 45:363–366.
9. Bromberg LE, Klibanov AM. 1995. Transport
of proteins dissolved in organic solvents across bio-
mimetic membranes. Proc Natl Acad Sci USA 92:
1262–1266.
10. Neubert R. 1989. Ion pair transport across mem-
branes. Pharm Res 6:743–747.
11. Neubert R, Fischer S. 1991. Influence of lipophilic
counter ions on the transport of ionizable hydro-
philic drugs. J Pharm Pharmacol 43:204–206.
12. Hatanaka T, Kamon T, Morigaki S, Katayama K,
Koizumi T. 2000. Ion pair skin transport of a zwit-
terionic drug, cephalexin. J Control Rel 66:63–71.
13. Valenta C, Siman U, Kratzel M, Hadgraft J. 2000.
The dermal delivery of lignocaine: influence of ion
pairing. Int J Pharm 197:77–85.
14. Aungst BJ, Hussain MA. 1992. Sustained propra-
nolol delivery and increased oral bioavailability in
dogs given a propranolol laurate salt. Pharm Res
9:1507–1509.
15. Seth V. 1990. Isomiazid—The pivot of chemother-
apy in tuberculosis. Indian Pediatr 27:119–123.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002

HYDROPHOBIC ION PAIRING OF ISONIAZID
1511
30. Dutt M, Khuller GK. 2001. Sustained release of
isoniazid from a single injectable dose of poly (DL-
lactide-co-glycolide) microparticles as a therapeutic
approach towards tuberculosis. Int J Antimicrob
Agents 17:115–122.
31. Logemann W. 1956. Derivatives of carboxylic acid
hydrazides and method of making them. U.S.
Patent No. 2,759,994
32. Dearden JC, Bresnen GM. 1988. The measure-
ments of partition coefficients. Quant Struct Act
Relat 7:133–144.
33. Kitamoto O, Okada H, Fukuhara Y, Takayama H,
Ishii S, Sakamoto T. 1953. Antituberculous pro-
perties of isonicotinyl hydrazide methansulfonate
(Na-salt). Jpn J Tuberculosis 1:92–97.
rabbit blood by high-performance liquid chromato-
graphy. J Chromatogr 493:305–312.
35. Adjei A, Rao S, Garren J, Menon G, Vadnere M.
1993. Effect of ion-pairing on 1-octanol-
water partitioning of peptide drugs. I: The non-
apeptide leuprolide acetate. Int J Pharm 90:141–
149.
´
´
´
36. Takacs-Novak K, Szasz G. 1999. Ion-pair
partition of quaternary ammonium drugs: The
influence of counter ions of different lipophili-
city, size, and flexibility. Pharm Res 16:1633–
1638.
37. Pitera D, Lengsfeld CS, Manning MC, Randolph
TW. 2002. Fundamental study of hydrophobic ion-
pairing: Pharmaceutical complex chemical nature
and kinetic behavior. Pharm Res, accepted for
publication.
34. Hsu KY, Ho Y. 1989. Determinutesation of iso-
niazid methanesulphonate and its metabolites in
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 6, JUNE 2002