Synthesis, chemical and enzymatic hydrolysis, and aqueous solubility of amino acid ester prodrugs of 3-carboranyl thymidine analogs for boron neutron capture therapy of brain tumors

Article abstract of DOI:10.1016/j.ejmech.2012.07.033

Various water-soluble l-valine-, l-glutamate-, and glycine ester prodrugs of two 3-Carboranyl Thymidine Analogs (3-CTAs), designated N5 and N5-2OH, were synthesized for Boron Neutron Capture Therapy (BNCT) of brain tumors since the water solubilities of the parental compounds proved to be insufficient in preclinical studies. The amino acid ester prodrugs were prepared and stored as hydrochloride salts. The water solubilities of these amino acid ester prodrugs, evaluated in phosphate buffered saline (PBS) at pH 5, pH 6 and pH 7.4, improved 48-6600 times compared with parental N5 and N5-2OH. The stability of the amino acid ester prodrugs was evaluated in PBS at pH 7.4, Bovine serum, and Bovine cerebrospinal fluid (CSF). The rate of the hydrolysis in all three incubation media depended primarily on the amino acid promoiety and, to a lesser extend, on the site of esterification at the deoxyribose portion of the 3-CTAs. In general, 3′-amino acid ester prodrugs were less sensitive to chemical and enzymatic hydrolysis than 5′-amino acid ester prodrugs and the stabilities of the latter decreased in the following order: 5′-valine > 5′-glutamate > 5′-glycine. The rate of the hydrolysis of the 5′-amino acid ester prodrugs in Bovine CSF was overall higher than in PBS and somewhat lower than in Bovine serum. Overall, 5′-glutamate ester prodrug of N5 and the 5′-glycine ester prodrugs of N5 and N5-2OH appeared to be the most promising candidates for preclinical BNCT studies.

Full text of DOI:10.1016/j.ejmech.2012.07.033

European Journal of Medicinal Chemistry 55 (2012) 325e334
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis, chemical and enzymatic hydrolysis, and aqueous solubility of amino
acid ester prodrugs of 3-carboranyl thymidine analogs for boron neutron capture
therapy of brain tumors
,
b
,
b,
Sherifa Hasabelnaby a , Ayman Goudah c, Hitesh K. Agarwal b, Mosaad S.M. Abd alla a,
*
,
Werner Tjarks b
**
a Division of Pharmaceutical Organic Chemistry, College of Pharmacy, Helwan University, Ain Helwan, Cairo, Egypt
b Division of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, The Ohio State University, 500 West 12th Avenue, Columbus, OH 43210, United States
c Division of Pharmacology, College of Veterinary Medicine, Cairo University, Giza, Egypt
a r t i c l e i n f o
a b s t r a c t
Article history:
Various water-soluble L-valine-, L-glutamate-, and glycine ester prodrugs of two 3-Carboranyl Thymidine
Received 27 April 2012
Received in revised form
5 July 2012
Accepted 18 July 2012
Available online 27 July 2012
Analogs (3-CTAs), designated N5 and N5-2OH, were synthesized for Boron Neutron Capture Therapy
(BNCT) of brain tumors since the water solubilities of the parental compounds proved to be insufficient in
preclinical studies. The amino acid ester prodrugs were prepared and stored as hydrochloride salts. The
water solubilities of these amino acid ester prodrugs, evaluated in phosphate buffered saline (PBS) at pH
5, pH 6 and pH 7.4, improved 48e6600 times compared with parental N5 and N5-2OH. The stability of
the amino acid ester prodrugs was evaluated in PBS at pH 7.4, Bovine serum, and Bovine cerebrospinal
fluid (CSF). The rate of the hydrolysis in all three incubation media depended primarily on the amino acid
promoiety and, to a lesser extend, on the site of esterification at the deoxyribose portion of the 3-CTAs. In
general, 30-amino acid ester prodrugs were less sensitive to chemical and enzymatic hydrolysis than 50-
amino acid ester prodrugs and the stabilities of the latter decreased in the following order: 50-valine > 50-
glutamate > 50-glycine. The rate of the hydrolysis of the 50-amino acid ester prodrugs in Bovine CSF was
overall higher than in PBS and somewhat lower than in Bovine serum. Overall, 50-glutamate ester pro-
drug of N5 and the 50-glycine ester prodrugs of N5 and N5-2OH appeared to be the most promising
candidates for preclinical BNCT studies.
Keywords:
3-Carboranyl thymidine analogs
Amino acid ester prodrugs
Boron neutron capture therapy
Glioblastoma multiforme
Ó 2012 Elsevier Masson SAS. All rights reserved.
1. Introduction
Boron Neutron Capture Therapy (BNCT) is a binary cancer
treatment modality that is based on the irradiation of boron-10
(
10B), a stable isotope, with low energy (thermal) neutrons. The
Abbreviations: 10B, boron-10; 3-CTAs, 3-carboranyl thymidine analogs; AUCs,
areas under the curves; BNCT, boron neutron capture therapy; Boc, butoxycarbonyl;
BSA, Bovine serum albumin; BuLi, butylithium; CSF, cerebrospinal fluid; DCC,
resulting capture reaction produces cytocidal high linear energy
transfer (LET) lithium- and helium nuclei ( -particles). These have
ranges of <10 m in biological tissue, which approximates one cell
a
dicyclohexylcarboiimide;
DCM,
dichloromethane;
DMAP,
4-
m
dimethylaminopyridine; DMF, dimethyl formamide; DMSO, dimethyl sulfoxide;
GBM, glioblastoma multiforme; HR-ESI, high resolution-electrospray ionization;
hTK1, human thymidine kinase 1; ISF, interstitial fluid; i.c., intracranial; i.t., intra-
tumoral; i.v., intravenous; KMT, kinase mediated trapping; LET, linear energy
transfer; OSU-CCIC, The Ohio State University Campus Chemical Instrumentation
Center; PBS, phosphate buffered saline; RP-18, reversed phase 18; rt, room
temperature; TBAF, tetrabutylammonium fluoride; TBDMS, tert.-butyldimethylsilyl;
TFA, trifluoroacetic acid; Thd, thymidine; THF, tetrahydrofurane; TK1, thymidine
kinase 1.
diameter. Thus, lethal damage caused by BNCT can be restricted
largely to cancerous tissue by selective delivery of 10B to cancer
cells. Approximately 20
m
g of 10B/g tumor and tumor-to-normal
tissue ratios larger than 3:1 at the time of neutron irradiation are
required for successful BNCT [1].
In recent years, 3-Carboranyl Thymidine Analogs (3-CTAs), such
as N5 and N5-2OH, (compounds 1 & 2, Fig. 1), have been a major
focus of compound development for BNCT of high-grade brain
tumors, such as glioblastoma multiforme (GBM) [2e5]. Human
thymidine kinase 1 (hTK1), an enzyme of the DNA salvage pathway
synthesis that is predominantly active in proliferating cells [6],
* Corresponding author. Division of Pharmaceutical Organic Chemistry, College of
Pharmacy, Helwan University, Ain Helwan, Cairo, Egypt.
** Corresponding author. Tel./fax: þ1 614 292 7624.
E-mail addresses: sherifajanaa@yahoo.com (S. Hasabelnaby), tjarks.1@osu.edu
(W. Tjarks).
0223-5234/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2012.07.033

326
S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
concentration in the interstitial fluid (ISF) of the brain appears to be
significantly lower than that in normal serum [10]. The increased
molecular weight and hydrophilicity of prodrugs of 1 and 2 may
prevent their effective passage through cell membranes and the
subsequent intracellular enzymatic removal of the promoiety.
Therefore, cleavage of such prodrugs following i.c. administration
may depend primarily on chemical rather than enzymatic hydro-
lysis in brain ISF.
Consequently, the ideal water-solubilizing promoieties for 1 and
2 should be rapidly cleavable by chemical hydrolysis. Phosphate-,
carbohydrate-, or sulfate esters are probably not suitable as pro-
moieties of 1 and 2 because they proved to be sensitive to enzy-
matic cleavage but were fairly resistant toward chemical hydrolysis
when used for other types of agents [11e13]. In contrast, amino acid
ester prodrugs of established drugs, such as the antibiotic/anti-
protozoal metronidazole [14], the anticancer nucleoside analogues
gemcitabine, floxuridine and cytarabine [15e18], and the widely
used antiviral nucleoside analog acyclovir [19,20] (compound 3,
Fig. 1) proved to be sensitive to both enzymatic and chemical
hydrolysis. In addition, valacyclovir (compound 4, Fig. 1), the
established valine ester prodrug of acyclovir, is about 130 times
better water-soluble than its parental nucleoside analog [21].
This paper describes the design, synthesis, and evaluation of
water-soluble amino acid ester prodrugs of 1 and 2. Based on the
aforementioned studies with metronidazole, gemcitabine, floxur-
Fig. 1. Structures of N5 (1), N5-2OH (2), acyclovir (3), and valacyclovir (4).
catalyzes the 50-monophosphorylation of these agents. Apparently,
this causes the selective accumulation of 3-CTAs in cancer cells by
a process that has been referred to as Kinase Mediated Trapping
(KMT) [5,7]. The in vitro uptake of 1 and 2 in TK1 (þ) L929 cells was
5e10 higher than in TK1 (ꢀ) L929 cells following 24 h incubation in
compound containing medium. In addition, both compounds were
retained only in the former cell line when incubation was
continued for 12 h in compound free medium [2]. After intra-
tumoral (i.t.) injection of 2 into subcutaneous TK1 (þ)- or TK1 (ꢀ)
L929 tumors implanted in nude mice, boron concentrations were
w3 times higher in the former tumor [3]. Survival times of rats
bearing intracranial RG2 glioma that had received 2 via intracranial
(i.c.) injection prior to BNCT were about 2 times longer compared
with rats that received BNCT without administration of 2 [3].
Compared with i.c. and i.t. injections, intravenous (i.v.) adminis-
tration of 2 appeared to be less effective [3].
These biological studies clearly established the therapeutic
potential of 1 and 2 but they also revealed a major limitation. Both
compounds are very lipophilic because of the presence of the car-
borane cluster [4] and the absence of any functional groups that can
be ionized under physiological conditions. Thus, both agents are
almost insoluble in water and aqueous solutions containing 30%e
70% dimethyl sulfoxide (DMSO) had to be used to solubilize 2 for
i.t. and i.c. injections into tumor-bearing rodents [3]. Although no
significant pathologic alterations were observed in rat brains
following i.c. injections [3], administration of 2 in DMSO via i.c.
route to the brain of human patients with intracranial tumors has
the potential to cause damage to healthy brain tissue. Even if such
injections should prove to be acceptable for humans, it may
significantly complicate approval of clinical trials by regulatory
agencies.
idine, cytarabine, and acyclovir, glycine-, L-valine-, and L-glutamate
were selected as promoieties. The chemical and enzymatic stabil-
ities of the synthesized amino acid ester prodrugs were evaluated
in phosphate buffered saline (PBS), Bovine serum, and Bovine
cerebrospinal fluid (CSF), which has a similar protein concentration
as brain ISF [10]. The solubilities of the synthesized prodrugs were
evaluated in PBS at pH 5, pH 6, and pH 7.4.
2. Results and discussion
2.1. Chemistry
The syntheses of N5 (1) and acetonide-protected N5-2OH (10)
are shown in Schemes 1 and 2. Both compounds served as starting
materials in the synthesis of the amino acid ester prodrugs of 1 and
N5-2OH (2), as shown in Scheme 3.
Previously reported syntheses for 1 utilized highly toxic and
explosive decaborane as the starting material [22,23]. Therefore, an
alternative synthesis was developed, which is shown in Scheme 1.
In this route, tert.-butyldimethylsilyl (TBDMS)-o-carborane (5) is
A successful approach to overcome water-insolubility of drugs
has been the preparation of carrier-linked prodrugs [8]. For the
design and synthesis of water-soluble amino acid ester prodrugs of
1 and 2, three important factors have to be considered: 1) The water
solubilities of the amino acid ester prodrugs of 1 and 2 should be in
the range of those of the parental compounds in aqueous DMSO
solutions. 2) The tumor selective uptake of 1 and 2 depends on TK1
depended KMT. Thus, cleavage of any promoiety attached to 1 and 2
should be rapid, as additional structural modification of these
structures, not only at the 50-position, will most likely abolish their
TK1 substrate characteristics [9]. 3) The protein (enzyme)
Scheme 1. Reagents and conditions: (a) n-BuLi, THF, 1,5-pentandiol ditosylate, 5 h,
80 ꢁC; (b) Thd, K2CO3, DMF/acetone (1/1), 6 h, 50 ꢁC; (c) TBAF, THF, 0.5 h, rt.

S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
327
ditosylate, as described previously by Byun et al. [25], was feasible
yielding 9 in 41% yield. The reaction of 9 with Thd in the presence of
potassium carbonatein a DMF/acetone mixture at 50 ꢁC produced 10
in 48% yield.
Compound 1 was treated with commercially available N-Boc
protected amino acids (N-Boc-
glutamic acid -tert.-butyl ester). As reported by Anand et al. [19]
and Katragadda et al. [20] for studies with acyclovir, glutamic
acid was attached to the 50-position of 1 via the
-carboxylic
L-valine, N-Boc-glycine, or N-Boc-L-
a
g
function. For reasons discussed in section 2.2, compound 10 was
only treated with N-Boc-glycine. All reactions were carried out in
the presence of dicyclohexylcarbodiimide (DCC) and a catalytic
amount of 4-dimethylaminopyridine (DMAP) in DMF for 24 h at
room temperature to give mixtures of 30-N-Boc-, 50-N-Boc-, and
30,50-di-N-Boc-amino acid ester prodrugs of 1 and 10 (Scheme 3).
Progress of the reactions was monitored by TLC and separation of
the Boc-protected amino acid ester prodrugs was possible by
column chromatography. Unfortunately, all of these isolated
intermediates remained contaminated with DCC (see
supplementary material for 1H NMR data and Rf values). Thus,
removal of the Boc groups from these partially purified interme-
diates was achieved with a 50% solution of trifluoroacetic acid (TFA)
in dichloromethane at room temperature for 2 h to give the 30-
monosubstituted amino acid ester prodrugs 11, 14, 17, and 20
(Scheme 3) in yields ranging from 1% to 5%, the 50-monosubstituted
amino acid ester prodrugs 12, 15, 18, and 21 (Scheme 3) in yields
ranging from 7% to 22%, and the 30,50-disubstituted amino acid ester
prodrugs 13, 16, 19, and 22 (Scheme 3) in yields ranging from 13% to
30%.
Scheme 2. Reagents and conditions: (a) n-BuLi, THF, 1,5-diiodopentane, 6 h, 80 ꢁC;
(b) Thd, K2CO3, DMF/acetone (1/1), overnight, 50 ꢁC.
used as the staring material, which is prepared from less toxic,
stable, and commercially-available o-carborane [24]. The reaction
of 5 with n-BuLi in THF followed by addition of 1,5-pentandiol
ditosylate [25] afforded compound 6 in 36% yield. Use of readily
available commercial 1,5-diiodopentane rather than 1,5-
pentanediol ditosylate resulted in complex reaction mixtures
from which the desired product could not be separated effectively.
Reacting 6 with thymidine (Thd) in the presence of potassium
carbonate in a DMF/acetone (1/1) mixture at 50 ꢁC gave compound
7 in 65% yield. Removal of the TBDMS protective group of
compound 7 was achieved with tetrabutylammonium fluoride
(TBAF) in THF at room temperature for 0.5 h to give target
compound 1 in 58% yield.
In the case of the glutamate ester prodrugs of N5 (17e19), the
reaction was carried for 4 h to assure removal of both the Boc- and
the
a-tert.-butyl protective group. All amino acid ester prodrugs
Compound 10 was prepared according to Scheme 2. In this
reaction sequence, treatment of compound 8 [25] with n-BuLi fol-
lowed by addition of 1,5-diiodopentane rather than 1,5-pentanediol
were obtained with wax-like consistency following purification by
preparative RP18-HPLC (Table 3) and were analyzed and identified
by 1H NMR, 13C NMR, and MS. The latter method was most useful to
Scheme 3. Reagents and conditions: (a) N-Boc-L-amino acids, DCC, DMAP, 24 h, rt; (b) Partial column chromatographic purification (ethyl acetate/hexanes, 8/5) of 30-N-Boc-, 50-N-
Boc-, and 30,50-di-N-Boc-amino acid esters of 1 and 10; (c) TFA:DCM, 1:1, 2 h e 4 h, rt; (d) 1 M HCl in diethyl ether, THF, overnight, rt.

328
S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
Table 1
Table 3
Prodrug half-lives (t1/2) in PBS (pH 7.4), Bovine serum, and Bovine CSF.
HPLC methods and retention times of compounds 1e4 and 11e22.
Compds
t1/2 (h)
Compds
Methoda
Retention time (min)
PBS
Bovine serum
Bovine CSF
30
50
30, 50
4
13.60 ꢃ 0.49
29.79 ꢃ 0.42
25.59 ꢃ 0.61
12.91 ꢃ 0.17
1.51 ꢃ 0.11
0.70 ꢃ 0.04
0.22 ꢃ 0.03
26.91 ꢃ 0.71
11.16 ꢃ 0.37
3.42 ꢃ 0.10
2.21 ꢃ 0.15
6.20 ꢃ 0.17
ND
7.59 ꢃ 0.12
ND
11e13
14e16
17e19
20e22
11e13
14e16
17e19
20e22
A1 ꢀ65:35 to 30:70 (H2O:MeCN)
over 27 min
18.82
18.09
13.91
11
12
13
14
15
16
17
18
19
21
22.21 ꢃ 0.52
25.47 ꢃ 0.10
A2 ꢀ70:30 to 30:70 (H2O:MeCN)
18.43
13.99
14.00
33.03
30.61
32.05
28.30
17.57
13.69
13.51
31.29
30.40
31.64
28.18
14.67
8.92
ND
ND
over 27 min
ND
ND
A3 ꢀ60:40 to 30:70 (H2O:MeCN)
over 27 min
0.69 ꢃ 0.02
0.96 ꢃ 0.04
ND
ND
A2 ꢀ70:30 to 30:70 (H2O:MeCN)
10.55
27.63
26.45
26.50
23.52
ND
ND
over 27 min
0.85 ꢃ 0.12
1.85 ꢃ 0.11
B1 ꢀ65:35 to 0:100 (H2O:MeOH)
over 40 min
ND
ND
1.51 ꢃ 0.04
2.11 ꢃ 0.14
B2 ꢀ70:30 to 0:100 (H2O:MeOH)
over 40 min
Standard deviations are based on three experiments; ND ¼ not determined.
B3 ꢀ60:40 to 0:100 (H2O:MeOH)
over 40 min
B2 ꢀ70:30 to 0:100 (H2O:MeOH)
distinguish between monosubstituted- and disubstituted amino
acid ester prodrugs, whereas 1H NMR spectroscopy was helpful to
distinguish between 30-monosubstituted- and 50-monosubstituted
over 40 min
Retention time (min)
25.45
1
2
3
4
A1 ꢀ65:35 to 30:70 (H2O:MeCN)
over 27 min
amino acid ester prodrugs based on the chemical shifts (d
) of the 30-
A2 ꢀ70:30 to 30:70 (H2O:MeCN)
18.72
14.35
19.43
and 50-protons. In general, amino acid ester substitution at the 30-
hydroxyl group caused a downfield shift for the signal of the 30-Hs
from 3.88 ppm and 3.95 ppm in the case of 1 and 10, respectively, to
5.04 ppm-5.54 ppm in the case of 11, 14, 17, and 20 whereas 50-
subsititution caused a downfield shift for the signal of the 50-Hs
from 3.74 ppm and 3.81 ppm in the case of 1 and 10, respectively, to
4.30 ppm - 4.74 ppm in the case of 12, 15, 18, and 21.
over 27 min
A4 - 100:0 to 85:15 (H2O:MeCN)
over 20 min
A4 - 100:0 to 85:15 (H2O:MeCN)
over 20 min
a
Solvent systems contained 0.1% TFA.
The low overall yields starting from 1 or 10 are a shortcoming of
the applied synthetic route. However, low yields are inherent to
this specific synthetic route and have also been reported previously
in the synthesis of various gemcitabine amino acid ester prodrugs
[15]. On the other hand, introduction of ester functions at different
hydroxyl groups is desirable to explore which ester site is most
susceptible to chemical or enzymatic cleavage (see section 2.2).
Once the optimal esterification site has been identified, appropriate
protective group strategies could be applied to improve yields.
Treatment of 11e22 (Scheme 3) with commercially available
anhydrous 1 M HCl gas in diethyl ether followed by stirring over-
night at room temperature gave the corresponding hydrochloride
salts 11ꢂHCle22ꢂHCl (Scheme 3) in yields ranging from 50% to 70%.
The formation of 11ꢂHCle22ꢂHCl was confirmed by elemental
analysis except for the 30-glycine ester prodrug of N5 (14) since the
low synthetic yield for this compound limited its availability.
These hydrochloride salts were synthesized for two reasons. 1)
They had crystalline consistency and were much easier to handle
than the wax-like free bases. 2) Various amino acid ester prodrugs
of acyclovir (3), AZT, and gemcitabine exhibited significant stability
in aqueous solution in the pH-range of 3e5 for several days
whereas fast chemical cleavage was observed in phosphate buffer
at pH 7.4 [15,19,26,27], which indicated preferential basic hydro-
lysis of the ester bond. Thus, the free bases are expected to be rather
sensitive toward traces of water during storage and handling since
the pH environment should be fairly neutral. In contrast, the
hydrochloride salts should generate a more acidic environment in
the presence of water traces and remain more stable.
2.2. Chemical and enzymatic stability studies
The stabilities of the synthesized amino acid ester prodrugs of
N5 (11e19) and the 50-glycine prodrug of N5-2OH (21) toward
hydrolysis were evaluated in three different media: 1) PBS to
examine their chemical hydrolysis, 2) Bovine serum to evaluate
their enzymatic hydrolysis, and 3) Bovine CSF. The protein
composition of the latter medium is similar to that of the ISF of the
brain [10,28]. Thus, it seems likely that the hydrolysis pattern of
amino acid ester prodrugs under conditions similar to those they
may encounter following i.c. injection can be established in CSF.
Valacyclovir (4) was evaluated as a reference compound in all three
media.
Table 2
pKa values and compound solubilities in PBS at pH 5, pH 6, and pH 7.4 at 37 ꢁC.
Compds pKa Valuesa PBS pH 5
PBS pH 6
0.003e0.005 g/L 0.003e0.005 g/L 0.003e0.005 g/L
0.007e0.01 g/L 0.007e0.01 g/L 0.007e0.01 g/L
72.70 ꢃ 0.40 g/L 13.43 ꢃ 0.26 g/L 1.91 ꢃ 0.06 g/L
PBS pH 7.4
2.2.1. Chemical (non-enzymatic) hydrolysis in PBS at pH 7.4
The hydrolytic half-lives (t1/2) of the prodrugs, measured by
using the linear regression of pseudo-first-order plots of the
remaining prodrug concentrations versus the hydrolysis times, are
shown in Table 1. The obtained results clearly indicate that the
hydrolysis rates of the amino acid ester prodrugs primarily depend
on the amino acid promoieties. The glycine ester prodrugs of N5
(14e16) and the 50-glycine ester prodrug of N5-2OH (21) are 20e25
times more sensitive to chemical hydrolysis than the corresponding
glutamate- (17e19) and valine ester prodrugs (11e13) of N5. The
sensitivity of 14e16 and 21 may be due to the presence of the small
glycine promoiety, which is sterically not interfering with the
nucleophilic hydrolytic attack at the ester bond.
1
2
4
e
e
1.90, 7.47
& 9.43b
w7.85
w7.30
w2.20
& w9.25
w7.30
12
15
18
26.30 ꢃ 0.90 g/L 3.70 ꢃ 0.31 g/L
10.99 ꢃ 0.20 g/L 2.06 ꢃ 0.11 g/L
1.31 ꢃ 0.05 g/L
0.19 ꢃ 0.01 g/L
1.77 ꢃ 0.08 g/L
1.16 ꢃ 0.04 g/L
1.13 ꢃ 0.06 g/L
21
11.18 ꢃ 0.21 g/L 3.60 ꢃ 0.12 g/L
0.62 ꢃ 0.03 g/L
Standard deviations are based on three experiments.
a
The pKa values for 12, 15, 18, and 21 are estimated based on calculations using
Advanced Chemistry Development (ACD/Labs) software within ScifinderScholarÒ
for the corresponding benzyl esters.
b
Measured values from reference [36].

S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
329
Although 4,11, and 12 are all monovaline esters, the half-Life of 4
is w2 times shorter than those of 11 and 12. This difference may be
due to the presence of the lipophilic carborane cage in 11 and 12,
which reduces their solubility, and hence, limits their accessibility
to chemical hydrolysis. Song et al. [15] reported a half-life of 17.15 h
for the chemical hydrolysis of 4, which is somewhat longer than the
value we found in our studies (13.6 h), but still comparable.
Compounds 17e19 appear to have intermediate stability toward
chemical hydrolysis is already so fast that the contribution of
enzyme activity becomes negligible.
The rate of the hydrolysis of the amino acid ester prodrugs in
serum was also influenced by the amino acid promoiety. In
particular 18 appears to depend significantly stronger on enzymatic
(protein-aided) cleavage than the other prodrugs. The enzymatic
cleavage of 4 is 2.6 times more rapid than that of 12. Song et al. [15]
reported a half-life of 5.2 h for the hydrolysis of 4 in human plasma,
which is comparable with the half-life of 6.2 h found in our studies
with Bovine serum. The difference between the half-lives of 12 in
PBS and Bovine serum (25.59 ꢃ 0.61 vs. 22.21 ꢃ 0.52) is relatively
small indicating that it is in contrast to 4 not a good substrate for
any degrading enzymes.
chemical hydrolysis. In previous studies [19], the g-glutamate ester
prodrug of acyclovir exhibited significant chemical stability with
a half-life of 28.2 h in PBS at pH 7.4, which is more than 2 times
longer than the half-life we found for 18 in our studies.
The site of esterification also influences the rate of the hydrolysis
of the amino acid ester prodrugs. Generally, 30-monosubstituted
amino acid ester prodrugs were more stable toward chemical
hydrolysis than the corresponding 50-monosubstituted amino acid
ester prodrugs, which is consistent with reported data for floxur-
idine amino acid ester prodrugs [29], but in contrast to those found
for amino acid ester prodrugs of gemcitabine [15]. The preference
for the cleavage of the 50-promoieties may be due to steric effects
since the 30-position is generally more crowded by the ribofuranose
ring thereby interfering with the rate of the hydrolysis.
2.2.2.2. Stability studies in CSF. Thehydrolysispatternsof4,12,15,18,
and 21 were also studied in Bovine CSF at 37 ꢁC for various time
intervals and the data are presented inTable 1. It appears that the rate
of the hydrolysis of the amino acid ester prodrugs in Bovine CSF is
overall faster than in PBS and somewhat slower than in Bovine serum.
The protein content in the Bovine CSF (Biochemed Services,
Winchester, Virginia) was determined by the Bradford protein assay
(seesection4.4.)andwasfoundtobe150timeslowerthantheprotein
content in Bovine serum, which is somewhat higher than previously
reported (260 times lower) [10]. Although a low protein (enzyme)
concentration is present in the CSF, enzymatic degradation may be
possible causing an increased rate of the hydrolysis in the CSF
comparedwithPBS.Thisisespeciallythecaseforthehydrolysisrateof
18, which is 6 times faster than in PBS. The dramatic increase in the
rate of the hydrolysis of 18 in the CSF may be due to the presence of
glutamate-specific enzymes in the CSF, as glutamate is the most
common neurotransmitters in the brain [30].
In general, the 30,50-disubstituted amino acid ester prodrugs of
N5 showed the shortest half-lives (Table 1). This result is consistent
with previously reported data for the amino acid ester prodrugs of
floxuridine [29] and gemcitabine [15]. This finding seems plausible
as their half-lives not only depend on the hydrolysis to the parental
compounds, but also on the conversion into the corresponding 30-
and 50-monosubstituted amino acid ester prodrugs. On the other
hand, the presence of two promoieties in theses structures
decreases, of course, the rate of production of parental compounds
by a factor of w2 compared with 50-monosubstituted amino acid
ester prodrugs (data not shown). In the case of 17e19, the half-life
of the 30,50-diglutamate ester prodrug of N5 (19) is significantly
shorter than those of the corresponding 30- (17) and 50-glutamate
(18) esters. This may be due to the observed high solubility of 19 in
PBS, which may increase access of the molecule to hydrolyzing
conditions.
Compound 12, as a representative of the very stable prodrugs,
18, as a representative of the medium stable prodrugs, and 15, as an
example of the more sensitive prodrugs, were chosen for further
studies in Bovine serum and Bovine CSF. Low synthetic yields of the
30-monosubstituted amino acid ester prodrugs (w2 times lower
than the corresponding 50-monosubstituted amino acid ester pro-
drugs) and the special characteristics of the 30,50-disubstituted
amino acid ester prodrugs were major reasons for their exclusion
from further stability assays. The same observations and the fact
that 14e16 were generally most sensitive to chemical hydrolysis
provided the rationale for limiting preparative efforts to the
synthesis of the 30-mono- (20), 50-mono- (21), and 30,50-diglycine
(22) ester prodrugs of N5-2OH (section 2.1.) and evaluating exclu-
sively the stability of 21 in all three test media.
2.2.3. Solubility studies
The solubilities of 1, 2, and 4 and the hydrochloride salts of 12,
15, 18, and 21 were determined in PBS at pH 5, pH 6 and pH 7.4 at
37 ꢁC (Table 2). The pHs of these PBS solutions were controlled
before and after the addition of the amino acid ester prodrugs with
pH-indicator strips (ColorpHastÒ, Merck, Darmstadt, Germany) and
no changes were observed. Overall, the obtained data are in
agreement with the estimated pKa values for the amino acid ester
prodrugs (Table 2). Compound 12 is w 1e7 times more soluble than
compounds 15 and 21, as the former has a slightly higher estimated
pKa value. On the other hand, 4 is w1.5e3 times more soluble than
12. The estimated water solubilities for 4 were 84 g/L at pH 5, 12 g/L
at pH 6, and 1.4 g/L at pH 7.4 (Advanced Chemistry Development
[ACD/Labs] software within Scifinder ScholarÒ), which are almost
identical to those we measured in our study. On the other hand,
Beauchamp et al. [21] determined the solubility of 4 in aqueous
solution with 174 mg/mL. Although 12 has a higher estimated pKa
value than 4 (7.85 vs 7.47, Table 2), the latter showed better solu-
bility. The differences in the solubility between 4 and 12 may be
due to the presence of the lipophilic carborane cage [31e34] in the
latter compound.
2.2.2. Enzymatic stability studies
Compound 21 showed slightly better solubilities than 15, which
was expected due to the presence of two additional hydroxyl groups
in the former. The measured solubilities for 18 were moderate at all
three pH levels and increased only slightly from 1.13 g/L at pH 5 to
1.77 g/L at pH 7.4. This is probably due to the presence of the
additional carboxylic function with an estimated pKa of 2.20
(Table 2), and thus, zwitterion formation in the pH range of 5e7.4.
Zwitterions can form strong intramolecular ioneion interactions
that interfere with the formation of ionedipole interactions with
water molecules, thereby reducing water-solubility [35].
2.2.2.1. Stability studies in Bovine serum. The in vitro enzymatic
hydrolysis pattern of the amino acid ester prodrugs was studied in
Bovine serum at 37 ꢁC. The hydrolytic half-lives of the prodrugs are
shown in Table 1. The hydrolysis rates of 4, 12, 18, and 21 in Bovine
serum are 2.2-, 1.2-, 13.1- and 1.5 times, respectively, faster than the
corresponding rates in PBS. These findings are consistent with
previously reported results for other amino acid ester prodrugs
[14,15,19]. They clearly indicate that the hydrolysis of the amino
acid ester prodrugs is generally facilitated by enzymatic catalysis or
protein-aided hydrolysis. However, the hydrolysis pattern of 15 in
PBS did not change in serum. This may be due to the fact that
All evaluated prodrugs showed significantly increased solubil-
ities compared with their parental drugs. Depending on the pH

330
S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
level in the test media, 12 is 327e6600 times more soluble than 1,
18 is 282e443 times more soluble than 1, and 15 and 21 are
Electrospray Ionization (HR-ESI) mass spectra were recorded on
a Micromass LCT Electrospray mass spectrometer at The Ohio State
University Campus Chemical Instrumentation Center (OSU-CCIC).
For all carborane-containing compounds, the mass of the most
intensive peak of the isotopic pattern was reported for an 80%
boron-11e20% boron-10 distribution. Compound visualization on
Silica Gel 60 F254 precoated TLC plates (0.25 mm layer thickness)
(Merck, Darmstadt, Germany) was attained by UV light and KMnO4
spray. Carborane-containing compounds were selectively visual-
ized by spraying a solution of 0.06% PdCl2/1% aqueous HCl on TLC
plates and subsequent heating to w120 ꢁC. Reagent grade solvents
were used for column chromatography using Silica gel 60, particle
48e2748 and 78e1400 times more soluble than
1 and 2,
respectively.
In preclinical BNCT studies with rats bearing intracerebral glioma
[3], 200 L of a solution of 2.5 g of 2 in 1 L of a mixture of 35% DMSO and
m
65% of water was injected i.c. The solubility data shown in Table 2
demonstrate that similar concentrations can be achieved success-
fully with all tested prodrugs underaqueous conditions. Compound 12
showed the highest solubility, in particular at pH 5. Unfortunately, this
amino acid ester prodrug does not seem to be a suitable candidate for
further consideration because of its stability to chemical and enzy-
matic hydrolysis. Compounds 15 and 21 showed adequate solubilities
at pH 5 and pH 6, whereas 18 had its highest solubility at pH 7.4. This
may allow for the formulation of the latter amino acid ester prodrug at
physiological pH for i.c. administration because it is fairly stable to
chemical hydrolysis at pH 7.4 (see Table 1). In contrast, 15 and 21
require slightly acidic pH for appropriate solubilization. The possibility
of administration at physiological pH may be an advantage of 18
compared with 15 and 21 because it may reduce toxic side effects.
size 63e200
mm (Dynamic Adsorbents, Inc, Georgia). Standard
reagent grade chemicals were obtained from commercial vendors
and used as such. o-Carborane was purchased from Katchem, Ltd,
Prague, Czech Republic. Valacyclovir was purchased from Toronto
Research Chemicals, Inc. Canada. Solvents were dried prior to use
following standard procedures. All chemical reactions were carried
out under argon atmosphere. Bovine serum and phosphate buff-
ered saline (PBS) were obtained from SigmaeAldrich. Bovine
cerebrospinal fluid (CSF) was obtained from Biochemed Services,
Winchester (Virginia). Filters (Corning, Germany) used in the
3. Summary and conclusions
stability studies had 0.2 mm pore size. The colorimetric measure-
Novel improved strategies were developed for the synthesis of 1
and 10. All proposed amino acid ester prodrugs of N5 (11e19) and N5-
2OH (20e22) were successfully prepared. Notable are the inherently
low overall yields for the synthesized prodrugs, inparticular those for
the 30-monosubstituted amino acid ester prodrugs.
ment of protein concentrations in Bovine serum and Bovine CSF
was carried out by using a Spectramax Plus Spectrophotometer
from Molecular Devices (Sunnyvale, California). The obtained data
were analyzed using SoftMax Pro 3.1.2 software.
Preparative HPLC was performed on a Gemini 5
(21.20 mm ꢄ 250 mm, 5 m particle size) from Phenomenex Inc. CA,
USA using a Hitachi HPLC system ( -2130) with a Windows based data
acquisition and a Hitachi Diode array detector ( -2455). Purification
m C18 column
The rate of the hydrolysis of the amino acid ester prodrugs of 1
and 2 in the three incubation media depends primarily on the amino
acid promoiety, and to lesser extend, on the site of esterification at
the deoxyribose portion. Generally, 30-monosubstituted amino acid
ester prodrugs were more stable to chemical hydrolysis than 50-
monosubstituted amino acid ester prodrugs. The 50-valine ester
prodrug 12 is the most stable both to enzymatic and non-enzymatic
hydrolysis, the 50-glycine acid ester prodrugs 15 and 21 are the most
sensitive under both conditions, whereas the 50-glutamate ester
prodrug 18 shows intermediate stability to chemical hydrolysis and
high sensitivity to cleavage in serum and CSF. The high sensitivity of
15 and 21 and the stability of the 12 to hydrolysis are in agreement
with the results of previous stability studies with prodrugs of
metronidazole [14] and gemcitabine [15]. In all three incubation
media, the hydrolysis of valacyclovir (4) was faster than that of 12. In
contrast to 4, however,12 is notably less susceptible to enzymatic or
protein-aided cleavage, indicating that it is not a good substrate for
any degrading enzymes. Compounds 15 and 21 appeared to be least
affected by enzymatic or protein-aided cleavage, as their half-lives
did not vary significantly in all three media. Overall, the observed
solubilities for all tested amino acid ester prodrugs at pH 5, 6, and 7.4
are significantly better than those of 1 and 2 and appeared to be
suitable for i.c. injection. As discussed earlier, the fundamental
prerequisite for amino acid ester prodrugs of 1 and 2 must be rapid
cleavage to enable 50-monophosphorylation. This requirement is
fulfilled in the cases of 15 and 21. Also, 18 seems to be a suitable
candidate for preclinical BNCT studies because it showed rapid
cleavage in CSF combined with moderate stability in PBS at pH 7.4.
m
L
L
was accomplished using a water (0.1% TFA)/methanol gradient at
7 mL/min flow rate (100:0 to 50:50 over 15 min then 50:50 to 15:85
[H2O:MeOH] over 70 min followed by 15:85 to 0:100 over 15 min).
Analytical HPLC for purity confirmation and stability analysis
was carried out with a Gemini 5
m
C18 110A Column (250 ꢄ 4.6 mm)
from Phenomenex Inc. CA, USA using the HPLC system above. Two
different solvent systems were used at 1 mL/min flow rate:
A ¼ water (0.1% TFA)/acetonitrile (0.1% TFA); B ¼ water (0.1% TFA)/
methanol (0.1% TFA). For each solvent system, different gradients
were used for compound analysis (Table 3, A1-A4, B1-B3). For all
Thd derivatives, a wavelength of 265 nm was used for detection,
whereas valacyclovir/acyclovir were detected at 254 nm.
4.1.1. 5-{2-(tert.-Butyldimethylsilyl)-o-carboran-1-yl}pentyl 4-
methylbenzenesulfonate (6)
To a solution of 5 [24] (0.52 g, 2 mmol) in THF (50 mL) was added
n-BuLi (0.9 mL, 2.36 mmol, 2.5 M solution in hexanes) at ꢀ78 ꢁC
over a period of 10 min. The solution was gradually warmed to rt
and then stirred for 1 h. Subsequently, the reaction mixture was
cooled to 0 ꢁC and 1,5-pentandiol ditosylate (1.1 g, 2.6 mmol) was
added dropwise. The mixture was refluxed at 80 ꢁC for 5 h. Distilled
water (20 mL) was added and excess THF was removed under
reduced pressure. The residue was extracted with ethyl acetate
(30 mL ꢄ 2). The combined organic layers were washed with
diluted HCl solution (2%, 30 mL) and brine (30 mL), and dried over
anhydrous MgSO4. After filtration and evaporation, the residue was
purified by silica gel column chromatography using hexanes/
dichloromethane, 8:2, as the eluent to give compound 6 (0.35 g,
4. Experimental section
4.1. Chemistry
36%). Rf 0.70; 1H NMR (CDCl3):
d 0.28 (s, 6H, (CH3)2Si), 1.01 (s, 9H,
(CH3)3C), 1.18e1.30 (m, 2H, C2H4CH2), 1.36e1.46 (m, 2H,
CH2CH2eCcarborane), 1.56e1.64 (m, 2H, OCH2CH2), 2.09 (t, J ¼ 7.7 Hz,
2H, CH2eCcarborane), 2.42 (s, 3H, CH3), 3.96 (t, J ¼ 6.1 Hz, 2H, OCH2),
7.32 (d, J ¼ 7.9 Hz, 2H, Ph), 7.74 (d, J ¼ 7.9 Hz, 2H, Ph); 13C NMR
NMR spectra were obtained on a Bruker Avance 400 at The Ohio
State University College of Pharmacy (400 MHz for 1H and 100 MHz
for 13C). Chemical shifts are reported in parts per million (ppm). The
coupling constants are reported in Hertz (Hz). High Resolution-
(CDCl3):
d
ꢀ1.53 ((CH3)2Si), 21.24 ((CH3)3C), 22.59 (CH3), 25.96

S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
331
(C2H4CH2), 28.45 ((CH3)3C), 29.34 (CH2CH2eCcarborane), 30.38
(OCH2CH2), 38.55 (CH2eCcarborane), 70.82 (CH2O), 77.12, 82.10
(Ccarborane), 128.76, 130.83, 133.84, 145.38 (Ph); MS (HR-ESI) calcd
for C20H42B10O3SSiNa (M þ Na)þ: 521.3525, found: 521.3434.
CHCH2eCcarborane), 3.17 (t, J ¼ 6.8 Hz, 2H, ICH2), 3.55 (dd, J ¼ 11.9
and 3.4 Hz, 1H, CH2O), 4.13 (dd, J ¼ 11.9 and 3.6 Hz, 1H, CHCH2O),
4.21e4.30 (m, 1H, CH2CHO); 13C NMR ((CD3)2CO):
d
6.92 (CH2I),
25.80
(CH3),
27.35
(CH3),
29.07
(C2H4CH2), 30.52
(CH2CH2eCcarborane), 33.20 (ICH2CH2), 35.32 (CH2eCcarborane), 39.92
(CH2eCcarborane), 69.51 (CH2O), 74.87 (CHO),77.44, 80.22 (Ccarborane),
110.02 (C(CH3)2); MS (HR-ESI) calcd for C13H31B10IO2Na (M þ Na)þ:
477.2270, found: 477.2257.
4.1.2. 3-(5-{2-[tert.-Butyldimethylsilyl]-o-carboran-1-yl}pentan-1-
yl)thymidine (7)
To a solution of compound 6 (0.35 g, 0.71 mmol) in a mixture of
DMF and acetone (30 mL, 1:1) were added Thd (0.86 g, 3.55 mmol)
and potassium carbonate (0.59 g, 4.26 mmol) and stirred at 50 ꢁC
for 6 h. The reaction mixture was filtered and the filtrate was
concentrated under reduced pressure. The residue was purified by
silica gel column chromatography using ethyl acetate/methanol,
20:1, as the eluent to give compound 7 (0.26 g, 65%). Rf 0.5; 1H NMR
4.1.5. 3-(5-{2-[(2,2-Dimethyl-1,3-dioxolane-4-yl)methyl]-o-
carboran-1-yl}pentan-1-yl)thymidine (10) [25]
To a solution of compound 9 (1.67 g, 3.68 mmol) in a mixture of
DMF and acetone (50 mL, 1:1) were added Thd (4.45 g, 18.4 mmol)
and potassium carbonate (3.05 g, 22.1 mmol) and stirred at 50 ꢁC
overnight. The reaction mixture was filtered and the filtrate was
concentrated under reduced pressure. The residue was purified by
silica gel column chromatography using ethyl acetate/methanol,
20:1, as the eluent to give compound 10 (1.0 g, 48%). Rf 0.29; 1H
(CD3OD):
2H, NC2H4CH2),
d
0.37 (s, 6H, Si(CH3)2), 1.09 (s, 9H, C(CH3)3), 1.27e1.37 (m,
1.54e1.65 (m, 4H, NCH2CH2, and
CH2CH2eCcarborane), 1.91 (s, 3H, CH3), 2.16e2.34 (m, 4H,
CH2eCcarborane, H-200 and H-20), 3.75 (dd, J ¼ 11.8, and 3.3 Hz, 1H, H-
500), 3.82 (dd, J ¼ 11.8, and 3.3 Hz,1H, H-50), 3.87e3.96 (m, 3H, CH2N,
and H-30), 4.39e4.45 (m, 1H, H-40), 6.31 (t, J ¼ 6.8 Hz, 1H, H-10), 7.85
NMR (CD3OD):
d 1.29e1.42 (m, 8H, (CH3)2C, NC2H4CH2), 1.57e1.69
(m, 4H, NCH2CH2, and CH2CH2eCcarborane), 1.92 (s, 3H, CH3),
2.17e2.40 (m, 4H, CH2eCcarborane, H-200, and H-20), 2.54 (dd, J ¼ 11.9,
and 3.4 Hz, 1H, CHCH2eCcarborane), 2.61 (dd, J ¼ 11.9, and 3.4 Hz, 1H,
CHCH2eCcarborane), 3.56 (t, J ¼ 6.8 Hz, 1H, CH2O), 3.74 (dd, J ¼ 11.9,
and 3.4 Hz, 1H, H-500), 3.82 (dd, J ¼ 11.9, and 3.4 Hz, 1H, H-50),
3.88e3.96 (m, 3H, CH2N, and H-30), 4.06e4.14 (m, 1H, CH2O),
4.20e4.32 (m, 1H, CH2CHO), 4.38e4.44 (m, 1H, H-40), 6.32 (t,
J ¼ 6.8 Hz, 1H, H-10), 7.86 (s, 1H, H-6).
(s, 1H, H-6), 13C NMR (CD3OD):
d
ꢀ2.94 (Si(CH3)2), 12.49 (CH3),
20.38 ((CH3)3C), 26.47 (NC2H4CH2), 27.25 ((CH3)3C), 30.14
(CH2CH2eCcarborane NCH2CH2), 38.01 (CH2eCcarborane), 40.50
,
(CH2N), 41.02 (C-20), 61.82 (C-50), 71.15 (C-30), 76.88, 82.24 (Ccarbor-
ane), 86.10 (C-10), 87.87 (C-40), 109.70 (C-5), 135.48 (C-6), 151.24 (C-
2), 164.30 (C-4); MS (HR-ESI) calcd for C23H48B10N2O5SiNa
(M þ Na)þ: 591.4233, found: 591.4304.
4.1.3. 3-(5-{o-Carboran-1-yl}pentan-1-yl)thymidine (1) [22,23]
To a solution of compound 7 (0.35 g, 0.62 mmol) in THF (30 mL)
was added a 1 M solution of TBAF (1.0 mL,1 mmol) in THF at ꢀ78 ꢁC.
The reaction mixture was stirred at rt for 0.5 h. Distilled water
(10 mL) was added and excess THF was removed under reduced
pressure. The residue was extracted with ethyl acetate (50 mL ꢄ 2),
the combined organic layers were washed with brine (10 mL) and
dried over anhydrous MgSO4. After filtration and evaporation, the
residue was subjected to silica gel column chromatography using
ethyl acetate/methanol, 20:1, as the eluent to give compound 1
4.1.6.a. General procedure for the synthesis of the amino acid ester
prodrugs of N5 (11e19) and the glycine ester prodrugs of N5-2OH
(20e22). Compound 1 (0.22 mmol) or 10 (0.18 mmol), N-Boc-
amino acid (Boc-L-valine, Boc-glycine, or Boc-L-glutamic acid a-
tert.-butyl ester) (0.44/0.36 mmol), DCC (0.44/0.36 mmol) and
DMAP (0.044/0.036 mmol) were dissolved in 10 mL of anhydrous
DMF. The reaction mixture was stirred at rt for 24 h. The progress
of the reaction was monitored by TLC (ethyl acetate:hexanes, 8:5).
The reaction yielded three products as determined by TLC. After
24 h, the reaction mixture was filtered and DMF was removed in
vacuum. The residue was dissolved in ethyl acetate (30 mL) and
washed with water (20 mL), 0.1 N HCl (20 mL), saturated NaHCO3
(20 mL), and brine (20 mL). The organic layer was dried over
anhydrous MgSO4. The three intermediates observed by TLC were
separated and purified partially by column chromatography (ethyl
acetate: hexane, 8:5, see supplementary material for 1H NMR data
and Rf values for Boc-protected intermediates). Individual Boc-
substituted amino acid ester prodrugs were added to 10 mL of
a mixture of trifluoroacetic acid (TFA) and dichloromethane (1:1)
(0.17 g, 58%). Rf 0.6; 1H NMR (CD3OD)
d 1.28e1.37 (m, 2H,
NC2H4CH2), 1.49e1.66 (m, 4H, NCH2CH2, and CH2CH2eCcarborane),
1.92 (s, 3H, CH3), 2.17e2.33 (m, 4H, CH2eCcarborane, H-200, and H-
20), 3.74 (dd, J ¼ 11.8, and 3.3 Hz, 1H, H-500), 3.81 (dd, J ¼ 11.8, and
3.3 Hz, 1H, H-50), 3.88e3.95 (m, 3H, CH2N, and H-30), 4.39e4.44 (m,
1H, H-40), 4.53 (s, 1H, carboranyl-CH), 6.31 (t, J ¼ 6.4 Hz, 1H, H-10),
7.85 (s, 1H, H-6); MS (HR-ESI) calcd for C17H34B10N2O5Na
(M þ Na)þ: 477.3369, found: 477.3546.
4.1.4. 4-(2-{5-Iodopentyl}-o-carboran-1-yl)methyl-2,2-dimethyl-
1,3-dioxolane (9)
and stirred at rt for 2 h (Boc-L-glutamic acid 5-tert.-butyl esters
were stirred for 4 h). The solvent was removed and the amino acid
ester prodrugs were purified by preparative HPLC to give products
11e22 as wax-like compounds.
To a solution of compound 8 [25] (2.58 g, 10 mmol) in THF
(50 mL) was added a solution of n-BuLi (4.8 mL, 12 mmol, 2.5 M
solution in hexanes) at ꢀ78 ꢁC. The reaction mixture was allowed to
warm up to 0 ꢁC and 1,5-diiodopentane (5.83 g, 18 mmol) was
added dropwise in rapid fashion. The reaction mixture was refluxed
at 80 ꢁC for 6 h. Distilled water (30 mL) was added and excess THF
was removed under reduced pressure. The residue was extracted
with ethyl acetate (50 mL ꢄ 2), and the combined organic layers
were washed with brine (10 mL) and dried over anhydrous MgSO4.
After filtration and evaporation, the residue was purified by silica
gel column chromatography using hexanes/dichloromethane, 7:4,
as the eluent to give compound 9 (1.85 g, 41%). Rf 0.3; 1H NMR
4.1.6.b. General procedure for the synthesis of the hydrochloride salts
of the amino acid ester prodrugs 11e13 and 15e22. To a solution of
the amino acid ester prodrug (11e13 and 15e22 [0.2 mmol]) in
diethyl ether/THF (1/1, 2 mL) was added 1 mL of a 1 M solution of
HCl gas in diethyl ether at 0 ꢁC and stirred overnight at rt. The
solvent was removed under reduced pressure and the oily
residue was solidified by adding 5 mL of a 0.1 M solution of HCl
gas in diethyl ether. The obtained solid was washed 3e4 times
with a 0.1 M solution of HCl gas in diethyl ether to give product
(11e13ꢂHCl and 15e22ꢂHCl) in 50%e70% yield.
((CD3)2CO):
d 1.31e1.45 (m, 8H, C(CH3)2, NC2H4CH2), 1.50e1.62 (m,
2H, CH2CH2eCcarborane), 1.78e1.87 (m, 2H, ICH2CH2), 2.16e2.33 (m,
2H, CH2eCcarborane), 2.39 (dd,
CHCH2eCcarborane), 2.42 (dd,
J
¼
11.9, and 3.4 Hz, 1H,
11.9, and 3.4 Hz, 1H,
4.1.6.1. 30-L-valyl-3-(5-{o-carboran-1-yl}pentan-1-yl)thymidine (11).
J
¼
Yield: 6 mg, 5%; 1H NMR (CD3OD):
d
1.11 (d, J ¼ 4.6 Hz, 6H,CH(CH3)2),

332
S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
1.28e1.38 (m, 2H, NC2H4CH2), 1.48e1.68 (m, 4H, NCH2CH2, and
CH2CH2eCcarborane), 1.92 (s, 3H, CH3), 2.21e2.50 (m, 5H,
CH2eCcarborane, H-200, H-20, and CH(CH3)2), 3.82e3.85 (m, 2H, H-500
and H-50), 3.90(t, J¼ 6.8 Hz, 2H, CH2N), 4.09(d, J ¼ 4.6 Hz,1H, CHNH2),
4.14e4.18 (m, 1H, H-40), 4.52 (s, 1H, carboranyl-CH), 5.48-5.52 (m,
1H, H-30), 6.34 (t, J ¼ 6.8 Hz, 1H, H-10), 7.86 (s, 1H, H-6); 13C NMR
4.1.6.5. 50-glycinyl-3-(5-{o-carboran-1-yl}pentan-1-yl)thymidine
(15). Yield: 8 mg, 7%; 1H NMR (CD3OD):
1.25e1.35 (m, 2H,
d
NC2H4CH2), 1.44e1.64 (m, 4H, NCH2CH2, and CH2CH2eCcarborane),
1.92 (s, 3H, CH3), 2.23e2.35 (m, 4H, CH2eCcarborane, H-200 and H-
20), 3.84e3.95 (m, 4H, CH2N, and CH2NH2), 4.04e4.10 (m, 1H, H-30),
4.36e4.40 (m, 1H, H-40), 4.44 (dd, J ¼ 11.7, and 3.6 Hz, 1H, H-500),
4.48e4.55 (m, 2H, carboranyl-CH, and H-50), 6.26 (t, J ¼ 6.7 Hz, 1H,
H-10), 7.46 (s, 1H, H-6); 13C NMR (CD3OD): 12.20 (CH3), 26.13
(NC2H4CH2), 26.92 (CH2CH2eCcarborane), 28.94 (NCH2CH2), 37.56
(CH2eCcarborane), 39.17 (NCH2), 39.96 (CH2NH2), 40.94 (C-20), 62.65
(Ccarborane), 65.81 (C-50), 71.05 (C-30), 76.25 (Ccarborane), 84.12 (C-10),
86.63 (C-40), 110.11 (C-5), 135.34 (C-6), 151.19 (C-2), 164.27 (C-4),
167.47 (COO); MS (HR-ESI) calcd for C19H38B10N3O6 (M þ H)þ:
512.3764, found: 512.3798; 15ꢂHCl: Anal. (C19H38B10ClN3O6) C, H, N.
(CD3OD):
d 12.69 (CH3), 17.87 (CH(CH3)2), 26.59 (NC2H4CH2), 27.39
(CH2CH2eCcarborane), 29.42 (NCH2CH2), 30.54 (CH(CH3)2), 37.89
(CH2eCcarborane), 38.04 (CH2N), 41.42 (C-20), 58.81 (CHNH2),
62.29 (Ccarborane), 63.14 (C-50), 76.75 (Ccarborane), 78.24 (C-30), 85.78
(C-10), 86.52 (C-40), 110.69 (C-5), 135.56 (C-6), 151.86 (C-2), 164.76
(C-4), 169.20 (COO); MS (HR-ESI) calcd for C22H44B10N3O6 (M þ H)þ:
554.4233, found: 554.4145; 11ꢂHCl: Anal. (C22H44B10ClN3O6) C, H, N.
4.1.6.2. 50-L-valyl-3-(5-{o-carboran-1-yl}pentan-1-yl)thymidine (12).
4.1.6.6. 30,50-di-glycinyl-3-(5-{o-carboran-1-yl}pentan-1-yl)thymi-
Yield: 12 mg, 10%; 1H NMR (CD3OD):
d 1.04e1.09 (m, 6H, CH(CH3)2),
dine (16). Yield: 19 mg, 15%; 1H NMR (CD3OD):
d
1.24e1.36 (m, 2H,
NC2H4CH2), 1.48e1.64 (m, 4H, NCH2CH2, and CH2CH2eCcarborane),
1.92 (s, 3H, CH3), 2.26 (t, 6.8 Hz, 2H, CH2eCcarborane),
1.25e1.34 (m, 2H, NC2H4CH2), 1.46-1.63 (m, 4H, NCH2CH2, and
CH2CH2eCcarborane), 1.91 (s, 3H, CH3), 2.23e2.36 (m, 5H,
CH2eCcarborane, H-200, H-20, and CH(CH3)2), 3.89 (t, J ¼ 7.2 Hz, 2H,
CH2N), 3.97 (d, J ¼ 4.4 Hz, 1H, CHNH2), 4.06-4.13 (m, 1H, H-30),
4.32e4.37 (m, 1H, H-40), 4.41 (dd, J ¼ 11.8, and 3.3 Hz, 1H, H-500), 4.52
(s,1H, carboranyl-CH), 4.56e4.64 (m,1H, H-50), 6.27 (t, J ¼ 6.8 Hz,1H,
J
¼
2.50e2.62 (m, 2H, H-200 and H-20), 3.89e3.95 (m, 6H, CH2N, and
2CH2NH2), 4.32e4.38 (m, 1H, H-40), 4.48e4.55 (m, 2H, carboranyl-
CH and H-500), 4.60 (dd, J ¼ 11.7, and 3.6 Hz, 1H, H-50), 5.46e5.49 (m,
1H, H-30), 6.25 (t, J ¼ 6.7 Hz, 1H, H-10), 7.48 (s, 1H, H-6); 13C NMR
H-10), 7.46 (s, 1H, H-6); 13C NMR (CD3OD):
d 12.68 (CH3), 17.65, 17.76
(CD3OD):
d
13.33
(CH3),
27.52
(NC2H4CH2),
28.04
(CH(CH3)2), 26.62 (NC2H4CH2), 27.41 (CH2CH2eCcarborane), 29.44
(NCH2CH2), 30.55 (CH(CH3)2), 38.05 (CH2eCcarborane), 39.59 (CH2N),
41.42 (C-20), 58.86 (CHNH2), 63.16 (Ccarborane), 66.56 (C-50), 71.61 (C-
30), 76.76 (Ccarborane), 84.64 (C-10), 87.37 (C-40), 110.58 (C-5), 135.92
(C-6), 151.66 (C-2), 164.75 (C-4), 169.55 (COO); MS (HR-ESI) calcd for
C22H44B10N3O6 (M þ H)þ: 554.4233, found: 554.4227, calcd for
C22H43B10N3O6Na (M þ Na)þ 576.4053 (M þ Na)þ, found 576.4076;
12ꢂHCl: Anal. (C22H44B10ClN3O6) C, H, N.
(CH2CH2eCcarborane), 30.08 (NCH2CH2), 37.17 (CH2eCcarborane), 38.68
(NCH2), 41.11, 41.26 (2CH2NH2), 42.09 (C-20), 63.80 (Ccarborane), 66.45
(C-50), 77.39 (Ccarborane), 77.49 (C-30), 82.93 (C-10), 88.20 (C-40),
111.45 (C-5), 136.52 (C-6), 152.28 (C-2), 165.33 (C-4), 168.52, 168.58
(2COO); MS (HR-ESI) calcd for C21H41B10N4O7 (M þ H)þ: 569.3978,
found: 569.3975; 16ꢂHCl: Anal. (C21H42B10Cl2N4O7) C, H, N.
4.1.6.7. 30-L-glutam-5-yl-3-(5-{o-carboran-1-yl}pentan-1-yl)thymi-
dine (17). Yield: 3 mg, 2%; 1H NMR (CD3OD):
d 1.28e1.35 (m, 2H,
4.1.6.3. 30,50-di-L-valyl-3-(5-{o-carboran-1-yl}pentan-1-yl)thymi-
NC2H4CH2), 1.47e1.64 (m, 4H, NCH2CH2, and CH2CH2eCcarborane), 1.92
(s, 3H, CH3), 2.15e2.40 (m, 6H, CH2eCcarborane, CH2CHNH2, H-200 and H-
20),2.64e2.69 (m, 2H, CH2COO), 3.81e3.83 (m, 2H, H-500 and H-50), 3.90
(t, J ¼ 6.8 Hz, 2H, CH2N), 4.05 (t, J ¼ 6.8 Hz, 1H, CHNH2), 4.10e4.13 (m,
1H, H-40), 4.52 (s,1H, carboranyl-CH), 5.34e5.39 (m, 1H, H-30), 6.31 (t,
dine (13). Yield: 18 mg, 13%; 1H NMR (CD3OD):
d
1.05 (d, J ¼ 4.4 Hz,
6H, CH(CH3)2), 1.10 (d, J ¼ 4.4 Hz, 6H, CH(CH3)2), 1.26e1.35 (m, 2H,
NC2H4CH2), 1.47e1.63 (m, 4H, NCH2CH2, and CH2CH2eCcarborane),
1.92 (s, 3H, CH3), 2.23-2.38 (m, 4H, CH2eCcarborane, H-200 and H-
20), 2.50e2.57 (m, 1H, CH(CH3)2), 2.61e2.70 (m, 1H, CH(CH3)2), 3.89
(t, J ¼ 6.8 Hz, 2H, CH2N), 3.98 (d, J ¼ 4.4 Hz, 1H, CHNH2), 4.02 (d,
J ¼ 4.4 Hz, 1H, CHNH2), 4.33e4.39 (m, 1H, H-40), 4.45e4.55 (m, 2H,
carboranyl-CH and H-500), 4.68e4.74 (m, 1H, H-50), 5.46e5.50 (m,
1H, H-30), 6.23 (t, J ¼ 6.8 Hz, 1H, H-10), 7.47 (s, 1H, H-6); 13C NMR
J ¼ 6.8 Hz,1H, H-10), 7.86 (s, 1H, H-6); 13C NMR (CD3OD):
d 12.19 (CH3),
25.45 (CH2CHNH2), 26.12 (NC2H4CH2), 26.91 (CH2CH2eCcarborane),
28.94 (NCH2CH2), 29.50 (CH2COO), 37.51 (CH2eCcarborane), 37.55
(NCH2), 40.94 (C-20), 52.00 (CHNH2), 61.90 (C-50), 62.66 (Ccarborane),
75.97 (C-30), 77.50 (Ccarborane), 85.65 (C-10), 85.98 (C-40), 110.17 (C-5),
135.12 (C-6), 151.38 (C-2), 164.28 (C-4), 170.37 (COOCH2), 172.15
(COOH). MS (HR-ESI) calcd for C22H41B10N3O8Na (M þ Na)þ: 606.3795,
found: 606.3813 (M þ Na)þ; 17ꢂHCl: Anal. (C22H42B10ClN3O8) calcd: C,
42.61; H, 6.83; N, 6.78, found: C, 41.95; H, 6.54; N, 6.62.
(CD3OD)
d 12.16 (CH3), 17.17, 17.24, 17.36, 17.39 (2CH(CH3)2), 26.09
(NC2H4CH2), 26.89 (CH2CH2eCcarborane), 28.91 (NCH2CH2), 30.04
(2CH(CH3)2), 35.96 (CH2eCcarborane), 37.55 (CH2N), 40.91 (C-20),
58.32 (2CHNH2), 62.69 (Ccarborane), 65.51 (C-50), 72.28 (Ccarborane),
76.31 (C-30), 81.89 (C-10), 87.71 (C-40), 110.32 (C-5), 135.72 (C-6),
151.14 (C-2), 164.13 (C-4), 168.83, 169.07 (2COO); MS (HR-ESI) calcd
for C27H53B10N4O7 (M þ H)þ: 653.4917, found: 653.4932; 13ꢂHCl:
Anal. (C27H54B10Cl2N4O7) C, H, N.
4.1.6.8. 50-L-glutam-5-yl-3-(5-{o-carboran-1-yl}pentan-1-yl)thymi-
dine (18). Yield: 17 mg, 13%; 1H NMR (CD3OD):
d 1.28e1.35 (m, 2H,
NC2H4CH2), 1.46e1.64 (m, 4H, NCH2CH2, and CH2CH2eCcarborane),
1.92 (s, 3H, CH3), 2.12e2.35 (m, 6H, CH2eCcarborane, CH2CHNH2, H-
200 and H-20), 2.66 (t, J ¼ 6.9 Hz, 2H, CH2COO), 3.89 (t, J ¼ 6.9 Hz, 2H,
CH2N), 4.01e4.11 (m, 2H, CHNH2, and H-30), 4.30 (dd, J ¼ 11.9 and
3.4 Hz, 1H, H-500), 4.33e4.38 (m, 1H, H-40), 4.41 (dd, J ¼ 11.9 and
3.4 Hz, 1H, H-50), 4.51 (s,1H, carboranyl-CH), 6.28 (t, J ¼ 6.8 Hz, 1H,
4.1.6.4. 30-glycinyl-3-(5-{o-carboran-1-yl}pentan-1-yl)thymidine (14).
Yield: 1 mg, 1%; 1H NMR (CD3OD):
d 1.22e1.34 (m, 2H, NC2H4CH2),
1.44e1.63 (m, 4H, NCH2CH2, and CH2CH2eCcarborane), 1.89 (s, 3H, CH3),
2.20e2.48 (m, 4H, CH2eCcarborane, H-200 and H-20), 3.87e3.94 (m, 6H,
CH2N, CH2NH2, H-500 and H-50), 4.12e4.17 (m, 1H, H-40), 4.49 (s, 1H,
carboranyl-CH, and), 5.45e5.55 (m, 1H, H-30), 6.32 (t, J ¼ 6.7 Hz, 1H,
H-10), 7.49 (s, 1H, H-6); 13C NMR (CD3OD):
d 12.29 (CH3), 25.56
(CH2CHNH2), 26.13 (NC2H4CH2), 26.91 (CH2CH2eCcarborane), 28.94
(NCH2CH2), 29.48 (CH2COO), 37.55 (CH2eCcarborane), 39.57 (NCH2),
40.93 (C-20), 51.99 (CHNH2), 62.64 (C-50), 64.56 (Ccarborane), 71.22 (C-
30), 76.26 (Ccarborane), 84.65 (C-10), 86.47 (C-40), 110.00 (C-5), 135.00
(C-6), 151.18 (C-2), 164.28 (C-4), 170.29 (COOCH2), 172.32 (COOH);
MS (HR-ESI) calcd for C22H41B10N3O8Na (M þ Na)þ : 606.3795,
found: 606.3793. 18ꢂHCl: Anal. (C22H42B10ClN3O8) calcd: C, 42.61;
H, 6.83; N, 6.78, found: C, 42.12; H, 6.74; N, 6.74.
H-10), 7.83 (s, 1H, H-6); 13C NMR (CD3OD):
d 13.07 (CH3), 26.99
(NC2H4CH2), 27.78 (CH2CH2eCcarborane), 29.82 (NCH2CH2), 38.27
(CH2eCcarborane), 38.43 (NCH2), 41.02 (CH2NH2), 41.82 (C-20), 62.71
(Ccarborane), 63.53 (C-50), 77.13 (Ccarborane), 78.49 (C-30), 86.22 (C-10),
86.82 (C-40), 111.08 (C-5), 135.92 (C-6), 152.25 (C-2), 165.13 (C-4),
168.17 (COO); MS (HR-ESI) calcd for C19H38B10N3O6 (M þ H)þ:
512.3764, found: 512.5839.

S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
333
4.1.6.9. 30,50-di-L-glutam-5-yl-3-(5-{o-carboran-1-yl}pentan-1-yl)
thymidine (19). Yield: 48 mg; 1H NMR (CD3OD):
1.28e1.35 (m, 2H,
(Ccarborane), 80.86 (Ccarborane), 81.83 (C-10), 87.16 (C-40), 110.33 (C-5),
135.39 (C-6), 151.19 (C-2), 164.18 (C-4), 167.41, 167.47 (2COO); MS (HR-
ESI) calcd for C24H47B10N4O9 (M þ H)þ: 643.4346, found: 643.4391;
22ꢂHCl: Anal. (C24H47B10ClN4O9) calcd: C, 40.28; H, 6.76; N, 7.83, found:
C, 40.36; H, 6.27; N, 7.53.
d
NC2H4CH2), 1.45e1.64 (m, 4H, NCH2CH2, and CH2CH2eCcarborane),
1.91 (s, 3H, CH3), 2.10e2.28 (m, 4H, CH2eCcarborane, H-200, and H-20),
2.35e2.50 (m, 4H, 2CH2CHNH2), 2.62e2.69 (m, 4H, 2CH2COO), 3.89
(t, J ¼ 6.8 Hz, 2H, CH2N), 3.93e4.01 (m, 2H, 2CHNH2), 4.26e4.31 (m,
1H, H-40), 4.34 (dd, J ¼ 11.9, and 3.4 Hz, 1H, H-500), 4.47 (dd, J ¼ 11.9,
and 3.4 Hz,1H, H-50), 4.51 (s,1H, carboranyl-CH), 5.29-5.33 (m,1H, H-
30), 6.24 (t, J ¼ 6.8 Hz, 1H, H-10), 7.49 (s, 1H, H-6), 13C NMR (CD3OD):
4.2. Stability studies
4.2.1. Stability studies in PBS at pH 7.4
Solutions of amino acid ester prodrugs 4, 11e19, and 21 in PBS
d
12.32 (CH3), 25.40, 25.54 (2CH2CHNH2), 26.12 (NC2H4CH2), 26.90
(2 mM, 2.7 mL, pH 7.4) wereincubated at 37 ꢁC for 6e48 h depending
(CH2CH2eCcarborane), 28.94 (NCH2CH2), 29.34, 29.52 (2CH2COO),
36.66 (CH2-Ccarborane), 37.55 (NCH2), 40.97 (C-20), 52.00 (2CHNH2),
62.66 (C-50), 64.36 (Ccarborane), 76.20 (C-30), 77.25 (Ccarborane), 82.65
(C-10), 86.59 (C-40),110.29 (C-5),134.77 (C-6),151.20 (C-2),164.21 (C-
4), 170.34, 170.38 (2COOCH2), 172.17, 172.31 (2COOH); MS (HR-ESI)
calcd for C27H48B10N4O11Na, (M þ Na)þ, 735.4220, found: 735.4203;
19ꢂHCl: Anal. (C27H50B10Cl2N4O11) C, H, N.
on the degradation time of the prodrug. Aliquots of 200
incubated solutions were withdrawn at various time intervals and
mixed with ice-cold TFA in acetonitrile solution (10%, 400 L). The
mixtures were passed through 0.2 m filters and analyzed by HPLC
mL from the
m
m
(Table 3) to determine the areas under the curves (AUCs) that cor-
responded to the prodrugs and their degradation products.
4.1.6.10. 30-glycinyl-3-[5-(2-{2,3-dihydroxyprop-1-yl}-o-carboran-1-
4.2.2. Stability studies in Bovine serum
yl)pentan-1-yl]thymidine (20). Yield: 4 mg, 4%; 1H NMR (CD3OD):
Solutions of amino acid ester prodrugs 4, 12, 15, 18, and 21 in
Bovine serum (2 mM, 2.7 mL) were incubated at 37 ꢁC for 6e48 h
depending on the degradation profiles of the prodrugs. Aliquots
d
1.28e1.38 (m, 2H, NC2H4CH2), 1.52e1.66 (m, 4H, NCH2CH2, and
CH2CH2eCcarborane), 1.89 (s, 3H, CH3), 2.17e2.58 (m, 6H,
CH2eCcarborane, CHCH2eCcarborane, H-200 and H-20), 3.33e3.49 (m, 2H,
CH2OH), 3.71e3.96 (m, 7H, CH2N, CH2NH2, H-500, H-50 and CH2CHO),
4.12e4.17 (m, 1H, H-40), 5.44e5.49 (m, 1H, H-30), 6.31 (t, J ¼ 6.7 Hz,
of 200
points and added to ice-cold TFA in acetonitrile solutions (10%,
400 L). The samples were centrifuged for 10 min at 3000 rcf, the
supernatants were passed through 0.2 m filters, and the filtrates
mL from the incubation media were taken at various time
m
1H, H-10), 7.83 (s, 1H, H-6), 13C NMR (CD3OD):
d 13.37 (CH3), 27.38
m
were analyzed by HPLC (Table 3) to obtain AUCs.
(NC2H4CH2), 28.00 (CH2CH2eCcarborane), 30.49 (NCH2CH2), 35.93
(CH2eCcarborane), 38.55 (CH2eCcarborane), 40.03 (NCH2), 41.32
(CH2NH2), 42.09 (C-20), 62.98 (C-50), 67.06 (CH2OH), 72.35 (CHOH),
78.75 (C-30), 80.57 (Ccarborane), 82.00 (Ccarborane), 86.47 (C-10), 87.08
(C-40), 111.34 (C-5), 136.16 (C-6), 152.51 (C-2), 165.40 (C-4), 16.46
(COO); MS (HR-ESI) calcd for C22H44B10N3O8 (M þ H)þ: 586.4132,
found: 586.4059; 20ꢂHCl: Anal. (C22H44B10ClN3O8) C, H, N.
4.2.3. Stability studies in Bovine CSF
Solutions of amino acid ester prodrugs 4, 12, 15, 18, and 21 in CSF
(2 mM, 2.7 mL) were incubated at 37 ꢁC for 6e48 h depending on
the prodrug degradation profiles. Aliquots of 200
experiment were taken at various time intervals and mixed with
ice-cold TFA in acetonitrile solution (10%, 400 L). The samples
were passed through 0.2 m filters and analyzed by HPLC (Table 3)
mL from each
m
4.1.6.11. 50-glycinyl-3-[5-(2-{2,3-dihydroxyprop-1-yl}-o-carboran-1-
m
yl)pentan-1-yl]thymidine (21). Yield: 23 mg, 22%; 1H NMR (CD3OD):
to give the corresponding AUCs.
d
1.24e1.37 (m, 2H, NC2H4CH2), 1.51e1.63 (m, 4H, NCH2CH2, and
4.2.4. Calculation of the degradation half-lives (t ) of the amino
acid ester prodrugs
½
CH2CH2eCcarborane), 1.90 (s, 3H, CH3), 2.18e2.38 (m, 4H, CH2e
Ccarborane, H-200 and H-20), 2.52 (d, J ¼ 4.4 Hz, 2H, CHCH2eCcarborane),
3.27e3.36 (m, 1H, CH2OH), 3.45 (dd, J ¼ 11.7, and 3.6 Hz, 1H,
CH2OH), 3.72e3.80 (m, 1H, H-30), 3.82e3.94 (m, 4H, CH2N, and
CH2NH2), 4.02e4.08 (m, 1H, CH2CHO), 4.31e4.37 (m, 1H, H-40), 4.41
(dd, J ¼ 11.7, and 3.6 Hz,1H, H-500), 4.49 (dd, J ¼ 11.7, and 3.6 Hz,1H, H-
50), 6.24 (t, J ¼ 6.7 Hz, 1H, H-10), 7.44 (s, 1H, H-6); 13C NMR (CD3OD):
The hydrolytic half-lives (t ) of the amino acid ester prodrugs 4,
½
11e19, and 21 were calculated by utilizing the integrated AUCs from
analytical HPLC data. The AUCs were converted into the remaining
concentrations of the amino acid ester prodrugs using generated
standard curves for the amino acid ester prodrugs. Standard curves for
the amino acid ester prodrugs were generated by plotting different
concentrations of the amino acid ester prodrug vs. the corresponding
AUCs values. The apparent first-order degradation rate constants of the
amino acid ester prodrugs 4, 11e19, and 21 were determined by plot-
ting the logarithm of the remaining concentrations of the amino acid
ester prodrug vs. time. The concentrationetime curves were analyzed
by a computerized a curve-stripping program (R Strip; Micromath
Scientific Software, Salt Lake City, UT, USA). The slopes of these plots
were related to the rate constant k given by the following relation:[15]
d
12.23 (CH3), 26.29 (NC2H4CH2), 26.91 (CH2CH2eCcarborane), 29.38
(NCH2CH2), 34.82 (CH2eCcarborane), 38.91 (CH2eCcarborane), 39.19
(NCH2), 39.99 (CH2NH2), 40.99 (C-20), 65.84 (C-50), 65.93 (CH2OH),
71.07 (CHOH), 71.22 (C-30), 79.45 (Ccarborane), 80.89 (Ccarborane), 84.14
(C-10), 86.67 (C-40), 110.12 (C-5), 135.35 (C-6), 151.12 (C-2), 164.28 (C-
4), 167.46 (COO); MS (HR-ESI) calcd C22H44B10N3O8 (M þ H)þ:
586.4132, found: 586.4107; 21ꢂHCl: Anal. (C22H44B10ClN3O8) C, H, N.
4.1.6.12. 30,50-di-glycinyl-3-[5-(2-{2,3-dihydroxyprop-1-yl}-o-car-
boran-1-yl)pentan-1-yl]-thymidine (22). Yield: 24 mg, 21%; 1H NMR
K ¼ 2:303*slopeðlogCvs:timeÞ
(CD3OD): d 1.25e1.38 (m, 2H, NC2H4CH2), 1.53e1.64 (m, 4H, NCH2CH2,
The hydrolytic half-lives were then determined by the following
equation:[15]
and CH2CH2eCcarborane), 1.91 (s, 3H, CH3), 2.20e2.39 (m, 4H,
CH2eCcarborane, H-200 and H-20), 2.46e2.61 (m, 2H, CHCH2eCcarborane),
3.45 (dd, J ¼ 11.7, and 3.6 Hz, 1H, CH2OH), 3.71e3.80 (m, 1H, CH2OH),
3.82e3.97 (m, 7H, CH2N, 2CH2NH2 and CH2CHO), 4.32e4.37 (m,1H, H-
40), 4.49 (dd, J ¼ 11.7, and 3.6 Hz,1H, H-500), 4.59 (dd, J ¼ 11.7, and 3.6 Hz,
1H, H-50), 5.42-5.47 (m,1H, H-30), 6.23 (t, J¼ 6.7 Hz,1H, H-10), 7.46 (s,1H,
t1=2 ¼ 0:693=k
4.3. Solubility studies
H-6);13C NMR (CD3OD):
d 12.21 (CH3), 26.25 (NC2H4CH2), 26.86
(CH2CH2eCcarborane), 29.35 (NCH2CH2), 34.79 (CH2eCcarborane), 36.09
(CH2eCcarborane), 38.89 (NCH2), 39.97, 40.15 (2CH2NH2), 40.79 (C-20),
65.41 (C-50), 65.93 (CH2OH), 71.21 (CHOH), 76.39 (C-30), 79.44
Quantities of 1 mg of the hydrochloric acid salts of the amino acid
ester prodrugs 4, 12, 15, 18, and 21 or their corresponding parental
drugs 1 and 2 were added to micro centrifuge tubes containing

334
S. Hasabelnaby et al. / European Journal of Medicinal Chemistry 55 (2012) 325e334
[11] F.P. Bonina, L. Arenare, F. Palagiano, A. Saija, F. Nava, D. Trombetta,
P.D. Caprariis, Synthesis, stability, and pharmacological evaluation of nipecotic
acid prodrugs, J. Pharm. Sci. 88 (1999) 561e567.
[12] D.B. Williams, S.A. Varia, V.J. Stella, I.H. Pitman, Evaluation of the prodrug
potential of the sulfate esters of acetaminophen and 3-hydroxymethyl-
phenytoin, Int. J. Pharm. 14 (1983) 113e120.
[13] G.L. Flynn, D.J. Lamb, Factors influencing solvolysis of corticosteroid-21-
phosphate esters, J. Pharm. Sci. 59 (1970) 1433e1438.
[14] H. Bundgaard, C. Larsen, P. Thorbek, Prodrugs as drug delivery systems XXVI.
Preparation and enzymatic hydrolysis of various water-soluble amino acid
esters of metronidazole, Int. J. Pharm. 18 (1984) 67e77.
either PBS (100 mL, pH 5), PBS (100 mL, pH 6) or PBS (100 mL, pH 7.4).
The micro centrifuge tubes were placed in a shaker for 4 h followed
by centrifugation at 3000 rcf for 10 min. The solutions were passed
through 0.2 mm filters and analyzed by HPLC (Table 3) to obtain the
AUC corresponding to amount of the drug soluble in the each buffer.
Standard curves for the each amino acid ester prodrug and its
parental compound were generated and utilized to calculate the
amino acid ester prodrug concentration corresponding to each AUC.
[15] X. Song, P.L. Lorenzi, C.P. Landowski, B.S. Vig, J.M. Hilfinger, G.L. Amidon,
Amino acid ester prodrugs of the anticancer agent gemcitabine: synthesis,
bioconversion, metabolic bioevasion, and hPEPT1-mediated transport, Mol.
Pharm. 2 (2005) 157e167.
4.4. Determination of the protein content of PBS, Bovine serum and
CSF using the Bradford protein assay
[16] Y. Tsume, B.S. Vig, J. Sun, C.P. Landowski, J.M. Hilfinger, C. Ramachandran,
G.L. Amidon, Enhanced absorption and growth inhibition with amino acid
monoester prodrugs of floxuridine by targeting hPEPT1 transporters, Mole-
cules 13 (2008) 1441e1454.
[17] B.S. Vig, P.J. Lorenzi, S. Mittal, C.P. Landowski, H. Shin, H.I. Mosberg,
J.M. Hilfinger, G.L. Amidon, Amino acid ester prodrugs of floxuridine:
synthesis and effects of structure, stereochemistry, and site of esterification
on the rate of hydrolysis, Pharm. Res. 20 (2003) 1381e1388.
Seven dilutions of the standard protein (Bovine serum albumin,
BSA) were required to set up a standard curve. Ten
concentration and 10 L of each sample (PBS, serum, CSF) were
added into separate plate wells of a 96 well plate. The quantity of
500 L of 5-fold diluted Bradford reagent (BioRAD, Philadelphia)
mL of each BSA
m
m
was added to each well. The plate was incubated in the dark at
room temperature for 5 min. Following incubation, the absorbances
of the different BSA concentrations and the samples were measured
at 595 nm. A standard curve was generated by plotting the seven
BSA concentrations vs. the corresponding absorbances. The stan-
dard curve was utilized to calculate the protein concentrations
corresponding to the measured absorbance of each test sample.
[18] Y. Sun, J. Sun, S.Y. Jing, S. Yin, Y. Chen, G. Li, Y. Xu, Z. He, Synthesis, transport
and pharmacokinetics of 5’-amino acid ester prodrugs of 1-
b-D-arabinofur-
anosylcytosine, Mol. Pharm. 6 (2009) 315e325.
[19] B. Anand, S. Katragadda, Y. Nashed, A. Mitra, Amino acid prodrugs of acyclovir
as possible antiviral agents against ocular HSV-1 infections: interactions with
the neutral and cationic amino acid transporter on the corneal epithelium,
Curr. Eye Res. 29 (2004) 153e166.
[20] S. Katragadda, R. Jain, D. Kwatra, S. Hariharan, A. Mitra, Pharmacokinetics of
amino acid ester prodrugs of acyclovir after oral administration: interaction
with the transporters on Caco-2 cells, Int. J. Pharm. 362 (2008) 93e101.
[21] L.M. Beauchamp, G.F. Orr, P. De Miranda, T. Burnette, T.A. Krenitsky, Amino
acid ester prodrugs of acyclovir, Antivir. Chem. Chemother. 3 (1992) 157e164.
[22] A.J. Lunato, J. Wang, J.E. Woollard, A.K.M. Anisuzzaman, W. Ji, F.G. Rong,
S. Ikeda, A.H. Soloway, S. Eriksson, D.H. Ives, T.E. Blue, W. Tjarks, Synthesis of
Acknowledgments
This work was supported by funds from The Ohio State
University College of Pharmacy, NIH Grant R01CA127935, and
a scholarship from the Scholar Exchange Program of the Egyptian
Education Ministry. The content is solely the responsibility of the
authors and does not necessarily represent the official views of the
National Institutes of Health.
5-(carboranylalkylmercapto)-2’-deoxyuridines
and
3-(carboranylalkyl)
thymidines and their evaluation as substrates for human thymidine kinases 1
and 2, J. Med. Chem. 42 (1999) 3378e3389.
[23] R. Tiwari, A. Toppino, H.K. Agarwal, T. Huo, Y. Byun, J. Gallucci, S. Hasabelnaby,
A. Khalil, A. Goudah, R.A. Baiocchi, M.V. Darby, R.F. Barth, W. Tjarks, Synthesis,
biological evaluation, and radioiodination of halogenated closo-carbor-
anylthymidine analogues, Inorg. Chem. 51 (2012) 629e639.
[24] F.A. Gomez, M.F. Hawthorne, A simple route to C-monosubstituted carborane
derivatives, J. Org. Chem. 5 (1992) 1384e1390.
Appendix A. Supplementary material
[25] Y. Byun, B.T.S. Thirumamagal, W. Yang, S. Eriksson, R.F. Barth, W. Tjarks,
Preparation and biological evaluation of 10B-enriched 3-[5-{2-(2,3-
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ejmech.2012.07.
033. Data include original MS-, 1H NMR-, and 13C NMR spectra,
HPLC chromatograms, additional NMR data, and figures related to
stability studies.
dihydroxyprop-1-yl)-o-carboran-1-yl}pentan-1-yl]thymidine
a new boron delivery agent for boron neutron capture therapy of brain
tumors, J. Med. Chem. 49 (2006) 5513e5523.
(N5-2OH),
[26] L. Colla, E. De Clercq, R. Busson, H. Vanderhaeghe, Synthesis and antiviral
activity of water-soluble esters of acyclovir [9-[(2-hydroxyethoxy)methyl]
guanine], J. Med. Chem. 26 (1983) 602e604.
[27] A.E. Thomsen, G.M. Friedrichsen, A.H. Sørensen, R. Andersen, C.U. Nielsen,
B. Brodin, M. Begtrup, S. Frokjaer, B. Steffansen, Prodrugs of purine and
pyrimidine analogues for the intestinal di/tri-peptide transporter PepT1:
affinity for hPepT1 in Caco-2 cells, drug release in aqueous media and in vitro
metabolism, J. Control. Release 86 (2003) 279e292.
[28] N.J. Abbott, Evidence for bulk flow of brain interstitial fluid: significance for
physiology and pathology, Neurochem. Int. 45 (2004) 545e552.
[29] C.P. Landowski, P.L. Lorenzi, X. Song, G.L. Amidon, Nucleoside ester prodrug
substrate specificity of liver carboxylesterase, J. Pharmacol. Exp. Ther. 316
(2006) 572e580.
References
[1] R.F. Barth, J.A. Coderre, M. Vicente, T.E. Blue, Boron neutron capture therapy of
cancer: current status and future prospects, Clin. Cancer Res. 11 (2005) 3987e4002.
[2] R.F. Barth, W. Yang, A.S. Al-Madhoun, J. Johnsamuel, Y. Byun, S. Chandra, D.R. Smith,
W. Tjarks, S. Eriksson, Boron-containing nucleosides as potential delivery agents for
neutron capture therapy of brain tumors, Cancer Res. 64 (2004) 6287e6295.
[3] R.F. Barth, W. Yang, G. Wu, M. Swindall, Y. Byun, S. Narayanasamy, W. Tjarks,
K. Tordoff, M.L. Moeschberger, S. Eriksson, P.J. Binns, K.J. Riley, Thymidine
kinase 1 as a molecular target for boron neutron capture therapy of brain
tumors, Proc. Nat. Acad. Sci. U. S. A. 105 (2008) 17493e17497.
[4] A.S. Al-Madhoun, J. Johnsamuel, R.F. Barth, W. Tjarks, S. Eriksson, Evaluation of
human thymidine kinase 1 substrates as new candidates for boron neutron
capture therapy, Cancer Res. 64 (2004) 6280e6286.
[5] W. Tjarks, R. Tiwari, Y. Byun, S. Narayanasamy, R.F. Barth, Carboranyl thymidine
analogues for neutron capture therapy, Chem. Commun. 47 (2007) 4978e4991.
[6] E.S. Arner, S. Eriksson, Mammalian deoxyribonuclease kinases, Pharmacol.
Ther. 67 (1995) 155e186.
[30] M. Nedergaard, T. Takano, A.J. Hansen, Beyond the role of glutamate as
a neurotransmitter, Nat. Rev. Neurosci. 3 (2002) 748e755.
[31] K. Yamamoto, Y. Endo, Bioorg. Utility of boron clusters for drug design.
Hansch-fujita hydrophobic parameters
p of dicarba-closo-dodecaboranyl
groups, Med. Chem. Lett. 11 (2001) 2389e2392.
[32] B.T.S. Thirumamagal, X.B. Zhao, A.K. Bandyopadhyaya, J. Johnsamuel, R. Tiwari,
D.W. Golightly, V. Patel, B.T. Jehning, M.V. Backer, R.F. Barth, R.J. Lee,
J.M. Backer, W. Tjarks, Receptor-targeted liposomal delivery of boron-
containing cholesterol mimics for boron neutron capture therapy (BNCT),
Bioconjug. Chem. 17 (2006) 1141e1150.
[33] J.L. Fauchere, K. Quang Do, P.Y.C. Jow, C. Hansch, Unusually strong lipophilicity
of ’fat’ or ’super’ amino-acids, including a new reference value for glycine,
Experientia 36 (1980) 1203e1204.
[34] S. Narayanasamy, B.T.S. Thirumamagal, J. Johnsamuel, Y. Byun, A.S. Al-Madhoun,
E. Usova, G.Y. Cosquer, J. Yan, A.K. Bandyopadhyaya, R. Tiwari, S. Eriksson, W. Tjarks,
Hydrophilically enhanced 3-carboranyl thymidine analogues (3CTAs) for boron
neutron capture therapy (BNCT) of cancer, Bioorg. Med. Chem. 14 (2006) 6886e6899.
[35] T.L. Lemke, D.A. Williams, Foye’s Principles of Medicinal Chemistry, sixth ed.,
Lippincott Williams & Wilkins, Philadelphia, 2007.
[7] S.E. Aspland, C. Ballatore, R. Castillo, J. Desharnais, T. Eustaquio, P. Goelet,
Z. Guo, Q. Li, D. Nelson, C. Sun, A.J. Castellino, M.J. Newman, Kinase-mediated
trapping of bi-functional conjugates of paclitaxel or vinblastine with thymi-
dine in cancer cells, Bioorg. Med. Chem. Lett. 16 (2006) 5194e5198.
[8] R.B. Silverman, The Organic Chemistry of Drug Design and Drug Action,
second ed., Elsevier Academic Press, San Diego, 2004.
[9] A.S. Al-Madhoun, W. Tjarks, S. Eriksson, The role of thymidine kinases in the
activation of pyrimidine nucleoside analogues, Mini-Rev. Med. Chem. 4 (2004)
341e350.
[36] G.E. Granero, G.L. Amidon, Stability of valacyclovir: Implications for its oral
bioavailability, Int. J. Pharm. 317 (2006) 14e18.
[10] H. Davson, M.B. Segal, Physiology of the CSF and Blood-Brain Barriers, CRC
Press, Boca Raton, 1996.