Anticancer a-helical peptides and their structure / function relationships underpinning their interactions with tumour cell membranes

Abstract

Over the last decade there has been an increase in premature deaths due to cancer. The identification of new chemotherapeutic agents has led to the discovery of toxic peptides with cancer cell specificity. Extensive research is therefore required to determine how these anticancer peptides (ACPs) function and how their structural characteristics relate to both specificity and efficacy.
An ACP database comprising of 158 peptides was sub-divided into 3 datasets as follows: ACP 1 (inactive) containing 21 peptides, ACPAO (anticancer only) containing 14 peptides and ACPT (toxins) containing 123 peptides. Multivariate analysis of the physiochemical properties and toxicity indicated that no significant difference (Kruskal-Wallis = 8.82; p = 0.18) was observed in the median toxicity of the ACP datasets against the cancerous and non-cancerous cell lines.
In contrast to the oligopeptides from Mtaldon and Argos control dataset, C residues were observed to be completely absent from all 3 ACP subsets with an abundance of the positively charged K residues (relative frequency >0.3) in the ACPAO and ACPTpeptides. This suggested that positive charge and the absence of C residues may facilitate peptide binding to anionic
cancer cell membranes. Furthermore, lower levels of K residues were observed in the ACP 1 datasets, indicating that positive net charge may also be important for efficacy. A combination of several other physiochemical properties was analysed using box-plot analysis, linear regression and frequency distribution analysis. It was found that whilst sequence length, net charge, mean hydrophobicity (<H>) and amphiphilicity (<I'H>) of the peptides did not show a linear correlation with the toxicity, they may help increase efficacy in the case of some tumour cell lines. Three-dimensional clustering using net charge, <H> and <p a> was carried out using the unweighted pair group method with arithmetic mean (UPGMA). The cluster analysis results showed that there was no optimal combination of these properties for high level efficacy and no
specific residue arrangement was found to affect ACP specificity for cancer cells. Therefore, inter-quartile ranges were identified for key parameters but it was observed that no single optimal values could be obtained, emphasizing the complexity of these biological systems and the interdependence of these parameters. Further work in this study therefore, involved the investigation of secondary structural elements such as hydrophobicity gradients generated by residue arrangements. It was observed that although surface activity and amphiphilicity were important parameters, hydrophobic gradients were not essential for anticancer activity or selectivity.