QSAR STUDY TO PREDICT ANTI-AMOEBIC ACTIVITIES OF PYRAZOLINE AND DIOXAZOLE DERIVATIVES WITH THE HELP OF PM5-BASED DESCRIPTORS

QSAR STUDY TO PREDICT ANTI-AMOEBIC ACTIVITIES OF PYRAZOLINE AND DIOXAZOLE DERIVATIVES WITH THE HELP OF PM5-BASED DESCRIPTORS

Anil K. Srivastava* and Ratna Gupta

Department of Chemistry, M.L.K. (P.G.) College, Balrampur, Uttar Pradesh, India

ABSTRACT

In quest of better anti-amoebic agents, quantitative structure-activity relationship (QSAR) studies were performed on a series of pyrazoline & dioxazoles derivatives with the help of PM5 calculations and geometry optimizations using CAChe software. Multiple Linear Regression (MLR) analysis was performed to derive QSAR models using the descriptors, molecular weight (M_W), conformation minimum energy (ɛ), HOMO energy (e_HOMO), shape index, basic kappa second order (k2), absolute hardness (h), electronegativity (c), electrophilicity index (ω), molar volume (MV), molar refractivity (MR), LogP (LP), parachor (Pc) and solvent accessibility surface area (SASA). The QSAR models equations of anti-amoebic agents have been developed by using maximum of seven descriptors, in which conformation minimum energy, shape index, molar volume and parchor were present have good predictive powers of correlation coefficients. These models can successfully predict the anti-amoebic activity of any newly discovered pyrazoline and dioxazole derivatives which can later be tested in laboratory.

QSAR models

INTRODUCTION: Parasitic infections such as amoebiosis and other protozooses are still major threats against public health, especially in developing countries and the intestinal protozoan Entamoeba histolytica is a major cause of morbidity and mortality  ^{1, 2}. Infection occurs through the oral uptake of the pathogen in its cyst form, with contaminated food or water.

Despite its socio-economic importance, intestinal and extra-intestinal amoebiasis is not yet officially listed among the “neglected infectious diseases”, obviously due to difficulties in developing effective control strategies like studies involving drug molecules and hygiene management. Amoebiasis is primarily treated with the drug metronidazole which has significant side-effects ^3-6.

Diloxanide furoate, a luminal amoebicide can be used for the treatment of oligosymptomatic and asymptomatic carriers of E. histolytica where as chloroquine is a useful support to other medications in the management of invasive amoebiasis ⁷.

The available anti-amoebic drugs have short-comings regarding tolerability and efficacy and the range of medicaments available for the treatment of amoebiasis has not changed over the past decade.

Recent studies tried to improve the treatment of this infection by developing anti-amoebic therapy ^{8, 9}, a set of dioxazoles derivatives showed better activity than the reference drug metronidazole; besides being non-toxic to human kidney epithelial cells. Recently QSAR studies have been quite helpful to identify important structural parameters responsible for anti-amoebic activity and a number of industrial research units are using classical as well as 3D QSAR techniques for contemporary drug design ^10-15.

The basis of QSAR method is use of molecular descriptors which represent the structural, stereochemical and topological features of the target molecule ^16-20. Recently our group is engaged in finding new drugs using QSAR study ^{21, 22}, herein we have taken, a series of 63 1-N-substituted thiocarbamoyl-3-phenyl-2-pyrazolines ²³ and 3, 5-substituted-1, 4, 2-dioxazoles ⁸ were subjected to QSAR study by choosing appropriate molecular descriptors incorporating important structural features of the target molecule.

A multiple linear regression (MLR) analysis was executed to obtain and select best models in the form of regression equations to predict the anti-amoebic activity of chosen molecules.

MATERIALS AND METHODS: The experimental IC₅₀ (μM) of anti-amoebic activities of 1-N-substituted thiocarbamoyl- 3- phenyl- 2- pyrazolines and 3, 5-substituted-1, 4, 2-dioxazoles are collected from recent publications ^{23, 8}. We have chosen the values of experimental observed activity and converted them into logarithmic scale of -logIC₅₀ and are placed in Tables 1-4.

–logIC₅₀ can be defined as, “It is negative of logIC₅₀ value and because of negative sign, its magnitude has an inverse relationship with the biological activity or drug potency of the selected molecules”. Consequently a low magnitude of -logIC₅₀ predicts a higher biological value and a high magnitude of -logIC₅₀ indicates lower potency.

QSAR studies of the compounds listed in Tables 1-4 have been made with the help of following quantum chemical and topological descriptors-

1. Molecular weight                                           M_W
2. Conformation minimum energy                     ɛ
3. HOMO energy                                                         e_HOMO
4. Shape index, basic kappa second order         k2
5. Absolute hardness                                           h
6. Electronegativity                                             c
7. Electrophilicity index                                      ω
8. Molar volume                                                 MV
9. Molar refractivity                                           MR
10. LogP                                                                LP
11. Parachor                                                         Pc
12. Solvent accessibility surface area               SASA

PM5 based calculations of the above descriptors have been made on the compounds listed in Tables 1-4 with the help of Cache Software and their relationship with the known activity of the anti-amoebic drugs have been studied by developing QSAR models. The values of the descriptors have been used to prepare Multiple Linear Regression (MLR) equations for predicted activities and compared with the known activity. The correlation coefficient and cross-validation coefficient have been evaluated to adjudge the quality of QSAR model and its predictive power.

RESULT AND DISCUSSION: Descriptors in different combinations have been used for Multiple Linear Regression (MLR) analysis. The predicted activity obtained by regression equation has been examined for selecting QSAR models, which have high degree of predictive power; the correlation coefficient and cross validation coefficient of all the regression equation have been evaluated.

The best QSAR model and the combination of descriptors providing that model have been identified. On the basis of such models new derivatives can be proposed which may have better anti-amoebic activity.

Cache software has been used for the calculation of descriptors of pyrazoline and dioxazole derivatives. At first, we have optimized the geometry by using PM5 Hamiltonian and then calculated the values of descriptors with the help of project leader associated with cache programme. Values of quantum chemical and topological descriptors of anti-amoebic agents are included in Table 5.

TABLE 1: PYRAZOLE DERIVATIVES AND THEIR OBSERVED ANTI-AMOEBIC ACTIVITIES –logIC₅₀

TABLE 2: DIOXAZOLE DERIVATIVES AND THEIR OBSERVED ANTI-AMOEBIC ACTIVITIES –logIC₅₀

We have also calculated the predicted activity of anti-amoebic agents PA1-PA5 by substituting the values of descriptors in MLR equations. These values are listed in Table 6.

Several QSAR models in different combination of descriptors have been tried and five models were chosen from best five equations, whose correlation coefficients values are above 0.80. The descriptors used in these models are presented in Table 7 and the QSAR model equations after the table numbered as 1, 2, 3, 4 and 5 and their graphs (1-5), respectively.

   TABLE 3: DIOXAZOLE DERIVATIVES AND THEIR OBSERVED ANTI-AMOEBIC ACTIVITIES –logIC₅₀

 TABLE 4: DIOXAZOLE DERIVATIVES AND THEIR OBSERVED ANTI-AMOEBIC ACTIVITIES –logIC₅₀

MOPAC 2000 engine was used for calculating the value of descriptors of pyrazoline and dioxazole derivatives after optimizing the geometry by using PM5 Hamiltonian. These values are presented in Table 5.

TABLE 5: THE VALUES OF QUANTUM CHEMICAL AND TOPOLOGICAL DESCRIPTORS FOR ANTI-AMOEBIC AGENTS

Com = Compound, ɛ = Conformation minimum energy (kcal/mole), LP = LogP, Mw = Molecular weight, k2 = Shape Index (basic kappa, order 2), ɛ_HOMO = HOMO energy, χ = Electronegativity, η = Absolute hardness, ω = Electrophilicity index, MR = Molar refractivity,  SASA = Solvent accessibility surface area, MV = Molar volume, Pc = Parachor

TABLE 6: CALCULATED PREDICTED ACTIVITIES FROM REGRESSION EQUATIONS PA1 TO PA5

PA = Predicted activity derived from various QSAR model equations

Examination of Table 6 suggests that compounds No. 1, 3, 7, 9, 10, 11, 17, 19, 20 and 21 of pyrazoline group show predicted activity almost at par with the observed activity in all the five selected models (PA1 to PA5).

The same is true for compounds No. 33, 35, 46, 47, 49, 55, 56, 57, 60 and 61 of dioxazole group which show highly comparable predicted activity with observed activity in all the selected five models (PA1 to PA5).

TABLE 7: VALUES OF CROSS VALIDATION AND CORRELATION COEFFICIENTS OF BEST FIVE QSAR MODELS

PAE = Predicted activity equations, rCV^2 = Cross validation coefficient, r^2 = Correlation coefficient

QSAR MODEL EQUATION-1

PA1= 0.00417292*ɛ+0.802645*χ+0.00181302*Mw-0.51571*k2+0.00797046*SASA+0.0951141*MV-0.0337697*Pc+4.22522

rCV^2 = 0.756404                                                                                                                                                               ----(1)

r^2 = 0.806883

GRAPH 1: CORRELATION BETWEEN OBSERVED ACTIVITY AND PREDICTED ACTIVITY DERIVED FROM REGRESSION MODEL PA1

QSAR MODEL EQUATION- 2

PA2 = 0.00447077*ɛ+12.1849*χ+6.49678*η+10.2269*ω-0.556798*k2+ 0.0870785*MV-0.0274321*Pc+3.43851

rCV^2 = 0.729243                                                                                                                                                                ---(2)

r^2 = 0.806268

GRAPH 2. CORRELATION BETWEEN OBSERVED ACTIVITY AND PREDICTED ACTIVITY DERIVED FROM REGRESSION MODEL PA2

QSAR MODEL EQUATION-3

PA3 = 0.00337618*ɛ+0.932314*χ-0.517113*k2-0.0115611*MR+0.0161658*SASA+ 0.0869679*MV-0.0292388*Pc+4.42091

rCV^2 = 0.747698                                                                                                                                                               ----(3)

r^2 = 0.804047

GRAPH 3. CORRELATION BETWEEN OBSERVED ACTIVITY AND PREDICTED ACTIVITY DERIVED FROM REGRESSION MODEL PA3

 QSAR Model Equation-4

PA4 = 0.00423129*ɛ+0.761142*χ+0.00256411*Mw-0.477313*k2+ 0.0198134*LP +0.0920914*MV-0.0321332*Pc+4.25876

rCV^2 = 0.760651                                                                                                                                                               ----(4)

r^2 = 0.803411

GRAPH 4. CORRELATION BETWEEN OBSERVED ACTIVITY AND PREDICTED ACTIVITY DERIVED FROM REGRESSION MODEL PA4

QSAR MODEL EQUATION-5

PA5 = 0.00379732*ɛ+0.824393*χ-0.508899*k2+0.0200405*LP+ 0.0118551*SASA+ 0.0895119*MV-0.0313529*Pc+4.05228

rCV^2 = 0.750508                                                                                                                                                               ---(5)

r^2 = 0.802723

GRAPH 5. CORRELATION BETWEEN OBSERVED ACTIVITY AND PREDICTED ACTIVITY DERIVED FROM REGRESSION MODEL PA5

These equations contain various descriptors in different combinations and each descriptor has a positive or negative co-efficient attached to it. These coefficients along with the value of descriptor have a significant role in deciding the overall biological activity of the molecule as discussed below. Examination of selected equation shows that coefficients of each parameter play an important role in deriving the biological activity. From the point of view of potency or biological activity of the drug molecule in terms of -logIC₅₀ values, the weight of a negative co-efficient is very significant because it contributes towards a decreased value of -logIC₅₀, meaning increased value of biological activity. So the parameters with a negative co-efficient are most important followed by parameters with low weight positive coefficients and lastly the parameters with high weight positive coefficients.

On the basis of values of these coefficients, the associated descriptors are arranged in a sequence pertaining to their contribution towards overall biological activity of the molecule, in following decreasing order of biological activity of anti amoebic agents;

Shape Index (k2) > Parachor (Pc) > Conformation Minimum Energy (ɛ) and/or Molecular Weight (Mw) and/or Molecular Refractivity (MR) > Solvent Accessibility Surface Area (SASA) > LogP > Molar Volume (MV) > Electronegativity (χ), eHOMO_, Electrophilicity Index (ω), Absolute Hardness (η)

CONCLUSION: The QSAR models developed by us in this paper represent some of the easiest ways of determining the biological activity of anti-amoebic agents. All the models are highly predictive and provides good values for cross validation coefficient (rCV^2) and correlation coefficient (r^2). Study and analysis of these models reveal that negative coefficients of regression model are most significant followed by positive coefficients of low weight and finally positive coefficients of high weight. The whole intention behind this was to facilitate the designing of new anti-amoebic drugs for the treatment against E. histolytica.

ACKNOWLEDGEMENT: We gratefully acknowledge to Dr. O. P. Mishra, Principal of M.L.K. (P.G.) College, Balrampur for providing research facilities and co-operation.

REFERENCES: 

1. Walsh JA, “Prevalence of Entamoeba histolytica infection” In Amoebiasis: Human Infection by Entamoeba histolytica (Ravdin, J. I., Ed.), 1988, pp. 93-105. Wiley Medical, New York, NY, USA.
2. Walsh JA, Reviews in Infectious Diseases. 1986; 8: 228.
3. Goldman P, Koch RL, Yeung TC et al., Biochemical Pharmacology 1986; 35: 43.
4. Knight RC, Skolioowski IM, Edwards DI. Pharmacology 1986; 27: 2089.
5. Koch CJ, Lord EM, Shapiro IM et al. Advances in Experimental Medicine and Biology 1997; 428: 585.
6. Kock RL, Beaulieu BB Jr, Chrystal EJT et al. Science 1981; 211: 398.
7. Bansal D, Sehgal R, Chawla Y, Chander MR, Malla N. Annals of Clinical Microbiology and Antimicrobials 2004; 3: 27.
8. Bhat AR, Athar F, Azam A. Eur. J. Med. Chem. 2009; 44: 926-936.
9. Bhat AR, Athar F, Azam A. Eur. J. Med. Chem. 2009; 44: 426-431.
10. Adhikari N, Maiti MK. J. Bioorg. Med. Chem. Lett. 2010; 20: 4021.
11. Mbarki S, Dguigui K, El Hallaoui M. J. Mater. Environ. Sci. 2011;2: 61-70.
12. Trinajstic N. Chemical Graph Theory 1992; 225-273.
13. a) Bazoui H, Zahouily M, Boulaajaj S, Sebti S, Zakarya D., SAR QSAR Environ. Res. 2002; 13: 567. b) Bazoui H, Zahouily M, Sebti S, Boulaajaj S, Zakarya DJ. Mol. Model. 2002; 8: 1-7. c) Agrawala VK, Singha J, Mishra KC, Khadikar PV, Jaliwalac YA. Arkivoc 2006; 162.
14. Kubinyi H. QSAR: Hansch analysis and related approaches. In: Mannhold R, Krogsgarrd Larsen P, Timmerman H (eds) Methods and principles in medicinal chemistry. Wiley, Weinheim; 1 (1993).
15. Kubinyi H. (ed) 3D QSAR in drug design: theory, methods and applications. ESCOM, Leiden (1993).
16. Todeschini R, Consonni V. Handbook of Molecular Descriptors; Wiley-VCH: (2000) Germany.
17. Karelson M. Molecular Descriptors in QSAR/QSPR; John Wiley & Sons: New York, (2000).
18. Diudea MV. Ed. QSPR/QSAR Studies by Molecular Descriptors; Nova Science: Huntington, NY, (2001).
19. Balaban AT. Ed. From Chemical Graphs to Three-Dimensional Geometry; (1997). New York.
20. Balaban A. SAR QSAR Environ. Res. 1998; 8: 1.
21. Srivastava AK, Gupta R. J. Chem. Pharm. Res. 2012; 4: 2228-2241.
22. Srivastava AK, Gupta R, Srivastava R, Mishra DK. Int. J. Pharm. Sci. Res. 2012; 8: 2648-2654.
23. Abid M, Bhat AR, Athar F, Azam A. Eur J. Med. Chem. 2009; 44: 417-425.
    

    ---

    How to cite this article:

    Srivastava AR and Gupta R: QSAR Study to predict Anti-Amoebic Activities of Pyrazoline and Dioxazole Derivatives with the help of Pm5-Based Descriptors. Int J Pharm Res Sci. 3(9); 3249-3258.

    ---