In-vitro hydrolysis, permeability, and ocular uptake of prodrugs of N-[4-(benzoylamino)phenylsulfonyl]glycine, a novel aldose reductase inhibitor

Article abstract of DOI:10.1211/0022357001774877

To enhance the ocular uptake of N-[4-(benzoylamino)phenylsulfonyl]glycine (BAPSG), two ester (methyl and isopropyl) prodrugs were synthesized and evaluated for their stability in various buffers (pH 1-9), hydrolysis in rabbit ocular tissues (cornea, conju

Full text of DOI:10.1211/0022357001774877

J. Pharm. Pharmacol. 2000, 52: 1113±1122
Received January 28, 2000
Accepted April 20, 2000
# 2000 J. Pharm. Pharmacol.
In-vitro Hydrolysis, Permeability, and Ocular Uptake of Prodrugs
of N-[4-(Benzoylamino)phenylsulfonyl]glycine, a Novel Aldose
Reductase Inhibitor
GANGADHAR SUNKARA, JACK DeRUITER*, C. RANDALL CLARK* AND UDAY B. KOMPELLA
College of Pharmacy, University of Nebraska Medical Center, Omaha, NE 68198-6025 and
*Department of Pharmacal Sciences, Auburn University, Auburn, AL 36849-5503, USA
Abstract
To enhance the ocular uptake of N-[4-(benzoylamino)phenylsulfonyl]glycine (BAPSG),
two ester (methyl and isopropyl) prodrugs were synthesized and evaluated for their
stability in various buffers (pH 1±9), hydrolysis in rabbit ocular tissues (cornea, con-
junctiva, iris-ciliary body, lens, aqueous humor, and vitreous humor), transport across
cornea and conjunctiva, and in-vivo uptake following topical administration.
Over the pH range of 1±9, the rate constants for degradation ranged from 5Á67 to
1
1
218Á9Â10 3 h for the methyl ester and from 3Á14 to 4Á45Â10 3 h for the isopropyl
ester. At all pH conditions, the isopropyl ester was more stable when compared with the
methyl ester. A change in buffer concentration at pH 7Á4 did not in¯uence the stability of
the prodrugs. The prodrugs were rapidly hydrolysed in the tissue homogenates, with the
rate constants for hydrolysis ranging from 1Á98 to 7Á2Â10 3 min 1 for the methyl ester and
3Á32 to 6Á53Â10 3 min for the isopropyl ester. The in-vitro permeability of the methyl
1
ester was less than the parent drug across cornea and conjunctiva. Isopropyl ester levels
were not detectable in the receiver chamber even at the end of the 4-h transport study.
Following topical administration of BAPSG and the two prodrugs at a dose of 60 mg=eye,
the lowest levels were seen in vitreous humor for parent compound and its methyl ester. In
general, the tissue uptake of methyl ester was less than BAPSG. Isopropyl ester levels were
below detection limits in all the ocular tissues.
Lipophilic ester prodrugs of BAPSG showed good aqueous solution stability in tissue
homogenates. However, these prodrugs lacking the free carboxylate anion exhibited
reduced in-vitro permeability and in-vivo uptake, suggesting the importance of free car-
boxylate anion in the delivery of BAPSG.
N-[4-(Benzoylamino)phenylsulfonyl]glycine (BA-
PSG; I, Figure 1), a selective aldose reductase
inhibitor (May®eld & DeRuiter 1987; DeRuiter et
al 1987, 1991), is of potential value in preventing
diabetic complications of the eye including retino-
pathy (Kinoshita 1965; Kador et al 1990), cataract
(Hotta 1997), and corneal epitheliopathy (Datiles
et al 1983). It is well established that the polyol
pathway, proceeding in two steps, plays a critical
role in diabetic complications (Van Heyningen
1959; Kinoshita 1990; Lee et al 1995; Hotta
1997). During hyperglycaemia, aldose reductase
(alditol±NADP oxidoreductase, EC 1.1.1.21)
‡
utilizes NADPH to reduce glucose to sorbitol, a
membrane impermeable polyol that accumulates in
cells, thereby leading to diabetic complications.
Inhibition of aldose reductase has been shown to
delay or prevent diabetic complications by rever-
sing sorbitol accumulation and the biochemical
changes associated with hyperglycaemic conditions
(Lightman 1993; Yabe-Nishimura 1998). Various
structurally diverse compounds obtained from
synthetic and natural origin have exhibited very
good in-vitro aldose reductase inhibitory activity.
However, none are being marketed due to side
effects associated with non-speci®c inhibition of
Correspondence: U. B. Kompella, UNMC College of Phar-
macy, 986025 University of Nebraska Medical Center, Omaha,
NE 68198-6025, USA.
E-Mail: ukompell@unmc.edu

1114
GANGADHAR SUNKARA ET AL
other members of the oxido-reductase family of
enzymes or due to lack of signi®cant clinical ef®-
cacy (Spielberg et al 1991; Engerman & Kern
1993). Such lack of clinical ef®cacy can be par-
tially attributed to poor ocular tissue penetration
and retention of the previously tested inhibitors.
The goal of our research is to identify novel aldose
reductase inhibitors capable of attaining therapeutic
levels in the ocular tissues. Topical administration
is likely to achieve signi®cant drug levels in the
ocular tissues while allowing systemic uptake of
the drug.
Among a series of compounds, BAPSG was
selected for further development based on its high
activity and its permeability across cornea and
conjunctiva (Kompella et al 1999). The empirical
indicators of ocular effectiveness and systemic
effectiveness, estimated as permeability coef®cient
across cornea=IC50 and permeability coef®cient
across conjunctiva=IC50, were high for BAPSG
compared with other compounds tested. Further-
more, BAPSG displays concentration-dependent
enzyme kinetic pro®les similar to well-established
aldose reductase inhibitors such as sorbinil and
tolrestat (DeRuiter & May®eld 1990). Competitive
and multiple enzyme inhibition studies have indi-
cated that BAPSG has unique aldose reductase
binding sites and that it selectively inhibits aldose
reductase (Davis et al 1993).
Following topical administration, less than 10% of
ocular drugs such as pilocarpine and ¯urbiprofen
reached intraocular tissues (Lee & Robinson 1986).
The bioavailability of a topically administered drug
depends on the physicochemical properties of the
drug including its partition coef®cient (log P) and
degree of ionization (pKa). The pKa, log P and IC50
of BAPSG are 3Á35, 1Á09 and 0Á4 mM, respectively.
The apparent permeability coef®cient of BAPSG
across cornea and conjunctiva was 7Á44Æ 1Á49
Figure 1. Scheme for the synthesis of ester prodrugs of
BAPSG.
(Â10 6 cm s ) and 15Á79Æ 0Á27 (Â10 6 cm s ),
respectively (Kompella et al 1999). BAPSG contains
a free carboxylic acid group that can be esteri®ed to
form lipophilic prodrugs. Lipophilic prodrugs are
bioreversible and are designated to improve drug
absorption through increased partition coef®cient
(Schoenwald & Ward 1978; Bundgaard et al 1986a).
The prodrug approach has already been used to
improve the ocular absorption of various drugs
including adrenaline (Bodor & Visor 1984), pilo-
carpine (Bundgaard et al 1985, 1986b), pros-
taglandins (Bito 1984), timolol (Chang et al 1988),
and acyclovir (Hughes & Mitra 1993). Such lipo-
philic prodrugs can reduce the topical dose or dosing
frequency, and hence, the risk of side effects.
1
1
and isopropyl ester prodrugs of BAPSG (methyl-
BAPSG (II) and isopropyl-BAPSG (III)) were
synthesized (Figure 1). Stability in aqueous solu-
tions and ocular tissue homogenates, corneal and
conjunctival permeability, and in-vivo uptake of
these prodrugs were assessed.
Materials and Methods
Materials
HPLC grade acetonitrile, methanol and glacial
acetic acid were obtained from Fisher Scienti®c
(Fair Lawn, NJ). Hydroxypropyl-b-cylclodextrin
(HP-b-CyD) was a gift from American Maize-
Products Co. (Hammond, IN). Other chemicals
We have tested the utility of lipophilic prodrugs
in enhancing BAPSG delivery to the eye. Methyl

OCULAR DELIVERY OF A NOVEL ALDOSE REDUCTASE INHIBITOR
1115
used for preparing buffer solutions and formula-
tions were obtained from Sigma Chemical Co. (St
Louis, MO). In-vitro and enzyme hydrolysis studies
were conducted using a Dubnoff metabolic shake
incubator (Precision Scienti®c Co., Chicago, IL).
Male New Zealand White rabbits (2Á5±3Á5 kg) were
obtained from Knapp Creek Rabbitry (Amana, IA).
All animals were handled in accordance with the
Guiding Principles in the Care and Use of Animals
(DHEW Publication, NIH 80-23).
Pallas software (Version 2.0, CompuDrug Interna-
tional, Inc., Budapest, Hungary).
HPLC assay
A reverse-phase HPLC method was used for
quantifying BAPSG, methyl-BAPSG, and iso-
propyl-BAPSG. A Waters HPLC system including
a solvent delivery pump (Waters TM 616), a con-
troller (Waters 600 S), an autoinjector (Waters 717
plus), and a PDA detector (Waters 996) were used.
The peak areas were integrated using Millennium
software (Version 2.15.01). A 25-cm long Micro-
sorb C-18 column (Rainin Instruments, Emeryville,
CA) with a particle diameter of 5 mm and a pore
Synthesis of BAPSG prodrugs
The general scheme for the synthesis of ester
prodrugs of BAPSG is shown in Figure 1. The
procedure for each ester is described below.
Ê
size of 100 A was used. The mobile phase for
BAPSG and methyl-BAPSG consisted of acetoni-
trile, methanol and an aqueous phase (6 mL glacial
acetic acid and 7Á5 g sodium sulphate per 500 mL
water) in a ratio of 20 : 20 : 60 (v=v). The ¯ow
Synthesis of methyl-BAPSG (II). A solution of
470 mg BAPSG (synthesized as reported by May-
®eld & DeRuiter (1987)) in 25 mL methanol with
one drop of concentrated sulphuric acid was heated
at re¯ux for three days. The solvent was removed
under reduced pressure and the resulting product
suspended in water, ®ltered, and washed with water
(2Â10 mL). The crude product was air-dried and
recrystallized from a minimum amount of boiling
methanol. The puri®ed product was isolated by
®ltration and oven-dried under low heat
(>100ꢀC). The structure and purity of the product
was con®rmed by proton NMR (methyl singlet
integrating for 3 protons at 3Á85 ppm), mass spec-
trometry (MW 348) and elemental analysis (theo-
retical: C, 55Á16, H, 4Á63, N, 8Á04; found: C, 55Á16,
H, 4Á69, N, 8Á03). The melting point of methyl-
BAPSG was 202±203ꢀC.
1
rate was maintained at 1Á0 mL min and the col-
umn ef¯uent was monitored at 240 nm for BAPSG
and 280 nm for methyl-BAPSG. 2-Nitro BAPSG
was used as an internal standard. The retention
times for BAPSG, methyl-BAPSG, and 2-nitro
BAPSG were 8Á0, 17Á2, and 6Á5 min, respectively.
The mobile phase for isopropyl-BAPSG consisted
of acetonitrile, methanol and an aqueous phase
(8 mL glacial acetic acid and 5 g sodium sulphate
per 400 mL water) in a ratio of 40 : 20 : 40 (v=v).
1
The ¯ow rate was maintained at 1Á0 mL min and
the column ef¯uent was monitored at 275 nm. The
retention times for isopropyl-BAPSG and BAPSG
were 7Á0 and 4Á5 min, respectively. For all three
compounds, the standard graphs ranged from 1Á5 to
1
250 mg mL for in-vitro studies and from 0Á15 to
1
Synthesis of isopropyl-BAPSG (III). A solution of
500 mg BAPSG in 40 mL isopropyl alcohol with
one drop of concentrated sulphuric acid was heated
at re¯ux for three days. The solvent was removed
under reduced pressure and the resulting product
suspended in water, ®ltered and washed with water
(2Â10 mL). The crude product was air-dried and
recrystallized from a minimum amount of boiling
isopropyl alcohol. The puri®ed product was iso-
lated by ®ltration and oven-dried under low heat
(>100ꢀC). The structure and purity of the product
were con®rmed by proton NMR (isopropyl methyl
doublet integrating for 6 protons centred at
1Á15 ppm), mass spectrometry (MW 376) and
elemental analysis (theoretical: C, 57Á43, H, 5Á36,
N, 7Á44; found: C, 57Á19, H, 5Á28, N, 7Á28). The
melting point of isopropyl-BAPSG was 169±
170ꢀC.
20 mg mL
for in-vivo uptake studies. Samples
with concentrations below the lowest limit of
standard graphs were designated as N.D. (not
detectable).
In-vitro chemical hydrolysis studies
The chemical stability of methyl-BAPSG and iso-
propyl-BAPSG was assessed in aqueous buffer
solutions at pH 1, 5, 6, 7Á4 and 9 at 37Æ 0Á5ꢀC, and
the chemical stability of BAPSG was studied in
aqueous buffers at pH 5, 7Á4 and 9 at 37Æ 0Á5ꢀC.
Furthermore, the stability studies for the two pro-
drugs were conducted at different buffer con-
centrations (0Á025±0Á2 M) at pH 7Á4 to understand
the in¯uence of buffer concentration on the stabi-
lity. Methyl-BAPSG and isopropyl-BAPSG were
expected to hydrolyse to BAPSG in aqueous solu-
tions, with the cleavage of the ester bond being the
rate-limiting step.
Log partition coef®cient (log P) values of
BAPSG and its prodrugs were calculated using

1116
GANGADHAR SUNKARA ET AL
Reactions were initiated by adding 250 mL of a
BAPSG) to 1Á8 mL tissue homogenate maintained
at 37ꢀC and pH 7Á4. Exactly 100-mL samples were
collected at predetermined time intervals (0, 10, 20,
30, 45, 60, and 90 min) into tubes containing
100 mL acetonitrile±methanol (50 : 50) to inactivate
the enzymatic reactions. The tubes were immedi-
ately centrifuged at 5000 g for 5 min, and the
supernatant was collected and stored at 20ꢀC until
assayed for the intact prodrug the next day.
10 mM prodrug solution to 4Á75 mL buffer in screw-
capped test tubes that were pre-equilibrated at
37Æ 0Á5ꢀC. BAPSG and methyl-BAPSG stocks
were prepared in distilled water. Isopropyl-BAPSG
stock was prepared in 50% acetonitrile due to its
limited solubility in water. The concentration of
acetonitrile in the ®nal reaction mixture was 2Á5%.
No cloudiness or haziness was observed upon the
addition of stocks to appropriate buffer solutions.
Unless otherwise stated, all the buffers prepared
were of 0Á05 M concentration and the ionic strength
was maintained at 0Á5 for each buffer by adding
calculated amounts of potassium chloride. For
BAPSG and isopropyl-BAPSG, 200 mL of the
sample was collected at predetermined time inter-
vals (0, 6, 12, 18, 24, 36, 48, 72, and 96 h). At the
end of each time-point, the reaction was stopped by
adding 200 mL acetonitrile±methanol (50 : 50) to
the samples. The samples were stored at 20ꢀC
until assayed. Exactly 50 mL of each sample was
injected onto the HPLC column. For methyl-
BAPSG, 200 mL of the sample was taken from
reaction vessels and added to 200 mL acetonitrile±
methanol (50 : 50) containing 5 mg 2-nitro BAPSG
as internal standard. All other steps were similar to
those used for BAPSG and isopropyl-BAPSG.
Following HPLC analysis, the rate constants for
prodrug hydrolysis were determined by linear
regression of the logarithm of percentage of the
prodrug remaining vs time plots. Triplicate samples
were analysed, and the mean rate constant was
calculated.
Solutions for in-vitro transport studies
Glutathione bicarbonate Ringer's solution (con-
taining in mM: 111Á5 NaCl, 4Á82 KCl, 0Á86
NaH2PO4, 29Á2 NaHCO3, 1Á04 CaCl2Á2H2O, 0Á74
MgCl2Á6H2O, 5 D-glucose, and 0Á3 reduced gluta-
thione) was used for the transport studies.
In-vitro transport studies
New Zealand White male rabbits were killed and
the rabbit cornea and conjunctiva were isolated and
mounted according to the procedure reported by
Kompella et al (1992, 1993). After mounting the
cornea or conjunctiva in modi®ed Ussing chambers
(Navicyte, Sparks, NV), the tissues were exposed to
glutathione bicarbonate Ringer's solution contain-
1
ing the drug (BAPSG, 1 mg mL ; methyl-BAPSG,
1
1
450 mg mL ; isopropyl-BAPSG, 300 mg mL ) on
the mucosal side and neat glutathione bicarbonate
Ringer's solution on the serosal side. The ¯uids
were maintained at 37ꢀC and at pH 7Á4 under 95%
air=5% CO2 aeration. At predetermined time
intervals (0, 30, 60, 120, 180, and 240 min), a
750-mL sample was collected from the serosal side
and the lost volume was compensated with neat
glutathione bicarbonate Ringer's solution pre-
equilibrated at 37ꢀC. The samples collected were
centrifuged and 50-mL supernatant samples were
directly injected onto the HPLC column for quan-
tifying the prodrug as well as the parent drug. The
Enzymatic hydrolysis of prodrugs in ocular tissue
homogenates
New Zealand White male rabbits were killed
1
by injecting 150 mg kg
sodium pentobarbital
(Sleepaway, Fort Dodge, IA) into the right mar-
ginal ear vein. The eyes were enucleated and rinsed
with cold saline to remove any traces of blood.
Firstly, the aqueous humor was collected using a
23-gauge  1Á5-inch needle attached to a 1-mL
tuberculin syringe. A small incision was made in
the eyeball in the conjunctival region to aspirate
vitreous humor. Subsequently, the cornea, iris-
ciliary body, lens, and conjunctiva were separated
in a sequence. All tissues were homogenized
(Bundgaard et al 1986a), the protein concentration
was estimated using a Bio-Rad protein assay kit
(Bio-Rad Laboratories, Hercules, CA), and the
1
apparent permeability coef®cient (Papp, cm s )
was calculated using the equation, Papp
ˆ
¯ux=(surface area of membrane  initial con-
centration), where ¯ux is the slope of the linear
portion of a plot of cumulative amount of drug
transported vs time.
In-vivo uptake studies
1
Isotonic solutions at pH 7Á1 containing 2 mg mL
drug, 5% w=v HP-b-CyD (for BAPSG and methyl-
BAPSG) or 10% w=v HP-b-CyD (for isopropyl-
BAPSG) were used as the dosing solutions. Thirty
1
protein concentration was adjusted to 0Á5 mg mL
in all the homogenates. Enzymatic hydrolysis
reactions were initiated by adding 200 mL prodrug
(10 mM methyl-BAPSG or 5 mM isopropyl-
1
microlitres of a dosing solution (2 mg mL ) con-
taining the parent drug or the prodrugs was instilled

OCULAR DELIVERY OF A NOVEL ALDOSE REDUCTASE INHIBITOR
1117
directly into the conjunctival cul-de-sac of each
eye. At 30 min post-dosing, the rabbit was killed
and the corneal and conjunctival surfaces were
thoroughly rinsed with normal saline and blotted
dry. The ocular tissues including cornea, con-
junctiva, sclera, iris-ciliary body, lens, aqueous
humor, and vitreous humor were collected in pre-
weighed microtubes, weighed and stored immedi-
ately at 70ꢀC. Aqueous humor (50 mL) was
directly injected onto the HPLC column. All other
tissues were minced, and to the ®nely ground tissue
mass 2 mL acetonitrile±methanol (50 : 50) was
added and placed in a shaker for 6 h at 20ꢀC (cold
extraction). The samples were then vortexed thor-
oughly for 5 min and centrifuged at 3000 g for
10 min. The supernatant (1Á5 mL) was collected in
test tubes and the organic layer was dried under N2
gas (N-EVAP, Organomotion Associates, Berlin,
MA). The dried supernatant was reconstituted with
100 mL mobile phase and 50 mL was injected onto
the HPLC column. Standard graphs were generated
using corneal tissue homogenates every day before
the analysis of the samples. The extraction recovery
performed at two concentrations (10 and
tested, methyl-BAPSG was less stable when com-
pared with isopropyl-BAPSG at all pH conditions
and buffer concentrations. This may be due to
greater stabilization of the ester bond with iso-
propyl substitution compared with methyl sub-
stitution. This difference in stability may also be
due in part to the differences in their solubility
behaviour. Isopropyl-BAPSG was less soluble in
aqueous solutions compared with methyl-BAPSG.
The greater stability of isopropyl-BAPSG was
consistent with previous reports when an increase
in chain length made ester prodrugs of 5-iodo-20-
deoxyuridine more resistant to chemical hydrolysis
(Narurkar & Mitra 1989). The rate of hydrolysis of
methyl-BAPSG
was
in
the
order:
pH 9>pH 6>pH 7Á4=pH 5>pH 1 (P < 0Á05), and no
signi®cant difference was observed in the rate
of hydrolysis at different buffer concentrations
at pH 7Á4. Methyl-BAPSG was highly unstable at
pH 9 compared with buffers at low pH (P < 0Á005),
suggesting that the hydrolysis of the methyl ester
1
100 mg mL ) indicated that the recovery was
>95% for the parent drug as well as the prodrugs.
Standard curves generated using corneal tissue
homogenates were used as a representative of other
tissues, which are softer than the cornea.
Mechanisms of BAPSG transport
Corneal permeability of BAPSG (1 mg mL ) was
1
evaluated in the mucosal to serosal direction in the
presence of 100 mM indomethacin (inhibitor of Na -
‡
monocarboxylate transporter) and 100 mM 2,4-dini-
trophenol (inhibitor of oxidative phosphorylation).
Statistical analysis
All values are expressed as meanÆ s.d. for n ˆ 3.
The statistical signi®cance of differences between
the means of the data was evaluated by means of the
unpaired Student's t-test. P < 0Á05 was considered
statistically signi®cant.
Results and Discussion
In-vitro stability
The percentage BAPSG and prodrugs remaining as
a function of time at various pH conditions is
shown in Figure 2. The rate constants for the
degradation of BAPSG and prodrugs in aqueous
solutions are shown in Table 1. Of the two esters
Figure 2. Percent (a) BAPSG, (b) methyl-BAPSG and (c)
isopropyl-BAPSG remaining following incubation in aqueous
solutions of different pH at 37ꢀC. The ionic strength of all
buffers was maintained at 0Á5. The data are expressed as
meanÆ s.d., n ˆ 3. r, pH 1Á0; j, pH 5Á0; m, pH 6Á0; Â,
pH 7Á4; d, pH 9Á0.

1118
GANGADHAR SUNKARA ET AL
Table 1. First-order rate constants for the hydrolysis of
10 mM BAPSG and BAPSG esters at various pH conditions
and buffer concentrations.
prodrugs of BAPSG were converted to BAPSG
in various ocular tissue homogenates (Figure 3).
Although the prodrugs were very stable in aqueous
solutions at pH 7Á4, they were rapidly hydrolysed to
the parent drug in tissue homogenates. This sug-
gests that enzymatic systems must have been
responsible for the ester hydrolysis. The prodrugs
tested in this study are carboxylic esters similar to
PGF2a ester prodrugs tested previously (Camber et
al 1986; Camber & Edman 1987).
Methyl and benzyl esters of PGF2a were shown
to be hydrolysed by butyrylcholinesterase but not
acetylcholinesterase or carbonic anhydrase (Horibe
et al 1998). The enzymatic hydrolysis rate con-
stants for methyl-BAPSG estimated in the cornea,
conjunctiva, lens, and vitreous humor (Table 2)
were not signi®cantly different. The rate of
hydrolysis of methyl-BAPSG in iris-ciliary body
was less compared with other tissues, contrary to
the earlier reports suggesting that hydrolysis of
ester prodrugs proceeded most readily in iris-ciliary
body (Bundgaard et al 1986b; Narurkar & Mitra
1989). There was no signi®cant difference in the
hydrolysis pattern of methyl-BAPSG and iso-
propyl-BAPSG in tissue homogenates of cornea,
lens, aqueous humor, and vitreous humor. The rate
of hydrolysis of isopropyl-BAPSG was about
pHa
Buffer
concn (M)
Apparent ®rst-order rate constant
1
(Â10 3 h
)
BAPSG
Methyl-
BAPSG
Isopropyl-
BAPSG
1
5
6
7Á4
0Á05
0Á05
0Á05
0Á025
0Á05
0Á1
N.D.
8Á63Æ 1Á3
N.D.
5Á67Æ 1Á15 4Á45Æ 0Á03
14Á97Æ 3Á68 3Á72Æ 0Á67
22Á02Æ 4Á38 3Á35Æ 0Á36
12Á44Æ 2Á53 3Á15Æ 0Á14
N.D.
2Á51Æ 0Á99 11Á75Æ 2Á53 3Á14Æ 1Á30
N.D.
N.D.
1Á93Æ 0Á93 218Á9Æ 20Á9 4Á03Æ 0Á14
12Á76Æ 1Á84 2Á01Æ 0Á14
0Á2
N.D.
2Á85Æ 0Á41
9
0Á05
aIonic strength was maintained at 0Á5. Data are expressed as
meanÆ s.d. for n ˆ 3. N.D., not determined.
was catalysed by alkaline conditions. This can be
explained by the general mechanism of hydrolysis
of carboxylic esters (Okuyama 1986).
The rate of hydrolysis of isopropyl-BAPSG in
aqueous solutions was in the order: pH 1=
pH 9>pH 5=pH 6=pH 7Á4 (P < 0Á05). The stability
of isopropyl-BAPSG under acidic and alkaline
conditions was probably due to the protection
offered by the bulky isopropyl group. At alkaline
pH, BAPSG was the most stable of the three
compounds. This can be explained in part by
assuming the presence of intramolecular hydrogen
bonding of BAPSG in aqueous solutions. Possibly
for this reason, we observed a change in the UV
spectrum of methyl and isopropyl esters of BAPSG
compared with BAPSG alone. From the results, it
can be observed that methyl and isopropyl group
substitution affected the aqueous stability of the
prodrug. Thus, ester hydrolysis reduced with
increasing chain length of esters. The ester bond
appeared to be sensitive at alkaline pH while the
parent compound degraded more at acidic pH.
Many ophthalmic prodrugs tested previously were
O-acyl derivatives that were less stable in aqueous
solutions compared with the carboxylic esters
tested in this study (Bundgaard et al 1986b; Chien
et al 1991).
Enzymatic hydrolysis
Esterases including acetyl cholinesterase and
butyryl cholinesterase are abundant in various tis-
sues including ocular tissues (Wistrand et al 1986).
Also, ocular tissues possess carbonic anhydrase, an
enzyme capable of hydrolysing ester bonds (Wis-
trand et al 1986; Elleby et al 1999). Consistent with
those earlier reports was our observation that ester
Figure 3. Percent (a) methyl-BAPSG and (b) isopropyl-
BAPSG remaining following incubation in various ocular
tissue homogenates at pH 7Á4 and 37ꢀC. The data are expressed
as meanÆ s.d., n ˆ 3. r, Cornea; j, conjunctiva; m, iris-
ciliary body; s, lens; Â, aqueous humor; d, vitreous humor.

OCULAR DELIVERY OF A NOVEL ALDOSE REDUCTASE INHIBITOR
1119
Table 2. First-order rate constants for the hydrolysis of
BAPSG esters in cornea, conjunctiva, iris-ciliary body, lens,
aqueous humor and vitreous humora.
BAPSG and BAPSG during the methyl-BAPSG
transport study is shown in Figure 4. Although
more lipophilic, the permeability coef®cients of
methyl-BAPSG were lower (cornea 2Á17Æ 1Á08 Â
Tissue
Apparent ®rst-order rate constant
1 b
1
10 6 and conjunctiva 4Á8Æ 0Á9 Â 10 6 cm s ). The
(Â10 3 min
)
effective permeability coef®cients of methyl-BAPSG,
estimated by combining the amounts of BAPSG and
methyl-BAPSG in the receiver compartments, were
3Á92Æ 1Á54 Â 10 6 and 6Á28Æ 3Á06 Â 10 6 cm s 1 for
cornea and conjunctiva, respectively. At the end of 4 h,
percent methyl-BAPSG transported was less than that
of BAPSG (Table 3).
The cumulative amount of BAPSG transported
when isopropyl-BAPSG was added to the donor
compartment is shown in Figure 5. While the pro-
drug levels were below detection limits in the
receiver chamber at the end of 4 h, BAPSG levels
(hydrolytic product of isopropyl-BAPSG) were
4Á5Æ 0Á005 mg (1Á01%) and 21Á9Æ 6Á8 mg (4Á91%)
across cornea and conjunctiva, respectively. These
levels were lower than those achieved during the
BAPSG transport study (Table 3).
Methyl-BAPSG
Isopropyl-BAPSG
Cornea
6Á25Æ 0Á69
7Á11Æ 0Á57
1Á98Æ 0Á76
7Á20Æ 1Á76
4Á13Æ 0Á79
6Á85Æ 0Á57
6Á26Æ 2Á04
4Á35Æ 0Á51
4Á01Æ 0Á37
6Á53Æ 1Á31
3Á32Æ 0Á88
5Á49Æ 0Á14
Conjunctiva
Iris-ciliary body
Lens
Aqueous humor
Vitreous humor
aExperiment was conducted at 37ꢀC and pH 7Á4. Protein
1
.
concentration was adjusted to 0Á5 mg mL
expressed as meanÆ s.d. for n ˆ 3.
bData are
2-fold that of methyl-BAPSG in iris-ciliary body
homogenates. In conjunctiva, the rate of hydrolysis
of methyl-BAPSG was about 1Á5-times that of
isopropyl-BAPSG. Esterases possess a hydrophobic
cavity in their active site (Camber et al 1986),
suggesting that the greater the lipophilicity of an
ester, the higher is the rate of hydrolysis. However,
an increase in the chain length by two units in our
study enhanced prodrug hydrolysis in iris-ciliary
body only. To understand more about the
mechanism of hydrolysis of these compounds,
enzyme kinetic studies need to be conducted with
speci®c esterases.
The calculated log P values for BAPSG, methyl-
BAPSG and isopropyl-BAPSG were 1Á09, 1Á53, and
2Á45, respectively. Since methyl-BAPSG and iso-
In-vitro transport
The percent of BAPSG, methyl-BAPSG and iso-
propyl-BAPSG remaining at 37ꢀC in pH 7Á4 buffer
after 4 h was 99%, 95Á4%, and 98Á7% (Figure 2),
respectively, suggesting that these drugs were
relatively stable during the course of in-vitro
transport studies. The permeability coef®cient of
Figure 4. Cumulative amount of methyl-BAPSG and BAPSG
transported across rabbit cornea and conjunctiva following
the parent drug (BAPSG) across cornea and con-
junctiva was 7Á44Æ 1Á5Â10
1
6
mucosal exposure to 450 mg mL methyl-BAPSG. The data
and 15Á8Æ 0Á27
are expressed as meanÆ s.d., n ˆ 3. u, Cornea (methyl-
BAPSG); j, conjunctiva (methyl-BAPSG); s, cornea
(BAPSG); d, conjunctiva (BAPSG).
Â10 6 cm s , respectively, as reported by Kom-
1
pella et al (1999). The appearance of methyl-
Table 3. Effective permeability coef®cients and percent transport of BAPSG, methyl-BAPSG, and isopropyl-BAPSG across
.
cornea and conjunctivaa, b
BAPSGc
Methyl-BAPSGd
Isopropyl-BAPSGd
Cornea
Conjunctiva
Cornea
Conjunctiva
Cornea
Conjunctiva
1
Papp (Â10 6 cm s
)
7Á44Æ 1Á49
6Á27Æ 2Á08
15Á79Æ 0Á27
14Á03Æ 1Á71
3Á92Æ 1Á54
2Á77Æ 0Á48
6Á28Æ 3Á06
5Á7Æ 0Á5
1Á31Æ 0Á23
1Á02Æ 0Á09
2Á7Æ 1Á2
% Transport in 4 h
4Á89Æ 1Á5
aData are expressed as meanÆ s.d. for n ˆ 3. bThe initial donor concentrations of BAPSG, methyl-BAPSG and isopropyl-
1
c
BAPSG, were 1 mg mL 1, 450 mg mL and 300 mg mL 1, respectively. From Kompella et al (1999). dEffective permeability
coef®cient estimated was based on the total amount of BAPSG and prodrug transported. Effective % transported was estimated
similarly.

1120
GANGADHAR SUNKARA ET AL
administration of BAPSG, the tissue levels of
BAPSG were in the order: conjunctiva > cornea >
aqueous humor > iris-ciliary body > sclera > vit-
reous humor. This seems justi®ed because cornea
and conjunctiva were the tissues directly exposed to
the formulation and the underlying tissues were
aqueous humor and iris-ciliary body. Vitreous
humor is a deeper tissue and so lower drug levels
were anticipated. With methyl-BAPSG, the intact
prodrug levels were higher in the aqueous humor
than cornea and conjunctiva. BAPSG was below
detection limits following methyl-BAPSG dosing.
Following isopropyl-BAPSG dosing, neither the
prodrug nor the drug levels were detectable in any
of the tissues. The low solubility of isopropyl-
BAPSG may be one reason for the low tissue
levels.
Figure 5. Cumulative amount of BAPSG transported across
rabbit cornea and conjunctiva following mucosal exposure to
300 mg mL isopropyl-BAPSG. No isopropyl-BAPSG was
detectable on the receiver side. The data are expressed as
meanÆ s.d., n ˆ 3. s, cornea; d, conjunctiva.
1
propyl-BAPSG are more lipophilic than BAPSG,
we expected higher permeability coef®cients for
these prodrugs similar to the ester prodrugs of
timolol and acyclovir (Chang et al 1987; Hughes &
Mitra 1993). On the contrary, the permeability was
lower for both prodrugs. In our previous studies,
lower tissue permeability was observed with phenyl
group substitution at the second carbon in the
glycine moiety of BAPSG suggesting the impor-
tance of unmodi®ed glycine moiety in the trans-
tissue permeability of BAPSG (Kompella et al
1999). The current study suggested that modi®ca-
tion of free COOH could reduce transport of
BAPSG.
With the increase in log P, we expected an
increase in uptake of the esters and=or parent drug
into deeper ocular tissues, especially aqueous
humor. To the contrary, we observed a decrease in
the uptake of ester prodrugs compared with parent
drug alone (Table 4). This can be explained based
on the physicochemical properties of BAPSG. As
the pKa of BAPSG is 3Á35 it predominantly exists
in its anionic form at physiological conditions. It is
possible that such anionic molecules could be
transported by carrier-mediated processes such as
anion transporters, known to be functional in ocular
tissues (Horibe et al 1998; Philip et al 1998).
Indeed, we observed that the transport of BAPSG
across cornea was reduced from 7Á44Æ 1Á49 Â
6
6
6
10
to 4Á07Æ 0Á7 Â 10
and 3Á57Æ 0Á4 Â10
1
cm s in the presence of indomethacin and dini-
In-vivo uptake
The ocular tissue levels of the prodrugs and parent
drug were estimated using an HPLC assay at
30-min post-dosing (Table 4). Following topical
trophenol, respectively, suggesting that BAPSG
Na -mono-
‡
was possibly transported by
a
carboxylate transporter in an energy-dependent
fashion. Based on this, we speculate that the
transport of BAPSG was facilitated by the presence
of a carboxylic anion, the lack of which reduced the
in-vitro transport and in-vivo uptake of lipophilic
prodrugs.
Table 4. Tissue concentrations of BAPSG, methyl-BAPSG,
and isopropyl-BAPSG following single administration at a
1
dose of 60 mg (30 mL drop of a 2 mg mL solution) to each
eye.
1 a
Tissue
Tissue concn (mg g )
Conclusions
BAPSG
Methyl-
Isopropyl-
BAPSGb
The amount of the drug reaching the intraocular
tissues following topical administration of ester
prodrugs depends on their tissue permeablility and
hydrolysis by tissue esterases. In-vitro stability
studies suggested that BAPSG and its prodrugs
were relatively stable in physiological buffer con-
ditions. The two esters were hydrolysed in ocular
tissues with half-lives ranging from 97 to 350 min.
We observed that BAPSG transport and uptake by
ocular tissues was reduced by lipophilic ester pro-
drugs. As the transport of BAPSG was reduced by
BAPSGb
Aqueous humor
Vitreous humor
Cornea
Conjunctiva
Iris-ciliary body
Sclera
3Á09Æ 0Á42
0Á01Æ 0Á007
5Á27Æ 2Á13
8Á73Æ 3Á00
2Á19Æ 0Á72
0Á30Æ 0Á05
3Á29Æ 0Á45
0Á04Æ 0Á03
0Á24Æ 0Á03
0Á31Æ 0Á13
0Á30Æ 0Á03
0Á25Æ 0Á21
N.D.c
N.D.
N.D.
N.D.
N.D.
N.D.
aData are expressed as meanÆ s.d. for n ˆ 3 (three animals,
b
six eyes). No parent drug detectable in ocular tissues follow-
ing prodrug administration. cN.D., not detected (below the
lower limit of the standard curve).

OCULAR DELIVERY OF A NOVEL ALDOSE REDUCTASE INHIBITOR
1121
‡
DeRuiter, J., Brubaker, A. N., Garner, M. A., Barksdale, J. M.,
May®led, C. A. (1987) In vitro aldose reductase inhibitory
activity of substituted N-benzenesulfonlyglycine deriva-
tives. J. Pharm. Sci. 76: 149±152
DeRuiter, J., Davis, R. A., Wandrekar, V. G., May®eld, C. A.
(1991) Relative structure-inhibition analyses of the N-
benzoyl and N-(phenylsulfonyl) amino acid aldose reduc-
tase inhibitors. J. Med. Chem. 34: 2120±2126
Elleby, B., Sjoblom, B., Lindskog, S. (1999) Changing the
ef®ciency and speci®city of the esterase activity of human
carbonic anhydrase II by site-speci®c mutagenesis. Eur. J.
Biochem. 262: 516±521
Engerman, R., Kern, T. S. (1993) Aldose reductase inhibition
fails to prevent retinopathy in diabetic and galactosemic
dogs. Diabetes 42: 820±825
Horibe, Y., Hosoya, K., Kim, K. J., Lee, V. H. L. (1998)
Carrier-mediated transport of monocarboxylate drugs in the
pigmented rabbit conjunctiva. Inves. Ophthalmol. Vis. Sci.
39: 1436±1443
an inhibitor of the Na -monocarboxylate trans-
porter, and as the modi®cation of COOH in gly-
cine moiety through prodrug derivatization reduced
transport and tissue uptake in this study, it appeared
that the COOH group facilitated the ocular
transport of BAPSG.
Acknowledgements
This work was supported by a grant (EY11777) from
the National Institutes of Health, Bethesda, MD. We
would like to acknowledge Kimberly Tyeryar for her
assistance in the synthesis of the prodrugs.
Hotta, N. (1997) Symposium on treatment of diabetic
complications. 1. Retinopathy. Intern. Med. 36: 136±
138
Hughes, P. M., Mitra, A. K. (1993) Effect of acylation on the
ocular disposition of acyclovir. II: Corneal permeability and
anti-HSV 1activity of 20-esters in rabbit corneal epithelial
keratitis. J. Ocul. Pharmacol. 9: 299±309
Kador, P. F., Akagi, Y., Takahashi, Y., Ikebe, H., Wyman, H.,
Kinoshita, J. H. (1990) Prevention of retinal vessel changes
associated with diabetic retinopathy in galactose-fed dogs by
aldose reductase inhibitors. Arch. Ophthalmol. 108: 1301±
1309
References
Bito, L. Z. (1984) Comparison of the ocular hypotensive
ef®cacy of eicosanoids and related compounds. Exp. Eye
Res. 38: 181±194
Bodor, N., Visor, G. (1984) Formation of adrenaline in the iris-
ciliary body from adrenalone diesters. Exp. Eye Res. 38:
621±626
Bundgaard, H., Falch, E., Larsen, C., Mosher, G. L., Mikkel-
son, T. J. (1985) Pilocarpic acid esters as novel sequentially
labile pilocarpine prodrugs for improved ocular delivery. J.
Med. Chem. 28: 979±981
Kinoshita, J. H. (1965) Cataracts in galactosemia. The Jonas S.
Friendenwal Memorial Lecture. Invest. Opthalmol. 4: 786±
799
Kinoshita, J. H. (1990) A thirty year journey in the polyol
pathway. Exp. Eye Res. 50: 567±573
Kompella, U. B., Kim, K. J., Lee, V. H. L. (1992) Paracellular
permeability of a chloride secreting epithelium. Proceed. Int.
Symp. Contr. Rel. Bioact. Mat. 19: 522±523
Bundgaard, H., Burr, A., Chang, S. C., Lee, V. H. L. (1986a)
Prodrugs of timolol for improved ocular delivery: synthesis,
hydrolysis kinetics and lipophilicity of various timolol
esters. Int. J. Pharm. 33: 15±26
Bundgaard, H., Falch, E., Larsen, C., Mosher, G. L., Millelson,
T. J. (1986b) Pilocarpine prodrugs I. Synthesis, physico-
chemical properties, and kinetics of lactonization of pilo-
carpic acid esters. J. Pharm. Sci. 75: 36±43
Kompella, U. B., Kim, K. J., Lee, V. H. L. (1993) Active
chloride transport in the pigmented rabbit conjunctiva. Curr.
Eye Res. 12: 1041±1048
Camber, O., Edman, P. (1987) Factors in¯uencing the corneal
permeability of prostaglandins F2a and its isopropyl ester
in-vitro. Int. J. Pharm. 37: 27±32
Kompella, U. B., Sunkara, G., Thomas, E. C., Clark, R. C.,
DeRuiter, J. (1999) Corneal and conjunctival permeability
of novel aldose reductase inhibitors: N-[[4-(benzoylamino)-
phenyl]sulfonyl]glycines and N-benzoyl-N-phenyl glycines.
J. Pharm. Pharmacol. 51: 921±927
Lee, A. Y., Chung, S. K., Chung, S. S. (1995) Demonstration
that polyol accumulation is responsible for diabetic cataract
by the use of transgenic mice expressing the aldose reduc-
tase gene in the lens. Proc. Natl Acad. Sci. USA 92: 2780±
2784
Lee, V. H. L., Robinson, J. R. (1986) Topical ocular drug
delivery: recent developments and future challenges. J.
Ocul. Pharmacol. 2: 67±108
Lightman, S. (1993) Does aldose reductase have a role in the
development of the ocular complications of diabetes? Eye 7:
238±241
May®eld, C. A., DeRuiter, J. (1987) Novel inhibitors of rat
lens aldose reductase: N-[[(substituted amino)phenyl]-
sulfonyl]glycines. J. Med. Chem. 30: 1595±1598
Narurkar, M. M., Mitra, A. K. (1989) Prodrugs of 5-iodo-20-
deoxyuridine for enhanced ocular transport. Pharm. Res. 6:
887±891
Okuyama, T. (1986) Mechanism of acid-catalyzed hydrolysis
of ketene dithioacetals: reversibility of carbon protonation.
Acc. Chem. Res. 19: 370±376
Camber, O., Edman, P., Olsson, L.-I. (1986) Permeability of
prostaglandin F2a and prostaglandin F2a esters across
cornea in-vitro. Int. J. Pharm. 26: 259±266
Chang, S. C., Bundgaard, H., Burr, A., Lee, V. H. L. (1987)
Improved corneal penetration of timolol by prodrugs as a
means to reduce systemic drug load. Invest. Ophthalmol.
Vis. Sci. 28: 487±491
Chang, S. C., Bundgaard, H., Buur, A., Lee, V. H. L. (1988)
Low dose O-butyryl timolol improves the therapeutic index
of timolol in the pigmented rabbit. Invest. Ophthalmol. Vis.
Sci. 29: 526±529
Chien, D. S., Sasaki, H., Bundgaard, H., Burr, A., Lee, V. H. L.
(1991) Role of enzymatic lability in the corneal and con-
junctival penetration of timolol ester prodrugs in the pig-
mented rabbit. Pharm. Res. 8: 728±733
Datiles, M. B., Kador, P. F., Fukui, H. N., Hu, T. S., Kinoshita,
J. H. (1983) Corneal re-epithelialization in galactosemic
rats. Invest. Opthalmol. Vis. Sci. 24: 563±569
Davis, R. A., May®eld, C. A., Aull, J. L., DeRuiter, J. (1993)
Enzyme selectivity analyses of arylsulfonylamino acid
aldose reductase inhibitors. J. Enzyme Inhib. 7: 87±96
DeRuiter, J., May®eld, C. A. (1990) Inhibitory activity and
mechanism of inhibition of the N-[[(4-benzoylamino)-
phenyl]sulfonyl] amino acid aldose reductase inhibitor.
Biochem. Pharmacol. 40: 2219±2226

1122
GANGADHAR SUNKARA ET AL
Philip, N. J., Yoon, H., Grollman, E. F. (1998) Monocar-
boxylate transporter MCT1 is located in the apical mem-
brane and MCT3 in the basal membrane of rat TRP.
Am. J. Physiol. 274: R1824±R1828
Schoenwald, R. D., Ward, R. L. (1978) Relationship between
steroid-permeability across rabbit cornea and octanol-water
partition coef®cient. J. Pharm. Sci. 67: 786±788
Spielberg, S. P., Shear, N. H., Cannon, M., Huston, N. J.,
Gunderson, K. (1991) In-vitro assessment of a hypersensi-
tivity syndrome associated with sorbinil. Ann. Intern. Med.
114: 720±724
Van Heyningen, R. (1959) Formation of polyols by the
lens of the rat with `sugar' cataract. Nature 184: 194±
195
Wistrand, P. J., Schanholm, M., Lonnerholm, G. (1986)
Carbonic anhydrase isoenzymes CA I and CA II in
the human eye. Invest. Ophthalmol. Vis. Sci. 27: 419±
428
Yabe-Nishimura, C. (1998) Aldose reductase in glucose toxi-
city: a potential target for the prevention of diabetic com-
plications. Pharmacol. Rev. 50: 21±33