Abstract
Background: In order to extend the search for tumour-targeting compounds, this study performed a quantitative structure–activity relationship (QSAR) analysis of 26 newly synthesised trihaloacetylazulenes. Materials and Methods: The value of 50% cytotoxic concentration (CC50) of trihaloacetylazulenes against human oral squamous cell carcinoma (HSC-2, HSC-3, HSC-4) and human promyelocytic leukaemia (HL-60) cell lines was calculated from the dose–response curve by the MTT method. CONFLEX/CAChe PM5 was used for the calculation of 11 physico-chemical features for each compound. Results: When all 26 compounds were analysed together, the CC50 values correlated well with the dipole moment, the lowest unoccupied molecular orbital energy and the reactivity index, and somewhat with the heat of formation, the stability of hydration and the absolute electron negativity, but not with other features. When these 26 compounds were separated into two groups with or without substituents of the 7-membered ring, the correlation coefficient was increased. When the 26 compounds were separated into a different set of two groups with different halogen atoms in the 5-membered ring, no improvement of correlation coefficient was observed. Conclusion: The present study suggests that appropriate grouping of test compounds may further increase the correlation coefficient. The QSAR approach was useful in the design of azulene compounds that are expected to show higher potency.
Quantitative structure–activity relationship (QSAR) represents the quantitative relationship between the structure and the biological (pharmaceutical or toxicological) activity of chemical entities. QSAR analyses are possible, based on previously published data, and provide a powerful tool for predicting the pharmacological potency of structurally related compounds (Figure 1).
Azulenes have shown antibacterial (1), anti-ulcer (2) and chemotherapeutic activities against mucous membrane diseases (3). The cytotoxicity and type of cell death induced by 20 trihaloacetylazulenes in human tumour cell lines was previously reported (4). Using a conventional semi-empirical molecular-orbital method (PM5) (5-7) and chemical hardness (8), QSAR analysis of these trihaloacetylazulenes has been recently performed (9). It was found that their 50% cytotoxic concentrations (CC50) showed good correlation with all physicochemical features, except for absolute hardness in human promyelocytic leukaemia HL-60 cells. CC50 was also well correlated with the highest occupied molecular orbital energy (EHOMO), the absolute electron negativity (χ), the reactivity index (ω), the surface area, the volume and the molecular weight in human oral squamous cell carcinoma cell lines (HSC-2, HSC-3 and HSC-4) (9).
As an extension of the search for azulene compounds that show highly selective actions against various human tumour cell lines, the present study performed a QSAR analysis of 26 newly synthesised trihaloacetylazulenes (Figure 2).
Materials and Methods
Assay for cytotoxicity. HL-60 and oral squamous cell carcinoma cell lines were cultured in RPMI-1640 or DMEM supplemented with 10% heat-inactivated foetal bovine serum, in a 5% CO2 atmosphere at 37°C. The cells were incubated for 48 h with various concentrations of 26 newly synthesised trihaloacetylazulenes (Figure 2) and the CC50 value was determined from the dose–response curves (9).
Flow chart of QSAR analysis.
Calculation. The most stable configuration of the 26 trihaloacetylazulenes was calculated by CONFLEX 5 (Conflex Co. Ltd., Tokyo). The optimisation of the structure was achieved using the PM5 method using a CAChe Worksystem 4.9 (MOPAC, PM5, non-COSMO, COSMO) (Fujitsu Co. Ltd., Tokyo) (Figure 1).
The following eleven chemical descriptors and one other parameter were used: heat of formation (COSMO, non-COSMO; kcal/mole); stability of hydration (=COSMO – nonCOSMO (ΔH); kcal/mole); dipole moment (D); hydrophobicity (log P); highest occupied molecular orbital energy (EHOMO; eV); lowest unoccupied molecular orbital energy (ELUMO; eV); absolute hardness [η= (EUMO−EHOMO)/2; eV)]; absolute electron negativity [χ=−(ELUMO + EHOMO)/2; eV]; reactivity index (ω=χ2/2η; eV); surface area of the molecule (Å2); volume of the molecule (Å3); and molecular weight (Figure 1).
The QSAR was investigated from each feature (determined from the molecular structure) and the CC50 values (plotted on a logarithmic scale) (9), using a CAChe Worksystem 4.9 project reader (Figure 1).
Results and Discussion
Most stable structure. The most stable configurations of trihaloacetylazulene compounds calculated with CONFLEX/PM5 are shown in Figure 3. The CC50 value and 11 chemical features and molecular weight of each compound are shown in Table I.
QSAR with unseparated compounds. When QSAR between the cytotoxicity (evaluated by CC50) and chemical feature was performed with all 26 compounds (Figure 2), the cytotoxicity showed fairly good correlation with dipole moment, ELUMO (r2 range, 0.490 to 0.609) and ω (r2 range, 0.471 to 0.581) (Figure 4, Exp. 1 in Table II). These ELUMO and ω values provide the information of QSAR regarding the relationship between molecular electric state and CC50. The ELUMO value is an indication of electron-withdrawing capability. The lower the ELUMO value, the greater the electron-capturing capability. The cytotoxicity showed some correlation with the heat of formation (r2 range, 0.280 to 0.435), ΔH (r2 range, 0.224 to 0.308), or χ (r2 range, 0.307 to 0.380), but almost no correlation with the other features (Exp. 1 in Table II). A previous study described the correlation between the biological activity and two chemical hardness (η, χ) of environmental hormones, as one of the QSAR analyses (8). In contrast, the present study could not demonstrate such correlation between CC50 and η.
Chemical structure of the 26 trihaloacetylazulenes used in this study.
QSAR with two separated groups of compounds. In order to obtain markedly improved QSAR, the 26 compounds were separated into two groups, those without (A-1) [1-14] and those with (A-2) substitutes in the 7-membered ring skeleton [15-26] (Figure 2), and subjected to QSAR analysis (Figure 5, Exp. 2 in Table II). In the A-1 group [1-14], the cytotoxicity showed good correlation, especially with ELUMO (r2 range, 0.570 to 0.878) and ω (r2 range, 0.492 to 0.815) (much higher r2 value). In the A-2 group [15-26], the cytotoxicity showed fairly good correlation with dipole moment (r2 range, 0.667 to 0.818), ω (r2 range, 0.319 to 0.580), surface area (r2 range, 0.296 to 0.560), volume (r2 range, 0.369 to 0.652) and molecular weight (r2 range, 0.245 to 0.573) of the molecule. This means that the higher the r2 value, the closer the estimated CC50 value to the regression line.
Cytotoxicity (CC50) and chemical descriptors for the 26 trihaloacetylazulenes used in this study.
Correlation coefficients (r2) between CC50 and each chemical descriptor in four different cell lines.
Most stable structure of the 26 trihaloacetylazulenes as determined by CAChe (PM5).
Correlation between CC50 value (log scale) and descriptors of all 26 trihaloacetylazulene derivatives.
Correlation between CC50 value (log scale) and descriptors of trihaloacetylazulene derivative groups A-1 [1-14] and A-2 [15-26].
When the 26 compounds were separated [by the difference in halogen element (fluoride or chloride) of the substituents] into another set of two groups, namely the COCF3 group (B-1) [1-7, 15-17, 19, 23, 25] and the COCl3 group (B-2) [8-14, 18, 20-22, 24, 26] in the 5-membered ring skeleton (Figure 1), no improvement in the correlation coefficient was observed (r2≤0.544 and r2≤0.595, respectively) (Figure 6, Exp. 3 in Table II). The lower correlation coefficient exhibited by the 13 compounds of group B-1 compared to the 13 compounds of group B-2 (Table I) may be explained by a reduction in the range of cytotoxic concentration of the test compounds.
Correlation between CC50 value (log scale) and descriptors of trihaloacetylazulene derivatives groups B-1 [1-7, 15-17, 19, 23, 25] and B-2 [8-14, 18, 20-22, 24, 26].
Conclusion
Based on the present QSAR analysis between the CC50 and 11 physico-chemical descriptors of newly synthesised trihaloacetylazulenes, it is suggested that the trihaloacetylazulenes should be separated into several groups before QSAR analysis in order to obtain a better correlation efficient. The chemical descriptors, determined by this CONFLEX/PM5, are useful as tools by which the effects of different chemical substituents on the biological activity can be analyzed. Furthermore, the resulting QSAR analysis may be useful to estimate the cytotoxicity of structurally related compounds and may provide the basis for the future design of more active compounds.
- Received January 25, 2011.
- Revision received January 25, 2011.
- Accepted January 25, 2011.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.