5′-O-D-valyl ara A, a potential prodrug for improving oral bioavailability of the antiviral agent vidarabine

Article abstract of DOI:10.1080/15257770802581757

In order to improve the oral bioavailability of Adenine 9-β-D-arabinofuranoside (Vidarabine, also called ara A), an antiviral drug which is active against herpes simplex and varicella zoster viruses and the first agent to be licensed for the treatment of

Full text of DOI:10.1080/15257770802581757

This article was downloaded by: [McMaster University]
On: 12 December 2014, At: 11:16
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Nucleosides, Nucleotides and Nucleic
Acids
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/lncn20
5′-O-D-Valyl ara A, A Potential Prodrug
for Improving Oral Bioavailability of the
Antiviral Agent Vidarabine
a
a
a
a
Wei Shen , Jae-Seung Kim , Stefanie Mitchell , Phil Kish , Paul
a
a
Kijek & John Hilfinger
a
TSRL Inc. , Ann Arbor, Michigan, USA
Published online: 04 Jan 2011.
To cite this article: Wei Shen , Jae-Seung Kim , Stefanie Mitchell , Phil Kish , Paul Kijek & John
Hilfinger (2009) 5′-O-D-Valyl ara A, A Potential Prodrug for Improving Oral Bioavailability of
the Antiviral Agent Vidarabine, Nucleosides, Nucleotides and Nucleic Acids, 28:1, 43-55, DOI:
10.1080/15257770802581757
To link to this article: http://dx.doi.org/10.1080/15257770802581757
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“
Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions

Nucleosides, Nucleotides and Nucleic Acids, 28:43–55, 2009
Copyright ꢀC Taylor & Francis Group, LLC
ISSN: 1525-7770 print / 1532-2335 online
DOI: 10.1080/15257770802581757
ꢀ
5
-O-D-VALYL ara A, A POTENTIAL PRODRUG FOR IMPROVING ORAL
BIOAVAILABILITY OF THE ANTIVIRAL AGENT VIDARABINE
Wei Shen, Jae-Seung Kim, Stefanie Mitchell, Phil Kish, Paul Kijek, and
John Hilfinger
TSRL Inc., Ann Arbor, Michigan, USA
2
In order to improve the oral bioavailability of Adenine 9-β-D-arabinofuranoside (Vidarabine,
also called ara A), an antiviral drug which is active against herpes simplex and varicella zoster
viruses and the first agent to be licensed for the treatment of systematic herpes virus infection in man,
ꢁ
the corresponding 5 -O-D-valyl ester derivative has been synthesized. Based on their physicochemical
ꢁ
properties, 5 -O-valyl ara A has emerged as a potential prodrug candidate to improve the oral
bioavailability of vidarabine. We describe in this paper a facile synthesis route for the prodrug and
its physicochemical properties.
Keywords Herpes virus; small pox virus; adenine 9-β-D-arabinofuranoside vidarabine;
ara A; prodrug; selective protection of secondary hydroxyl groups; levulinate ester
INTRODUCTION
1-β-D-Arabinofuranosyladenine (vidarabine or ara A) is an antiviral drug
with activity against herpes viruses, poxviruses, and certain rhabdoviruses,
[
1–4]
hepadnarviruses, and RNA tumor viruses.
However, it is more toxic
and less metabolically stable than other current antivirals such as acyclovir
and ganciclovir. It is readily deaminated by adenosine deaminase to ara-
hypoxanthine (ara-H).[ This metabolite possesses weak antiviral activity,
5]
[
6]
which is at least 10-fold less potent than vidarabine. Our recent interest
in prodrugs of vidarabine was triggered by the discovery that vidarabine was
3
- to 5-fold more active against vaccinia and cow pox viruses than cidofovir
[
7]
[8–11]
in plaque reduction assays.
Cidofovir is an anti-pox virus agent
with serious limitations.[
12,13–15]
Furthermore, vidarabine’s activity against
these viruses was enhanced approximately 10-fold when combined with
Received 23 June 2008; accepted 21 October 2008.
Address correspondence to Wei Shen, TSRL Inc, Ann Arbor, MI 48108, USA. E-mail:
wshen@tsrlinc.com
4
3

4
4
W. Shen et al.
ꢁ
2
-deoxycoformycin (pentostatin, a potent inhibitor of adenosine deami-
nase; ADA), thus providing significant superiority to cidofovir. Based on
ꢁ
[16]
these results and earlier studies on 5 -substituted vidarabine analogs,
we hypothesized that minimizing the conversion of vidarabine to its hy-
poxanthine analog could yield a significantly more potent anti-pox virus
agent. With this goal in mind, we have developed a prodrug strategy that
ꢁ
protects the vidarabine from metabolic conversion by making 5 -D- valyl
esters derivative of vidarabine. We report here the chemical syntheses of
ꢁ
the prodrug, as well as some physicochemical characteristics of 5 -O-D-valyl
ara A.
CHEMISTRY
ꢁ
In order to efficiently synthesize 5 -O-amino acid ester derivative of
ꢁ
vidarabine in scalable synthetic process, it is crucial to selectively protect 2
and 3 hydroxyl groups of vidarabine. Such blocking groups had to be easily
ꢁ
and quickly removed under non-basic conditions (otherwise concomitant
cleaving of the acyl groups in the amino acid ester moiety would be ex-
pected). After several unsuccessful attempts with various protecting groups,
ꢁ
ꢁ
the final candidate for protection of the 2 , 3 hydroxyl positions was the
levulinate group. The levulinate group can survive the synthesis conditions
for these prodrugs while it can be easily removed by treating with 2M
[
17a,b]
hydrazine in pyridine-acetic acid buffer for 10 minutes.
Normal esters
[
17c,d]
are not cleaved under these conditions.
In addition, the levulinate is
[
18]
less prone to migrate than the benzoate and acetate.
is illustrated in Scheme 1.
The synthesis route
EXPERIMENTAL
General Procedures
Reagents, solvents and solutions: All reagents employed were of Amer-
ican Chemical Society (ACS) grade or finer and were used without fur-
ther purification unless otherwise stated. All chemicals were obtained
from Sigma-Aldrich Co. (St. Louis, MO, USA). For anhydrous reactions,
solvents were dried according to Vogel’s Textbook of Practical Organic
[
28]
Chemistry.
Removal of solvent was performed under reduced pressure
using a B u¨ chi rotary evaporator, followed by evacuation (<0.1 mm Hg)
to constant sample weight. Deionized water was obtained from a Milli-Q
reagent water system (Millipore Co., Milford, MA, USA). All reactions were
carried out under argon atmosphere in dried glassware.

ꢁ
5
-O-D-Valyl ara
45
ꢁ
SCHEME 1 Synthesis of 5 -O-D-valyl ara A.
Purification Techniques
All reactions and fractions from column chromatography were moni-
tored by thin layer chromatography (TLC) using plates with a UV fluores-
cent indicator (normal SiO , Merck 60 F254). One or more of the following
2
methods were used for visualization: UV absorption by fluorescence quench-
ing; iodine staining. Flash chromatography was performed according to the
method of Still[ using Merck type 60, 230–400 mesh silica gel.
29]
Instrumentation for Compound Characterization
Low resolution FAB mass spectra (FAB-MS) were recorded on a Quattro
II mass spectrometer (Waters Corp., Milford, MA, USA). Nuclear magnetic
resonance (NMR) spectra were obtained on Bruker 400 spectrometer
(
(
Bruker BioSpin Ltd., Billerica, MA, USA) from Spectral Data Service
Champaign, IL, USA). 1H NMR chemical shifts are reported in parts
per million (ppm) downfield relative to tetramethylsilane (TMS) using
the residual proton resonance of solvents as reference: D O δ 4.8 and
2
DMSO-d6 δ 2.50. 13C NMR chemical shifts are reported relative to TMS

4
6
W. Shen et al.
or DMSO-d6 δ 39.5. 31P NMR chemical shifts are reported relative to
trimethylphosphate δ 3.0. 1H NMR data are reported in the following order:
number of protons, multiplicity (s, singlet; d, doublet; t, triplet; q, quartet;
qn, quintet and m, multiplet), and coupling constant (J) in Hertz (Hz).
When appropriate, the multiplicity is preceded by br, indicating that the
signal was broad. Elemental analysis (C, H, N) were performed by the
Robertson Microlit Laboratories (Madison, NJ, USA). Yields refer to isolated
yields of compounds estimated to be > 95% pure as determined by 1H and
13C NMR, analytical HPLC and elemental analysis.
9
-[5-O-(tert -Butyldimethylsilyl)-β-D-arabinofuranosyl]adenine (2): The
[
19]
described procedure
was modified. To a stirred suspension of 2.0 g (7.5
mmol) of dry 9-(β-D-arabinofuranosyl)adenine (1) in 20 mL of anhydrous
N,N-dimethylformamide was added 1.2 g (18 mmol) of imidazole, followed
by 1.13 g (7.48 mmol) of tert-butyldimethylchlorosilane. The mixture was
stirred with protection of argon at room temperature for 20 hours, at the
◦
end of which time the solvent was removed at 50 C in vacuo. The residue
was dissolved in 50 mL of ethyl acetate, the solution was washed with two
times of 40 mL of water, one time of 30 mL brine and the extract was dried
(
anhydrous Na SO ) and evaporated to give a white solid which is pure
2 4
enough for next step. Yield: 2.76 g (96%).
H NMR (DMSO-d6) δ 0.100 (6H, 2s, SiMe2), 0.926 (9H, s, t-butylSi);
arabino: 3.822–3.915 (3H, m), 4.193–4.206 (1H, m), 4.238–4.251 (1H, m),
1
ꢁ
ꢁ
3
ꢁ
5
7
.5–6.0 (2H, br, OH2 and OH3 ), 6.305 (1H, d, J = 5.35Hz, H1 ); Adenine:
.298 (2H, s, NH2), 8.168 (1H, s, H2 of adenine), 8.179 (1H, s, H8 of
adenine). 13C NMR (DMSO-d6) δ -5.351(Me from SiMe2), 18.125 (Qua-
ternary Carbon from t-butylSi), 25.855 (Me from t-butylSi); arabino: 62.478
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
(
1
C5 ), 74.262 (C3 ), 75.627 (C2 ), 83.000 (C4 ), 83.056 (C1 ); adenine:
21.681 (C5), 138.128 (C8), 139.909 (C4), 152.437 (C2), 155.882 (C6).
Mass spectrum: calculated for C16H27N5O4Si: 381.5. FAB-MS: m/z 382.30
M+1). Anal. Calcd. For C16H27N5O4Si: C, 50.37; H, 7.13; N, 18.36. Found:
C, 50.67; H, 7.33; N, 18.26
-[5-O-(tert -Butyldimethylsilyl)-2,3-O-dilevulinyl-β-D-arabinofuranosyl]-
(
9
adenine (3): Levulinic acid (3.77 g, 3.33 mL, 32 mmole) was dissolved in
anhydrous ethyl acetate (30 mL), and solution of DCC (6.7 g, 32.6 mmol) in
1
0 mL anhydrous ethyl acetate was added. The reaction mixture was stirred
under protection of argon at room temperature for 1.5 hours. The mixture
was then filtered through a filter tube into a suspension of 2 (2.56 g, 6.7
mmol) and DMAP (0.26 g, 2.1 mmol) in anhydrous ethyl acetate (20 mL),
washing the filter pad (DCU) with more anhydrous ethyl acetate (20 mL).
This reaction was stirred under protection of argon at room temperature
for another 1.5 hours after which time anhydrous EtOH (4 mL) was added.
After an additional 30 minutes, the precipitate (DCU) was filtered off, and
solvent was removed from filtrate in vacuo. The residue was redissolved in 50
ml ethyl acetate and extracted with saturated ammonium chloride aqueous

ꢁ
5
-O-D-Valyl ara
47
solution 40 mL once, water 40 mL twice and 40 ml brine once, the organic
layer was dried with anhydrous Na SO before being evaporated to dryness.
2
4
The resulted syrup was further purified by silica flash chromatograph (6%
MeOH in DCM as eluent) to obtain 3.48 g pure product (89.7%).
1H NMR (DMSO-d6) δ 0.050 (6H, 2s, SiMe2), 0.880 (9H, s, t-butylSi);
Levulinate: 1.946 (3H, s CH3), 2.119 (3H, s, CH3), 2.345–2.379 (2H, t,
CH2), 2.500–2.547 (4H, m, two CH2), 2.744–2.777 (2H, t, CH2); arabino:
3
ꢁ
3
.875–4.121 (3H, m), 5.557–5.633 (2H, m), 6.486 (1H, d, J = 5.19Hz, H1 );
adenine: 7.333 (2H, s, NH2), 8.148 (1H, s, H2 of adenine), 8.190 (1H, s, H8
of adenine).
1
3C NMR (DMSO-d6) δ-5.869 (Me from SiMe2), 18.273 (Quaternary
Carbon from t-butylSi), 26.027 (Me from t-butylSi); Levulinate: 28.463 and
8.624 (two CH3), 29.532 and 29.847 (two CH2 α to the ester bonds), 37.183
2
and 37.526 (two CH2 β to the ester bonds), 170.829 and 171.947 (two ester
carbonyl carbon), 206.308 and 207.027 (two ketone carbo) arabino 61.386
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
(
1
C5 ), 74.037 (C3 ), 75.290 (C2 ), 82.068 (C4 ), 82.296 (C1 ); adenine:
20.385 (C5), 138.776 (C8), 149.269 (C4), 152.832 (C2), 155.711 (C6).
Mass spectrum: calculated for C26H39N5O8Si: 577.7. FAB-MS: m/z 579.13
M+1). Anal. Calcd. For C26H39N5O8Si: C, 54.06; H, 6.80; N, 12.12. Found:
C, 54.07; H, 7.03; N, 12.07.
-[2,3-O-dilevulinyl-β-D-arabinofuranosyl]adenine (4): The foam of 3.0
g (5.21 mmol) of 3 was dissolved in a solution of 1 M TBAF in THF
23 ml) containing glacial acetic acid (2 mL). The reaction mixture was
(
9
(
stirred under protection of argon at room temperature for 1.5 hours
and was then directly filtered through a silica gel plug (4 × 5 cm) with
THF as eluent. The collected fractions were evaporated to dryness in
vacuo. The residue was redissolved in DCM and purified by silica gel flash
chromatography with 3% methanol in DCM as eluent to obtain 2.1 g (87%
yield) gum like product, which was crystallized from ethyl acetate as white
solid.
1H NMR (DMSO-d6) δ Levulinate: 1.916 (3H, s CH3), 2.127 (3H, s,
CH3), 2.380–2.413 (2H, t, CH2, 3J = 6.7), 2.532–2.564 (2H, t, CH2, 3J =
6
.4Hz), 2.748–2.780 (2H, t, CH2, 3J = 6.6Hz), 3.167–3.177 (2H, d, CH2,);
ꢁ
arabino: 3.659–3.730 (2H, m), 4.071–4.099 (1H, m), 5.163–5.192 (1H, t, 5
3
3
ꢁ
OH, J = 5.51), 5.499–5.5.528 (2H, m), 6.475 (1H, d, J = 5.29Hz, H1 );
adenine: 7.324 (2H, s, NH2), 8.144 (1H, s, H2 of adenine), 8.287 (1H, s, H8
of adenine).
13C NMR (DMSO-d6) δ Levulinate: 27.243 and 27.592 (two CH3),
29.195 and 29.466 (two CH2 α to the ester bonds), 36.925 and 37.383
(
two CH2 β to the ester bonds), 170.538 and 171.541 (two ester carbonyl
ꢁ
carbon), 205.990 and 206.745 (two ketone carbo); arabino 60.307 (C5 ),
ꢁ
ꢁ
ꢁ
ꢁ
7
(
4.406 (C3 ), 74.866 (C2 ), 81.185 (C4 ), 81.743 (C1 ); adenine: 118.286
C5), 139.610 (C8), 149.040 (C4), 152.688 (C2), 155.972 (C6). Mass spec-
trum: calculated for C20H25N5O8: 463.44. FAB-MS: m/z 464.86 (M+1).

4
8
W. Shen et al.
Anal. Calcd. For C20H25N5O8: C, 51.83; H, 5.44; N, 15.11. Found: C, 51.81;
H, 5.47; N, 15.31.
9
-[5-O-D-valyl-β-D-arabinofuranosyl]adenine (7): 435 mg (2 mmole) of
α
N-Boc-D-valine, 24.5 mg (0.2 mmole) of DMAP and 310 mg (0.669 mmole)
of compound 4 was dried in a 250 ml single-necked round bottom flask
under high vacuum overnight. 5 ml anhydrous DMF was added. After stired
◦
for 5 minutes at 0 C, 414 mg (2.00 mmole) of DCC in 5 ml DMF was
◦
added dropwise at 0 C. After 1.5 hours, TLC indicated completion of the
reaction. 5 ml anhydrous ethanol was added and the reaction mixture was
stirred overnight. Then DMF was removed under high vacuum by rotavapor
◦
at 40 C. The residue was dissolved in 50 ml ethyl acetate. The insoluble
white precipitate was removed by filtration. The filtrate was consecutively
washed with 20 ml saturated ammonium chloride, 20 ml water, and 20
ml brine. After the solvent was dried with anhydrous sodium sulfate, it
was removed in vacuo. The residue was dissolved in dichloromethane and
eluted from a silica gel column using 92:8 dichloromethane and methanol
as the eluting solvent. Proper fractions were combined and evaporated to
dryness. The fully blocked compound 5 obtained was treated with 1 ml 2 M
hydrazine hydrate in pyridine-acetic acid buffer for 10 minutes to produce
an uncharged compound 6. The excess hydrazine was eliminated through
reaction with pentane-2,4-dione for another 10 minutes. After evaporating
the volatile components, the residue was dissolved in EtOAc, washed with
2
0 ml saturated ammonium chloride, water, and brine. The organic layer
was dried over anhydrous Na SO , filtered and evaporated to dryness. The
2
4
residue was dissolved in dichloromethane and eluted from a silica gel
column using 85:15 dichloromethane and methanol as the eluting solvent.
Proper fractions were combined and evaporated to dryness. 6 ml of 1:1 TFA
and dichloromethane was added to the residue and stirred for 2 hours.
Solvent was removed in vacuo. The residue was washed with cold diethyl
ether. The resulting white solid was purified by reverse phase preparative
HPLC to obtain 240 mg compound 7 (60% from compound 4). HPLC
condition: gradient, 0 to 85% methanol in water containing 0.02% TFA
within 30 minutes.
1
H NMR (D2O) δ Valyl: 0.881–0.912 (6H, t, CH3,), 2.205–2.286 (1H,
3
m, CH), 3.949–3.961 (1H, d, CH, J = 4.63Hz); arabino: 4.185–4.225 (1H,
m), 4.289–4.315 (1H, m), 4.448–4.501 (2H, m), 4.590–4.650 (1H, m), 6.405
ꢁ
(
1H, d, 3J = 5.25Hz, H1 ); adenine: 8.315 (1H, s, H2 of adenine), 8.331 (1H,
s, H8 of adenine).
3C NMR (DMSO-d6) δ Valyl: 16.827 and 17.031 (two CH3), 29.017
1
ꢁ
(
7
CH), 58.163 (CH), 169.394 (ester carbonyl carbon); arabino: 65.42 (C5 ),
5.024 (C3 ), 75.193 (C2 ), 80.481 (C4 ), 84.868 (C1 ); adenine: 114.694
ꢁ
ꢁ
ꢁ
ꢁ
(C5), 143.480 (C8), 144.368 (C4), 148.106 (C2), 149.897 (C6).
Mass spectrum: calculated for C15H22N6O5: 366.37. FAB-MS: m/z
3
67.95 (M+1)

ꢁ
5
-O-D-Valyl ara
49
TABLE 1 Antiviral activity of amino acid prodrugs of vidarabine
EC50a (µM)
CC50b (µM)
Promoiety
(
Compound #)
Cowpox
Vaccinia
HFF
Vidarabinec
12.6
1.4
35
6.2
1.0
8.3
30
>100
90
>100
>100
c
Vidarabine + dCF
ꢁ
5
-O-D-valyl ara A
Cidofovir
26
aFifty percent inhibitory concentration by plaque assay against
the two viruses grown in human foreskin fibroblasts (HFFs).
b
Cytotoxicity determined in HFFs not affected by virus replica-
tion expressed as 50% inhibitory concentrations.
c
Data for vidarabine and vidarabine plus 1 µM deoxycoformycin
(dCF) are averages from three to five experiments. All other data
are averages of at least two experiments.
Anal. Calcd. For C15H22N6O5•2TFA: C, 38.39; H, 4.07; N, 14.14.
Found: C, 38.65; H, 4.15; N, 14.36.
PHYSICOCHEMICAL PROPERTIES
ꢁ
Aqueous solubility of 5 -O-D-valyl ara A has been determined in com-
ꢁ
parison with the solubility of the parent nucleoside ara A. 5 -O-D-valyl ara
A is highly soluble in aqueous media (>10mg/ml) while ara A has aqueous
solubility ranging from 0.143 mg/ml (pH = 7) to 0.605 mg/ml (pH = 2).
ꢁ
Stability of 5 -O-D-valyl ara A in the presence of adenosine deaminase (ADA)
[
7]
is also improved.
vidarabine and 5 -O-D valyl ara A are illustrated in details below.
The different physicochemical properties in between
ꢁ
Antiviral Activity of Vidarabine and Its Prodrugs
We examined the antivirus activity of vidarabine and D-valyl ara A
using two small pox viruses, vaccinia, and cowpox virus. Drug effects were
calculated as a percentage of the reduction in plaque number in the
presence of each drug concentration compared to the numbers obtained
in the absence of drug. As seen in Table 1, vidarabine has significant anti-
pox virus activity with EC50 values that are 4–5 times lower than cidofovir,
the positive control for this virus. It is intriguing that coadministration of
ꢁ
the vidarabine with the adenosine deaminase inhibitor 2 -deoxycoformycin
(at a concentration of 1 µM) gave EC50 values that are >5-(vaccinia) to
1
0- (cowpox) fold lower than those seen with vidarabine alone. At this
ꢁ
low level (1 uM), the 2 -deoxycoformycin alone had no effect on virus
replication; however, with a Ki of ∼50 nM as an ADA inhibitor, its effect is
[
20]
likely at the level of adenosine deaminase.
The D-valyl ester prodrugs of

5
0
W. Shen et al.
SCHEME 2 Enzymatic hydrolysis of vidarabine to its hypoxanthine analog.
vidarabine showed EC50 values comparable to vidarabine alone. It is unclear
at this stage whether the antiviral activity of the prodrugs resulted from the
prodrug itself or vidarabine derived by the hydrolysis of the amino acid
ester moiety. However, as discussed below, these compounds show enhanced
enzymatic stability and high potential for oral delivery.
Stability Studies with Adenosine Deaminase
One aspect of our prodrug strategy for vidarabine is to minimize the
activity of adenosine deaminase (ADA) activity that converts the vidarabine
to 9-(β-D-arabinofuranosyl)hypoxanthine (ara-H), which is at least 10-fold
less potent than vidarabine (Scheme 2). Significant levels of this enzyme are
found in the intestinal mucosa,[
20,21]
which limits the effectiveness of oral
ꢁ
vidarabine delivery. Previous studies suggested that 5 -hydroxy substituted
vidarabine analogs were resistant to adenosine deaminase.
[
22]
Thus, we
examined the stability of vidarabine and the prodrugs in the presence of
adenosine deaminase to evaluate the effect of the amino acid promoieties
on deamination by the enzyme. Briefly, 230 µL adenosine deaminase 1
(0.2 mg/mL) were placed in wells of 96-well plates, the reactions initiated
by the addition of substrates (200 µM) in pH 7.4 phosphate buffer, and
◦
incubated at 37 C for 90 minutes. A minimum of three wells were used for
each compound tested. At various time points, 40 µL aliquots were removed
and added to 40 µL 10% cold TFA. The mixtures were centrifuged and
◦
filtered through a 0.45 µm filter for 10 minutes at 2000 rcf and 4 C. The
filtrate was then analyzed by HPLC-Mass. The disappearance of vidarabine
or prodrug as well as the appearance of the deaminated product was
monitored to determine stability profiles. As seen in Figure 1, vidarabine
was rapidly hydrolyzed with a commensurate rise in the concentration of the
ꢁ
deaminated product, ara-H (Figure 1, top panel). However, the 5 -D-valyl-ara

ꢁ
5
-O-D-Valyl ara
51
FIGURE 1 Deamination of vidarabine and its prodrugs by adenosine deaminase1. Top: Concentration
µM) profiles of the disappearance of vidarabine and the appearance of Ara-H in the presence of
(
adenosine deaminase. Bottom: Percentage of vidarabine or prodrug remaining following incubation
with adenosine deaminase1 for 90 minutes.
A was resistant to the enzyme, with >90% remaining after a 90-minute
incubation with excess enzyme (Figure 1, bottom panel). These results
indicate that the prodrug was capable of withstanding the metabolic effects
of the ADA.
Caco-2 Monolayer Transport Studies
Our next step was to look at the transepithelial transport of the prodrugs
through Caco-2 monolayers. The permeability of vidarabine and the 5-D-
[
23]
valyl ara A was evaluated as previously described.
In these experiments,
transport studies were performed 21 days post-seeding. The assay was

5
2
W. Shen et al.
ꢁ
TABLE 2 Transport of vidarabine and 5 -O-D valyl ara A in
Caco-2 cell monolayer
Compound
Permeability (× 10−7, cm/s)
Vidarabine
1.70
6.00
ꢁ
5
-O-D-valyl ara A
initiated by adding drug transport solution (0.2 mM drug in MES buffer, pH
.0 containing 5 mM D-glucose, 5 mM MES, 1 mM CaCl , 1 mM MgCl , 150
6
2
2
mM NaCl, 3 mM KCl, 1 mM NaH PO ) to the apical chamber of the Caco-2
2
4
insert. Two hundred microliter aliquots were withdrawn from the basolateral
chamber at predetermined intervals and replaced with fresh HEPES pH
7
.4 buffer. The epithelial integrity of representative cell monolayers was
assessed by monitoring transepithelial resistance. As seen in Table 2, the
ꢁ
permeabilities of the 5 -O- D-valyl ara A were enhanced by >3-fold.
In Vivo Absorption Studies
One important aspect of our prodrug strategies is to improve the
absorption of the drug through the intestinal tract. In previous work,
we have shown that amino acid ester prodrugs of a variety of antiviral
and anticancer agents is improved over that of the parent compound, in
part because the transport of the prodrug is mediated by nutrient trans-
porters (e.g., the dipeptide transporter PepT1) that reside in the intestinal
membrane.[
24–27]
We examined the absorption of the 5 -O-D- valyl-ara A in
ꢁ
rats. As a test of intestinal uptake without the complication of gastric dilution
or degradation, the duodenal intestinal segment of an anesthetized rat
was exposed and the prodrug or vidarabine parent compound was directly
injected into the intestinal lumen. Blood was sampled from a jugular vein
cannula over a 4-hour period and analyzed for the prodrug, vidarabine, and
the metabolite ara-H. In the case of ara-A (vidarabine) dosing (Figure 2,
top panel), over 95% of the drug was converted to ara-H. However, for the
D-valyl ara A (Figure 2, bottom panel), significant levels of vidarabine were
detected. Approximately 13% of the total dose was present as vidarabine,
1
2% of the total dose was found to be ara-H, and 75% was the prodrug,
ꢁ
TABLE 3 AUC0-4hr values for vidarabine and 5 -O-D-valyl ara A (n = 3)
AUC0–4hr values (ng/mL·hr)
Compounds
Vidarabine
Ara-H
Prodrug
Vidarabine
D-Val-ara A
8.3
109 ± 13
181.5
98 ± 16
—
637 ± 74

ꢁ
5
-O-D-Valyl ara
53
FIGURE 2 Plasma profile of vidarabine (top) or D-valy-Ara A (bottom) after duodenal injection in
the rat. Animals were dosed via duodenal injection into anesthetized rats (n = 3 per compound)
ꢁ
with either Vidarabine (ara A) or 5 -O-D valyl ester of vidarabine prodrug. Plasma samples were
taken at 0, 0.25, 0.5, 1, 2, and 4 hours from the jugular vein and were assayed for prodrug (ꢀ),
vidarabine (ꢁ), and ara H (ꢂ) by LC/MS/MS. The dose for vidarabine (1.5 mg/rat) and D-valyl
vidarabine (2.05 mg/rat) were such that the same molar amount of vidarabine was delivered in both
groups.

5
4
W. Shen et al.
ꢁ
based on AUC0–4hr calculations (Table 3). The 5 -O-D-valyl ara A resulted in
1
3-fold increase in circulating vidarabine levels over dosing with Vidarabine.
CONCLUSION
ꢁ
Our results have shown that 5 -O-D-valyl ester prodrugs of vidarabine
maintain antiviral activity against the surrogate pox virus, vaccinia. Further,
using purified enzyme and intact cell systems, we show that the prodrugs are
resistant to inactivation by deamination. The prodrug also shows enhanced
transport potential. In our experiments, delivery of the prodrug to the
intestine results in a 20-fold increase in the vidarabine plasma levels when
compared with intestinal delivery of vidarabine.
REFERENCES
1
. De Clercq, E. Antiherpes drugs: Promises and pitfalls. Eur. J. Clin. Microbiol. Infect. Dis. 1984, 3(2),
6–107.
9
2. Whitley, R.J.; Soong, S.-J.; Dolin, R.; Galasso, G.J.; Ch’ien, L.T.; Afford, C.A. The Collaborative Study
Group: Adenine arabinoside therapy of biopsy-proved herpes simplex encephalitis. N. Eng. J. Med.
1977, 297, 289–294.
3
. Whitley, R.J.; Soong, S.-J.; Hirsch, M.S.; Karehmer, A.W.; Dolin, R.; Galasso, G.; Dunnick, J.K.; Alford,
C.A. The NIAID Collaborative Antiviral Study Group: Herpes simplex encephalitis. Vidarabine
therapy and diagnostic problems. N. Eng. J. Med. 1981, 304, 313–318.
4. Whitley, R.J.; Nahmias, A.J.; Soong, S.-J.; Galasso, G.G.; Fleming, C.L.; Alford, C.A. Vidarabine
therapy of neonatal herpes simplex virus infection. Pediatrics 1980, 66, 495–501.
5
. Whitley, R.J.; Tucker, B.C.; Kinkel, A.W.; Barton, N.H.; Pass, R.F.; Whelchel, J.D.; Cobbs, C.G.;
Diethelm, A.G.; Buchanan, R.A. Pharmacology, tolerance, and antiviral activity of vidarabine
monophosphate in humans. Antimicrob. Agents Chemother. 1980, 18(5), 709.
6
. a) Plunkett, W.; Cohen, S.S. Two approaches that increase the activity of analogs of adenine
nucleosides in animal cells. Cancer Res. 1975, 35, 1547–1554; (b) Plunkett, W. Cohen S.S. Increased
toxicity of 9-β-D-arabinofuranosyladenine in the presence of an inhibitor of adenosine deaminase.
Ann. NY Acad. Sci. 1977, 284, 91–102.
7
8
9
. Hilfinger, J.; Wu, Z.; Kim, J.-S.; Mitchell, S.; Breitenbach, J.; Amidon, G.; Drach, J. Vidarabine
prodrugs as anti-pox virus agents. Antiviral Res. 2006, 70, A14, 138.
. De Clercq, E. Antiviral drug discovery and development: Where chemistry meets with biomedicine.
Antiviral Res. 2005, 67(2), 56–75.
. Magee, W.C.; Aldern, K.A.; Hostetler, K.Y.; Evans, D.H. Cidofovir and (S)-9-[3-hydroxy-(2-
phosphonomethoxy)propyl]adenine are highly effective inhibitors of vaccinia virus DNA poly-
merase when incorporated into the template strand. Antimicrob. Agents Chemother. 2008, 52(2),
586–597.
10. Goff, A.; Twenhafel, N.; Garrison, A.; Mucker, E.; Lawler, J.; Paragas, J. In vivo imaging of cidofovir
treatment of cowpox virus infection. Virus Res. 2007, 128(1–2), 88–98.
11. Tempesta, M.; Camero, M.; Bellacicco, A.L.; Thiry, J.; Crescenzo, G.; Neyts, J.; Thiry, E.; Buonavoglia,
C. Cidofovir is effective against caprine herpesvirus 1 infection in goats. Antiviral Res. 2007, 74(2),
138–141.
1
1
1
2. Mercorelli, B.; Sinigalia, E.; Loregian, A.; Palu, G. Human cytomegalovirus DNA replication: antiviral
targets and drugs. Rev. Med.Virol. 2008, 18(3), 177–210.
3. McKenna, C.E.; Kashemirov, B.A.; Eriksson, U.; Amidon, G.L.; Kish, P.E.; Mitchell, S.; Kim, J.-S.;
Hilfinger, J.M. Cidofovir peptide conjugates as prodrugs. J. Organo. Chem. 2005, 690(10), 2673–2678.
4. Borthwick, A.D. Design of translactam HCMV protease inhibitors as potent antivirals. Med. Res. Rev.
2005, 25(4), 427–452.

ꢁ
5
-O-D-Valyl ara
55
1
5. Buller, R.M.; Owens, G.; Schriewer, J.; Melman, L.; Beadle, J.R.; Hostetler, K.Y. Efficacy of oral active
ether lipid analogs of cidofovir in a lethal mouse pox model. Virology, 2004, 318(2), 474–481.
6. Shen, W.; Hilfinger, J.; Kim, J.-S.; Laviolette, S.; Zhang, J.; Mitchell, S.; Kish, P.; Kijek, P.; Breitenbach,
J.M.; Gentry, B.G.; Drach, J.C. Design and synthesis of 1-Beta-D-arabinofuranosyladenine prodrugs
as antiviral agents. Abstracts of Papers, 236th ACS National Meeting, Philadelphia, PA; August 17–21,
1
2008.
1
7. a). Hassner, A.; Strand, G.; Rubinstein, M.; Patchornik, A. Levulinic esters, Alcohol protecting group
applicable to some nucleosides. J. Am. Chem. Soc. 1975, 97(6), 1614; b). van Boom, J.H.; Burgers,
P.M.J. Use of levulinic acid in the protection of oligonucleotides via the modified phosphotriester
method: Synthesis of decaribonucleotide U-A-U-A-U-A-U-A-U-A. Tetrahedron Lett. 1976, 4875; c) Jeker,
N.; Tamm, C. Verrucarins and roridins. 46. Synthesis of new, unnatural macrocyclic trichothecenes:
ꢁꢁ
3
-isoverrucarin A ((1 -O)(3→ 4) abeo-verrucarin A), verrucinol, and verrucene. Helv. Chim. Acta
1988, 1895; d). Jeker, N.; Tamm, C. Verrucarins and roridins. 47. Synthesis of new unnatural
macrocyclic trichothecenes: 4-epiverrucarin A. Helv. Chim. Acta 1988, 1904.
1
8. a) Rej, R.N.; Glushka, J.N.; Chew, W.; Perlin, A.S. Chromic acid oxidation in the synthesis of uronic
acids. Use of the O-levulinoyl group to minimize acyl migration. Carbohydr. Res. 1989, 135–48; b).
Glushka, J.N.; Perlin, A.S. Formation of disaccharides related to heparin and heparan sulfate by
chemical modification of maltose. Carbohydr. Res. 1990, 305–321.
1
2
2
2
2
2
2
9. Baker, D.C.; Haskell, T.H.; Putt, S.R.; Sloan, B.J. Prodrugs of 9-(.beta.-D-arabinofuranosyl)adenine.
ꢁ
ꢁ
ꢁ
ꢁ
2. Synthesis and evaluation of a number of 2 ,3 - and 3 ,5 -di-O-acyl derivatives. J. Med. Chem. 1979,
22(3), 273–279.
0. Singhal, D., and Anderson, B.D. Optimization of the local inhibition of intestinal adenosine
deaminase (ADA) by erythro-9-(2-hydroxy-3-nonyl)adenine: enhanced oral delivery of an ADA-
activated prodrug for anti-HIV therapy. J. Pharm. Sci. 1998, 87(5), 578–585.
1. Singhal, D.; Ho, N.F.; and Anderson, B.D. Absorption and intestinal metabolism of purine
ꢁ
ꢁ
dideoxynucleosides and an adenosine deaminase-activated prodrug of 2 ,3 -dideoxyinosine in the
mesenteric vein cannulated rat ileum. J. Pharm. Sci. 1998, 87(5), 569–577.
2. Lipper, R.A.; Machkovech, S.M.; Drach, J.C.; and Higuchi, W.I. Inhibition of drug metabolism by a
ꢁ
prodrug: 9-beta-D-arabinofuranosyladenine 5 -valerate as an inhibitor of adenosine deaminase. Mol.
Pharmacol. 1978, 14(2), 366–369.
3. Han, H.; de Vrueh, R.L.; Rhie, J.K.; Covitz, K.M.; Smith, P.L.; Lee, C.P.; Oh, D.M.; Sadee, W., and
ꢁ
Amidon, G.L. 5 -Amino acid esters of antiviral nucleosides, acyclovir, and AZT are absorbed by the
intestinal PEPT1 peptide transporter. Pharm. Res. 1998, 15(8), 1154–1159.
4. Song, X.; Lorenzi, P.L.; Landowski, C.P.; Vig, B.S.; Hilfinger, J.M., and Amidon, G.L. Amino acid
ester prodrugs of the anticancer agent gemcitabine: synthesis, bioconversion, metabolic bioevasion,
and hPEPT1-mediated transport. Mol. Pharm. 2005, 2(2), 157–167.
5. Vig, B.S.; Lorenzi, P.J.; Mittal, S.; Landowski, C.P.; Shin, H.C.; Mosberg, H.I.; Hilfinger, J.M.,
and Amidon, G.L. Amino acid ester prodrugs of floxuridine: Synthesis and effects of structure,
stereochemistry, and site of esterification on the rate of hydrolysis. Pharm. Res. 2003, 20(9),
1381–1388.
2
6. Song, X.; Vig, B.S.; Lorenzi, P.L.; Drach, J.C.; Townsend, L.B.; and Amidon, G.L. Amino acid
ester prodrugs of the antiviral agent 2-bromo-5,6-dichloro-1-(beta-D-ribofuranosyl)benzimidazole as
potential substrates of hPEPT1 transporter. J. Med. Chem. 2005, 48(4), 1274–1277.
7. Landowski, C.P.; Song, X.; Lorenzi, P.L.; Hilfinger, J.M.; and Amidon, G.L. Floxuridine amino acid
ester prodrugs: enhancing Caco-2 permeability and resistance to glycosidic bond metabolism. Pharm.
Res. 2005, 22(9), 1510–1518.
2
28. Vogel, A.I.; Tatchell, A.R.; Furnis, B.S.; Hannaford, A.J.; and Smith, P.W.G. Vogel’s textbook of
practical organic chemistry, 5th ed., 1989.
29. Still, W.C.; Kahn, M.; and Mitra, A. Rapid chromatographic technique for preparative separations
with moderate resolution. J. Org. Chem. 1978, 43(14), 2923.