Effect of mono- and di-acylation on the ocular disposition of ganciclovir: Physicochemical properties, ocular bioreversion, and antiviral activity of short chain ester prodrugs

Article abstract of DOI:10.1002/jps.10072

A series of short-chain carboxylic mono- and diesters of ganciclovir were synthesized in our laboratory. Physico-chemical properties, i.e., solubility (pH 4.2), partition coefficient in 1-octanol/phosphate buffer (pH 7.4), aqueous stability at various pH values, bioreversion kinetics in various ocular homogenates and effectiveness against various Herpes viruses in vitro were determined. The compounds exhibited a decrease in solubility as the ester length ascended with a corresponding increase in the octanol/buffer partition coefficient values. All of the prodrugs exhibit stability profiles typical of a carboxylic ester with maximum stability at neutral or slight acidic pH (4.07.0). Apparent first-order rate constants associated with prodrug to drug hydrolysis in the ocular homogenates varied depending on the size of the promoiety, lipophilicity of the compound, and the ocular tissue studied. The acetyl and butyryl mono and diesters were screened against various Herpes viruses. The monobutyrate ester of ganciclovir exhibits excellent activity against HSV-2 and VZV and provides a very high selectivity index against most of the viruses studied.

Full text of DOI:10.1002/jps.10072

Effect of Mono- and Di-Acylation on the Ocular Disposition
of Ganciclovir: Physicochemical Properties, Ocular
Bioreversion, and Antiviral Activity of Short Chain
Ester Prodrugs
CLAPTON S. DIAS, BANMEET S. ANAND, ASHIM K. MITRA
Division of Pharmaceutical Sciences, School of Pharmacy, University of Missouri±Kansas City, 5005 Rockhill Road,
Kansas City, Missouri 64110-2499
Received 26 February 2001; revised 5 November 2001; accepted 7 November 2001
ABSTRACT: A series of short-chain carboxylic mono- and diesters of ganciclovir were
synthesized in our laboratory. Physico-chemical properties, i.e., solubility (pH 4.2),
partition coef®cient in 1-octanol/phosphate buffer (pH 7.4), aqueous stability at various
pH values, bioreversion kinetics in various ocular homogenates and effectiveness
against various Herpes viruses in vitro were determined. The compounds exhibited a
decrease in solubility as the ester length ascended with a corresponding increase in the
octanol/buffer partition coef®cient values. All of the prodrugs exhibit stability pro®les
typical of a carboxylic ester with maximum stability at neutral or slight acidic pH (4.0±
7.0). Apparent ®rst-order rate constants associated with prodrug to drug hydrolysis in
the ocular homogenates varied depending on the size of the promoiety, lipophilicity of
the compound, and the ocular tissue studied. The acetyl and butyryl mono and diesters
were screened against various Herpes viruses. The monobutyrate ester of ganciclovir
exhibits excellent activity against HSV-2and VZV and provides a very high selectivity
index against most of the viruses studied. ß 2002 Wiley-Liss, Inc. and the American
Pharmaceutical Association J Pharm Sci 91:660±668, 2002
Keywords: mono- and di-acylation; ocular disposition; ganciclovir
INTRODUCTION
implants, and intravitreal injections of ganciclo-
vir or foscarnet. Ganciclovir (GCV), an acyclogua-
nosine analog, was the ®rst drug to show ef®cacy
in HIV-associated CMV retinitis, and hence, is
still the most commonly used anti-CMV drug for
treatment.2 GCV was primarily administered
systemically, but high systemic doses (5±10 mg/
kg) lead to severe side effects like neutropenia,
thrombocytopenia, and abnormal liver function.
GCV, being a hydrophilic molecule, is occluded
out by the Blood±Retinal Barrier (BRB), which
prevents the entry of GCV into the eye after
systemic administration. Oral delivery of GCV
produces very low bioavailability (6%), and hence,
is not appropriate for the treatment in the in-
duction phase of treatment, which requires high
levels of GCV in the eye. Hence, local intravitreal
injections (0.2±0.4 mg), which minimize systemic
Human Cytomegalovirus (CMV) retinitis is the
most common ocular opportunistic infection
occurring in HIV-infected patients. It occurs in
about 20±40% of AIDS patients, especially in
individuals with blood CD4 cell counts less than
50/mL.1 Focal yellowish-white granular patches,
large areas of retinal hemorrhage with necrosis,
diffuse edema are symptoms of advanced CMV
retinitis, and if left untreated can lead to blind-
ness. Many agents are available to treat this
infection, including intravenous ganciclovir, fos-
carnet and cidofovir, slow release ganciclovir
Correspondence to: Ashim K. Mitra (Telephone: 816-235-
1615; Fax: 816-235-5190; E-mail: amitra@ctr.umkc.edu)
Journal of Pharmaceutical Sciences, Vol. 91, 660±668 (2002)
ß 2002 Wiley-Liss, Inc. and the American Pharmaceutical Association
660
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002

OCULAR DISPOSITION OF GANCICLOVIR
661
toxicities, have become the route of choice as the
drug is delivered directly to the target area. But
this mode of delivery has its own limitations. GCV
is rapidly eliminated from the vitreous (elimina-
tion t1/2 is approx.13 h in humans3), and hence,
repeated intravitreal injections are necessary
leading to side effects like endophthalmitis, re-
tinal detachment, vitreal hemorrhage, and optic
atrophy in the long run. Many delivery systems
like intraocular implants, liposomes, and micro-
spheres have been developed to reduce repeated
administration of GCV with varying degrees of
success. Intraocular implants now serve as a
mode for maintenance therapy, with intravitreal
injections still being the route of choice for the
induction phase of treatment.2
Scheme 1. Structures of various GCV ester prodrugs
synthesized.
Lipophilic chemical modi®cation has been
widely used for various hydrophilic drugs to
improve their ocular bioavailability.4±8 The lipo-
philic prodrug modi®cation may release GCV
continuously in the eye, depending on the rate of
hydrolysis and elimination of the pro moiety. Also,
if lipophilic prodrugs of GCV are delivered
directly into the eye, the levels of GCV may be
maintained for a prolonged period of time com-
pared to GCV. Prolonged intravitreal GCV levels
may be attributed to two factors, i.e., the slow
hydrolysis of the prodrug to regenerate GCV and
melanin binding to ocular tissues such as the
retina and iris ciliary body. Unpublished in vitro
studies in our laboratory have shown that as the
lipophilicity of the pro moiety is ascended, binding
to natural melanin progressively increases. To
this end, lipophilic prodrug moieties ranging
from the monoacetate to the monovalerate and
from the diacetate to di-isobutyrate have been
synthesized in our laboratory (Scheme 1). This
research article focuses on these short-chain
aliphatic mono- and di-ester prodrugs of GCV.
Physico-chemical characterization, ocular tissue
hydrolysis, and antiviral screening have been
carried out to determine the optimal prodrug or
prodrugs from the eight ester prodrugs studied.
The objective of this research is to optimize the
physico-chemical properties and enzymatic hy-
drolysis of GCV prodrugs for increased intraocu-
lar bioavailability.
procedures.9 All reagents used were of HPLC
grade. New Zealand albino male rabbits weighing
between 2±2.5 kg were obtained from Myrtle's
Rabbitry, Thompson Station, TN.
Physicochemical Properties
Stability Studies
Six pH values, 1.2, 2.6, 4.2, 5.6, 7.4, and 9.0 were
selected. HCl (pH 1.2and 2.6), phthalate (pH 4.2
and 5.6), phosphate (pH 7.4), and boric acid (pH
9.0) buffers were prepared and ionic strength
was adjusted at 0.1 M. The electrode was cali-
brated at 258C, and pH of the buffers was measur-
ed before each study. Methanolic stock solutions
of the prodrugs (1±3 mg/mL) were prepared and
used immediately for stability studies. Aliquots
(9.8 mL) of the buffer were placed in a screw
capped vial and allowed to equilibrate at 378C.
Prodrug stock solution (0.2mL) was subsequently
added to the buffer, thus adjusting the total
methanol concentration to 2% (v/v). Controls con-
taining methanol and the prodrug were also
carried out simultaneously to determine any de-
gradation due to methanol. The vials were placed
in a constant shaker bath set at 378C and 60 rpm.
Samples (0.2mL) were collected at appropriate
time intervals for up to 24 h. The samples were
immediately stored at À808C. During analysis
the samples (one by one) were thawed and imme-
diately analyzed. All experiments were conducted
at least in triplicate.
MATERIALS AND METHODS
Solubility Studies
GCV was obtained as a gift from Hoffman La
Roche (Nutley, NJ). All the prodrugs were syn-
thesized in our laboratory according to published
Solubility determinations were carried out at pH
4.2, where the prodrugs exhibited maximal
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002

662
DIAS, ANAND, AND MITRA
stability. To establish equilibration time, a time-
dependent study was carried out in the case of
GCV monoacetate (GCVMA). For the rest of the
prodrugs excess amount of the prodrug was
placed in 3-mL buffer in a screw-capped vial.
The vials were placed in a constant temperature
shaker bath at 378C and 60 rpm for 24 h. At the
end of this time period, the samples were ®ltered
through 0.45-mM Nalgene syringe ®lter mem-
branes. The initial 1 mL of the ®ltrate was
discarded. All experiments were conducted in
triplicate.
logical saline and blotted dry. About 150 mL of
aqueous humor was aspirated using a 27-gauge
needle. An incision was made at the back of the
eye, and the vitreous humor was withdrawn with
a tuberculin syringe. The cornea, lens, and iris
cilary body were sequentially removed after cut-
ting along the scleral±limbus junction. Finally,
the retina±choroid layer was removed and all tis-
sues were stored at À808C prior to use. The
tissues were homogenized in 5-mL chilled (48C)
IPBS for about 4 min with a tissue homogenizer in
an ice bath. The homogenate was centrifuged at
12,500 rpm for 25 min at 48C to remove debris,
and the supernatant obtained was used for
hydrolysis studies. Protein content of each super-
natant was determined by using a BioRad assay
with bovine serum albumin as the standard.
Partition Coef®cient Determination
Partitioning studies were performed with n-
octanol and pH 7.4 Isotonic Phosphate Buffer
(IPBS). The two phases were presaturated with
each other for 48 h. After presaturation, 5 mL
aliquots of octanol-saturated IPBS were placed in
screw-capped vials and 0.2-mL stock solution of
the prodrug prepared in IPBS was added. The
mixture was vortexed and an initial sample
(0.2mL) was collected to determine the initial
aqueous phase concentration Caqi. Subsequently,
5 mL of IPBS-saturated octanol was added. The
vials were then placed in a constant-temperature
shaker bath (60 rpm) at 378C for 24 h. After 24 h,
the 0.2-mL sample was collected and analyzed to
determine the ®nal aqueous phase concentration
Caqf. Because these studies were carried out at pH
7.4 loss of drug from the aqueous phase could also
be in part due to chemical degradation. Hence,
controls containing the prodrug solution in IPBS
were carried out. The ®nal aqueous concentration
was corrected for any degradative losses to
generate Caqf*, which is the concentration in the
aqueous phase after partitioning. The Partition
Coef®cient was determined by the ratio of Coct/Caq
given by eq. 1:
Hydrolysis Procedure
The supernatant was equilibrated at 378C for
about 30 min prior to an experiment. The study
was initiated by the addition of 0.2mL of an
(0.1 mM) IPBS solution of a given prodrug to
0.8 mL of the supernatant. The control consisted
of 0.8 mL of IPBS instead of the supernatant.
Samples (50 mL) were withdrawn at appropriate
time intervals for up to 24 h. The samples were
immediately diluted with 50-mL chilled methanol
to terminate the reaction and stored at À808C
until further analysis. The samples were thawed
and centrifuged at 10,000 rpm for 10 min prior to
analysis by HPLC for the intact ester prodrug and
the regenerated parent drug GCV. Apparent ®rst-
order rate constants were calculated and cor-
rected for any chemical hydrolysis observed with
the control.
Antiviral Testing
The in vitro potency of selected prodrugs namely
GCVMA, GCVMB, GCVDA, and GCVDB were
determined against various Herpes viruses. The
compounds were screened against HSV-1, HSV-2,
EBV, VZV, and HCMV. Low-passage human
®broblast foreskin cells (HFF) were used at a
concentration of 2.5 Â 106 cells per mL in 0.1 mL
of minimum essential medium supplemented
with 10% fetal bovine serum were used. For
HSV-1 and HSV-2, 1000 plaque-forming units
(PFU) per well were used. CMV, VZV, and EBV
were used at a concentration of 2500 PFU per
well. All studies were conducted at NIAID by Dr.
Christopher Tseng under the NIAID antiviral
evaluation program.
ꢀCaqi À CÃaqf †=CÃaqf
ꢀ1†
All samples obtained from the above physico-
chemical studies were analyzed by HPLC assay as
described in the Analytical section.
Ocular Tissue Hydrolysis Studies
Preparation of Ocular Tissues
New Zealand albino male rabbits were used for
this study. Animals were euthanized by a lethal
injection of sodium pentobarbital to the marginal
ear vein. Each eye was immediately enucleated,
and the ocular surface was rinsed with physio-
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002

OCULAR DISPOSITION OF GANCICLOVIR
663
Analytical Procedures
®rst method. For the diesters, a step gradient had
to be established after the monoester and GCV
eluted, as the diesters were very lipophilic com-
pared to the monoester and the GCV. The details
of the method and the retention times have been
summarized in Table 1. Samples were analyzed
by either ¯uorescence or UV (l ˆ 254 nm) detec-
tion. It should be noted that the mono-isobutyrate
prodrug synthesis was not successful, and hence,
there was no standard available for identifying
this compound with the HPLC assay.
All samples obtained from the various studies
were assayed using HPLC. The system comprised
of a Waters 600 gradient pump, HP 1100 series
¯uorescence detector, and Waters 2487 UV
detector.
HPLC Assay I
This method was used to analyze samples ob-
tained from the different physico-chemical stu-
dies. The system consisted of a single C18 Luna
column 4.6 Â 250 mm (Phenomenex), and the
mobile phase was composed of a mixture of 0.01
M Ammonium phosphate buffer (pH 2.5) and an
organic modi®er, which consisted of a 1:1 mixture
of methanol and acetonitrile. The percentage of
organic phase was varied to elute the different
prodrugs. The samples were detected using
¯uorescence at ex l ˆ 285 nm and em l ˆ 370 nm
(Table 1). This method provided rapid and re-
producible results.
RESULTS AND DISCUSSION
Physico-Chemical Properties
Aqueous Stability
The stability of the eight prodrugs was examined
within the pH range pH 1.2to pH 9.0. The effect of
pH on the stability of these prodrugs was
examined. Figure 1 exhibits the GCVMP stability
pro®le. The observed pro®le exhibits a typical V-
shape characterization of a carboxylic ester with
maximal stability occurring between pH 4.0 and
5.0. All of the prodrugs studied showed similar
behavior in the pH range studied with varying
rate constants of degradation for speci®c acid and
base catalyzed hydrolysis. The prodrugs followed
apparent ®rst-order degradation kinetics at all
pH values (Figure 2). Lipophilicity of the prodrug
does not appear to correlate with the stability
of the prodrug. The esters exhibit maximum
stability between pH 4 and 5.0 with moderate
stability at neutral pH. Considering this fact the
solubility studies were carried out at pH 4.2.
HPLC Assay II
For tissue hydrolysis studies it is imperative that
a method be developed, which would quantify the
diester, monoester, and the parent moiety GCV
simultaneously. The method should be capable of
separating the analyte peaks from any homoge-
nate peaks. To resolve this problem a dual column
method consisting of a cation exchange SCX
column 4.6 Â 250 mm from Alltech Associates
connected in series with a Luna C18 column
(Phenomenex) 4.6 Â 250 mm was employed. The
mobile phase was similar in composition to the
Table 1. Structures of Various GCV Ester Prodrugs Synthesized
HPLC Assay I (Single
C18 Column)
HPLC Assay II (SCX Column in Series With a C18 Column)
Mobile Phase A
(Aq:Org)
(First 15 min)
Mobile Phase B
(ACN:MeOH)
After 15 min
Retention
Time
GCV-
Monoester
GCV-
Diester
Prodrug
Aq:Org
GCV
GCVMA
GCVMP
GCVMB
GCVMV
GCVDA
87:13
80:20
70:30
70:30
75:25
5.287:13
4.9
5.1
8.1
7.9
Ð
1131.3.2
Ð
Ð
Ð
Ð
80:20
70:30
70:30
87:13
8.9
8.5
7.7
10.9
12.4
10.7
Ð
Ð
Ð
0±35%
11.213.3 5.1
2
(15±35 min)
GCVDP
GCVDB
GCVDIB
65:35
55:45
55:45
9.2
11.9
12.7
80:20
70:30
70:30
''
''
''
8.9
8.5
8.5
10.9
12.4
Ð
27.3
36.9
39.1
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002

664
DIAS, ANAND, AND MITRA
GCVMP has solution half lives of approximately
43 days, 9 days, 5 days, and half a day at pH
values 4.2, 5.6, 7.4, and 1.2, respectively. The
results indicate that the prodrug does not undergo
signi®cant degradation during a short stability
study. But, to account for any chemical hydrolysis
control studies were also carried out.
The pro®le in Figure 1 shows a typical V-shape
pro®le. The smooth line ®tted to the data shows a
curvature close to the pka of the free amino group
(pka ˆ 9.2) on the purine ring indicating faster
hydrolysis of the protonated form compared to the
unprotonated species. This observation is consis-
tent with a previously published report on
Acyclovir ester prodrugs.4 Speci®c acid and base
catalysis as well as water-catalyzed spontaneous
hydrolysis were also evident.
Figure 1. pH stability pro®le for GCVMP.
Solubility and Partition Coef®cient
solubility decreases, the lipophilicity is enhanced
to a larger extent with increasing carbon bulk in
the ester moiety. Partition coef®cient values
increased by about 80-fold upon di-substitution
with the butyrate pro-moiety. The results indicate
that branching of the ester moiety enhances its
partition into the octanol phase because of in-
creased surface area of the carbon chain. Partition
studies done previously in our laboratory have
determined the log P value for GCV to be À1.55.10
Thus, by esteri®cation the lipophilicity of GCV
has been enhanced to approximately 100 times
with the monovalerate ester and upto 425 times
with the dibutyrate ester. The rate-limiting
barrier to transport of hydrophilic drugs is the
lipophilic bilayer of the cell membrane. It may be
hypothesized that the enhanced lipophilicity of
Aqueoous solubility of these prodrugs was deter-
mined at pH 4.2(phthalate buffer) because the
prodrugs exhibited maximal stability at this pH.
A solubility time study was conducted for GCVMA
to establish equilibration period, which was
determined to be 24 h. All solubility determina-
tions were carried out after 24-h equilibration.
The determined values for solubility and partition
coef®cient have been provided in Table 2. As the
number of carbons in the promoiety is gradually
ascended both in the mono- and di-substituted
prodrugs, lipophilicity gradually increases and
solubility decreases. The solubility decrease is
about ®vefold for the monovalerate ester com-
pared to the monoacetate ester, and about sixty
times with the di-isobutyrate ester. Although the
Table 2. Physio-Chemical Properties for the
Ganciclovir Ester Prodrugs
Prodrug
Solubilitya
Log Pb
GCVMA
GCVMP
GCVMB
GCVMV
GCVDA
GCVDP
GCVDB
GCVDIB
15.15 (Æ 2.55)
11.89 (Æ 0.73)
8.35 (Æ 0.23)
4.11 (Æ 0.48)
5.3 (Æ 0.25)
À1.08 (Æ 0.24)
À0.92( Æ 0.05)
À0.30 (Æ 0.07)
À0.07 (Æ 0.01)
À0.78 (Æ 0.03)
À0.23 (Æ 0.07)
0.59 (Æ 0.03)
0.83 (Æ 0.09)
3.26 (Æ 0.19)
0.61 (Æ 0.09)
0.26 (Æ 0.06)
Values are mean (Æ SD).
aSolubility determination carried out in pH 4.2phthalate
buffer at 378C.
bPartition coef®cient values (P) were carried out in
1-octanol/phosphate buffer(7.4) at 378C.
Figure 2. First-order chemical hydrolysis of GCVMV
at pH 1.2, 4.2, 7.4, and 9.0.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002

OCULAR DISPOSITION OF GANCICLOVIR
665
Figure 4. Hydrolysis of GCV monoesters in rabbit
lens homogenate at 378C.
Figure 3. Regeneration of GCV from GCVMA upon
hydrolysis in isolated rabbit vitreous humor.
the synthesized prodrugs may increase ocular
bioavailability on systemic administration. The
blood retinal barrier (BRB) like the blood±brain
barrier prevents the entry of relatively hydro-
philic molecules. Enhanced lipophilicity will help
overcome this barrier, and may increase the
ocular availability of GCV when these prodrugs
are delivered systemically. In addition, the lipo-
philicity of the prodrugs may lead to enhanced
melanin binding due to increased hydrophobic
interactions between the prodrug and melanin,
leading to formation of a reservoir depot in the
retinochoroidal and iris±ciliary body.
drug hydrolysis reported previously.11 The results
indicate that increased lipophilicity of the sub-
strate leads to an increased interaction with the
enzyme-active site. It has been hypothesized that
the enzyme-active site is a lipophilic pocket.4 As
the chain length is ascended from the acetate to
the valerate the rate of hydrolysis increases in
most tissues. Also, the diesters cleave at a faster
rate compared to the monoesters, indicating that
the increased bulk at the two esteri®cation sites
did not hinder enzyme activity. However, values
for the di-isobutyrate ester are less compared to
the dibutyrate ester, indicating that branching of
the side chain slows down the enzymatic cleavage
by the various ocular tissues. This may be attri-
buted to the increased bulk of the side chains
leading to steric hindrance. The esterase activity
may decrease with further increase in carbon
chain because of the bulk of the pro moiety.11
Majority of the esterase activity exhibited in
ocular tissues is due to the presence of acetylcho-
line esterase and butyrylcholine esterases. The
esterase activity of a tissue is a function of the
activity exhibited by both enzymes. The propor-
tion of each enzyme present plays an important
role, as these enzymes are speci®c, and will have
different af®nities for the different pro-moieties.
Generally, in most of the ocular tissues it has
been shown that butyrylcholine esterases form
the major portion of the esterases present in
ocular tissues.12 This could also be one important
factor leading to increased hydrolysis of pro-
moieties with a carbon chain length greater or
equal to butyrate. Values in Table 3 indicate that
the retina±choroid and iris±ciliary body show
Ocular Homogenate Studies
Apparent ®rst-order hydrolytic rate constants
associated with GCV mono- and di-esters in
aqueous humor, vitreous humor, lens, iris±cili-
ary, cornea and retina±choroid have been pre-
sented in Table 3. The values in the table have
been obtained after correcting for any chemical
degradation due to IPBS. Loss of the prodrug was
accompanied by regeneration of GCV in the case
of monoesters (Figure 3), whereas the diester ®rst
hydrolyzes to generate the monoester, which is
sequentially hydrolyzed to GCV. The combined
molar amounts of the prodrug and regenerated
parent compound remained constant throughout
the experiment. The formation of no other
metabolites like guanine indicates only esterase
activity in the tissues. Ocular tissue hydrolysis
shows dependence on ester chain length
(Figure 4), and this ®nding is consistent with
the chain-length dependence of ocular ester pro-
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002

666
DIAS, ANAND, AND MITRA
highest esterase activity compared to other ocular
tissues studied. This result is expected, as blood
capillaries supply the retina and the iris±ciliary
body and the amount of esterase enzymes present
in these tissues is higher than in avascular
tissues. Vitreous humor, which will be the biolo-
gical environment once the drug is administered
intravitreally, also cleaves the prodrug moiety,
albeit to a much lesser extent compared to other
tissues.
Antiviral Testing
The acetyl and butyryl prodrugs of GCV were
screened against various HSV strains. The results
have been summarized in Table 4. The stability of
the prodrugs was not measured in the media used
for carrying out the antiviral activity determina-
tions. These prodrugs are susceptible to chemical
hydrolysis, and will de®nitely hydrolyze to give
GCV to a certain extent. Also, the cellular
enzymes present will hydrolyze the prodrugs to
generate the parent drug GCV, which will elicit
its antiviral activity. The activity noted for the
monoester prodrugs is indeed the activity exerted
by the regenerated GCV. GCV monobutyrate has
similar activity as ACV and GCV against most of
the strains studied. The monobutyrate has the
most activity and the best selectivity index (CC50/
EC50) compared to the other esters studied. The
diester prodrugs of GCV had very little activity
compared to the monoesters. A plausible explana-
tion for this could be the fact that for antiviral
activity a free hydroxyl group is required.13,14 The
diester prodrugs though lipophilic lack a free
hydroxyl group, and hence, intact diester would
result in little or no activity. The monobutyryl
prodrug shows more activity compared to the
monoacetyl prodrug, owing to the fact that the
cellular esterases as in the ocular tissues may
have more af®nity for it. In similar studies done
with ACV ester prodrugs the butyrate ester show-
ed good activity against these herpes strains.4
GCV monobutyrate appears to be an ideal candi-
date as an ocular prodrug for antiviral activity.
The af®nity of the enzymes for this promoiety and
its suf®cient lipophilicity makes it a valuable
prodrug for use in ocular therapeutics.
CONCLUSIONS
The ideal ocular prodrug should be the one that
possesses suf®cient lipophilicity to cross the BRB
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002

OCULAR DISPOSITION OF GANCICLOVIR
667
Table 4. Antiviral Screening of GCV Ester Prodrugs for Various Herpes Viruses
Entity
HSV-1 mg/mL
HSV-2 mg/mL
HCMV mg/mL
MCMV mg/mL
VZV mg/mL
ACV
GCV
GCVMA
EC50 ˆ 0.14
Ð
EC50 ˆ 1.8
Ð
EC50 ˆ 2.7
CC50 > 100
SI > 37
EC50 ˆ 1.7
CC50 > 100
SI > 58.8
EC50 ˆ 15.8
CC50 > 100
SI > 6.3
EC50 ˆ 100
CC50 > 100
SI > 0
Ð
Ð
EC50 ˆ 0.24
Ð
EC50 ˆ 0.16
EC50 ˆ 1.2
CC50 > 100
SI > 83.3
EC50 ˆ 0.53
CC50 > 100
SI > 189
EC50 ˆ 0.54
CC50 > 100
SI > 185
EC50 ˆ 1.3
CC50 > 100
SI > 77
EC50 ˆ 1.1
EC50 ˆ 3.4
CC50 > 70
SI > 20.6
EC50 ˆ 7.7
CC50 > 100
SI > 13
EC50 ˆ 9.8
CC50 > 100
SI > 10.2
EC50 ˆ 4.4
CC50 > 100
SI > 23
EC50 ˆ 1.5
CC50 > 100
SI > 66.6
EC50 ˆ 0.5
CC50 > 100
SI > 200
EC50 ˆ 2.8
CC50 > 100
SI > 35.7
EC50 ˆ 2.4
CC50 > 100
SI > 41.6
EC50 ˆ 0.6
CC50 > 100
SI > 169
GCVMB
GCVDA
GCVDB
EC50 ˆ 0.2
CC50 > 100
SI > 370
EC50 ˆ 0.6
CC50 > 100
SI > 164
EC50 ˆ 1.1
CC50 > 100
SI > 91
CC50 (cytotoxic concentration 50) ˆ concentration required to inhibit 50% stationary cells to take up neutral red.
EC50 (effective concentration 50) ˆ concentration required to inhibit cell growth by 50%. SI (selective index) ˆ CC50/EC50
.
and possess suf®cient solution stability for it to be
formulated into an appropriate ophthalmic
dosage form. The ester prodrugs must exhibit
the desired hydrolytic kinetics when delivered
into the eye. The prodrug must hydrolyze to
generate the therapeutically active parent com-
pound at the target site. From the various studies
performed on the prodrugs synthesized so far it
can be concluded that these compounds show the
desired physico-chemical properties, namely
enhanced lipophilicity and suf®cient solution
stability, between pH 4.0 and 7.0. Ocular tissue
hydrolysis indicates that these prodrugs are
bioconverted to the parent compound GCV, which
is the therapeutically active drug. The prodrugs
particularly monoesters have also shown the
ability to inhibit viral growth in vitro, which adds
to their utility in the antiviral therapeutics.
In summary, the prodrugs investigated thus
far show positive contributions towards the
development of an ideal prodrug for enhancing
the ocular bioavailability of GCV. It would be
interesting to see how longer chained esters and
charged esters of GCV perform under similar
experimental conditions. The longer chained
esters may release the parent moiety very slowly
in the eye owing to their size. This may prove
to be useful for long-term therapeutics. Melanin
binding capacity of these esters and the
esters studied here needs to be investigated.
All these studies will be the topic of investiga-
tion in our future article with these series of
compounds.
ACKNOWLEDGMENTS
This work was supported by grants from the
National Eye Institute EY09171 and EY10659.
We would like to acknowledge Dr. HongWu Gao
for synthesizing the various ester prodrugs in our
laboratory, and Dr. Christopher Tseng from the
NIAID for conducting all of the in vitro antiviral
screening.
REFERENCES
1. Holland GN. 1999. New strategies for the manage-
ment of AIDS-related CMV retinitis in the era of
potent antiretroviral therapy. Ocular Immunol
In¯am 7(3±4):179±188.
2. Walmsley S, Tseng A. 1999. Comparative toler-
ability of therapies for cytomegalovirus retinitis.
Drug Safety Sep 21(3):203±224.
3. Ashton P, Brown JD, Pearson PA, Blandford SD,
Smith TJ, Anand R, Nightingale SD, Sanborn GE.
1992. Intravitreal ganciclovir pharmacokinetics in
rabbits and man. J Ocular Pharmacol 8(4):343±
347.
4. Hughes PM, Krishnamoorthy R, Mitra AK. 1993.
Effect of acylation on the ocular disposition of
acyclovir I: Synthesis, physicochemical properties
and antiviral activity of 20 esters.
Pharmacol 9(4):287±297.
5. Yaun SS, Bodor N. 1976. Synthesis and activity of
(R)-(À)-m-trimethylacetoxy-a-(methylamino)methyl]
benzyl alcohol chloride: A prodrug form of (R)-(À)
phenylephrine. J Pharm Sci 65:929±931.
J Ocular
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002

668
DIAS, ANAND, AND MITRA
6. McClure DA. 1975. Effect of a prodrug of epinephr-
ine (dipivalyl epinephrine) in glaucoma-general
pharmacology, and clinical experience. ACS sym-
posium series 14. Washington, DC: American
Chemical Society. pp 224±236.
7. Narurkar M, Mitra AK. 1989. Prodrugs of 5-iodo-20-
deoxyuridine for enhanced ocular transport. Phar-
maceut Res 67:887±891.
8. Narurkar M, Mitra AK. 1988. Synthesis, physico-
chemical properties and cytotoxicity of a series of 50
prodrugs of 5-iodo-20-deoxyuridine. Pharmaceut
Res 5:734±737.
9. Gao H, Mitra AK. 2000. Regioselective synthesis of
various prodrugs of ganciclovir. Tetrahedron Lett
41:1131±1136.
acycloguanosine analogs. Ph.D. Dissertation Part
2. Purdue University.
11. Chang SC, Lee VHL. 1983. In¯uence of chain
length on the in vitro hydrolysis of model ester
prodrugs by ocular esterases. Curr Eye Res
2(10):651±656.
12. Lee VHL, Chang SC, Oshiro CM, Smith RE. 1985.
Ocular esterase composition in albino and pigmented
rabbits: Possible implications in ocular prodrug des-
ign and evaluation. Curr Eye Res 4(11):1117±1125.
13. Chu CK, Cutler SJ. 1986. Chemistry and antiviral
activities of acyclonucleosides. J Heterocyc Chem
23:189±319.
14. Harnden MR, Wyatt PG, Boyd MR, Sutton D. 1990.
Synthesis and antiviral activity of 9-alkoxypurines. 1.
9-(3-hydroxypropoxy)- and 9-[3-hydroxy-2-hydroxy-
methyl) propoxy] purine. J Med Chem 33:187±196.
10. Krishnamoorthy R. 1995. Ocular pigmentation
effects on the intravitreal disposition of novel
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 91, NO. 3, MARCH 2002