Transport of acyclovir ester prodrugs through rabbit cornea and SIRC-rabbit corneal epithelial cell line

Article abstract of DOI:10.1002/jps.1101

The purpose of this study is to assess the permeability of acyclovir (ACV) prodrugs through the rabbit corneal cell line (SIRC) as well as the cornea, and characterize the SIRC cell line for transport and metabolism studies of ester prodrugs. Prodrug derivatization of an acycloguanosine antiviral agent, acyclovir, was employed to improve its permeability across the cornea. New Zealand albino rabbits were used as an animal model for corneal studies. The SIRC cell line grown on polyester membranes was used for transport of these prodrugs. SIRC cells grown on the membrane support for 10 days developed four to six layers of epithelial cells, and this is comparable to the normal rabbit corneal epithelial layer. Transport experiments were conducted across the rabbit cornea and confluent SIRC cells using side-by-side diffusion-cell apparatus. Enzymatic hydrolysis of these compounds was evaluated in SIRC cell lysates. Appropriate reversed phase HPLC method(s) were employed for quantitation of both the prodrug and ACV simultaneously. Corneal permeabilities of some of these prodrugs (Malonyl ACV and Acetyl ACV) were higher relative to ACV. The SIRC cell line permeability values of all the prodrugs were higher compared to that of the intact cornea. The total amount of ACV-prodrugs transported, i.e., unhydrolyzed prodrugs and regenerated ACV, across the SIRC cell line was more relative to ACV. Hydrolytic studies in the SIRC cell line homogenate demonstrated the bioreversion potential of the prodrugs and the presence of enzymes, particularly the cholinesterase in the SIRC cell line. It may be concluded that the SIRC cell line is leakier compared to the cornea. Keeping in mind the limitations, the SIRC cell line after further characterization may be used for transport and metabolism studies of ester prodrugs.

Full text of DOI:10.1002/jps.1101

Transport of Acyclovir Ester Prodrugs Through Rabbit Cornea
and SIRC-Rabbit Corneal Epithelial Cell Line
RAHUL V. TAK, DHANANJAY PAL, HONGWU GAO, SURAJIT DEY, ASHIM K. MITRA
Division of Pharmaceutical Sciences, School of Pharmacy, University of Missouri±Kansas City, 5005 Rockhill Road,
Kansas City, Missouri 64110
Received 5 May 2000; revised 6 March 2001; accepted 16 March 2001
ABSTRACT: The purpose of this study is to assess the permeability of acyclovir (ACV)
prodrugs through the rabbit corneal cell line (SIRC) as well as the cornea, and
characterize the SIRC cell line for transport and metabolism studies of ester prodrugs.
Prodrug derivatization of an acycloguanosine antiviral agent, acyclovir, was employed
to improve its permeability across the cornea. New Zealand albino rabbits were used as
an animal model for corneal studies. The SIRC cell line grown on polyester membranes
was used for transport of these prodrugs. SIRC cells grown on the membrane support
for 10 days developed four to six layers of epithelial cells, and this is comparable to
the normal rabbit corneal epithelial layer. Transport experiments were conducted
across the rabbit cornea and con¯uent SIRC cells using side-by-side diffusion-cell
apparatus. Enzymatic hydrolysis of these compounds was evaluated in SIRC cell
lysates. Appropriate reversed phase HPLC method(s) were employed for quantitation of
both the prodrug and ACV simultaneously. Corneal permeabilities of some of these
prodrugs (Malonyl ACV and Acetyl ACV) were higher relative to ACV. The SIRC cell
line permeability values of all the prodrugs were higher compared to that of the intact
cornea. The total amount of ACV-prodrugs transported, i.e., unhydrolyzed prodrugs and
regenerated ACV, across the SIRC cell line was more relative to ACV. Hydrolytic studies
in the SIRC cell line homogenate demonstrated the bioreversion potential of the
prodrugs and the presence of enzymes, particularly the cholinesterase in the SIRC cell
line. It may be concluded that the SIRC cell line is leakier compared to the cornea.
Keeping in mind the limitations, the SIRC cell line after further characterization may be
used for transport and metabolism studies of ester prodrugs. ß 2001 Wiley-Liss, Inc. and
the American Pharmaceutical Association J Pharm Sci 90:1505±1515, 2001
Keywords: rabbit cornea; SIRC; ACV; prodrugs; permeability
INTRODUCTION
simplex virus is among the most common patho-
gens to affect them. Primary herpetic keratitis is
2
Herpes simplex keratitis is the leading infectious
cause of blindness in the United States. Ocular
self-limiting, and leaves no scarring. HSV-1 can
remain latent in the trigeminal ganglion, and
shed periodically, producing clinical manifesta-
tions of secondary herpetic keratitis. The treat-
ment of ocular HSV-1 infection is a major concern
of ophthalmologists.
1
Herpes Simplex Virus (HSV-1) infections are
increasing in individuals whose immune system
is compromised. Approximately, 12% of human
immunode®ciency virus (HIV) infected patients
present with opportunistic infection and herpes
ACVÐa synthetic purine nucleoside analog of
guanine, is clinically used in the treatment of
herpes simplex virus infections because of its
Correspondence to: A. K. Mitra (Telephone: 816-235-1615;
Fax: 816-235-5190; E-mail: mitraa@umkc.edu)
3
af®nity for the viral thymidine kinase. However,
the treatment of these infections has become
problematic because of the inability of the active
Journal of Pharmaceutical Sciences, Vol. 90, 1505±1515 (2001)
ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001
1505

1
506
TAK ET AL.
drug to achieve desirable concentrations in ocular
tissues requiring frequent administrations of a
topical dose. ACV has low ocular bioavailability
because of its hydrophilic nature and rapid
precorneal drainage, upon topical administration,
relative to its transcorneal ¯ux. A lipophilic
prodrug approach can be utilized to overcome
the low penetration of ACV, thereby achieving
optimum therapeutic concentrations in the cor-
nea as well as the internal ocular tissues.
Triangle Park, NC). Heptanesulfonic acid,
acid anhydrides, and 4-dimethylaminopyridine
(DMAP) were obtained from Aldrich Chemical
Company (Milwaukee, WI), 1-octanol and dime-
thyl formamide (DMF) were provided by Fisher
Scienti®c Company (Fairlawn, NJ). Other re-
agents were of analytical grade, and were used
as received. Distilled, deionized water was used
in the preparation of all buffers and mobile
phases. The SIRC cell line was obtained from
American Type Culture Collection (ATCC).
Ketamine hydrochloride was obtained from the
animal lab, University of Missouri±Kansas City.
Sodium pentobarbital was obtained from the
school of Pharmacy, UMKC stock, and used under
supervision.
4
Hughes and Mitra showed that a series of
0
aliphatic 2 -esters of ACV signi®cantly increased
the corneal permeability of the ACV.
In this decade signi®cant efforts have been
dedicated to establish cell culture as an ef®cient
screening tool to assess the transport and meta-
bolism processes of compounds. A number of cell
lines have been characterized, and are used
routinely in pharmaceutical research and devel-
Methods
5
opment. Development of a cell line model to
Synthesis
assess the corneal transport and metabolism
characteristics of compounds would be desirable.
Such a model will also be economical, as it will
minimize the use of animal tissues. The SIRC-
rabbit corneal epithelial cell line has been used for
in vitro studies of corneal physiology, immunol-
The simple alkyl esters of [9-(2-hydroxyethoxy-
methyl) guanine] (ACV) were prepared in a single
step reaction, using a modi®ed procedure of Shao
et al. A solution of ACV in dimethyl formamide
was treated with acid anhydride in the presence of
catalytic amounts of DMAP (Figure 1). Diacyl
derivatives of ACV were observed to be the major
product using this literature protocol. Reducing
the acid anhydride amounts (from 10 to 3
equivalents) and reaction time (from 2±3 days to
1
3
6
±10
ogy, and toxicology.
elongated epithelial-like cell layers. Hutak et al.
SIRC cells form multiple
1
1
established a growth pattern of SIRC in microwell
inserts. SIRC cells grown for 10 days on the
surface of microwell inserts showed four to six
layers of cells. The number of cell layers remained
stable of average four-cell layers up to 14 days.
This number gradually decreased during 21 days
of culture. There was a decrease in the nuclear-to-
cytoplasmic ratioÐa sign of aging and an exfolia-
tion of surface cells was also evident. Very
1.5 days) produced more satisfactory results. ACV
pivalate is an exception because of the steric
hindrance effect of the t-butyl group. The diacetyl
derivatives were separated as by-products in
about 4 to 12% yields.
Valcyclovir, 2-[(2-Amino-1, 6-dihydro-6-oxo-
9H-purin-9-yl) methoxy]ethyl L-valinate hydro-
1
2
recently Goskonda et al. studied the permeabil-
ity characteristics of novel mydriatic agents in
SIRC rabbit corneal cells. However, the SIRC cell
line has not been characterized yet to assess
permeability of a wide range of compounds. The
objectives of this work are to assess the perme-
ability of ACV prodrugs through the rabbit cornea
as well as the SIRC corneal epithelial cell line,
and to evaluate the suitability of the SIRC cell line
for transport and metabolism of ester prodrugs.
EXPERIMENTAL SECTION
Materials
ACV, [9-(2-hydroxyethoxymethyl)guanine] was a
gift from Burroughs Wellcome Co. (Research
Figure 1. Scheme for the synthesis of ester prodrugs
of ACV.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

ACYCLOVIR ESTER PRODRUGS IN RABBIT CORNEA
1507
Phosphate-Buffer Saline (DPBS) pH 7.4 for tran-
sport studies with SIRC cell line.
Analytical Procedure
A high-pressure liquid chromatographic (HPLC)
method was developed for the analyses of ACV
and its monoesters (Table 1). The system com-
Figure 2. Scheme for synthesis of L-valyl ester of
ACV.
prised of a Waters 515 HPLC pump, a Rheodyne
1
4
chloride, was synthesized by Beauchamp et al.
method (Figure 2). Step I: Acyclovir (2.0 g,
.88 mM) was dissolved in dry DMF (150 mL) by
warming on a steam bath. Successively, DMAP
0.154 g, 1.25 mM), N-cbz-L-valine (3.01 g,
2.0 mM), and DCC (3.0 g, 14.4 mM) were added
TM
injector, and a Waters 486 tunable absorbance
detector. For the esters of acyclovir, a dual-
column isocratic HPLC method was used. A
8
2
50 Â 4.6 mm i.d. reversed phase C , 5 mm
1
8
(
1
particles, connected in series with a cation ex-
change column, Partisil 10 SCX, 5 mm, 250 Â 4.6
mm i.d. was employed. A mobile phase consisting
of 0.005 M ammonium phosphate buffer pH 2.2
to the cooled solution. The solution was stirred
under nitrogen atmosphere at ambient tempera-
ture for 18 h. The mixture was recharged with
same amounts of DMAP, N-cbz-L-valine, and
DCC, and stirring was continued at ambient
temperature for 2 days. The mixture was ®ltered
and the DMF was removed by treating the
mixture with water and ethyl acetate under acidic
conditions. The intermediate partitions into the
ethyl acetate layer and the DMF is miscible in the
water layer. The ethyl acetate fraction was
separated and rotavaped. The residue was chro-
matographed on silica gel, using 1:4 methanol
MeOH), CH Cl as the eluant, to generate 3.5 g of
2 2
the desired intermediate, 2-[(2-amino-1, 6-dihy-
dro-6-oxo-9H-purin-9-yl) methoxy] ethyl N-Cbz-
L-valinate, as a white solid. The H spectra was
consistent with the desired structure. Step II:
solution of the intermediate in MeOH (130 mL),
(
60%) and organic phase (40%) comprising of
methanol was employed. In case of ACV±acetate
and ACV±malonate, a 250 Â 4.6 mm i.d. reversed-
1
TM
Ê
phase C18 Microsorb-MV 100 A (Varian) con-
nected to Beckman System Gold 128 solvent
module with ¯uorescence detection (Schoeffel
Instr.-Corp). A mobile phase comprising of ammo-
nium phosphate 0.0025 M pH 2.2 (90%) and an
organic phase (10%) comprising of MeOH and
acetonitrile (1:1) was used. For guanosine and
guanosine monophosphate, a similar reverse-
phase C18 column connected to a 515 Waters
pump with a UV detection was used. A mobile
phase consisting of ammonium phosphate 0.002-
M pH 5.6, and a varying proportion of methanol
was used. This method was capable of detecting
guanosine, guanosine monophosphate, guanosine
diphosphate, and guanosine triphosphate in a
single run. Data acquisition and chromatographic
analysis was carried out using a Hewlett Packard
integrator (HP 3396 series III). For all analyses, a
(
1
THF (tetrahydrofuran, 65 mL), and H O (25 mL)
2
was added to 0.5 N aqueous HCl (18 mL) and 377
mg of 5% palladium on charcoal. The mixture was
shaken in a Parr apparatus under an initial
pressure of 50-psi hydrogen at ambient tempera-
ture for 18 h. The mixture was ®ltered, the
catalyst was washed with MeOH, and the com-
bined washings and ®ltrate were evaporated in
¯ow rate of 1 mL/min was maintained.
Transport Through Cornea
1
vacuo at a bath (< 608C) temperature. The H-
NMR spectra was satisfactory for the desired
A typical side-by-side diffusion apparatus (type-
VSC-1, Crown Glass Company Inc.) was used
for performing transport experiments. Cornea
obtained from New Zealand albino rabbits weigh-
ing 2.0 to 3.0 kg was used for transport study.
The animals were euthanized by an overdose of
pentobarbital through a marginal ear vein. Eyes
were proptosed and carefully enucleated, and
washed with ice-cold isotonic phosphate buffer
(IPB) pH 7.4 to remove any traces of blood. After a
small incision to the sclera, vitreous humor was
aspirated by using a 1-mL. syringe. Cornea was
carefully excised, leaving some scleral portion
structure.
Preparation of Drug Solutions
A fresh, 1±2 mM aqueous solution of the ester
prodrugs (butyryl-ACV, iso-butyryl-ACV, valyl-
ACV, 1 mM; malonyl-ACV, acetyl-ACV, 2 mM;
pivaloyl-ACV, 0.5 mM) was prepared in isotonic
phosphate buffer pH 7.4 for corneal transport
studies. A 1-mM aqueous solution of the ester
prodrugs was prepared in Dulbecco's Modi®ed
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

1
508
TAK ET AL.
Table 1. HPLC Conditions for Ester Prodrugs of ACV, Guanosine, and GMP
Compounds
Mobile Phase
Detection
UV lmax 249 nm
Butyryl-ACV, isobutyryl-ACV,
pivaloyl-ACV
Ammonium phosphate (0.005 M,
pH 2.5): methanol 60:40
Acetyl-ACV, malonyl-ACV
Ammonium phosphate (0.0025 M,
pH 2.2): methanol: acetonitrile
Fluorescence lex 249 nm lem 370 nm
9
0:5:5
L Valyl-ACV
Guanosine
GMP
Ammonium phosphate (0.0025 M,
pH 2.2)
Ammonium phosphate (0.002 M,
pH 5.6): methanol 95:5
Ammonium phosphate (0.002 M,
pH 5.6)
Fluorescence lex 249 nm lem 370 nm
UV lmax 254 nm
UV lmax 254 nm
attached to the cornea. The scleral part attached
to the cornea helps to hold the cornea in place
in between the half-cells during the transport
experiment. The lens was removed, and the iris-
ciliary body was separated from the cornea. The
cornea was washed immediately with ice-cold IPB
pH 7.4 and mounted on a side-by-side diffusion
half-chamber. The temperature was maintained
at 348C by circulating water through the jacketed
chambers of the diffusion apparatus. Aqueous
drug solutions (3 mL) were added on the epithelial
side of the cornea (donor chamber). In the other
half-chamber (receiver chamber), 3.4 mL of IPB
pH 7.4 was added. Both the chambers were
stirred continuously by using magnetic stirrer
bars. Receiver chamber volume of IPB added was
little more than that of the donor chamber to
maintain the curvature of the cornea throughout
the experiment. Sink conditions were maintained
during the experiment. One hundred microliter
aliquots were removed from the receiver chamber
at appropriate intervals, and were replaced with
an equal volume of IPB pH 7.4. The samples were
either immediately analyzed by HPLC or stored
at � 808C until further analysis. All experiments
were done in triplicate.
(100 mg/mL). The cells were incubated at 378C in
humidi®ed atmosphere of 5% CO /95% air. The
2
2
cells were plated at a density of 33,000 cells/cm
either on tissue culture-treated plastic dishes or
collagen±®bronectin-coated polyester clear mem-
brane (pore size 0.4 mm). The cells used for
the transport study were grown for 10 days, and
the medium was changed every alternate day.
Before the transport study the cells were rinsed
with Dulbecco's modi®ed phosphate-buffer saline
(DPBS) containing 1 mM CaCl , 0.74 mM MgSO ,
2
4
5.3-mM glucose, and pH 7.4. The cells grown on
the membrane support were mounted on side-by-
side diffusion chambers, and the drug solution
was added on the apical side of the membrane
(donor chamber). DPBS was added to the other
half-chamber (receiver chamber). One hundred
microliters of aliquots were removed from the
receiver chamber at predetermined time points
and analyzed by HPLC. The receiver was replaced
with equal volume of buffer after removing the
1
4
sample. Transport of the hydrophilic marker
C
mannitol was carried out in triplicate for each
compound used for the transport study to assess
the integrity of cells during the length of the
experiments. A similar sampling procedure was
14
used for C mannitol transport studies. The
samples were analyzed by a scintillation counter
Transport Through SIRC Rabbit Corneal
Epithelial Cell Line
(
Beckman, LS 6500). The temperature was main-
tained at 348C through circulated water from a
water bath, and the solutions were stirred by
using magnetic stirrer bars throughout the length
of the experiment. Data was obtained by perform-
ing experiments in triplicate or quadruplicate.
The SIRC cells grown on a polyester membrane
support (Whatman) were used for transport stud-
ies. The rabbit corneal epithelial cell line (SIRC)
was received from the American Type Culture
1
Collection (ATCC). The cells were cultured
according to the guidelines by ATCC with Mini-
mum Essential Medium (pH 7.4) supplemented
with 10% calf serum, lactalbumin (1.76 mg/mL),
HEPES (1.3 mg/mL), and Penicillin-streptomycin
Transmission Electron Microscopy
Transmission electron microscopy (TEM) was
used to determine the number of layers in the
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

ACYCLOVIR ESTER PRODRUGS IN RABBIT CORNEA
1509
SIRC cell line. The cells grown for 10 days were
washed two times with DPBS at 48C. The cells
were then washed with cacodylate buffer (48C) for
were washed three times with DPBS (pH 7.4). Ten
milliliters of buffer was added to a ¯ask and the
cells were scraped with the help of a rubber
policeman (cell scraper). The cell suspension was
centrifuged at 1500 rpm for 10 min. The pellet was
resuspended in 2 mL ice cold DPBS, and homo-
genized by a tissue homogenizer (Tissue Tearer
Model 985-370 Type 2, Biospace Products, Inc.)
for 1 min. The homogenate was centrifuged
(Beckman TL-100 Ultracentrifuge) at 10,000-
rpm, at 48C for 30 min. The supernate obtained
after centrifugation was the cytosol fraction and
used for hydrolysis studies. Protein content of the
5
min. Then the cells were ®xed with 2% glutar-
aldehyde in 0.15 M Na-cacodylate buffer for
0 min. The cells were washed several times with
6
cacodylate buffer and post®xed with osmium
tetraoxide for 30 min. Then the cells were pro-
cessed for thin sectioning using a routine TEM
procedure and photographed using a JEOL
Electron Microscope (Model JEM 1200 EX II) at
1
00 kV.
1
Preparation of Ocular Tissues for Enzymatic
Hydrolysis Studies
cytosol was estimated by BioRad Protein Assay
method. Hydrolysis experiments were carried
out as mentioned before. All experiments were
carried out in triplicate.
New Zealand albino rabbits were used as an
animal model for metabolism studies. Animals
were euthanized, and eyes were enucleated as
described earlier. The eyes were washed in ice-
cold IPB pH 7.4 to remove any traces of blood.
Aqueous humor was removed by inserting a 27-1/
Cholinesterase Assay in SIRC Cell Homogenate
Calibration was done according to the method
1
given in Sigma Diagnostics procedure No. 420
2
gauge needle through the sclera into the
(
catalog No. 420-MC) cholinesterase kit. Choli-
anterior chamber. Vitreous humor was collected
by incision in the sclera. Part of the sclera was
removed to expose the lens. The lens was separ-
ated followed by the iris-ciliary body. The cornea
was cut carefully to leave no traces of scleral
tissue attached to it. All the tissues were imme-
diately washed with ice-cold IPB pH 7.4 and
stored in ice. The procedure takes approximately
nesterase catalyzes the hydrolysis of cholinesters
of various short-chain organic acids, including
acetylcholine, which is the substrate used in this
Sigma Diagnostic procedure. The reaction was
conducted in the presence of an acid-base indi-
cator, such as m-nitrophenol. The acetic acid prod-
uced lowers the pH, causing a loss of color. This
color change is proportional to the cholinesterase
activity present. A serum ``blank'' was prepared
by inactivation at 608C for 10 min to compensate
for background absorbance contributed by the
sample. Absorbance of both ``blank'' and ``test'' was
read at 420 nm, and the ``difference'' was used to
estimate the cholinesterase level. Calf serum was
used instead of pooled serum. Sodium chloride
solution (0.2 mL) and SIRC cell homogenate were
taken in two separate tubes and then mixed. One
tube (blank) was placed in a 608C water bath for
1
0±15 min. The tissues were homogenized imme-
diately using a homogenizer (Tissue Tearer Model
85-370 Type 2, Biospace Products, Inc.) in 2 mL
9
ice-cold IPB pH 7.4 and the supernate was stored
at � 808C for further studies. The homogenate was
centrifuged at 50,000 rpm for 30 min using an
ultracentrifuge (Beckman TL-100). The super-
nate obtained was used for hydrolysis studies.
Reactions were initiated by adding 100±200 mL
(
7
10±50 mg/mL) of drug solution made in IPB pH
.4 to 1.0 to 1.2 mL of supernate. The reactions
10 min to inactivate the enzymes. The tube was
were carried out in a water bath at 348C. Samples
were taken at appropriate intervals and analyzed
by HPLC. Equal volume of methanol was added
to stop the reaction. Experiments were done
in triplicate. Protein concentration was estimated
heated thoroughly to ensure that all enzymes
have been inactivated. It was removed and cooled
in tap water. To both ``blank'' and ``test'' tubes,
3
0
.0 mL water, 2.0 mL nitrophenol solution, and
.2 mL acetylcholine chloride solution was added.
1
by using BioRad assay method for protein
All these reagents were brought to room tempera-
ture before making the additions. Exact time of
addition of acetylcholine chloride solution was
recorded, and both the ``blank'' and ``test'' tubes
were placed in water bath at 258C. Exactly 30 min
later, absorbance (A) was recorded by using a
Beckman UV at 420 nm and water as a reference.
estimation.
Preparation of SIRC Cell Homogenate
for Enzymatic Hydrolysis Studies
2
The SIRC cells were grown in a 75-cm ¯ask for 10
days, the medium was removed, and the cells
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

1
510
TAK ET AL.
Calculations (DA ˆ DBLANK � DTEST
)
Using DA, the cholinesterase activity of the SIRC
homogenate was determined in Rappaport Units
from the calibration curve. One Rappaport unit is
the amount of cholinesterase that liberates 1 mm
of acetic acid from acetylcholine in 30 min at 258C
at pH 7.4 under test conditions.
Data Analysis
Flux values for ACV and ester prodrugs were
calculated by dividing the slope (obtained by
plotting a cumulative amount of the compound
permeated versus time) with the area available
2
for diffusion, i.e., 0.636 cm . The permeability
values were calculated by dividing the ¯ux values
with the concentration of the prodrug used in the
donor. Statistical tests were performed to com-
pare the differences between the corneal and
the SIRC permeability values using one-way
ANOVA. A p-value less than 0.05 was considered
as signi®cant.
Apparent ®rst-order rate constants for enzy-
matic hydrolysis was calculated by linear regres-
sion of the line obtained by plotting log percent
remaining versus time pro®les. Triplicate sam-
ples were analyzed, and a mean rate constant was
calculated. The rate constants were expressed as
per milligram of protein.
Figure 3. Transmission electron microscopic photo-
micrograph of SIRC cells exhibiting six layers after 10
days of culture.
lectrical epithelial resistance (TEER value < 100
2
O cm ) indicative of the leaky nature of the cells.
During the 10 days of our culture, no exfoliation of
SIRC cells was observed. These studies also
con®rmed that the SIRC cell line grown for 10
days differentiated to a multicell layer epithelial
membrane.
RESULTS AND DISCUSSION
Electron Microscopy
The TEM photomicrograph depicts six layers of
SIRC cells (Figure 3). This observation is in
11
agreement with Hutak et al. that the SIRC cells
grown on the membrane support for 10 days
developed four to six layer of cells and comparable
to the normal corneal epithelial layer. The TEM
photomicrograph shows an elongated epithelial
structure and has a ®broblast like appearance.
There are no microvilli and microplicae on the
surface. There is a sparse distribution of the inter-
cellular space and the cells appear to be fairly
tight. However, the cellular tight-junction compo-
nents such as desmosomes are lacking. Our
TEM results of SIRC cells are similar with the
previous reports that SIRC cells lack desmo-
somes, cytoplasmic ®laments, and cytokeratinÐ
the structures that are characteristic of corneal
Enzymatic Hydrolysis in Ocular Tissues and SIRC
Cell Line Homogenate
The apparent ®rst-order rate constants for the
hydrolysis of the ACV esters in SIRC cells have
been reported in Table 5. Loss of ACV ester was
accompanied by the formation of ACV. A mass
balance with respect to total ACV concentration
was maintained throughout the experiment.
Controls with no homogenate and containing only
the DPBS (pH 7.4) and the respective esters
showed no quanti®able degradation over the time
course of the study. Cytosolic fraction of the SIRC
cell line was used for the hydrolysis studies. The
amount of cholinesterase in the SIRC cell line
homogenate was found to be 38 Rappaport Units/
mL. The esterase activity in the epithelium of the
6
epithelial cells. The SIRC cells grown on six-well
1
Transwell ®lters did not develop enough transe-
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

ACYCLOVIR ESTER PRODRUGS IN RABBIT CORNEA
1511
rabbit cornea is higher than that of stroma±
endothelium, and this esterase activity varies
1
5
with rabbits's age and strain.
The hydrolysis of ACV prodrugs and formation
of ACV demonstrated the presence of esterase
activity. The hydrolytic rate constant for butyryl±
ACV in SIRC homogenate studies was found to be
the highest ( p < 0.05). The concentration versus
time pro®le for isobutyryl±ACV is shown in
16
Figure 5. Percent log versus time pro®le for acetyl±
ACV in the SIRC cell lysate.
Figure 4. Lee et al. demonstrated that there
exists a chain length at which a prodrug would be
hydrolyzed equally well by both cholinesterases.
This could be attributed to the higher hydrolytic
rate for the butyryl±ACV. A lipophilic pocket at
the active site of the ocular esterase has been
1
4
the transport of a paracellular marker
C
mannitol for each set of experiments. All experi-
1
4
ments were conducted in triplicate, with
C
1
7
suggested. The isobutyryl and pivaloyl esters
showed low enzymatic hydrolytic rate constants
in the SIRC cell homogenate, most likely due to a
steric hindrance at the enzyme active site. The
hydrolysis studies of acetyl ester in SIRC cell line
homogenate also showed regeneration of ACV
mannitol throughout the length of the experi-
ments for each prodrug used. Mannitol transport
was found to be between 2±4% per hour. Perme-
ability values for mannitol for each prodrug are
summarized in Table 3. To monitor transcellular
14
permeability, transport of
C diazepum was
(
Figure 5). The SIRC cell line enzymatic data
conducted across SIRC and the freshly excised
1
4
con®rms the presence of esterase activity and
thereby bioreversibility of the ACV prodrugs. The
apparent ®rst-order rate constants for acetyl±
ACV and malonyl±ACV in ocular tissue homo-
genates is illustrated in Table 6. The rate of
hydrolysis of acetyl±ACV in ocular tissues was
highest in the iris±ciliary body, followed by in the
cornea and in the aqueous humor ( p < 0.05). For
malonyl ester, the highest rate constant was also
found in the iris±ciliary body ( p < 0.05). The
amount of esterases present in the iris±ciliary
body was signi®cantly higher (almost twice) than
in the cornea.
cornea (Figure 7). The permeability of
C
diazepum across SIRC was more than double
�
�
5
(8.37 Â 10
cm/s) compared to the cornea
5
(3.43 Â 10 cm/s).
All the ACV prodrugs, except ACV±acetate and
ACV±malonate, were partially hydrolyzed to
ACV during transport across the SIRC cell layers.
Enzymatic activity of the SIRC cell line is evident
from the regeneration of the parent compound
ACV during transport (Figure 6). Permeability
values for the prodrugs through SIRC were found
to be higher than that of across the intact cornea
( p < 0.05). From the electron microscopic data it
was observed that SIRC cells grown for 10 days
developed four to six cell layers thick. Normal
rabbit corneal tissue consists of four to seven
layers of epithelial cells, followed by stroma and
endothelium. The top two super®cial layers offer
the primary resistance to transport of drugs.
Although SIRC cells grown for 10 days were four
to six layers thick, but they did not produce
Transport of Acyclovir Esters Across SIRC
Cell Line and Rabbit Cornea
Corneal and SIRC cell line permeability values for
the prodrugs and ACV are reported in Table 2.
Integrity of the SIRC cell line was determined by
enough resistance, having a poor TEER value
2
(
< 100 O cm ). The TEER value for rabbit cornea
2
is > 1500 O cm . Thus, SIRC cell layers are leakier
than the cornea, and hence, the higher perme-
ability values result for the ACV prodrugs.
The amounts of ACV generated during trans-
port across the SIRC cell line are reported in
1
8
Table 4. Hughes and Mitra
reported 50%
hydrolysis of isobutyryl±ACV and about 80%
hydrolysis of other esters in the homologous series
for transport experiment in the cornea. However,
Figure 4. Concentration versus time pro®le for iso-
butyryl±ACV in the SIRC cell lysate.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

1
512
TAK ET AL.
Table 2. Permeability Values of ACV Prodrugs, Guanosine, and GMP in Cornea and SIRC Cells
Prodrugs
Cornea P cm/s
SIRC P cm/s
Ratio: SIRC/Cornea
�
6 a
� 5
� 5
� 5
� 5
ACV
3.65 Â 10
2.6 Æ 0.34 Â 10
7.12
1.72
1.37
2.44
3.47
�
5
5
� 5
Malonyl-ACV
Acetyl-ACV
Butyryl-ACV
Isobutyryl-ACV
Pivaloyl-ACV
L-Valyl-ACV
Guanosine
GMP
2.33 Æ 0.42 Â 10
4.0 Æ 1.4 Â 10
�
1.83 Æ 1.2 Â 10
2.5 Æ 0.21 Â 10
1.25 Æ 0.11 Â 10
1.36 Æ 0.12 Â 10
1.67 Æ 0.10 Â 10
1.20 Æ 0.33 Â 10
9.1 Æ 0.24 Â 10
5.06 Æ 0.13 Â 10
�
�
6 a
5.12 Â 10
3.92 Â 10
6 a
�
5
n/d
�
�
6
6
6
� 5
� 6
� 6
5.8 Æ 0.95 Â 10
2.07
1.14
2.14
7.95 Æ 0.21 Â 10
2.36 Æ 0.14 Â 10�
a
Data obtained from ref. 18.
b
n/d ˆ not detected.
in SIRC transport studies, isobutyryl±ACV ester
hydrolyzed to about 5%. Other ACV esters except
acetyl±ACV and malonyl±ACV were hydrolyzed
to about 10±13% during transport. In the SIRC
homogenate study the apparent ®rst-order rate
constant for the hydrolysis of isobutyryl±ACV
was also low compared to other ACV esters in the
study. Moreover, hydrolysis of acetyl±ACV and
malonyl±ACV was also observed in the SIRC cell
line homogenate study. The epithelial layer
contributes approximately 70% of the enzymatic
activity of the cornea. The SIRC cell line consists
of only epithelial layers, and these cells might not
be expressing the esterases at the same level,
which are responsible for hydrolysis of ester
prodrugs. Hence, a lower percentage of hydrolysis
for the ACV±esters in the SIRC cell line was
observed during transport.
prodrugs except malonyl±ACV and acetyl±ACV
were lower compared to that of ACV across the
SIRC cell line ( p < 0.05). These prodrugs were
hydrolyzed during transport, and the total
amount of ACV transported was more than that
of ACV, resulting in increased amounts of ACV
transported across the SIRC cell line. It is well
known that the optimum log partition coef®cient
values for the compounds to diffuse across cornea
are between 1 and 2. However, higher log
partition coef®cient values would compromise
aqueous solubility of prodrugs. Thus, water-
soluble, solution stable, L-valyl ester of ACV was
synthesized. The permeability value for valacy-
clovir was comparable to other simple ester
prodrugs.
Higher corneal permeability values for Mal-
onyl±ACV and Acetyl±ACV were observed com-
pared to corneal permeability of ACV ( p < 0.05).
ACV was regenerated from acetyl prodrug in a
consistent manner during transport through the
cornea. Hydrolysis of the malonyl prodrug was
observed during transport across the cornea to a
lesser extent and in an inconsistent manner
compared to other prodrugs. The permeability
value for valacyclovir across the cornea was also
more than that of ACV ( p < 0.05). Its mechanism
Higher permeability for acetyl±ACV might be
attributed to its inability to undergo hydrolysis
during transport across the SIRC cell line.
However, the relationship observed by Hughes
1
2
and Mitra
for log partition coef®cient and
permeability may not be predicted for these
compounds. Similarly, a higher permeability
value was observed for malonyl±ACV ( p < 0.05).
The permeability values for the unhydrolyzed
1
4
Table 3.
C Mannitol Permeability Values Across
SIRC Cells for Each Compound
Permeability
�
6
Compounds
(cm/s) Â 10
Butyryl, iso-butyryl,
pivaloyl ACV
Valyl ACV
Guanosine
GMP
1.12 Æ 0.10
2.02 Æ 0.50
1.41 Æ 0.34
Figure 6. Receptor cell concentrations of ACV, total
ACV, and the unhydrolyzed butyryl±ACF. (a) Each
point is a mean of n ˆ 3.
1.20 Æ 0.047
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

ACYCLOVIR ESTER PRODRUGS IN RABBIT CORNEA
1513
Table 5. Apparent First-Order Rate Constants for
ACV Prodrugs in SIRC Cell Lysate
�
1
K min /mg of
4
Prodrugs
Protein  10�
Malonyl ACV
Acetyl ACV
Butyryl ACV
Iso-butyryl ACV
Pivaloyl ACV
2.41 Æ 0.5
2.9 Æ 0.13
14.4 Æ 0.83
1.6 Æ 0.13
3.12 Æ 0.17
1
4
Figure 7. Transport of C diazepam (a transcellular
marker) across the SIRC cell line and freshly excised
rabbit cornea. (a) Each point is a mean of n ˆ 4.
GMP through the cornea. No phosphate metabo-
lites were found during transport through the
SIRC cell line. The permeability value for guano-
sine across the SIRC was higher than that of
across the cornea ( p < 0.05). A possible reason
might be the lower resistance offered by SIRC
epithelial cell layers to diffusion of guanosine.
Moreover, absence of hydrolysis might have
resulted in higher transport. Guanosine is poorly
water soluble (less than 1 mg/mL) at pH 7.4,
although it has relatively lower log partition
coef®cient. ACV shares some of the physicochem-
ical characteristics of guanosine; hence, it was
selected as a model for antiviral nucleosides.
Similar results were obtained for GMP. GMP is
fairly hydrophilic compared to guanosine, and
diffuses through the cornea as well as the SIRC
cell line very poorly. The permeability values for
GMP were found to be lower compared to that of
guanosine ( p < 0.05). This also suggests that the
epithelial layer offers resistance to transport of
hydrophilic compounds. It can also be concluded
that the topmost two layers of cornea present a
considerable barrier to permeation of hydrophilic
compounds. The SIRC permeability values for
guanosine and GMP were 10 and 25% higher than
that of guanosine and GMP through the cornea,
respectively. Although SIRC expresses cholines-
terase, it may not be expressing the wide variety
of kinases present in the cornea as no metabolism
(phosphorylation) of guanosine and GMP was
observed during transport across the SIRC.
Characterization of the SIRC cell line for other
enzyme systems needs to be performed. The
permeability values obtained for these compounds
across the SIRC cell line were suf®ciently close to
predict their permeabilities across the cornea.
However, caution must be exercised while inter-
preting transport data for compounds, consider-
ing the limitations of the SIRC cell line.
of transport across the cornea is not clear, and
needs to be studied further in detail. In a recent
19
study Goskonda et al. have shown good correla-
tion between permeability coef®cients and lipo-
philicities of drugs. The ester prodrugs, such as
phenylacetyl ester, isovaleryl ester, and pivalyl
ester, having higher lipophilicities, showed a four-
to sixfold increase in corneal permeability com-
pared to phenylephrine HCl. The permeability of
ester prodrugs varied, depending on the pH of the
transport medium, because the degree of ioniza-
tion is dependent on the pH of the medium.
Permeability values for transport of guanosine
and guanosine monophosphate (GMP) are
reported in Table 2. Metabolism of guanosine
was found to occur during diffusion through the
cornea. These metabolites remain to be identi®ed.
However, no metabolism was observed during
transport through the SIRC cell line. GMP was
used as a model for antiviral nucleoside monopho-
sphate to investigate whether it gets converted
to higher phosphates, namely guanosine dipho-
sphate, and guanosine triphosphate, by the
cellular kinases during transport through the
cornea. These studies show that there was no
such conversion of GMP to higher phosphates.
However, three metabolites, which are yet to be
identi®ed, were found during the transport of
Table 4. Amount of ACV Generated During
Transport Through SIRC Cells
Prodrugs
Amount of ACV (mg/mL)
Butyryl ACV
Iso-butyryl ACV
Pivaloyl ACV
Valyl ACV
0.823 Æ 0.020
0.426 Æ 0.049
0.300 Æ 0.064
0.410 Æ 0.021
From the results of in vitro transport and
metabolism studies in the cornea and SIRC cells,
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

1514
TAK ET AL.
Table 6. Apparent First-Order Rate Constants for ACV Prodrugs in Ocular Tissue Homogenates
ICB
AQ
�
1
Prodrugs
Cornea
K min per mg of Protein
4.61 Æ 0.16 Â 10�
4.61 Æ 0.14 Â 10�
5
5
9.5 Æ 0.20 Â 10
1.26 Æ 0.11 Â 10
� 5
� 4
0.150 Æ 0.08 Â 10
� 5
� 4
Acetyl-ACV
Malonyl-ACV
1.00 Æ 0.02 Â 10
some encouraging observations were obtained.
However, a better correlation between the corneal
and the SIRC permeability values could be ob-
tained by further characterization of the SIRC cell
line in detail. Optimizing the growth conditions
using additional growth-promoting agents [pitui-
tary extracts, insulin-transferin, epidermal
growth factor (EGF)] in the medium, air interface
the desirable characteristics of a cell line is to
express the enzymes present in the original
animal tissues. Ester prodrugs must also display
the appropriate hydrolysis kinetics by the
esterases present in the cell line model similar
to that of the corneal tissues obtained from
animals. The prodrug must be metabolized to
the parent compound in the ocular tissues
including the cornea. Based on the experimental
data obtained in this study, it is concluded that
the SIRC cell line expresses the enzyme cholines-
terases, which are responsible for hydrolysis of
ester prodrugs. The bioreversion of ACV prodrugs
and the esterase determination con®rms the
presence of esterases in the SIRC cell line.
From the SIRC transport and metabolism data
butyryl±ACV may be chosen as the best prodrug.
Based on the corneal data for acetyl±ACV, it can
be considered as a prodrug with favorable trans-
port and metabolism characteristics in addition
to butyryl±ACV. The water-soluble amino acid
prodrug valacyclovir can also be considered
as a potential prodrug for topical delivery of
ACV, although it needs further evaluation.
In conclusion, the SIRC cell, upon further
characterization, may be used for assessing the
permeabilities and metabolism of ester pro-
drugs, thereby eliminating the need of animal
tissues. Further characterization of the SIRC
cell line using different growth supplements
(pituitary extracts, EGF, insulin-transferin, etc.)
and air interface culture is in progress in
this laboratory. However, considering the limita-
tions of in vitro cell culture model, caution must
be exercised to avoid over interpretation of
results.
2
0,21
culture,
etc., would obtain a cell line with
consistent characteristics and thus minimize the
variability in the data. This would also yield four
to seven layers with tight junctions with suf®cient
TEER value to offer resistance to transport of
compounds similar to the cornea. The ACV
prodrugs were hydrolyzed to ACV in the cornea
and the SIRC cell line. However, hydrolysis of the
malonyl and acetyl prodrug was not observed
during transport through the SIRC cell line
although these prodrugs were hydrolyzed to
ACV in the SIRC cell lysates. Further study of
the enzyme systems present in the SIRC cell line
would enhance the understanding of hydrolysis of
these prodrugs. Moreover, qualitative and quan-
titative characterization of the SIRC cell line
for the enzyme systems would facilitate the eva-
luation of a wide variety of prodrugs. Transport of
valacyclovir was increased signi®cantly relative
to ACV in the cornea and the SIRC cell line. Its
mechanism of transport could be studied in detail
by investigating the carrier systems present in
the cornea as well as the SIRC cell line, which can
be achieved by evaluating the transport of
valacyclovir and amino acid prodrugs.
CONCLUSIONS
An ideal cell line will have the similar anatomical
and physiological characteristics as that of the
tissue of interest. Such a cell line will be an
ef®cient tool in screening the compounds for
assessment of transport characteristics of com-
pounds. Based on our studies, it can be concluded
that SIRC cells form multilayers, but are leaky in
nature compared to the cornea. Moreover, one of
ACKNOWLEDGMENTS
We would like to thank Dr. Paul Brown for his
help in Mass Spectrometry, and Greg Williams
for electron microcopy. This work was supported
by NIH Grants RO1 EY 09171-05 and RO1 EY
10659-04.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001

ACYCLOVIR ESTER PRODRUGS IN RABBIT CORNEA
1515
REFERENCES
11. Hutak CM, Barletta MA. 1997. Growth pattern of
SIRC rabbit corneal cells in microwell inserts. J
Toxicol Cut Ocular Toxicol 16:145±156.
1
2
3
4
. Green MT, Dunkel E. 1985. Herpes simplex virus
infections. In: Darell RW, editor. Viral diseases of
the eye. Philadelphia: PA: Lea and Febiger. pp 30±
12. Goskonda VR, Khan MA, Hutak CM. 1999. Perme-
ability characteristics of novel mydriatic agents
using an in vitro cell culture model that utilizes
SIRC rabbit corneal cells. J Pharm Sci 88:180±184.
13. Shao Z, Park GB, Krishnamoorthy R, Mitra AK.
1994. The physicochemical properties, plasma
enzymatic hydrolysis, and nasal absorption of
3
5.
. Cotran RS, Kumar V, Robbins S. 1994. Diseases of
immunity. In: Cotran K, Kumar V, Robbins S,
editors. Robins pathologic basis of disease, 5th ed.
Saint Louis: W.B. Saunders Co. pp 229.
. Douglas RG Jr. 1986. Anti-microbial agents. In:
Gilman AG, Rall TW, Nies AS, Taylor P, editors.
The pharmacological basis of therapeutics, 8th ed.
New York: MacMillan. pp 1182.
0
acyclovir and its 2 -ester prodrugs. Pharm Res
11:237±242.
14. Beauchamp LMO, de Miranda GF, Burnette P,
Krenitsky TA. 1992. Amino acid ester prodrugs of
acyclovir. Antiviral Chem Chemother 3:157±164.
15. Lee VHL, Morimoto KW, Stratford RE. 1982.
Esterase distribution in the rabbit cornea and its
implications in the ocular drug bioavailablity.
Biopharmacol Drug Dispos 3:291±300.
16. Lee VH, Robinson JR. 1986. Topical ocular drug
delivery: Recent developments and future chal-
lenges. J Ocul Pharmacol 2:67±108.
. Hughes PM, Krishnamoorthy R, Mitra AK. 1993.
Effect of acylation on the ocular disposition of
acyclovir. I: Synthesis, physicochemical properties,
0
and antiviral activity of 2 -esters. J Ocul Pharmacol
9
:287±297.
. Audus KL, Bartel RL, Hidalgo IJ, Borchardt RT.
990. The use of cultural epithelial and endothelial
cells for drug transport and metabolism. Pharm
Res 7:435±451.
. Niederkorn JY, Meyer DR, Ubelaker JE, Martin
JH. 1990. Ultrastural and immunological charac-
terization of the SIRC corneal cell line. In Vitro Cell
Dev Biol 26:923±930.
. North-Root H, Yackovich F, Demetrulias J, Gacula
M, Heinz JE. 1982. Evaluation of an in vitro cell
toxicity test using rabbit corneal cells to predict the
eye irritation potential of surfactants. Toxicol Lett
5
6
7
1
17. Dixon M. 1979. In: Boyer PD, editor. Enzymes 3rd
ed. New York: Academic Press. p 45.
18. Hughes PM, Mitra AK. 1993. Effect of acylation on
the ocular disposition of acyclovir. II: Corneal
0
permeability and anti-HSV 1 activity of 2 -esters
in rabbit epithelial keratitis. J Ocul Pharmacol
9:299±309.
19. Goskonda VR, Hill RA, Khan MA, Reddy IK. 2000.
Permeability of chemical delivery systems across
rabbit corneal (SIRC) cell line and isolated corneas:
A comparative study. Pharm Dev Technol 5:409±
416.
20. Chang J, Basu SK, Lee VHL. 2000. Air±interface
condition promotes the formation of tight corneal
epithelial cell layer for drug transport studies.
Pharm Res 17:670±676.
1
4:207±212.
. Jacaruso RB, Barletta MA, Carson S, Hardig W.
985. An in vitro method for assessing corneal
opaci®cation potential using a rabbit corneal cell
line. J Toxicol Cut Ocular Toxicol 4:49±58.
8
9
1
. Korbmacher C, Helbig H, Forster C, Wiederholt M.
‡ ‡
1
988. Characterization of Na /H exchange in a
rabbit corneal epithelial cell line (SIRC). Biochem
Biophys Acta 943:405±410.
0. Jacaruso RB, Carson S, Barletta MA. 1986. The use
of cell lysis as an index of ocular irritation potential.
J Toxicol Cut Ocular Toxicol 5:143±161.
21. Lentz KA, Hayashi J, Lucisano LJ, Polli JE. 2000.
Development of a more rapid, reduced serum
culture system for Caco-2 monolayers and applica-
tion to the biopharmaceutics classi®cation system.
Int J Pharm 200:41±51.
1
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 10, OCTOBER 2001