1. Introduction
Nucleoside analogues are among the first chemotherapeutic agents for the medical
treatment of cancer.1,2 These chemicals include several analogues of physiological
pyrimidine or purine nucleosides and nucleobases, which are thought to interact
with various intracellular targets involved in the metabolism of physiological nucle-
osides and DNA synthesis.3 Troxacitabine, (–)-2 -deoxy-3-oxacytidine (Fig. 1), is a
novel unnatural L-nucleoside analogue with in vitro and in vivo antitumor activity
against human leukemias and a variety of solid tumors.4–6 It is phosphorylated to
947
November 16, 2007 17:48 WSPC/178-JTCC 00342
NH2
OH
O N
O
O
QSAR and Action Mechanism of Troxacitabine Prodrugs with Antitumor Activity 949
2. Calculations
2.1. Computational methods
The molecular structures used in model building are given in Fig. 2. The molecular
mechanics MM2 method in Chem3D software16 was applied to search for lower
energy conformations for each molecule. Then the stable geometric structures of
these compounds were further fully optimized using the DFT calculations at the
level of Becke’s three-parameter hybrid functional (B3LYP),17 together with 6-31G
basis set. There was no imaginary frequency for all configurations at the energy
minima by the frequency calculations, showing that the optimized stable structures
HN R
OH
O N
O
O
are reasonable and reliable. Atomic charges were computed using the natural popu-
lation analysis (NPA).18 All of the quantum chemistry calculations were performed
by using the Gaussian 03 package of programs.19 The Hyperchem software20 was
also employed to calculate the following parameters of molecular properties from the
energy-minimized structures: surface area (S), volume (V ) of the whole molecule,
hydrophobic coefficient (log PR ), and molar refractivity (MRR ) of the substituent
(R), and so on. In addition, the calculated hydrophobic coefficient (C log P ) and
calculated molar refractivity (CMR) of the whole molecule were obtained by Chem-
Draw software.16
QSAR and Action Mechanism of Troxacitabine Prodrugs with Antitumor Activity 951
6l 22.00
6ma 8.80
6n 14.00
6o 13.67
6p F 15.33
6qa Cl 2.77
6r Br 8.87
6s OMe 10.03
Cl
6t 14.33
Cl
QSAR and Action Mechanism of Troxacitabine Prodrugs with Antitumor Activity 953
Table 2. Calculated results for 16 compounds in the training set using Eqs. (1) and (2).
for Eq. (1) for Eq. (1) for Eq. (2) Dev.b for
Comp. (C log P )2 C log P QFR (a.u.) EHOMO (a.u.) pIC50 (expt.) Eq. (2)
6a 2.587 −1.609 −0.7974 −0.2512 4.971 −0.23 −0.53 −0.20 −0.49
6b 0.3031 −0.5506 −0.5671 −0.2500 5.093 0.07 0.08 0.05 0.06
6c 0.0005 −0.0216 −0.5644 −0.2499 5.04 0.52 0.60 0.51 0.59
6d 1.074 1.036 −0.5634 −0.2497 5.62 0.62 0.74 0.63 0.75
6f 6.882 2.624 −0.5634 −0.2496 7.398 −0.43 −0.58 −0.41 −0.55
November 16, 2007 17:48 WSPC/178-JTCC
a The regression (reg.) deviations (dev.) for all compounds, reg. dev. = pIC50 (reg.) −pIC50 (expt.).
b The prediction (pred.) deviation (dev.) for compound i(i = 1, . . . , n), pred. dev. = pIC 50 (pred.)i − pIC50 (expt.)i , here pIC50 (pred.)i is
calculated by the reg. equation obtained after leaving compound i out.
November 16, 2007 17:48 WSPC/178-JTCC 00342
QSAR and Action Mechanism of Troxacitabine Prodrugs with Antitumor Activity 955
Table 4. Computational results for four compounds in the test set using Eq. (1).
Compound (C log P )2 C log P QFR1 (a.u.) pIC50 (expt.) pIC50 (calc.)a Devb
6e 4.387 2.094 −0.5636 6.559 6.766 0.21
6i 47.00 6.856 −0.5634 6.903 7.159 0.26
6m 0.0186 −0.1366 −0.3583 5.056 5.028 −0.03
6q 0.3821 0.6181 −0.1601 5.558 5.112 −0.45
7.5
7.0
6.5
Predicted pIC50
6.0
5.5
5.0
4.5
4.0
Fig. 3. Plot of predicted activities versus experimental ones based on Eq. (1), in which 16 com-
pounds in the training set are expressed by dots and four compounds in the test set are expressed
by triangles.
C log P value (5.798) possesses the highest activity (pIC50 = 7.699) among this
data set, whereas compound 6a with smaller C log P value (−1.609) has lower activ-
ity (pIC50 = 4.971); compound 6k with larger C log P value (8.972) also has lower
activity (pIC50 = 6.046). Such a result suggests that a very low hydrophobicity may
November 16, 2007 17:48 WSPC/178-JTCC 00342
8.0
7.5
7.0
Experimental pIC50
6.5
6.0
5.5
5.0
4.5
-2 0 2 4 6 8 10
ClogP
restrict the transport of pharmaceutical molecule into active body tissues through
membranes, whereas a very high hydrophobicity may be disadvantageous to their
hydrolysis processes, which are necessary procedures resulting in a ultima specimen
with anticancer activity for these prodrugs.11 In addition, the diffusion of dots in
Fig. 4 implies that there are also other fators influencing the cytotoxicity.
On the other hand, the electronic parameter QFR exhibits negative correla-
tion coefficient with pIC50 (R = −0.486), so the more negative the QFR value,
the higher the activity of the compound is. Such a regularity can be easily found
through comparisons between compounds 6b and 6c, 6n and 6o, and other ones.
The more negative charges at the first atom of substituent R mean the stronger
electron-withdrawing ability of this atom and then make the positive charge of C
atom of adjacent carbonyl increase, so it must be advantageous to the nucleophilic
attack of water molecules for this C atom in active body tissues, and thus advan-
tageous to its hydrolysis processes resulting in a ultima specimen with anticancer
activity. Such a result is in good agreement with the hypothesis mechanism pro-
posed by Radi et al. who suggested that the amidase-catalyzed hydrolysis of the
prodrugs might play an important role, allowing the release of troxacitabine into
the intracellular compartment for triphosphate formation.11 In this QSAR model,
the selected descriptor QFR is closely correlative with the anticancer activity of
these compounds, and thus further supports Radi’s hydrolysis action mechanism of
the prodrugs, at least, in the preliminary procedure.
November 16, 2007 17:48 WSPC/178-JTCC 00342
QSAR and Action Mechanism of Troxacitabine Prodrugs with Antitumor Activity 957
4. Conclusions
The QSAR of this series of troxacitabine prodrugs with antitumor activity against
non-small-cell lung cancer cell line SW1573 has been studied by using the DFT,
MM2, and statistical methods. It is very interesting to find that the cytotoxicity
of this kind of compound mainly depends on two factors: the calculated hydropho-
bic coefficient (C log P ) and the net charge variable (QFR ) of the first atom of
substituent R. The established QSAR shows not only acceptable regressive perfor-
2
mance (RA = 0.867) but also considerable predictive ability (q 2 = 0.807). Moreover,
the predicted pIC50 values for the four compounds in test set are in a considerable
agreement with corresponding experimental ones. Such a result further confirms
this established QSAR having excellent predictive ability. Based on this QSAR
model, some useful insights are obtained as follows: (1) considering a suitable range
of lipophilicity (C log P of ca. 5.6) is very important to obtain the higher antitumor
activities for this series of prodrugs. (2) Selected another descriptor, i.e. the net
charges of the first atom of substituent R (QFR ) further supports the hydrolysis
action mechanism of the prodrugs, at least, in the preliminary procedure. These
theoretical results can be used as references for understanding the action mecha-
nism and directing the molecular design of this kind of compound with antitumor
activity.
Acknowledgments
The financial supports of the National Natural Science Foundation of China and
the Natural Science Foundation of Guangdong Province of China are gratefully
acknowledged.
References
1. Jordheim L, Galmarini CM, Dumontet C, Lancet Oncol 3:415, 2002.
2. Jordheim L, Galmarini CM, Dumontet C, Curr Drug Targets 4:443, 2003.
3. Helguera AM, Rodrı́guez-Borges JE, Garcı́a-Mera X, Fernández F, Cordeiro
MNatáliaDS, J Med Chem 50:1537, 2007.
4. Gourdeau H, Bibeau L, Ouellet F, Custeau D, Bernier L, Bowlin T, Cancer Chemother
Pharmacol 47:236, 2001.
5. Rabbani SA, Harakidas P, Bowlin T, Attardo G, Cancer Res 58:3461, 1998.
6. Weitman S, Marty J, Jolivet J, Locas C, Von Hoff DD, Clin Cancer Res 6:1574, 2000.
7. Kukhanova M, Liu SH, Mozzherin D, Lin TS, Chu CK, Cheng YC, J Biol Chem
270:23055, 1995.
8. Gunzburg WH, Salmons B, Trends Mol Med 7:30, 2001.
9. Gourdeau H, Clarke ML, Ouellet F, Mowles D, Selner M, Richard A, Lee N, Mackey
JR, Young JD, Jolivet J, Lafreniere RG, Cass CE, Cancer Res 61:7217, 2001.
10. De Bono JS, Stephenson J, Baker SD, Hidalgo M, Patnaik A, Hammond LA, Weiss
G, Goetz A, Siu L, Simmons C, Jolivet J, Rowinsky EK, J Clin Oncol 20:96, 2002.
11. Radi M, Adema AD, Daft JR, Cho JH, Hoebe EK, Alexander LEMM, Peters GJ,
Chu CK, J Med Chem 50:2249, 2007.
12. Isayev O, Rasulev B, Gorb L, Leszczynski J, Mol Divers 10:233, 2006.
November 16, 2007 17:48 WSPC/178-JTCC 00342
13. Chen JC, Qian L, Wu WJ, Chen LM, Zheng KC, J Mol Struct Theochem 756:167,
2005.
14. Wu WJ, Chen JC, Qian L, Zheng KC, J Theor Comput Chem 2:223, 2007.
15. Lameira J, Alves CN, Moliner V, Silla E, Eur J Med Chem 41:616, 2006.
16. CambridgeSoft Corp., 100 Cambrige Park, MA, 02140-2317, USA.
17. Becke AD, J Chem Phys 98:1372, 1993.
18. Carpenter JE, Weinhold F, J Mol Struct Theochem 169:41, 1988.
19. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Mont-
gomery JA, Vreven TJr, Kudin KN, Burant JC, Millam JM, Iyengar SS, Tomasi J,
Barone V, Mennucci B, Cossi M, Scalmani G, Rega N, Petersson GA, Nakatsuji H,
Hada M, Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda
Y, Kitao O, Nakai H, Klene M, Li X, Knox JE, Hratchian HP, Cross JB, Bakken V,
Adamo C, Jaramillo J, Gomperts R, Stratmann RE, Yazyev O, Austin AJ, Cammi R,
Pomelli C, Ochterski JW, Ayala PY, Morokuma K, Voth GA, Salvador P, Dannen-
berg JJ, Zakrzewski VG, Dapprich S, Daniels AD, Strain MC, Farkas O, Malick DK,
Rabuck AD, Raghavachari K, Foresman JB, Ortiz JV, Cui Q, Baboul AG, Clifford S,
Cioslowski J, Stefanov BB, Liu G, Liashenko A, Piskorz P, Komaromi I, Martin RL,
Fox DJ, Keith T, Al-Laham MA, Peng CY, Nanayakkara A, Challacombe M, Gill
PMW, Johnson B, Chen W, Wong MW, Gonzalez C, Pople J, Gaussian 03, Revision
D.1, A. Gaussian, Inc., Wallingford, CT, 2005.
20. HyperChem, release 6.03; Hypercube Inc., http://www.hyper.com/, 2001.
21. SPSS, vs 9.0; SPSS Inc., Chicago, IL.
22. De K, Sengupta C, Roy K, Bioorg Med Chem 12:3323, 2004.
23. Tetko IV, Tanchuk VY, Villa AE, J Chem Inf Comput Sci 41:1407, 2001.
24. Golbraikh A, Tropsha A, J Mol Graph Model 20:269, 2002.
25. Zhang S, Wei L, Bastow K, Zheng W, Brossi A, Lee KH, Tropsha A, J Comput Aided
Mol Des 21:97, 2007.
26. Roy DR, Sarkar U, Chattaraj PK, Mitra A, Padmanabhan J, Parthasarathi R, Sub-
ramanian V, Van Damme S, Bultinck P, Mol Divers 10:119, 2006.
27. Topliss JG, Costello RJ, J Med Chem 10:1066, 1972.