Synthesis and biological evaluation of 3'-carboranyl thymidine analogues.

Article abstract of DOI:10.1016/S0960-894X(02)00357-8

Boron neutron capture therapy (BNCT) is a chemoradio-therapeutic method for the treatment of cancer. It depends on the selective targeting of tumor cells by boron-containing compounds. One category of BNCT agents with potential to selectively target tumor

Full text of DOI:10.1016/S0960-894X(02)00357-8

Bioorganic & Medicinal Chemistry Letters 12 (2002) 2209–2212
Synthesis and Biological Evaluation of 30-Carboranyl
Thymidine Analogues
Junhua Yan,a Charlotta Naeslund,b Ashraf S. Al-Madhoun,c Jianghai Wang,c
Weihua Ji,a Guirec Y. Cosquer,a Jayaseharan Johnsamuel,a Stefan Sjoberg,b
Staffan Erikssonc and Werner Tjarksa,*
aCollege of Pharmacy, Ohio State University, Columbus, OH 43210, USA
bDepartment of Organic Chemistry, Institute of Chemistry, Uppsala University, Box 531, S-75121 Uppsala, Sweden
cDepartment of Veterinary Medical Chemistry, University of Agricultural Sciences, The Biomedical Centre, Box 575,
S-75123 Uppsala, Sweden
Received 17 January 2002; accepted 22 April 2002
Abstract—Boron neutron capture therapy (BNCT) is a chemoradio-therapeutic method for the treatment of cancer. It depends on
the selective targeting of tumor cells by boron-containing compounds. One category of BNCT agents with potential to selectively target
tumor cells may be thymidine derivatives substituted at the 30-position with appropriate boron moieties. Thus, several thymidine
analogues were synthesized with a carborane cluster bound to the 30-position either through an ether or a carbon linkage. The latter
are the first reported carborane-containing nucleosides in which the carboranyl entity is directly linked to the carbohydrate portion
of the nucleoside by a carbon–carbon bond. Low but significant phosphorylation rates in the range of 0.18% that of thymidine
were observed for the carbon-linked 30-carboranyl thymidine analogues in phosphoryl transfer assays using recombinant prepara-
tions of thymidine kinases 1 (TK1) and thymidine kinases 2 (TK2). Some of the ether-linked 30-carboranyl thymidine analogues
appeared to be slightly unstable under acidic as well as phosphoryl transfer assay conditions and were, if at all, poor substrates for
TK1. # 2002 Elsevier Science Ltd. All rights reserved.
Boron neutron capture therapy (BNCT) is a chemoradio-
therapeutic method for the treatment of cancer that is
based on the capability of 10B to undergo fission after
the capture of external low energy neutrons generating
highly cell toxic He2+ and Li3+ ions. Cells that are
exposed to these high linear energy transfer (LET) fission
products will be killed with high probability. However,
both ions can travel only approximately one cell-diameter
thus damaging only cells in which they are generated.
The prerequisite for successful BNCT is therefore the
selective targeting of tumor cells by boron-containing
compounds.1,2
proliferating tumor cells since the expression of TK1 is
tightly regulated during the cell cycle and the active
enzyme is found only in S-phase cells.3 TK1 is therefore
widely distributed and expressed in all proliferating neo-
plastic cells but it is virtually absent in all non-proliferating
normal tissue.3
4
7
For many years, TK1 was considered to have the most
stringent substrate specificity among all nucleoside kinases
allowing only phosphorylation of native thymidine/
20-deoxyuridine and, to a limited extent, analogues
with minor modifications at either the 5-position or the
3-position.3 Recently, however, we have evaluated several
thymidine and 20-deoxyuridine analogues substituted with
bulky carboranylalkyl groups at N-3 and C-5, respectively,
in phosphoryl transfer assays with recombinant human
TK1. The obtained results indicated that 3-substituted
carboranyl thymidines but not 5-substituted carboranyl
20-deoxyuridines were good substrates for TK1.4
Boronated nucleosides may be promising candidates for
the selective targeting of tumor cells for BNCT because
of their potential metabolic fate. For example, the key
step for the intracellular entrapment of a boronated
thymidine derivative may be the phosphorylation by
cytosolic thymidine kinase (TK1) to the corresponding
50-monophosphate. This would occur primarily in
In order to determine to which degree the TK1 sub-
strate characteristic of thymidine is preserved when its
30-position is substituted with a bulky carborane cluster,
*Corresponding author. Tel.: +1-614-292-7624; fax: +1-614-292-
2435; e-mail: tjarks.1@osu.edu
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00357-8

2210
J. Yan et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2209–2212
we have synthesized several 30-substituted carboranyl
thymidine analogues with varying physicochemical
properties and dimensions and evaluated their potential
as substrates for TK1 in phosphoryl transfer assays
using recombinant enzyme preparations. The substrate
specificities of TK1 and mitochondrial thymidine kinase
(TK2) differ significantly.3 Therefore, we have also
determined the phosphorylation rates of the synthesized
boronated nucleoside analogues with recombinant pre-
parations of the latter enzyme. A thymidine derivative
with a cyanoborane substituent at the 30-position (Fig.
1) has been described previously by Burnham et al.5 and
its phosphorylation rates with TK1 and TK2 were also
determined.
Lebreton et al.12 to give the dibromo-olefin 8 in 74%
yield. Compound 8 was refluxed for 12 h in the presence
of potassium hydroxide in isopropyl alcohol/water (4:1)
to give 30-ethynylthymidine (9) in 39% yield. Under
these reaction conditions, the 50-TBDPS-protective
group and both bromine atoms of 8 were removed and/
or eliminated. The analytical data13 of 9 were identical
to those of the 30-ethynylthymidine prepared previously
by Sahlberg14 using a different synthetic route. The 50-
hydroxyl group of compounds 9 and 10 were then pro-
tected by an acetyl group using acetic anhydride in pyr-
idine/dichloromethane (1/10) for 12 h to provide
compounds 11 and 12 in 85 and 79% yield, respectively.
Compound 11 and 12 were refluxed with a decaborane-
acetonitrile complex7 in benzene for 4 h. Benzene
was evaporated and the residues stirred at room tem-
perature in methanol/H2O in the presence of K2CO3 for
4 h to yield both compounds 13 and 14 in ꢁ10% over-
all. It was difficult to purify compound 14 by column
chromatography because the reaction mixture con-
tained a minor amount of compound 10 and the Rf
values of both compounds were similar in various sol-
vent systems [e.g., 10=0.22; 14=0.28, 9=0.20;
13=0.40 (acetonitrile:toluene=1:1)]. When compound
12 was reacted with a decaborane–acetonitrile complex
and the subsequent the removal of the acetyl group was
carried for 12 h under reflux condition, the nido-car-
boranyl thymidine derivative 15 was obtained in 44%
overall yield.
Selective O-alkylation of the 30-position of compound 26
using propargyl bromide and sodium hydride in refluxing
THF under N2 overnight gave compound 3 in 89%
yield. Compound 4 was obtained by refluxing 3 in toluene
for 6 h with a decaborane–diethyl sulfide complex1 that
was formed in situ. Toluene was evaporated and the
remaining residue was stirred in methanol containing
80% aqueous acetic acid for 30 min at room temperature
to remove the monomethoxytrityl (mMTr) group.
Compound 4 was obtained in 36% overall yield. Base
promoted degradation7 of the closo-carborane cage of
compound 4 using pyrrolidine in dichloromethane at
0 ꢀC for 1 h yielded compound 5 which was precipitated
from aqueous solution as the tetraphenylphosphonium
salt in 76% yield. For phosphoryl transfer assays, com-
pound 5 was converted to the sodium salt using Dowex
50X8–100 (Na+ form) as described previously.7 Target
compound 6 was obtained in 76% yield by selective
30-O-alkylation of 2 using iodopropyl-p-carborane8 fol-
lowed by acidic hydrolysis of the 50-mMTr-protective
group. Reactions conditions applied for the synthesis of
6 were identical to those applied for the synthesis of 3
and 4 (Scheme 1).
The 30-ether linkage in 4 and 5, but not 6, appeared to
be slightly unstable during reactions, workup, and
phosphoryl transfer assays since the appearance of small
quantities of thymidine/thymidine-50-monophosphate
could be detected during these processes. Consequently
the syntheses of compounds 12, 13, and 14 was under-
taken to eliminate the risk of hydrolysis by linking the
carboranyl entity directly to the ribose moiety through a
carbon–carbon bond (Scheme 2). Compounds 7 and 10
were synthesized via multistep syntheses as described by
Sanghvi et al.10 and Fandor et al.,11 respectively. Com-
pound 7 was added to 6-fold excess of the Wittig
reagent (Ph3P¼CBr2) in dichloromethane and reacted
for 24 h in accordance to a procedure described by
Scheme 1. (a) NaH, 80% propargylbromide, THF, reflux, 12 h, 89%;
(b) decaborane (CAUTION),9 diethyl sulfide, toluene, reflux, 6 h; (c)
80% AcOH, MeOH, 40 ꢀC, 4 h; (d) pyrrolidine, CH2Cl2, 0 ꢀC, 1 h,
76%; (e) NaH, iodopropyl-p-carborane, reflux, 12 h.
Figure 1.

J. Yan et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2209–2212
2211
Phosphoryl transfer assays using recombinant human
TK1 and TK2 were carried out as reported previously4
with 1, 4–6, and 13–15 (Table 1). Significant phosphoryl-
ation rates with TK1 could be observed for compounds
1 and 13 indicating that the 30-position of thymidine is
a preferred site for structural modifications that are
tolerated by this important kinase. Significant phosphoryl-
ation was also observed for compound 4. Unfortunately,
the already indicated instability of the ether linkage in 4
prevented an exact quantification of its TK1 phosphoryl-
ation rate. Compound 5, the negatively charged nido-
counterpart of 4, and compound 6, containing a slightly
larger and more lipophilic para-carboranylpropylether
at C-30 instead of the ortho-carboranylmethylether in 4,
were apparently not substrates for TK1. Compound 13
was clearly the best substrate and the activity decreased
drastically with one carbon extension (14) or with the
negatively charged nido form (15) of the carboranyl
substitution. This clearly demonstrates the narrow window
for any type of modification that TK1 can accept. The
TK2 phosphorylation rates for 13 and 14 were slightly
higher than those observed previously for 3-carboranyl
thymidine analogues.4
In conclusion, we have synthesized six new carboranyl
thymidine derivatives. Compounds 4 and 5 appeared to
be hydrolytically cleaved to some extend at the 30-ether
linkage during reactions, work-up, and phosphoryl
transfer assays. This prompted the synthesis of com-
pounds 13, 14, and 15 in which the carboranyl entities
are linked to the ribose moiety through carbon–carbon
bonds. These compounds are the first reported carborane-
containing nucleosides with a carbon linkage between
the carborane cluster and the carbohydrate portion of
the nucleosides. The results of TK1 phosphoryl transfer
assays with compounds 1, 4–6, and 13–15 indicated that
this enzyme, presumably the therapeutically relevant
thymidine kinase isoform, can tolerate bulky and highly
lipophilic carboranyl substituents at the 30-position of
thymidine. Our results suggest that the synthesis and
evaluation of 30-substituted boronated thymidines with
further improved steric and physico-chemical properties
as potential delivery agents for clinical BNCT is war-
ranted.
Acknowledgements
This work was supported by the U.S. Department of
Energy grants DE-FG02–90ER60972 and DE-FG02-
ER61138, and the Swedish Cancer Foundation. The
authors thank Dr. Bernard F. Spielvogel for providing
compound 1.
Scheme 2. (a) Ph3P¼CBr2, CH2Cl2, 25 ꢀC, 12 h, 74%; (b) KOH, i-
PrOH/H2O (4/1), reflux, 12 h, 39%; (c) Ac2O, pyridine/CH2Cl2 (1/10),
25 ꢀC,12 h; (d) decaborane (CAUTION),9 CH3CN, benzene, reflux, 4
h; (e) K2CO3, methanol/H2O, 25 ꢀC, 1 h,ꢁ10%; (f) (only for 12)
K2CO3, methanol/H2O, reflux, 12 h.
Table 1. Phosphorylation of 30-carboranyl thymidine analogues
References and Notes
Compd
TK1
TK2
1. Soloway, A. H.; Tjarks, W.; Barnum, B. A.; Rong, F.-G.;
Barth, R. F.; Codogni, I. M.; Wilson, J. G. Chem. Rev. 1998,
98, 1515.
2. Carlsson, J.; Sjoberg, S.; Larsson, B. Acta Oncol. 1992, 31,
803.
Thd
1
4
5
6
100
5.6 (ꢂ0.3)a
+
100
<0.1
nd
nd
nd
<0.1
<0.1
3. Arner, E. S. J.; Eriksson, S. Pharmacol. Ther. 1995, 67, 155.
4. Lunato, A. J.; Wang, J.; Wollard, J. E.; Anisuzzaman,
A. K. M.; Ji, W.; Rong, F.-G.; Ikeda, S.; Soloway, A. H.;
Eriksson, S.; Ives, D. H.; Blue, T. E.; Tjarks, W. J. Med.
Chem. 1999, 42, 3373.
5. Burnham, B. S.; Chen Shang, Y.; Sood, A.; Spielvogel, B. F.;
Miller, M. C., III; Hall, I. H. Anticancer Res. 1995, 15, 951.
6. Smith, M.; Rammler, D. H.; Goldberg, I. H.; Khorana,
H. G. J. Am. Chem. Soc. 1962, 84, 430.
13
14
15
8.1 (ꢂ0.3)
0.8 (ꢂ0.3)
0.1 (ꢂ0.02)
1.2 (ꢂ0.3)
0.5 (ꢂ0.2)
<0.1
Detailed experimental conditions for the phosphoryl transfer assays
have been described previously.4 Obtained values for thymidine (Thd)
were set to 100. The final DMSO assay concentration was 1%, sub-
strate and ATP concentrations were 100 mM.
aValues are means of three experiments, standard deviation is given in
parentheses. In the cases of 4, 5, and 6 only one measurement was
carried out. nd, Not determined; +, a phosphorylation product could
be detected but it could not be quantified accurately because 4 was
unstable under the assay conditions.
7. Tjarks, W.; Anisuzzaman, A. K.; Liu, L.; Soloway, A. H.;
Barth, R. F.; Perkins, D. J.; Adams, D. M. J. Med. Chem.
1992, 35, 1628.

2212
J. Yan et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2209–2212
8. Tjarks, W.; Ghaneolhosseini, H.; Henssen, C. L. A.;
Malmquist, J.; Sjoberg, S. Tetrahedron Lett. 1996, 37, 6905.
9. Decaborane is a highly toxic, impact sensitive compound,
which forms explosive mixtures especially with halogenated
materials. A careful study of the MSDS is advisable before
usage.
92.28, 111.18, 127.93, 128.70, 130.02, 130.08, 132.50, 132.91,
135.06, 135.40, 135.56, 136.76, 150.38, 163.61; m/z (HRESI)
671.0345, calcd for C28H32N2O4Br2Si (M+Na)+ 671.0378 11:
1H NMR (250 MHz, CDCl3) d 1.90–1.91 (d, 3H), 2.11 (s, 3H),
2.19 (s, 1H), 2.41–2.54 (m, 2H), 2.86–3.12 (m, 1H), 4.11–4.16
(m, 1H), 4.30–4.44 (m, 2H), 6.08–6.12 (t, 1H), 7.28 (s, 1H); 13
C
10. Sanghvi, Y. S.; Ross, B.; Bharadwaj, R.; Debart, F.; De
Mesmaeker, A. Synthesis 1994, 11, 1163.
NMR (62.5 MHz, CDCl3) d 12.48, 20.83, 29.80, 39.17, 62.19,
72.02, 80.33, 82.78, 85.49, 110.80, 135.21, 150.18, 163.79,
170.30; m/z (HREI) 292.1053, calcd for C14H16N2O5 (M)+
292.1059. 12: 1H NMR (250 MHz, CD3OD) d 1.84–1.85 (d,
3H), 2.04 (s, 3H), 2.09 (s, 1H), 2.25–2.37 (m, 5H), 3.95–4.10
(m, 1H), 4.26–4.39 (m, 2H), 5.99–6.04 (t, 1H), 7.51–7.56 (q,
1H); 13C NMR (62.5 MHz, CD3OD) d 12.57, 20.77, 20.98,
30.67, 38.38, 65.17, 71.87, 81.76, 83.65, 86.53, 111.29, 137.73,
152.27, 166.42, 172.36; m/z (HRESI) 307.1272, calcd for
C15H18N2O5 (M+H)+ 307.1294. 13: 1H NMR (250 MHz,
CD3OD) d 1.78 (s, br, 3H), 2.34–2.46 (m, 2H), 3.16–3.25 (m,
1H), 3.72–3.85 (m, 2H), 3.98–4.04 (m, 1H), 4.62 (s, 1H), 5.96
(t, 1H), 7.84 (q, 1H); 13C NMR (62.5 MHz, CD3OD) d 12.4,
41.60, 43.38, 61.95, 64.16, 77.28, 83.48, 86.42, 111.02, 138.19,
152.27, 166.47; m/z (HRESI) 391.2–656, calcd for
C12H24N2O4B10 (M+Na)+ 391.2643. 14: 1H NMR (250 MHz,
CD3OD) d 1.83 (s, br, 3H), 2.08–2.18 (m, 1H), 2.29–2.50 (m,
5H), 3.54–3.82 (m, 3H), 4.53 (s, 1H), 5.98 (t, 1H), 7.81 (q, 1H);
13C NMR (62.5 MHz, CD3OD) d 12.48, 38.63, 40.26, 41.52,
61.45, 64.29, 75.90, 86.50, 111.19, 138.27, 152.31, 166.49; m/z
(HRESI) 405.2837, calcd for C13H26N2O4B10 (M+Na)+
405.2801. 15: 1H NMR (250 MHz, CD3OD) 1.59–1.62 (m,
2H), 1.83 (q, br, 1H), 1.83–1.94 (m, 1H), 1.96–2.56 (m, 5H),
3.56–3.82 (m, 3H), 5.95–5.99 (t, 1H), 7.90–7.93 (q, 1H); 13C
NMR (62.5 MHz, CD3OD) d 12.53, 40.27, 41.17, 43.29, 61.54,
62.68, 86.49, 110.86, 138.51, 152.47, 166.60; m/z (HRESI)
372.2805, calcd for C13H26N2O4B9 (M)ꢃ 372.2777.
11. Fandor, J.; Tam, S. Y. Tetrahedron Lett. 1990, 31, 597.
12. Lebreton, J.; De Mesmaecker, A.; Waldner, A. Synlett
1994, 1, 54.
1
13. 4: H NMR (300 MHz, CD3OD) 1.87 (s, 3H), 2.24–2.26
(m, 1H), 2.49–2.51 (m, 1H), 3.73 (m, 2H), 3.96–4.08 (m, 3H),
4.21 (m, 1H), 4.62 (s, br, 1H), 6.19 (dd, 1H), 7.75 (s, 1H); 13C
NMR (75 MHz, CD3OD) d 12.5, 38.0, 61.1, 63.0, 71.5, 74.7,
82.2, 86.2, 111.8, 138.0, 152.4, 166.3. m/z (HR-FABꢃ)
397.2753, calcd for C13H26B10N2O5(MꢃH)ꢃ 397.2774. 5: 1H
NMR (300 MHz, CD3OD) d ꢃ2.80 (s, br, 1H), 1.87 (s, 3H),
2.14–2.35 (m, 2H), 3.75 (m, 2H), 3.96–4.02 (m, 3H), 4.19 (m,
1H), 6.21 (dd, 1H), 7.48–7.95 (m, 21H); 13C NMR (75 MHz,
CD3OD) d 12.5, 38.5, 63.5, 78.1, 79.2, 86.3, 86.7, 111.6, 118.5,
120.3, 131.6, 136.0, 136.7, 138.3, 152.4, 166.4. m/z (HR-FABꢃ)
388.2736, calcd for C13H26B9N2O5 (M)ꢃ 388.2726. 6: 1H
NMR (300 MHz, CDCl3) d 1.40–1.48 (m, 2H), 1.62–1.71 (m,
2H), 1.91 (s, 3H), 2.25–2.31 (m, 2H), 2.64 (s, br, 1H), 3.21–
3.29 (m, 2H), 3.69–3.76 (m, 1H), 3.88–3.95 (m, 1H), 3.99 (q,
1H), 4.06 (m, 1H), 6.03 (dd, 1H), 7.32 (s, 1H); 13C NMR
(75 MHz, CDCl3) d 12.51, 29.42, 35.54, 36.92, 58.09, 62.84,
68.46, 79.05, 84.98, 87.46, 111.09, 137.04, 150.16, 163.40; m/z
(HR-FABꢃ) 425.3100, calcd for C15H30B10N2O5 (MꢃH)ꢃ
425.3088; 8: 1H NMR (250 MHz, CDCl3) 1.08 (s, br, 9H), 1.60
(d, 3H), 2.26–2.33 (m, 2H), 3.35–3.41 (m, 1H), 3.72–3.86 (m,
2H), 4.00–4.07 (m, 1H), 6.15–6.20 (t, 1H), 6.32–6.35 (d, 1H),
7.34–7.69 (m, 11H), 9.21 (s, 1H); 13C NMR (62.5 MHz,
CDCl3) d 12.19, 19.40, 27.02, 37.70, 42.18, 63.28, 84.32, 84.53,
14. Sahlberg, C. Tetrahedron Lett. 1992, 33, 679.