Models for the Prediction of Receptor Tyrosine Kinase Inhibitory Activity of Substituted 3-Aminoindazole Analogues

Monika Gupta; Harish Dureja; Anil Kumar Madan

doi:10.3797/scipharm.1102-08

. 2011 Apr 28;79(2):239–257. doi: 10.3797/scipharm.1102-08

Models for the Prediction of Receptor Tyrosine Kinase Inhibitory Activity of Substituted 3-Aminoindazole Analogues

Monika Gupta ¹, Harish Dureja ¹, Anil Kumar Madan ^2,^*

PMCID: PMC3134851 PMID: 21773063

Abstract

The inhibition of tumor angiogenesis has become a compelling approach in the development of anticancer drugs. In the present study, topological models were developed through decision tree and moving average analysis using a data set comprising 42 analogues of 3-aminoindazoles. A total of 22 descriptors (distance based, adjacency based, pendenticity and distance-cum-adjacency based) were used. The values of all 22 topological indices for each analogue in the dataset were computed using an in-house computer program. A decision tree was constructed for the receptor tyrosine kinase KDR (kinase insert domain receptor) inhibitory activity to determine the importance of topological indices. The decision tree learned the information from the input data with an accuracy of 88%. Three independent topological models were also developed for prediction of receptor tyrosine kinase inhibitory (KDR) activity using moving average analysis. The models developed were also found to be sensitive towards the prediction of other receptor tyrosine kinases i.e. FLT3 (fms-like tyrosine kinase-3) and cKIT inhibitory activity. The accuracy of classification of single index based models using moving average analysis was found to be 88%. The performance of models was assessed by calculating precision, sensitivity, overall accuracy and Mathew’s correlation coefficient (MCC). The significance of the models was also assessed by intercorrelation analysis.

Keywords: Topological indices, Receptor tyrosine kinase inhibitors, 3-Aminoindazoles, Decision tree, Moving average analysis

Introduction

Cancer is a leading cause of death worldwide and accounted for 7.9 million deaths (around 13% of all deaths) in 2007. Moreover, the deaths resulting from cancer are expected to rise continuously with an estimated 12 million deaths in 2030 [1]. Cancer is thought to reflect a multistep process, resulting from an accumulation of inherited, acquired or both defects in genes involved in the positive or negative regulation of cell proliferation and survival. Activation or inactivation of just four or five different genes may be required for the development of clinically recognizable human cancer [2]. Despite advances in diagnosis and treatment, overall survival of patients still remains poor. Surgery, chemotherapy, radiotherapy, and endocrine therapy have been the standard options available for treatment of cancer patients. This has improved survival in several types of solid tumors, but treatment-related toxicity and emergence of drug resistance have been the major causes of morbidity and mortality [3]. Consequently, there is an urgent need to develop newer more effective therapies to improve patient outcomes. Blood is essential for solid tumors to manage nutritional supplies and waste removal to manage the tumor growth and metastasis. Angiogenesis is a process in which new blood vessels are formed from pre-existing vasculatures [4]. It has been reported that the angiogenesis is a rate limiting step in tumor development. Tumors that lack adequate vasculature become necrotic or apoptotic and don’t grow beyond a limited size [5, 6]. Inhibition of tumor angiogenesis has become a compelling approach in development of anticancer agents [7, 8]. Vascular endothelial growth factor (VEGF) is the primary endothelial cell specific angiogenic factor [9]. VEGF activity is mediated by three higher affinity receptors belonging to the class-V subfamily of receptor tyrosine kinases (RTKs). These are widely known for regulating angiogenesis, vasculogenesis, and lymphangiogenesis. The VEGFR family includes VEGFR-I/FLT-1 (fms-like tyrosine kinase-I), VEGFR-2/FLK-I (fetal liver kinase-I)/KDR (kinase inert domain containing receptor) and VEGFR-3/FLT-4 (fms-like tyrosine kinase-4). VEGFR-I is required for endothelial organization during vascular development while VEGFR-2 is required for formation of blood islands and also hematopoiesis [3, 5]. VEGFR-3 plays a significant role in VEGF-C and VEGF-D-mediated lymphangiogenesis. Over activation of KDR by VEGFs has been linked to progression of variety of human cancers. VEGF mediated KDR signaling induces a series of endothelial responses such as proliferation, migration and survival which ultimately leads to new vessel formation and maturation [10–13]. The immature tumor vasculature lacking a close association with pericytes appears to be most susceptible to inhibition of VEGFR signaling [14] and resistance to continued inhibition of VEGFR signalling has been reported [15]. Evidence increasingly points to a key role for PDGFR (platelet derived growth factor receptor tyrosine kinsases), located on the pericytes in the tumour stroma, in angiogenesis and vessel maturation and highlights its potential as a therapeutic target [16]. PDGFR is class III subfamily of RTKs including five receptors i.e. PDGFR-α, PDGFR-β, CSF-IR (colony stimulating factor 1 receptor), cKIT and FLT3 involved in cellular growth, differentiation, cytokine vascular regulation, gliomas and leukemia [3, 17]. FLT3 is shown to be commonly overexpressed in most B lineage acute lymphocytic leukemia (ALL), acute myeloid leukemias (AMLs) and chronic myeloid leukemias (CML) [18]. Preclinical studies have already shown the benefits of combining VEGFR and PDGFR inhibition with respect to tumour response [19]. Due to vital role that RTKs play in tumor angiogenesis, inhibition of these may prove to be an effective therapeutic intervention and potential inhibitors can be explored [2, 3].

The drug research and development is comprehensive, expensive, time-consuming and full of risk [20]. The traditional approach of drug discovery involves target identification, validation, lead search and optimization followed by clinical development phases [21]. The experimental search for better activities in drug discovery is commonly carried out in the laboratory by optimizing the structure–activity relationship (SAR) of the functional groups present in a leading structure in terms of their biological endpoint. However, an interesting alternative to this trial-error based procedure that constitutes an active field in complex biochemical phenomena are the analysis through Quantitative Structure–Activity/Property/Toxicity Relationships (QSAR/QSPR/QSTR) [22]. Quantitative structure-activity relationship (QSAR) represents an attempt to correlate structure descriptors of compounds with their biological activity [23, 24]. An important aspect of this method is the use of good structural descriptors that represent the molecular features responsible for the relevant biological activity [25]. The chemical graph theory is largely applied to the quantitative characterization of molecular structures for predicting physicochemical, pharmacological and toxicological properties using graph theoretical invariants [26, 27].The graph theoretical invariants have been termed as topological indices [28, 29]. The computation of TI is very swift and the TIs have the advantage of being true structural Invariants, which means that their values are independent of molecular conformations [25]. In last few decades, the topological indices have emerged as powerful tools for predicting biological activity of molecules, and lead identification forming an integral part of new molecular research [30–33]. In the present study, relationship of topological descriptors with KDR inhibitory activities of 3-aminoindazoles has been investigated using decision tree and moving average analysis. The proposed models were also evaluated for the prediction of FLT3 and cKIT inhibitory activities.

Methodology

Dataset

A dataset comprising of 42 analogues of substituted 3-aminoindazoles [34] was selected for the present investigations. The basic structure for these analogues is depicted in Fig. 1. and various substituents are enlisted in Tab. 1.

Fig. 1. — Basic structures of 3-aminoindazole analogues [34]

Tab. 1.

Relationship between topological indices and KDR inhibitory activity

Cpd. No.	Basic Structure	R	R₁	M₂^C	W_c	^A_ξ^c^₂	KDR inhibitory activity
							Predicted using MAA models			Reported
							M₂^C	W_c	^A_ξ^c^₂	Reported
1	A		⁻	86.79	494.18	4.11	⁻	⁻	⁻	⁻
2	B		⁻	123.29	1674.22	2.02	⁻	⁻	⁻	⁻
3	B		⁻	164.96	2093.78	2.15	⁻	⁻	⁻	⁻
4	B		⁻	129.46	1949.52	1.76	⁻	⁻	⁻	⁻
5	C		⁻	135.45	2172.95	1.81	⁻	⁻	⁻	⁺
6	C		⁻	135.46	1750.1	1.98	⁻	⁻	⁻	⁻
7	C		⁻	125.46	1729.09	1.98	⁻	⁻	⁻	⁻
8	C		⁻	123.29	1674.22	2.02	⁻	⁻	⁻	⁻
9	D	2–F	⁻	136.20	2160.53	1.82	⁻	⁻	⁻	⁻
10	D	3–F	⁻	137.78	2180.53	1.81	⁻	⁻	⁻	⁻
11	D	4–F	⁻	137.79	2200.53	1.58	⁻	⁻	⁻	⁻
12	D	2–Me	⁻	134.45	2152.95	1.82	⁻	⁻	⁻	⁻
13	D	4–Me	⁻	135.46	2192.95	1.58	⁻	⁻	⁻	⁻
14	D	3–Et	⁻	139.46	2423.38	1.62	⁺	⁺	⁺	⁺
15	D	3–Cl	⁻	143.29	2198.40	1.81	⁺	⁻	⁻	⁺
16	D	3–Br	⁻	158.12	2246.62	1.81	⁻	⁻	⁻	⁻
17	D	3–CF₃	⁻	152.70	2955.6	1.72	⁻	⁻	⁺	⁺
18	D	3–OH	⁻	136.79	2177.28	1.81	⁻	⁻	⁻	⁻
19	D	2–F–5–Me	⁻	141.20	2389.25	1.87	⁺	⁺	⁻	⁺
20	D	3–Me–4–F	⁻	144.37	2427.25	1.63	⁺	⁺	⁺	⁺
21	D	3–F–4–Me	⁻	144.37	2427.25	1.63	⁺	⁺	⁺	⁻
22	D	2–F–5–CF₃	⁻	153.70	3191.65	1.78	⁻	⁻	⁻	⁻
23	E	3–Me	–Me	141.62	2383.54	1.73	⁺	⁺	⁺	⁺
24	E	3–Me		147.1	2635.78	1.67	⁻	⁻	⁺	⁻
25	E	3–Me		144.63	2391.41	1.73	⁺	⁺	⁺	⁺
26	E	2–F–5–Me		170.90	2460.04	1.73	⁻	⁺	⁺	⁺
27	F	3–Me	–Me	165.41	3909.63	1.43	⁻	⁻	⁻	⁻
28	F	3–Me	–OMe	173.57	4669.82	1.39	⁻	⁻	⁻	⁻
29	F	3–Me	–F	179.73	4664.31	1.41	⁻	⁻	⁻	⁻
30	F	3–Me	–Br	190.45	5587.48	1.28	⁻	⁻	⁻	⁻
31	F	3–Me		192.51	5554.98	1.36	⁻	⁻	⁻	⁻
32	F	3–Me		161.78	3572.18	1.42	⁻	⁻	⁻	⁻
33	F	3–Me		175.23	4126.54	1.5	⁻	⁻	⁻	⁻
34	F	3–Me		173.85	4123.64	1.5	⁻	⁻	⁻	⁻
35	F	3–Me		141.96	2409.05	1.54	⁺	⁺	⁻	⁻
36	F	3–Me		150.18	2924.31	1.42	⁻	⁻	⁻	⁻
37	F	3–Me		179.93	4595.26	1.28	⁻	⁻	⁻	⁻
38	F	3–Me		161.09	3501.97	1.54	⁻	⁻	⁻	⁻
39	F	3–Cl		169.62	3602.53	1.42	⁻	⁻	⁻	⁻
40	F	2–F–5–Me		167.53	3873.60	1.45	⁻	⁻	⁻	⁻
41	F	2–F–5–Me		179.32	5030.37	1.41	⁻	⁻	⁻	⁻
42	F	2–F–5–Me		192.26	5466.49	1.37	⁻	⁻	⁻	⁻

Open in a new tab

⁺

Active analogue;

⁻

Inactive analogue.

Kinase enzymatic assays of al the 42 analogues were preformed by Dai et al. [34] utilizing the homogeneous time-resolved fluorescence (HTRF) protocol. Peptide substrate at 4 μM, 1 mM ATP, enzyme and inhibitors (3.2 nM to 50 μM) were incubated for 1 hour at ambient temperature in 50mM NaOH (pH 7.5), 10mM MgCl₂, 2mM MnCl₂, 2.5 mM DTT, 0.1 mM orthovandate and 0.01% bovine serum albumin. The reactions were stopped with 0.5 M EDTA and then 75 μL buffer containing detecting agents (streptividine-allphycocyanin and PT66 antibody europium cryptate) was added. The plates were read from 1to 4 hour for time-resolved fluorescence. The inhibition was calculated using control and background reading. Each IC₅₀ determination was preformed with seven concentrations and each assay point was reportedly determined in duplicate [34].

Subsequently, based on the results of KDR enzymatic assay, the potent inhibitors were reportedly characterized by cellular assay using 3T3 – murine fibroblasts cells [34].

The in vivo activity of compounds with potent cellular activity was also reportedly carried out using an estradiol-induced mouse uterine edema (UE) model. The said assay served as a valuable tool for a rapid and preliminary evaluation of KDR inhibitor’s oral activity.

The dataset comprised of variable degree of activities. Compounds having reported IC₅₀ values of ≤ 10 nM were considered to be active while those possessing IC₅₀ values >10 nM were treated to be inactive for the purpose of present study.

Topostructural and topochemical indices

Twelve topostructural and ten topochemical indices [30, 35–55] used for the present study are presented in Tab. 2. The distance based topological descriptor (Wiener index, Balaban index), adjacency based descriptors (Zagreb group parameter, M₁ and M₂, molecular connectivity index) and distance cum adjacency based topological descriptors (eccentric adjacency index, augmented eccentric connectivity index, superadjacency index, eccentric connectivity index, connective eccentric index, superaugmented eccentric connectivity index-2) and pendenticity based descriptor (superpendentic index) were calculated using an in-house computer program. The topochemical versions of topostructural descriptors were calculated from distance and adjacency matrices weighed by molecular mass with respect to that of carbon atom.

Tab. 2.

Topostructural and topochemical indices

Code	Index	Refe.
A1	Molecular connectivity topochemical index	35,36
A2	Eccentric adjacency topochemical index	37
A3	Augmented eccentric connectivity topochemical index	38
A4	Superadjacency topochemical index	39
A5	Eccentric connectivity topochemical index	40
A6	Connective eccentricity topochemical index	41
A7	Zagreb topochemical index, M₁^C	42
A8	Zagreb topochemical index, M₂^C	42
A9	Wiener’s topochemical index	43
A10	Superaugmented eccentric connectivity topochemical index-2	44
A11	Molecular connectivity index	30
A12	Eccentric adjacency index	45
A13	Augmented eccentric connectivity index	46
A14	Superadjacency index	39
A15	Eccentric connectivity index	47
A16	Connective eccentricity index	48
A17	Zagreb index, M₁	49,50
A18	Zagreb index, M₂	49,50
A19	Wiener’s index	51,52
A20	Superaugmented eccentric connectivity index-2	53
A21	Balaban mean square distance index	54
A22	Superpendentic index	55

Open in a new tab

Decision tree

Decision tree provides a useful solution for many problems of classification where large datasets are used and the information contained is complex. A decision tree consisting of nodes and branches represents a collection of rules with each terminal node corresponding to a specific decision rule [56]. Decision trees are constructed beginning with the roots of tree and proceeding down to its leaves. In terms of ability, decision trees are a rapid and effective method of classifying data set entries and can provide good decision support capabilities [57, 58]. Applications of classification-based decision tree methods predominate in science and medicine [59]. It has been applied to some bioinformatics and cheminformatics problems, such as characterizations of tumor [60, 61], prediction of drug response [62], classification of antagonist of receptors [63] and identification of DNA sections coding exons [64].

In present study, decision tree was grown to identify the importance of topological indices. In a decision tree, the molecules at each parent node are classified, based upon the index value, into two child nodes. The prediction for molecule reaching a given terminal node is obtained by majority vote of molecules reaching the same terminal node in the training set. In this study, R program (version 2.1.0) along with the RPART library was used to grow decision tree. The active compounds were labeled as “A” (n = 9) and the inactive compounds were labeled “B” (n = 33). Each analogue was assigned a biological activity which was then compared with the reported KDR inhibitory activity [34].

Moving average analysis

Moving average analysis of correctly predicted compounds is the basis of development of single topological index based model [45, 65]. For the selection and evaluation of range specific features, exclusive activity ranges were discovered from the frequency distribution of response level and subsequently identify the active range by analyzing the resulting data by maximization of the moving average with respect to active compounds (<35% inactive, 35–65 % transitional, > 65 % active) The KDR inhibitory activity assigned to each compound was compared with reported biological activity. The various ranges obtained were also studied for the cKIT and FLT3 inhibitory activities. The average IC₅₀ (nm) for each range and activity was also calculated.

Data analysis

The sensitivity and specificity values were calculated which represents the classification accuracies for the active and inactive compounds, respectively. The randomness of model was also predicted by calculating Mathew’s correlation coefficient (MCC). The MCC values ranging between −1 to +1 indicates the potential of model. MCC took both sensitivity and specificity into account and it is generally used as a balanced measure in dealing with data imbalance situation [66]. The intercorrelation between M₂^C and W_c, and ^Aξ^c^₂ was also investigated. The degree of correlation was appraised by correlation coefficient ‘r’. Pairs of indices with r ≥ 0.97 were considered highly inter-correlated, those with 0.90 ≤ r ≤ 0.97 were appreciably correlated, those with 0.50 ≤ r ≤ 0.89 were weakly correlated and finally the pairs of indices with r < 0.50 were not intercorrelated [67].

Results and Discussion

In the present study, decision tree was built from a set of 22 topological indices. The index at root node is most important and the importance of index decreases as the length of tree increases. The classification of 3-aminoindazoles analogues as inactive and active using a single tree, based on Zagreb topochemical index A8 is shown in Fig. 2.

Fig. 2. — A decision tree for distinguishing active analogue (A) from inactive analogue (B); A8- Zagreb topochemical index, M₂^C

The decision tree identified the Zagreb topochemical index A8 as the most important index. The decision tree classified the analogues with an accuracy of 88%. The precision and sensitivity of inactive analogues was of the order of 91.11% and 93.93%, whereas the precision and sensitivity of active analogues was of the order of 75% and 66.6% respectively (Tab. 3).

Tab. 3.

Confusion matrix for KDR inhibitory activity using models based on decision tree

Ranges	Number of cpds. predicted using decision tree		Precision (%)	Sensitivity (%)	MCC
Ranges	Active	Inactive	Precision (%)	Sensitivity (%)	MCC
Active	6	3	75	66.6	0.63
Inactive	2	31	91.11	93.93

Open in a new tab

Three independent moving average analysis (MAA) based topological models were developed (Tab. 4.). The topological index A8, Zagreb topochemical index identified as most important index by decision tree was used to construct model for the prediction of KDR inhibitory activity. Two more indices i.e. Wiener’s topochemical index, A9 and superaugmented eccentric connectivity index-2, A20 were also used to develop the models for predicting KDR inhibitory activity. The methodology used in the present study aims at development of suitable models for providing lead molecules through exploitation of the active ranges in the proposed models based on topological indices. Proposed models are unique and differ widely from the conventional QSAR models. Both systems of modelling have their own advantages and limitations. In the instant case, the modelling system adopted has distinct advantage of identification of narrow active range, which may be erroneously skipped during routine regression analysis in conventional QSAR modelling [44]. Since the ultimate goal of modelling is to provide lead structures, therefore, these active ranges can play vital role in lead identification.

Tab. 4.

MAA based models for the prediction of receptors tyrosine inhibitory activity

Model Index	Nature of range in proposed model	Index value	Number of analogues falling in the range		Overall accuracy of predict. (%)	KDR Aver. IC₅₀ (nM)	*FLT3 Aver.^ IC₅₀ (nM)**	*cKIT Aver.^ IC₅₀ (nM)**
Model Index	Nature of range in proposed model	Index value	Total	Correct	Overall accuracy of predict. (%)	KDR Aver. IC₅₀ (nM)	*FLT3 Aver.^ IC₅₀ (nM)**	*cKIT Aver.^ IC₅₀ (nM)**
M₂^C	Lower inactive	<139.46	13	12		1379.77	84.54	1629.62
	Active	139.46–144.63	8	6	88.09	9.63	19.86	19.71
	Upper inactive	>144.63	21	19		557.76	189.31	893.14

W_c	Lower inactive	<2383.54	16	14		1412.56	80.53	2029.19
	Active	2383.54–2460.04	8	6	88.09	9.87	21.29	23.43
	Upper inactive	>2460.04	18	17		391.5	207.86	108.73

^Aξ^c^₂	Lower inactive	<1.62	18	18		389.5	188.36	101.4
	Active	1.62–1.73	8	6	88.09	12.5	25	31
	Upper inactive	>1.73	16	13		1413.5	97	2037.25

Open in a new tab

Average in a range is taken only for the reported IC₅₀ values;

^#Values in brackets are based upon correctly predicted analogues in the particular range.

Retrofit analysis of the data reveals the following information with regards to different models used in this study. The biological activity was assigned to 42 analogues in both active and inactive ranges, out of which activity of 37 analogues was correctly predicted resulting in 88.09% accuracy with regard to KDR inhibition using model based on Zagreb topochemical index (A8).

31 out of 34 compounds (91%) in both the inactive ranges were predicted correctly. The average IC₅₀ value of all the correctly predicted analogues in both the inactive ranges was 1494.5 nM and 615.421 nM respectively (Fig. 3). The average IC₅₀ of correctly predicted analogues in the active range was found to be only 5 nM with regard to KDR inhibitory activity. Such a low average IC₅₀ value signifies high potency of the active range. The said active range also exhibited significant FLT3 activity with average IC₅₀ value of 19.86 nM (Fig. 4) and cKIT activity with average IC₅₀ value of 19.714 nM (Fig. 5). The precision and sensitivity of inactive analogues was found to be 93.93% and 91.11%, whereas the precision and sensitivity of active analogues was of the order of 66.6% and 75% respectively. The 3-aminoindazoles analogues were correctly classified as active or inactive using Wiener’s topochemical index with an accuracy of 88%. The biological activity was assigned to 42 analogues in both active and inactive ranges, out of which 37 analogues were correctly predicted. 31 out of 34 compounds (91.1%) in both the inactive ranges were predicted correctly. The average IC₅₀ value of all the correctly predicted analogues in both the inactive ranges was 1613.57 nM and 413.94 nM, respectively whereas the average IC₅₀ of correctly predicted analogues in the active range was found to be only 5.33 nM (Fig. 3). Such a low average IC₅₀ value signifies high potency of the active range.

Fig. 3. — Average IC₅₀ (nM) values of 3-aminoindazoles for KDR inhibitory activity in various ranges of topological models derived through moving average analysis

Fig. 5. — Average IC₅₀ (nM) values of 3-aminoindazoles for cKIT inhibitory activity in various ranges of topological models derived through moving average analysis

The above active range also exhibited significant FLT3 activity with average IC₅₀ value of 21.29 nM (Fig. 4) and cKIT activity with average IC₅₀ value of 23.43 nM (Fig. 5).

The precision and sensitivity of inactive analogues was found to be 93.93% and 91.11%, whereas the precision and sensitivity of active analogues was of the order of 66.6% and 75% respectively.

Superaugmented eccentric connectivity-2, classified the 3-aminoindazoles analogues as active and inactive with an accuracy of 88%. The biological activity was assigned to 42 analogues in both active and inactive ranges, out of which 37 analogues were correctly predicted. 31 out of 34 compounds (91.1%) in both the inactive ranges were predicted correctly.

The average IC₅₀ value of all the correctly predicted analogues in both the inactive ranges was 389.5 nM and 841.75 nM, respectively whereas the average IC₅₀ of correctly predicted analogues in the active range was found to be only 6.33 nM (Fig. 3). Such a low average IC₅₀ value signifies high potency of the active range. The said active range also exhibited significant FLT3 activity with average IC₅₀ value of 25 nM (Fig. 4) and cKIT activity with average IC₅₀ value of 31 nM (Fig. 5).

The overall accuracy of prediction was found to be 88%. The precision and sensitivity of inactive analogues was found to be 93.93% and 91.11%, whereas the precision and sensitivity of active analogues was of the order of 66.6% and 75% respectively.

The MCC value was found to be 0.633 (Tab. 3.) suggesting the randomness in the distribution of data. The result of intercorrelation analysis (Tab. 5) reveals that the ^Aξ^c^₂ was not correlated with M₂^C and W_c while the pair M₂^C and W_c was found to be appreciably correlated

Tab. 5.

Intercorrelation matrix

	M₂^C	W_c	^Aξ^c₂
M₂^C	1	0.92	−0.73
W_c		1	−0.68
^Aξ^c₂			1

Open in a new tab

Conclusion

Models based on all the three topological descriptors exhibited high degree of predictability with regard to KDR inhibitory activity using decision tree and moving average analysis. Moreover, the active ranges of the proposed models also exhibited significant FLT3 and cKIT inhibitory activities. A combination of VEGFR and PDGFR inhibitory activities will naturally be more beneficial for the treatment of tumors. High degree of predictability of the proposed models can provide valuable lead structures for the development of potent receptor tyrosine kinase inhibitors (RTKs).

Footnotes

This article is available from: http://dx.doi.org/10.3797/scipharm.1102-08

Authors’ Statement

Competing Interests

The authors declare no conflict of interest.

References

[1].Basha MR, Baker CH, Abdelrahim M. Biomarkers Clinical Relevance in Cancer: Emphasis on Breast Cancer and Prostate Cancer. Curr Trends Biotechnol Pharm. 2009;3:1–7. [Google Scholar]
[2].Kolibaba KS, Druker BJ. Protein Tyrosine Kinase and Cancer. Biochim Biophys Acta. 1997;1333:F217–248. doi: 10.1016/S0304-419X(97)00022-X. [DOI] [PubMed] [Google Scholar]
[3].Madhusudan S, Ganesan TS. Tyrosine Kinase Inhibitors in Cancer Therapy. Clin Biochem. 2004;37:618–635. doi: 10.1016/j.clinbiochem.2004.05.006. [DOI] [PubMed] [Google Scholar]
[4].Zetter BR. Angiogenesis and Tumor Metastasis. Ann Rev Med. 1998;49:407–424. doi: 10.1146/annurev.med.49.1.407. [DOI] [PubMed] [Google Scholar]
[5].Mustonen T, Alitalo K. Endothelial Receptor Tyrosine Kinases Involved in Angiogenesis. J Cell Biol. 1995;129:895–898. doi: 10.1083/jcb.129.4.895. [DOI] [PMC free article] [PubMed] [Google Scholar]
[6].Carmeliet P, Jain RK. Angiogenesis in Cancer and Other Diseases. Nature. 2000;407:249–257. doi: 10.1038/35025220. [DOI] [PubMed] [Google Scholar]
[7].Westwell AD. A New Era in Cancer Therapeutics. Drug Discov Today. 2003;8:64–65. doi: 10.1016/S1359-6446(02)02575-8. [DOI] [PubMed] [Google Scholar]
[8].Ferrara N, Kerbel RS. Angiogenesis as a Therapeutic Target. Nature. 2005;438:967–974. doi: 10.1038/nature04483. [DOI] [PubMed] [Google Scholar]
[9].Yancopoulos GD, Davis S, Gale NW, Rudge JS, Wiegand SJ, Holash J. Vascular-Specific Growth Factors and Blood Vessel Formation. Nature. 2000;407:242–248. doi: 10.1038/35025215. [DOI] [PubMed] [Google Scholar]
[10].Hubbard SR, Till JH. Protein Tyrosine Kinase Structure and Function. Annu Rev Biochem. 2000;69:373–398. doi: 10.1146/annurev.biochem.69.1.373. [DOI] [PubMed] [Google Scholar]
[11].Shibuya M. Role of VEGF-FIt Receptor System in Normal and Tumor Angiogenesis. Adv Cancer Res. 1995;67:281–316. doi: 10.1016/S0065-230X(08)60716-2. [DOI] [PubMed] [Google Scholar]
[12].Gschwind A, Fischer OM, Ullrich A. The Discovery of Receptor Tyrosine Kinases: Targets For Cancer Therapy. Nat Rev Cancer. 2004;4:361–370. doi: 10.1038/nrc1360. [DOI] [PubMed] [Google Scholar]
[13].Folkman J. Angiogenesis in Cancer, Vascular, Rheumatoid and Other Disease. Nat Med. 1995;1:27–31. doi: 10.1038/nm0195-27. [DOI] [PubMed] [Google Scholar]
[14].Benjamin LE, Golijanin D, Itin A, Pode D, Keshet E. Selective Ablation of Immature Blood Vessels in Established Human Tumors Follows Vascular Endothelial Growth Factor Withdrawal. J Clin Invest. 1999;103:159–165. doi: 10.1172/JCI5028. [DOI] [PMC free article] [PubMed] [Google Scholar]
[15].Glade Bender J, Cooney EM, Kandel JJ, Yamashiro DJ. Vascular Remodeling and Clinical Resistance to Antiangiogenic Cancer Therapy. Drug Resist Updat. 2004;7:289–300. doi: 10.1016/j.drup.2004.09.001. [DOI] [PubMed] [Google Scholar]
[16].Östman A. PDGF Receptors-Mediators of Autocrine Tumor Growth and Regulators of Tumor Vasculature and Stroma. Cytokine Growth Factor Rev. 2004;15:275–286. doi: 10.1016/j.cytogfr.2004.03.002. [DOI] [PubMed] [Google Scholar]
[17].Iwamoto H, Nakamuta M, Tada S, Sugimoto R, Munechika E, Nawata H. Platelet-Derived Growth Factor Receptor Tyrosine Kinase Inhibitor AG 1295 Attenuates Rat Hepatic Stellate Cell Growth. J Lab Clin Med. 2000;135:406–412. doi: 10.1067/mlc.2000.105974. [DOI] [PubMed] [Google Scholar]
[18].Chu SH, Small D. Mechanisms of Resistance to FLT3 Inhibitors. Drug Resist Updat. 2009;12:8–16. doi: 10.1016/j.drup.2008.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
[19].Potapova O, Laird AD, Nannini MA, Barone A, Li G, Moss KG, Cherrington JM, Mendel DB. Contribution of Individual Targets to the Antitumor Efficacy of the Multitargeted Receptor Tyrosine Kinase Inhibitor SU11248. Mol Cancer Ther. 2006;5:1280–1289. doi: 10.1158/1535-7163.MCT-03-0156. [DOI] [PubMed] [Google Scholar]
[20].Tang Y, Zhu W, Chen K, Jiang H. New Technologies in Computer-Aided Drug Design: Toward Target Identification and New Chemical Entity Discovery. Drug Disc Today: Technologies. 2006;3:307–313. doi: 10.1016/j.ddtec.2006.09.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
[21].Pârvu L. QSAR – A Piece of Drug Design. J Cell Mol Med. 2003;7:333–335. doi: 10.1111/j.1582-4934.2003.tb00235.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
[22].Duchowicz PR, Vitale MG, Castro EA, Fernàndez M, Caballero J. QSAR Analysis For Heterocyclic Antifungals. Bioorg Med Chem. 2007;15:2680–2689. doi: 10.1016/j.bmc.2007.01.039. [DOI] [PubMed] [Google Scholar]
[23].Hansch C, Leo A, Hoekman D. Exploring QSAR: Hydrophillic, Electronic and Steric Constants. American Chemical Society; Washington DC, USA: 1995. [Google Scholar]
[24].Chen SW, Li Z-R, Li XY. Prediction of Antifungal Activity by Support Vector Machine Approach. J Mol Struct (Theochem) 2005;731:73–81. doi: 10.1016/j.theochem.2005.06.032. [DOI] [Google Scholar]
[25].Mahmoudi N, Juliàn-Ortiz JV, Ciceron L, Gàlvez J, Mazier D, Danis M, Derouin F, Domenech RG. Identification of New Antimalarial Drugs by Linear Discriminant Analysis and Topological Virtual Screening. J Antimicrob Chemother. 2006;57:489–497. doi: 10.1093/jac/dki470. [DOI] [PubMed] [Google Scholar]
[26].Ivanciuc O, Ivanciuc T, Klein DJ, Seitz WA, Balaban AT. Wiener Index Extension by Counting Even/Odd Graph Distances. J Chem Inf Comput Sci. 2001;41:536–549. doi: 10.1021/ci000086f. [DOI] [PubMed] [Google Scholar]
[27].Garcia-Domenech R, Galvez J, de Jullian-Ortiz JV, Pogliani L. Some New Trends in Chemical Graph Theory. Chem Rev. 2008;108:1127–1169. doi: 10.1021/cr0780006. [DOI] [PubMed] [Google Scholar]
[28].Ivanciuc O, Dellivers J, Balaban AT. Topological Indices and Related Descripors in QSAR and QSPR. Gordon and Breach Science Publishers; The Netherlands: 1999. pp. 697–777. [Google Scholar]
[29].Estrada E, Uriarte E. Recent Advances on the Role of Topological Indices in Drug Discovery Research. Curr Med Chem. 2001;8:1573–1588. doi: 10.2174/0929867013371923. [DOI] [PubMed] [Google Scholar]
[30].Randic M. On Characterization of Molecular Branching. J Am Chem Soc. 1975;97:6609–6615. doi: 10.1021/ja00856a001. [DOI] [Google Scholar]
[31].Needham DE, Wei IC, Seybold PG. Molecular Modeling of the Physical Properties of the Alkanes. J Am Chem Soc. 1988;110:4186–4194. doi: 10.1021/ja00221a015. [DOI] [Google Scholar]
[32].Basak SC. Use of Molecular Complexity Indices in Predictive Pharmacology and Toxicology: A QSAR Approach. Med Sci Res. 1987;15:605–609. [Google Scholar]
[33].Estrada E, Patlewicz G, Uriate E. From Molecular Graphs to Drugs. A Review on the Use of Topological Indices in Drug Design and Discovery. Indian J Chem. 2003;42A:1315–1329. [Google Scholar]
[34].Dai Y, Hartandi K, Ji Z, Ahmed AA, Albert DH, Bauch JL, Bouska JJ, Bousquet PF, Cunha GA, Glasar KB, Harris CM, Hickman D, Guo J, Li J, Marcotte PA, Marsh KC, Moskey MD, Martin RL, Olson AM, Osterling DJ, Pease LJ, Soni NB, Stewart KD, Stoll VS, Tapang P, Reuter DR, Davidsen SK, Michaelides MR. Discovery of N-(4-(3-Amino-1H-Indazol-4-yl)Phenyl)-N-(2-Fluoro-5-Methylphenyl) Urea (ABT-869), A 3-Aminoindazole-Based Orally Active Multitargeted Receptor Tyrosine Kinase Inhibitor. J Med Chem. 2007;50:1584–1597. doi: 10.1021/jm061280h. [DOI] [PubMed] [Google Scholar]
[35].Goel A, Madan AK. Structure-Activity Study on Anti-Inflammatory Pyrazole Carboxylic Acid Hydrazide Analogs Using Molecular Connectivity Indices. J Chem Inf Comput Sci. 1995;35:510–514. doi: 10.1021/ci00025a019. [DOI] [PubMed] [Google Scholar]
[36].Dureja H, Madan AK. Topochemical Models for Prediction of Cyclin-Dependent Kinase 2 Inhibitory Activity of Indole-2-ones. J Mol Mod. 2005;11:525–531. doi: 10.1007/s00894-005-0276-3. [DOI] [PubMed] [Google Scholar]
[37].Gupta S, Singh M, Madan AK. Novel Topochemical Descriptors for Predicting Anti-HIV Activity. Indian J Chem. 2003;42A:1414–1425. [Google Scholar]
[38].Bajaj S. Study on Topochemical Descriptors For the Prediction of Physicochemical and Biological Properties of Molecules. 2005. Ph.D. Thesis, Guru Gobind Singh Indraprastha University, India.
[39].Bajaj S, Sambhi SS, Madan AK. Prediction of Carbonic Anhydrase Activation by Tri-/Tetrasubstituted-Pyridinium-Azole Drugs: A Computational Approach Using Novel Topochemical Descriptor. QSAR Comb Sci. 2004;23:506–514. doi: 10.1002/qsar.200439999. [DOI] [Google Scholar]
[40].Kumar V, Sardana S, Madan AK. Predicting Anti-HIV Activity of 2,3-Diaryl-1,3-Thiazolidin-4-ones: Computational Approach Using Reformed Eccentric Connectivity Index. J Mol Mod. 2004;10:399–407. doi: 10.1007/s00894-004-0215-8. [DOI] [PubMed] [Google Scholar]
[41].Gupta S. Application and Development of Graph Invariants of Drug Design. 2002. Ph.D. Thesis, Punjabi University, Patiala, India.
[42].Bajaj S, Sambhi SS, Madan AK. Prediction of Anti-Inflammatory Activity of N-Arylanthranilic Acids: Computational Approach Using Redefined Zagreb Indices. Croat Chem Acta. 2005;78:165–174. [Google Scholar]
[43].Bajaj S, Sambhi SS, Madan AK. Predicting Anti-HIV Activity of Phenenthylthiazolethiourea (PETT) Analogs: Computational Approach Using Wiener’s Topochemical Index. J Mol Struct (Theochem) 2004;684:197–203. doi: 10.1016/j.theochem.2004.01.052. [DOI] [Google Scholar]
[44].Dureja H, Gupta S, Madan AK. Predicting Anti-HIV-1 Activity of 6-Arylbenzonitriles: Computational Approach Using Superaugmented Eccentric Connectivity Topochemical Indices. J Mol Graph Mod. 2008;26:1020–1029. doi: 10.1016/j.jmgm.2007.08.008. [DOI] [PubMed] [Google Scholar]
[45].Gupta S, Singh M, Madan AK. Predicting Anti-HIV Activity: Computational Approach Using A Novel Topological Descriptor. J Comput Aided Mol Des. 2001;15:671–678. doi: 10.1023/A:1011964003474. [DOI] [PubMed] [Google Scholar]
[46].Bajaj S, Sambhi SS, Madan AK. Model For Prediction of Anti-HIV Activity of 2-Pyridinone Derivatives Using Novel Topological Descriptor. QSAR Comb Sci. 2006;25:813–823. doi: 10.1002/qsar.200430918. [DOI] [Google Scholar]
[47].Sharma V, Goswami R, Madan AK. Eccentric Connectivity Index: A Novel Highly Discriminating Topological Descriptor For Structure-Property and Structure-Activity Studies. J Chem Inf Comput Sci. 1997;37:273–282. doi: 10.1021/ci960049h. [DOI] [Google Scholar]
[48].Gupta S, Singh M, Madan AK. Connective Eccentricity Index: A Novel Topological Descriptor For Predicting Biological Activity. J Mol Graph Mod. 2000;18:18–25. doi: 10.1016/S1093-3263(00)00027-9. [DOI] [PubMed] [Google Scholar]
[49].Gutman I, Ruscic B, Trinajstic N, Wicox CF. Graph Theory and Molecular Orbitals XII Acyclic Polyenes. J Chem Phys. 1975;62:3399–3405. doi: 10.1063/1.430994. [DOI] [Google Scholar]
[50].Gutman I, Randic M. Algebric Characterization of Skeletal Branching. Chem Phys Lett. 1977;47:15–19. doi: 10.1016/0009-2614(77)85296-2. [DOI] [Google Scholar]
[51].Wiener H. Correlation of Heats of Isomerization and Differences in Heat of Vaporization of Isomers Among the Paraffin Hydrocarbons. J Am Chem Soc. 1947;69:2636–2638. doi: 10.1021/ja01203a022. [DOI] [Google Scholar]
[52].Wiener H. Influence of Interatomic Forces on Paraffin Properties. J Chem Phys. 1947;15:766. doi: 10.1063/1.1746328. [DOI] [Google Scholar]
[53].Dureja H, Madan AK. Eccentric Connectivity Indices: New Generation Highly Discriminating Topological Descriptors for QSAR/QSPR Modeling. Med Chem Res. 2007;16:331–346. doi: 10.1007/s00044-007-9032-9. [DOI] [Google Scholar]
[54].Balaban AT. Topological Indices Based on Topological Distances in Molecular Graph. Pure Appl Chem. 1983;55:199–206. doi: 10.1351/pac198855020199. [DOI] [Google Scholar]
[55].Gupta S, Singh M, Madan AK. Superpendentic Index: A Novel Topological Descriptor For Predicting Biological Activity. J Chem Inf Comput Sci. 1999;39:272–277. doi: 10.1021/ci980073q. [DOI] [PubMed] [Google Scholar]
[56].Myles AJ, Feudale RN, Liu Y, Woody NA, Brown SD. An Introduction to Decision Tree Modeling. J Chemomet. 2004;18:275–285. doi: 10.1002/cem.873. [DOI] [Google Scholar]
[57].Kim H, Koehler GJ. Theory and Practice of Decision Tree Induction Omega. Int J Mgmt Sci. 1995;23:637–652. doi: 10.1016/0305-0483(95)00036-4. [DOI] [Google Scholar]
[58].Sprogar M, Kokol P, Zorman M, Podgorelec V, Yamamoto R, Masuda G, Sakamoto N. Supporting Medical Decisions with Vector Decision Trees. Stud Health Technol Inform. 2001;84:552–556. [PubMed] [Google Scholar]
[59].Kuo WJ, Chang RF, Chen DR, Lee CC. Data Mining with Decision Trees For Diagnosis of Breast Tumor in Medical Ultrasonic Images. Breast Cancer Res Treat. 2001;66:51–57. doi: 10.1023/A:1010676701382. [DOI] [PubMed] [Google Scholar]
[60].Decaestecker C, Remmelink M, Salmon I, Camby I, Goldschmidt D, Petein M, Vanham P, Pasteels JL, Kiss R. Methodological Aspects of Using Decision Trees to Characterize Leiomyomatous Tumors. Cytometry. 1996;24:83–92. doi: 10.1002/(SICI)1097-0320(19960501)24:1<83::AID-CYTO10>3.0.CO;2-R. [DOI] [PubMed] [Google Scholar]
[61].Wellman MP, Eckman MH, Fleming C, Marshall SL, Sonnenberg FA, Pauker SG. Automated Critiquing of Medical Decision Trees. Med Decis Making. 1989;9:272–284. doi: 10.1177/0272989X8900900407. [DOI] [PubMed] [Google Scholar]
[62].Sabbagh A, Darlu P. Data-Mining Methods As Useful Tools For Predicting Individual Drug Response: Application to CYP2D6 Data. Hum Hered. 2006;62:119–134. doi: 10.1159/000096416. [DOI] [PubMed] [Google Scholar]
[63].Kim HJ, Choo H, Cho YS, Koh HY, No KT, Pae AN. Classification of Dopamine, Serotonin, and Dual Antagonists by Decision Trees. Bioorg Med Chem. 2006;14:2763–2770. doi: 10.1016/j.bmc.2005.11.059. [DOI] [PubMed] [Google Scholar]
[64].Aitkenhead MJ. A Co-evolving Decision Tree Classification Method. Expert Syst Appl. 2008;34:18–25. doi: 10.1016/j.eswa.2006.08.008. [DOI] [Google Scholar]
[65].Dureja H, Gupta S, Madan AK. Topological Models For Prediction of Pharmacokinetic Parameters of Cephalosporins Using Random Forest, Decision Tree and Moving Average Analysis. Sci Pharm. 2008;76:377–394. doi: 10.3797/scipharm.0803-30. [DOI] [Google Scholar]
[66].Han L, Wang Y, Bryant SH. Developing and Validating Predictive Decision Tree Models From Mining Chemical Structural Fingerprints and High Throughput Screening Data in Pubchem. BMC Bioinformatics. 2008;9:401. doi: 10.1186/1471-2105-9-401. [DOI] [PMC free article] [PubMed] [Google Scholar]
[67].Nikolic S, Kavacevic G, Milicevic A, Trinanjstic N. The Zagreb Indices 30 Years After. Croat Chem Acta. 2003;76:113–124. [Google Scholar]

[b1-scipharm-2011-79-239] [1].Basha MR, Baker CH, Abdelrahim M. Biomarkers Clinical Relevance in Cancer: Emphasis on Breast Cancer and Prostate Cancer. Curr Trends Biotechnol Pharm. 2009;3:1–7. [Google Scholar]

[b2-scipharm-2011-79-239] [2].Kolibaba KS, Druker BJ. Protein Tyrosine Kinase and Cancer. Biochim Biophys Acta. 1997;1333:F217–248. doi: 10.1016/S0304-419X(97)00022-X. [DOI] [PubMed] [Google Scholar]

[b3-scipharm-2011-79-239] [3].Madhusudan S, Ganesan TS. Tyrosine Kinase Inhibitors in Cancer Therapy. Clin Biochem. 2004;37:618–635. doi: 10.1016/j.clinbiochem.2004.05.006. [DOI] [PubMed] [Google Scholar]

[b4-scipharm-2011-79-239] [4].Zetter BR. Angiogenesis and Tumor Metastasis. Ann Rev Med. 1998;49:407–424. doi: 10.1146/annurev.med.49.1.407. [DOI] [PubMed] [Google Scholar]

[b5-scipharm-2011-79-239] [5].Mustonen T, Alitalo K. Endothelial Receptor Tyrosine Kinases Involved in Angiogenesis. J Cell Biol. 1995;129:895–898. doi: 10.1083/jcb.129.4.895. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b6-scipharm-2011-79-239] [6].Carmeliet P, Jain RK. Angiogenesis in Cancer and Other Diseases. Nature. 2000;407:249–257. doi: 10.1038/35025220. [DOI] [PubMed] [Google Scholar]

[b7-scipharm-2011-79-239] [7].Westwell AD. A New Era in Cancer Therapeutics. Drug Discov Today. 2003;8:64–65. doi: 10.1016/S1359-6446(02)02575-8. [DOI] [PubMed] [Google Scholar]

[b8-scipharm-2011-79-239] [8].Ferrara N, Kerbel RS. Angiogenesis as a Therapeutic Target. Nature. 2005;438:967–974. doi: 10.1038/nature04483. [DOI] [PubMed] [Google Scholar]

[b9-scipharm-2011-79-239] [9].Yancopoulos GD, Davis S, Gale NW, Rudge JS, Wiegand SJ, Holash J. Vascular-Specific Growth Factors and Blood Vessel Formation. Nature. 2000;407:242–248. doi: 10.1038/35025215. [DOI] [PubMed] [Google Scholar]

[b10-scipharm-2011-79-239] [10].Hubbard SR, Till JH. Protein Tyrosine Kinase Structure and Function. Annu Rev Biochem. 2000;69:373–398. doi: 10.1146/annurev.biochem.69.1.373. [DOI] [PubMed] [Google Scholar]

[b11-scipharm-2011-79-239] [11].Shibuya M. Role of VEGF-FIt Receptor System in Normal and Tumor Angiogenesis. Adv Cancer Res. 1995;67:281–316. doi: 10.1016/S0065-230X(08)60716-2. [DOI] [PubMed] [Google Scholar]

[b12-scipharm-2011-79-239] [12].Gschwind A, Fischer OM, Ullrich A. The Discovery of Receptor Tyrosine Kinases: Targets For Cancer Therapy. Nat Rev Cancer. 2004;4:361–370. doi: 10.1038/nrc1360. [DOI] [PubMed] [Google Scholar]

[b13-scipharm-2011-79-239] [13].Folkman J. Angiogenesis in Cancer, Vascular, Rheumatoid and Other Disease. Nat Med. 1995;1:27–31. doi: 10.1038/nm0195-27. [DOI] [PubMed] [Google Scholar]

[b14-scipharm-2011-79-239] [14].Benjamin LE, Golijanin D, Itin A, Pode D, Keshet E. Selective Ablation of Immature Blood Vessels in Established Human Tumors Follows Vascular Endothelial Growth Factor Withdrawal. J Clin Invest. 1999;103:159–165. doi: 10.1172/JCI5028. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b15-scipharm-2011-79-239] [15].Glade Bender J, Cooney EM, Kandel JJ, Yamashiro DJ. Vascular Remodeling and Clinical Resistance to Antiangiogenic Cancer Therapy. Drug Resist Updat. 2004;7:289–300. doi: 10.1016/j.drup.2004.09.001. [DOI] [PubMed] [Google Scholar]

[b16-scipharm-2011-79-239] [16].Östman A. PDGF Receptors-Mediators of Autocrine Tumor Growth and Regulators of Tumor Vasculature and Stroma. Cytokine Growth Factor Rev. 2004;15:275–286. doi: 10.1016/j.cytogfr.2004.03.002. [DOI] [PubMed] [Google Scholar]

[b17-scipharm-2011-79-239] [17].Iwamoto H, Nakamuta M, Tada S, Sugimoto R, Munechika E, Nawata H. Platelet-Derived Growth Factor Receptor Tyrosine Kinase Inhibitor AG 1295 Attenuates Rat Hepatic Stellate Cell Growth. J Lab Clin Med. 2000;135:406–412. doi: 10.1067/mlc.2000.105974. [DOI] [PubMed] [Google Scholar]

[b18-scipharm-2011-79-239] [18].Chu SH, Small D. Mechanisms of Resistance to FLT3 Inhibitors. Drug Resist Updat. 2009;12:8–16. doi: 10.1016/j.drup.2008.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b19-scipharm-2011-79-239] [19].Potapova O, Laird AD, Nannini MA, Barone A, Li G, Moss KG, Cherrington JM, Mendel DB. Contribution of Individual Targets to the Antitumor Efficacy of the Multitargeted Receptor Tyrosine Kinase Inhibitor SU11248. Mol Cancer Ther. 2006;5:1280–1289. doi: 10.1158/1535-7163.MCT-03-0156. [DOI] [PubMed] [Google Scholar]

[b20-scipharm-2011-79-239] [20].Tang Y, Zhu W, Chen K, Jiang H. New Technologies in Computer-Aided Drug Design: Toward Target Identification and New Chemical Entity Discovery. Drug Disc Today: Technologies. 2006;3:307–313. doi: 10.1016/j.ddtec.2006.09.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b21-scipharm-2011-79-239] [21].Pârvu L. QSAR – A Piece of Drug Design. J Cell Mol Med. 2003;7:333–335. doi: 10.1111/j.1582-4934.2003.tb00235.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b22-scipharm-2011-79-239] [22].Duchowicz PR, Vitale MG, Castro EA, Fernàndez M, Caballero J. QSAR Analysis For Heterocyclic Antifungals. Bioorg Med Chem. 2007;15:2680–2689. doi: 10.1016/j.bmc.2007.01.039. [DOI] [PubMed] [Google Scholar]

[b23-scipharm-2011-79-239] [23].Hansch C, Leo A, Hoekman D. Exploring QSAR: Hydrophillic, Electronic and Steric Constants. American Chemical Society; Washington DC, USA: 1995. [Google Scholar]

[b24-scipharm-2011-79-239] [24].Chen SW, Li Z-R, Li XY. Prediction of Antifungal Activity by Support Vector Machine Approach. J Mol Struct (Theochem) 2005;731:73–81. doi: 10.1016/j.theochem.2005.06.032. [DOI] [Google Scholar]

[b25-scipharm-2011-79-239] [25].Mahmoudi N, Juliàn-Ortiz JV, Ciceron L, Gàlvez J, Mazier D, Danis M, Derouin F, Domenech RG. Identification of New Antimalarial Drugs by Linear Discriminant Analysis and Topological Virtual Screening. J Antimicrob Chemother. 2006;57:489–497. doi: 10.1093/jac/dki470. [DOI] [PubMed] [Google Scholar]

[b26-scipharm-2011-79-239] [26].Ivanciuc O, Ivanciuc T, Klein DJ, Seitz WA, Balaban AT. Wiener Index Extension by Counting Even/Odd Graph Distances. J Chem Inf Comput Sci. 2001;41:536–549. doi: 10.1021/ci000086f. [DOI] [PubMed] [Google Scholar]

[b27-scipharm-2011-79-239] [27].Garcia-Domenech R, Galvez J, de Jullian-Ortiz JV, Pogliani L. Some New Trends in Chemical Graph Theory. Chem Rev. 2008;108:1127–1169. doi: 10.1021/cr0780006. [DOI] [PubMed] [Google Scholar]

[b28-scipharm-2011-79-239] [28].Ivanciuc O, Dellivers J, Balaban AT. Topological Indices and Related Descripors in QSAR and QSPR. Gordon and Breach Science Publishers; The Netherlands: 1999. pp. 697–777. [Google Scholar]

[b29-scipharm-2011-79-239] [29].Estrada E, Uriarte E. Recent Advances on the Role of Topological Indices in Drug Discovery Research. Curr Med Chem. 2001;8:1573–1588. doi: 10.2174/0929867013371923. [DOI] [PubMed] [Google Scholar]

[b30-scipharm-2011-79-239] [30].Randic M. On Characterization of Molecular Branching. J Am Chem Soc. 1975;97:6609–6615. doi: 10.1021/ja00856a001. [DOI] [Google Scholar]

[b31-scipharm-2011-79-239] [31].Needham DE, Wei IC, Seybold PG. Molecular Modeling of the Physical Properties of the Alkanes. J Am Chem Soc. 1988;110:4186–4194. doi: 10.1021/ja00221a015. [DOI] [Google Scholar]

[b32-scipharm-2011-79-239] [32].Basak SC. Use of Molecular Complexity Indices in Predictive Pharmacology and Toxicology: A QSAR Approach. Med Sci Res. 1987;15:605–609. [Google Scholar]

[b33-scipharm-2011-79-239] [33].Estrada E, Patlewicz G, Uriate E. From Molecular Graphs to Drugs. A Review on the Use of Topological Indices in Drug Design and Discovery. Indian J Chem. 2003;42A:1315–1329. [Google Scholar]

[b34-scipharm-2011-79-239] [34].Dai Y, Hartandi K, Ji Z, Ahmed AA, Albert DH, Bauch JL, Bouska JJ, Bousquet PF, Cunha GA, Glasar KB, Harris CM, Hickman D, Guo J, Li J, Marcotte PA, Marsh KC, Moskey MD, Martin RL, Olson AM, Osterling DJ, Pease LJ, Soni NB, Stewart KD, Stoll VS, Tapang P, Reuter DR, Davidsen SK, Michaelides MR. Discovery of N-(4-(3-Amino-1H-Indazol-4-yl)Phenyl)-N-(2-Fluoro-5-Methylphenyl) Urea (ABT-869), A 3-Aminoindazole-Based Orally Active Multitargeted Receptor Tyrosine Kinase Inhibitor. J Med Chem. 2007;50:1584–1597. doi: 10.1021/jm061280h. [DOI] [PubMed] [Google Scholar]

[b35-scipharm-2011-79-239] [35].Goel A, Madan AK. Structure-Activity Study on Anti-Inflammatory Pyrazole Carboxylic Acid Hydrazide Analogs Using Molecular Connectivity Indices. J Chem Inf Comput Sci. 1995;35:510–514. doi: 10.1021/ci00025a019. [DOI] [PubMed] [Google Scholar]

[b36-scipharm-2011-79-239] [36].Dureja H, Madan AK. Topochemical Models for Prediction of Cyclin-Dependent Kinase 2 Inhibitory Activity of Indole-2-ones. J Mol Mod. 2005;11:525–531. doi: 10.1007/s00894-005-0276-3. [DOI] [PubMed] [Google Scholar]

[b37-scipharm-2011-79-239] [37].Gupta S, Singh M, Madan AK. Novel Topochemical Descriptors for Predicting Anti-HIV Activity. Indian J Chem. 2003;42A:1414–1425. [Google Scholar]

[b38-scipharm-2011-79-239] [38].Bajaj S. Study on Topochemical Descriptors For the Prediction of Physicochemical and Biological Properties of Molecules. 2005. Ph.D. Thesis, Guru Gobind Singh Indraprastha University, India.

[b39-scipharm-2011-79-239] [39].Bajaj S, Sambhi SS, Madan AK. Prediction of Carbonic Anhydrase Activation by Tri-/Tetrasubstituted-Pyridinium-Azole Drugs: A Computational Approach Using Novel Topochemical Descriptor. QSAR Comb Sci. 2004;23:506–514. doi: 10.1002/qsar.200439999. [DOI] [Google Scholar]

[b40-scipharm-2011-79-239] [40].Kumar V, Sardana S, Madan AK. Predicting Anti-HIV Activity of 2,3-Diaryl-1,3-Thiazolidin-4-ones: Computational Approach Using Reformed Eccentric Connectivity Index. J Mol Mod. 2004;10:399–407. doi: 10.1007/s00894-004-0215-8. [DOI] [PubMed] [Google Scholar]

[b41-scipharm-2011-79-239] [41].Gupta S. Application and Development of Graph Invariants of Drug Design. 2002. Ph.D. Thesis, Punjabi University, Patiala, India.

[b42-scipharm-2011-79-239] [42].Bajaj S, Sambhi SS, Madan AK. Prediction of Anti-Inflammatory Activity of N-Arylanthranilic Acids: Computational Approach Using Redefined Zagreb Indices. Croat Chem Acta. 2005;78:165–174. [Google Scholar]

[b43-scipharm-2011-79-239] [43].Bajaj S, Sambhi SS, Madan AK. Predicting Anti-HIV Activity of Phenenthylthiazolethiourea (PETT) Analogs: Computational Approach Using Wiener’s Topochemical Index. J Mol Struct (Theochem) 2004;684:197–203. doi: 10.1016/j.theochem.2004.01.052. [DOI] [Google Scholar]

[b44-scipharm-2011-79-239] [44].Dureja H, Gupta S, Madan AK. Predicting Anti-HIV-1 Activity of 6-Arylbenzonitriles: Computational Approach Using Superaugmented Eccentric Connectivity Topochemical Indices. J Mol Graph Mod. 2008;26:1020–1029. doi: 10.1016/j.jmgm.2007.08.008. [DOI] [PubMed] [Google Scholar]

[b45-scipharm-2011-79-239] [45].Gupta S, Singh M, Madan AK. Predicting Anti-HIV Activity: Computational Approach Using A Novel Topological Descriptor. J Comput Aided Mol Des. 2001;15:671–678. doi: 10.1023/A:1011964003474. [DOI] [PubMed] [Google Scholar]

[b46-scipharm-2011-79-239] [46].Bajaj S, Sambhi SS, Madan AK. Model For Prediction of Anti-HIV Activity of 2-Pyridinone Derivatives Using Novel Topological Descriptor. QSAR Comb Sci. 2006;25:813–823. doi: 10.1002/qsar.200430918. [DOI] [Google Scholar]

[b47-scipharm-2011-79-239] [47].Sharma V, Goswami R, Madan AK. Eccentric Connectivity Index: A Novel Highly Discriminating Topological Descriptor For Structure-Property and Structure-Activity Studies. J Chem Inf Comput Sci. 1997;37:273–282. doi: 10.1021/ci960049h. [DOI] [Google Scholar]

[b48-scipharm-2011-79-239] [48].Gupta S, Singh M, Madan AK. Connective Eccentricity Index: A Novel Topological Descriptor For Predicting Biological Activity. J Mol Graph Mod. 2000;18:18–25. doi: 10.1016/S1093-3263(00)00027-9. [DOI] [PubMed] [Google Scholar]

[b49-scipharm-2011-79-239] [49].Gutman I, Ruscic B, Trinajstic N, Wicox CF. Graph Theory and Molecular Orbitals XII Acyclic Polyenes. J Chem Phys. 1975;62:3399–3405. doi: 10.1063/1.430994. [DOI] [Google Scholar]

[b50-scipharm-2011-79-239] [50].Gutman I, Randic M. Algebric Characterization of Skeletal Branching. Chem Phys Lett. 1977;47:15–19. doi: 10.1016/0009-2614(77)85296-2. [DOI] [Google Scholar]

[b51-scipharm-2011-79-239] [51].Wiener H. Correlation of Heats of Isomerization and Differences in Heat of Vaporization of Isomers Among the Paraffin Hydrocarbons. J Am Chem Soc. 1947;69:2636–2638. doi: 10.1021/ja01203a022. [DOI] [Google Scholar]

[b52-scipharm-2011-79-239] [52].Wiener H. Influence of Interatomic Forces on Paraffin Properties. J Chem Phys. 1947;15:766. doi: 10.1063/1.1746328. [DOI] [Google Scholar]

[b53-scipharm-2011-79-239] [53].Dureja H, Madan AK. Eccentric Connectivity Indices: New Generation Highly Discriminating Topological Descriptors for QSAR/QSPR Modeling. Med Chem Res. 2007;16:331–346. doi: 10.1007/s00044-007-9032-9. [DOI] [Google Scholar]

[b54-scipharm-2011-79-239] [54].Balaban AT. Topological Indices Based on Topological Distances in Molecular Graph. Pure Appl Chem. 1983;55:199–206. doi: 10.1351/pac198855020199. [DOI] [Google Scholar]

[b55-scipharm-2011-79-239] [55].Gupta S, Singh M, Madan AK. Superpendentic Index: A Novel Topological Descriptor For Predicting Biological Activity. J Chem Inf Comput Sci. 1999;39:272–277. doi: 10.1021/ci980073q. [DOI] [PubMed] [Google Scholar]

[b56-scipharm-2011-79-239] [56].Myles AJ, Feudale RN, Liu Y, Woody NA, Brown SD. An Introduction to Decision Tree Modeling. J Chemomet. 2004;18:275–285. doi: 10.1002/cem.873. [DOI] [Google Scholar]

[b57-scipharm-2011-79-239] [57].Kim H, Koehler GJ. Theory and Practice of Decision Tree Induction Omega. Int J Mgmt Sci. 1995;23:637–652. doi: 10.1016/0305-0483(95)00036-4. [DOI] [Google Scholar]

[b58-scipharm-2011-79-239] [58].Sprogar M, Kokol P, Zorman M, Podgorelec V, Yamamoto R, Masuda G, Sakamoto N. Supporting Medical Decisions with Vector Decision Trees. Stud Health Technol Inform. 2001;84:552–556. [PubMed] [Google Scholar]

[b59-scipharm-2011-79-239] [59].Kuo WJ, Chang RF, Chen DR, Lee CC. Data Mining with Decision Trees For Diagnosis of Breast Tumor in Medical Ultrasonic Images. Breast Cancer Res Treat. 2001;66:51–57. doi: 10.1023/A:1010676701382. [DOI] [PubMed] [Google Scholar]

[b60-scipharm-2011-79-239] [60].Decaestecker C, Remmelink M, Salmon I, Camby I, Goldschmidt D, Petein M, Vanham P, Pasteels JL, Kiss R. Methodological Aspects of Using Decision Trees to Characterize Leiomyomatous Tumors. Cytometry. 1996;24:83–92. doi: 10.1002/(SICI)1097-0320(19960501)24:1<83::AID-CYTO10>3.0.CO;2-R. [DOI] [PubMed] [Google Scholar]

[b61-scipharm-2011-79-239] [61].Wellman MP, Eckman MH, Fleming C, Marshall SL, Sonnenberg FA, Pauker SG. Automated Critiquing of Medical Decision Trees. Med Decis Making. 1989;9:272–284. doi: 10.1177/0272989X8900900407. [DOI] [PubMed] [Google Scholar]

[b62-scipharm-2011-79-239] [62].Sabbagh A, Darlu P. Data-Mining Methods As Useful Tools For Predicting Individual Drug Response: Application to CYP2D6 Data. Hum Hered. 2006;62:119–134. doi: 10.1159/000096416. [DOI] [PubMed] [Google Scholar]

[b63-scipharm-2011-79-239] [63].Kim HJ, Choo H, Cho YS, Koh HY, No KT, Pae AN. Classification of Dopamine, Serotonin, and Dual Antagonists by Decision Trees. Bioorg Med Chem. 2006;14:2763–2770. doi: 10.1016/j.bmc.2005.11.059. [DOI] [PubMed] [Google Scholar]

[b64-scipharm-2011-79-239] [64].Aitkenhead MJ. A Co-evolving Decision Tree Classification Method. Expert Syst Appl. 2008;34:18–25. doi: 10.1016/j.eswa.2006.08.008. [DOI] [Google Scholar]

[b65-scipharm-2011-79-239] [65].Dureja H, Gupta S, Madan AK. Topological Models For Prediction of Pharmacokinetic Parameters of Cephalosporins Using Random Forest, Decision Tree and Moving Average Analysis. Sci Pharm. 2008;76:377–394. doi: 10.3797/scipharm.0803-30. [DOI] [Google Scholar]

[b66-scipharm-2011-79-239] [66].Han L, Wang Y, Bryant SH. Developing and Validating Predictive Decision Tree Models From Mining Chemical Structural Fingerprints and High Throughput Screening Data in Pubchem. BMC Bioinformatics. 2008;9:401. doi: 10.1186/1471-2105-9-401. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b67-scipharm-2011-79-239] [67].Nikolic S, Kavacevic G, Milicevic A, Trinanjstic N. The Zagreb Indices 30 Years After. Croat Chem Acta. 2003;76:113–124. [Google Scholar]

PERMALINK

Models for the Prediction of Receptor Tyrosine Kinase Inhibitory Activity of Substituted 3-Aminoindazole Analogues

Monika Gupta

Harish Dureja

Anil Kumar Madan

Abstract

Introduction