ABSTRACT
The Quantitative structure-activity relationship (QSAR) analyses were carried out for series of 45 naphthylisoquinoline and 22 bisbenzylisoquinoline alkaloids to find out the structural requirements of their antimalarial activities as well as the cytotoxicity using molecular descriptors. All the structures before the calculation of the descriptors were fully optimised in Spartan 14 v.1.1.0 using density functional B3LYP and the standard Pople‘s basis set of 6-311G*. The combination of the Genetic algorithm (GA) and Multiple Linear Regression (MLR) analyses methods were applied to derive QSAR models from the descriptors. For the selection of the best descriptors, the elimination selection stepwise regression method was utilised. The statistically significant models with parameters; R2 = 0.7483, Q2 = 0.6042 for activity of naphthylisoquinoline and R2 = 0.6256, Q2 = 0.7865 for the cytotoxicity were obtained. Furthermore, the models for activity and cytotoxicity of bisbenzylisoquinoline give R2 = 0.8845, Q2 = 0.7942 and R2 = 0.8742, Q2 = 0.7115 statistical parameters, respectively. The activity models generated, revealed descriptors BCUTc-1l (encoding connectivity information and atomic properties of the molecule), SssCH2 (the electro-topological state indices for number of –CH2 group connected with two double single bonds), minHsOH (electrotopological state atom type descriptor), Wlambda2.unity (encoding molecular 3D information regarding molecular size, shape, symmetry), μ, dipole moment and logP, lipophilicity, to be responsible for the activities of naphthylisoquinoline, while from the cytotoxicity models, MDEC-33 descriptor, which stand for molecular distance edge between all tertiary and quaternary carbons, determined the cytotoxicity. In the bisbenzylisoquinoline alkaloids, the descriptors responsible for the activities are BCUTw-1h (encode atomic properties relevant to intermolecular interactions), CrippenlogP, lipophilicity and maxssO (electrotopological state atom type descriptor) while ATSc4 (autocorrelation
vii
descriptor, weighted by charges), VCH-6 (describing the overall topology of the molecules) and SHBint9 (strength for potential hydrogen bonds of path length 9) descriptors causes cytotoxicity. The accuracy of the proposed MLR models were illustrated using the following evaluation techniques: cross validation, validation through an external test sets, and Y-randomisation. Furthermore, the domain of applicability which indicates the area of reliable predictions was defined. The good correlation between experimental and predicted biological activity/cytotoxicity for the four selected models further highlights the reliability of the constructed QSAR models.
TABLE OF CONTENTS
Cover page i
Title pageii
Declarationiii
Certificationiv
Acknowledgementsv
Dedicationvi
Abstractvii
Table of contents
List of Tables xii
List of Figures xiv
CHAPTER ONE
1.0 INTRODUCTION1
1.1 Background of the Study1
1.2 The Research Problem2
1.3 Justification of the Study2
1.4 Research Questions 3
1.5 Aim and Objectives of the Study4
1.6 Research Hypotheses 5
1.7 Significance of the Study5
1.8 Scope and Limitation of the Study5
CHAPTER TWO
2.0 LITERATURE REVIEW 7
2.1 Malaria7
2.2 Approaches to Antimalarial Drug Discovery 9
2.2.1 Optimisation of therapy with available drugs 9
ix
2.2.2 Development of analogues of existing agents 10
2.2.3 Discovery of natural antimalarial products.10
2.2.4 Investigation of compounds that were originally developed to treat other diseases 10
2.2.5 Evaluation of drug resistance reversers 11
2.2.6 Exploiting compounds active against new targets. 12
2.3 Naphthylisoquinoline Alkaloids12
2.4 Bisbenzylisoquinoline Alkaloids13
2.5 Theoretical Framework of the Study 14
2.6 Molecular Descriptors16
2.7 Quantitative-Structure Activity/Toxicity Relationship. 17
CHAPTER THREE
3.0 MATERIALS AND METHODS21
3.1 Materials21
3.1.1Instrumentation 21
3.2 Methodology 21
3.2.1 Biological data35
3.2.2 Molecular structure sketching 35
3.2.3 Geometry optimisation and calculation of some quantum-chemical descriptors39
3.2.4 Theoretical background for calculation of the quantum descriptors 44
3.2.5 Calculation of 0- 3D descriptors from PaDEL. 48
3.2.6 Correlation matrix 54
3.2.7 Variable elimination and selection 64
3.2.8 Data standardisations 71
3.2.9 Division of data into training and test sets71
3.2.10 Development of a quantitative structure- activity relationship model 74
3.2.11 Descriptor selection and model development 75
3.2.12 Model validation 84
x
3.2.12.1 Internal validation84
3.2.12.2 External validation 86
3.2.13 Y-Randomisation 95
3.2.14 Applicability domain 103
3.2.15 Generation of the Williams plot 105
CHAPTER FOUR
4.0 RESULTS 124
4.1 Results of Optimisation and Calculation of Quantum-chemical Descriptors 125
4.2 Results of the Calculation of 0D, 1D, 2D and 3D Descriptors using Padel 125
4.3 Results of the Variable Elimination and Selection126
4.4 Results of Antimalarial Activity (pIC50) of Naphthylisoquinoline Derivatives Against K1 Strain of Plasmodium falciparum126
4.4.1 Evaluation of model 4.4.4 135
4.5 Results of Antimalarial Activity (pIC50) of Bisbenzylisoquinoline Derivatives Against K1 Strain of Plasmodium falciparum136
4.5.1 Evaluation of model 4.5.1 137
4.6 Results of Cytotoxicity (pIC50) Study of Naphthylisoquinoline Derivatives Against L6 Cell Line. 144
4.6.1 Evaluation of model 4.6.1 145
4.7 Results of Cytotoxicity (pIC50) Study of Bisbenzylisoquinoline Derivatives Against KB Cell Line. 145
4.7.1 Evaluation of model 4.7.3 153
CHAPTER FIVE
5.0 DISCUSSION 161
5.1 Interpretation of the Descriptors in Model 4.4.2 161
5.2 Interpretation of the Descriptors in Model 4.5.1 163
5.3 Interpretation of the Descriptors in Model 4.6.1 164
xi
5.4 Interpretation of the Descriptors in Model 4.7.1 164
CHAPTER SIX
6.0 SUMMARY, CONCLUSION AND RECOMMENDATION 167
6.1 Summary167
6.2 Conclusion 167
6.3 Recommendation168
References 169
CHAPTER ONE
1.0 INTRODUCTION 1.1 Background of the Study Malaria remains one of the most calamitous diseases of all the disease causing parasites. It is the most dangerous and destructive infectious agent in the developing world (Greenwood et al., 2005 and Winter et al., 2006). Malaria is usually believed to be caused by poverty, but it is also a cause of poverty and a major factor militating against economic development. An estimated case of malaria is about 247 million with 3.3 billion people at risk in 2006, causing deaths in approximately millions, with children under the ages of 5 years the most affected. It is widespread in tropical and subtropical regions, including parts of the Americas, Asia, and Africa. About 109 countries were estimated to be endemic for malaria in 2008, out of which 45 are in the WHO African region (WHO, 2008).
Malaria is caused by a parasite called plasmodium that is transmitted from one human to another by the bite of infected Anopheles mosquitoes. Although four species of the genus Plasmodium cause human malaria, Plasmodium falciparum is the deadliest. The traditional remedies are no longer effective and the incidence of malaria byP.falciparum, the most dangerous species of parasite,Plasmodium, continues to grow, while some traditional drugs such as chloroquine and its congeners are losing their activity due to multi drug resistance. New antimalarial drugs are urgently needed. Not only should these drugs be efficacious against P.falciparum strains, but to ensure good compliance, they should provide a cure within a reasonable length of time, they should be safe and low cost, and they should be available in an appropriate formulation for oral use. Natural antiplasmodial compounds such as naphthylisoquinoline and bisbenzylisoquinoline alkaloids exist as cited by Mambu and Grellier(2008). These compounds were found to possess manifold activities against the genius
2
plasmodium. Antimalarial drug discovery is achieved through different approaches, all of which take into account the cost of drug discovery and the cost of drug itself. Quantitative structure-activity relationship (QSAR) methodologies save resources and expedite the process of the development of new molecules and drugs andhence play an important role to minimise trial and error in designing new antimalarial drugs. 1.2 The Research Problem In 2012, about630,000 people died from malaria (WHO,2013), with pregnant women and children under the age of five being the most vulnerable to the infection (Shetty, 2012). The parasite developed resistance to a number of antimalarial drugs such as chloroquine and its derivatives which is the most widely used treatment drug for malaria. Many countries now rely on therapies that are expensive to slow down the development of resistance. Various antimalarial drugs currently in use are chemically related or have a similar mode of action, thus increasing the risk of cross-resistance and clinical failure of newly introduced therapies. There is, therefore, the need for the design of a more potent drug for malaria. 1.3 Justification of the Study
During the 1990s, child deaths caused by malaria increased by up to two-folds in some parts of sub-Saharan Africa. The disease also re-emerged in several countries in central Asia, Eastern Europe and South-East Asia. The majority of the cases in 2006 were estimated as Africa (86%), followed by South-East Asia (9%) and3% in Eastern Mediterranean region (Meneguzziet al., 2009).Although malaria is a curable and preventable disease, its prevalence increased in the 1980s and 1990s as the parasites
3
developed resistance to the most frequently used antimalarial drugs and the vectors became resistant to insecticides. Increasing resistance of P. falciparumagainst traditional antimalarialdrugs like chloroquine or the antifolates makes the developmentof new, highly efficient, and easily available agents anurgent task.Nowadays, drug development is too expensive to be guided by trial anderror. Quantitative structure-activity relationship (QSAR), molecular modelling, and protein crystallography are important andvaluable tools in computer-assisted drug design. In view of the above fact,QSARhas evolved as excellent optimisation technique which involves the derivation of a mathematical formula, which relates the biological activities of a group of compounds to their physico-chemical properties. 1.4 Research Questions The questions that this study tried to answer were
i. Are there correlations between the molecular descriptors and the activity/cytotoxicity of naphthylisoquinoline derivatives?
ii. What structural features of the molecules are responsible for the activity/cytotoxicity of naphthylisoquinoline derivatives?
iii. Are there correlations between the molecular descriptors and the activity/cytotoxicity of bisbenzylisoquinoline derivatives?
iv. What structural features of the molecules are responsible for the activity/cytotoxicity of bisbenzylisoquinoline derivatives?
v. Will the generated models be used to predict the activity/cytotoxicity of all the derivatives of naphthylisoquinoline/bisbenzylisoquinoline alkaloids?
4
1.5 Aim and Objectives of the Study The aim of this research was to carry out QSAR analysis and develop predictive QSAR models that could explain the antimalarial activities of naphthylisoquinoline and bisbenzylisoquinoline alkaloids using theoretical molecular descriptors. The aim of the research will be achieved through the following objectives:
i. Collection of naphthylisoquinoline and bisbenzylisoquinoline compounds with their activity and cytotoxicity values from published articles.
ii. Optimisation of these compounds so that equilibrium structures of the compounds with lower energy will be obtained for descriptors calculation.
iii. Determination of molecular descriptors using PaDEL software.
iv. Treatment of the calculated descriptors by correlating them with the experimental activity/cytotoxicity thereby removing the not so informative descriptors and also removing descriptors that are highly correlated with another descriptor (s) but less correlated with the activity/cytotoxicity.
v. Division of the data set into training and test sets. The training set is for constructing the model while the test set is for validation of the model.
vi. Performing a multiple linear regression (MLR) analysis using BuildQSAR modelling software.
vii. Validation of the predictive model using external method of validation such as calculating Y-randomisation and R2pred.
viii. Checking the applicability domain of the model which serves as a kind of quality control procedure in the prediction of new compounds.
5
1.6 Research Hypothesis The working hypotheses of the study are: a) The null hypothesis (Ho) which states that:
i. There is no correlation between the antimalarial activity/cytotoxicity of naphthylisoquinolines derivatives and their physio-chemical properties.
ii. There is no correlation between the antimalarial activity/cytotoxicity of bisbenzylisoquinoline alkaloids derivatives and their physico-chemical properties.
b) The alternate hypothesis (H1), which states that:
i. There is positive correlation between the antimalarial activity/cytotoxicity of naphthylisoquinolines derivatives and their physicochemical properties.
ii. There is positive correlation between the antimalarial activity/cytotoxicity ofbisbenzylisoquinoline alkaloids derivatives and their physico-chemical properties.
1.7 Significance of the Study The study would avail us with the various structural information responsible for the observed antimalarial activity/cytotoxicity of naphthylisoquinoline and bisbenzylisoquinoline derivatives which can be exploited by the synthetic chemist to develop antimalarial drugs with better activity/cytotoxicity. 1.8 Scope and Limitation of the Study
6
The study was aimed at quantitatively determined the relationship between antimalarial activity/cytotoxicity with physico-chemical parameters of antimalarial compounds. There are concerns over limited availability of data set. Those that are available may be uneven and may contain outliers. Due to financial constraint and other factors, the activities of these compounds will not be determined in vivo on P. falciparum and also there will be no experimental validation.
7
IF YOU CAN'T FIND YOUR TOPIC, CLICK HERE TO HIRE A WRITER»
