Journal of Pharmaceutical Research International

A QSAR Study of New Caffeine Derivatives with Epithelial Anticancer Activity

Luana K. S. Gonçalves *

Department of Biological Sciences and Health, Laboratory of Modeling and Computational Chemistry, Federal University of Amapá, Macapá, Brazil.

Josinete B. Viera

Department of Biological Sciences and Health, Laboratory of Modeling and Computational Chemistry, Federal University of Amapá, Macapá, Brazil.

Nayara S. R. Silva

Department of Biological Sciences and Health, Laboratory of Modeling and Computational Chemistry, Federal University of Amapá, Macapá, Brazil.

César F. Santos

Department of Biological Sciences and Health, Laboratory of Modeling and Computational Chemistry, Federal University of Amapá, Macapá, Brazil.

Francinaldo S. Braga

Department of Biological Sciences and Health, Laboratory of Modeling and Computational Chemistry, Federal University of Amapá, Macapá, Brazil.

Josivan S. Costa

Department of Biological Sciences and Health, Laboratory of Modeling and Computational Chemistry, Federal University of Amapá, Macapá, Brazil and Postgraduate Program in Biotechnology and Biodiversity-Network BIONORTE, Federal University of Amazonas, Manaus, Brazil.

Williams J. C. Macêdo

Postgraduate Program in Biotechnology and Biodiversity-Network BIONORTE, Federal University of Amazonas, Manaus, Brazil and V

Carlos H. T. P. Silva

Faculty of Pharmaceutical Sciences of Ribeirão Preto, University of São Paulo, Ribeirão Preto, São Paulo, Brazil.

Lorane I. S. Hage-Melim

Department of Biological Sciences and Health, Laboratory of Modeling and Computational Chemistry, Federal University of Amapá, Macapá, Brazil.

Cleydson Breno R. Santos *

Department of Biological Sciences and Health, Laboratory of Modeling and Computational Chemistry, Federal University of Amapá, Macapá, Brazil and Postgraduate Program in Biotechnology and Biodiversity-Network BIONORTE, Federal University of Amazonas, Manaus, Brazil and Faculty of Pharmaceutical Sciences of Ribeirão Preto, University of São Paulo, Ribeirão Preto, São Paulo, Brazil.

*Author to whom correspondence should be addressed.

---

Abstract

Aims: To study and propose new caffeine derivatives with epithelial anticancer activity using quantum chemistry methods and multivariate analysis (PCA, HCA, PLS and PCR).
Place and Duration of Study: Laboratory of Modeling and Computational Chemistry at Federal University of Amapá (UNIFAP), Macapá, Brazil, between March 2014 and February 2015.
Methodology: Caffeine and 31 derivatives with epithelial anticancer activity were selected from the literature, and modeled with the GaussView 3.0 program. The optimization was performed using the DFT method and B3LYP/6-31G** base set implemented in the Gaussian 03 program. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) were employed to select the molecular descriptors related to epithelial anticancer activity. The Pearson correlation between activity and important descriptors were used for the regression partial least squares (PLS) and principal component regression (PCR) models built, and these models were used to predict the anticancer activity of fourteen new caffeine derivatives (test set) with unknown activity.
Results: The PCA results showed that the descriptors related to the compounds with anticancer activity were: area (A²), distance radical 1 (dR1), distance radical 3 (dR3), radical partition coefficient 1 (logPR1) and radical partition coefficient 3 (logPR3). HCA showed similar results obtained with PCA, and the compounds were grouped in accordance with their biological activities. The results obtained with the PLS and PCR models were close, with variation between the models of R²=±0.005, R²ajust=±0.1998, s=±0.0053, F_(5,27)=±49.2261, Q²=±0.012, SEV=±0.0117, PRESS=±0.0473 and S_PRESS=±0.0087. The PLS model showed that eight compounds of the test set (37 and 40-46) are predicted to be more active, and they had values of ICT₅₀<0.48mM. However, the PCR model only seven compounds of all test sets (33-36, 38, 39 and 42) were predicted as most active, which showed values of ICT₅₀≥0.48mM, a total of 7 compounds proposed as less active of fourteen suggested compounds. However, compounds 37, 40, 41, 43, 44, 45 and 46 were the ones that had values of ICT₅₀<0.48mM in both PLS and PCR models, suggesting that these new compounds in the two models are more potent than caffeine may be tested for epithelial anticancer activity.
Conclusion: The PLS and PCR models showed good predictive ability. The test set showed for seven new caffeine compounds satisfactory results for epithelial anticancer activity. This strategy is fundamental for use in experimental syntheses and biological evaluation, and to understand the structural requirements for designing new ligands as anticancer agents.

Keywords: Caffeine, epithelial cancer, molecular modeling, B3LYP/6-31G**, QSAR

---

---