Selective cytotoxicity of PAMAM G5 core–PAMAM G2.5 shell tecto-dendrimers on melanoma cells

BACKGROUND The controlled introduction of covalent linkages between dendrimer building blocks leads to polymers of higher architectural order known as tecto-dendrimers. Because of the few simple steps involved in their synthesis, tecto-dendrimers could expand the portfolio of structures beyond commercial dendrimers, due to the absence of synthetic drawbacks (large number of reaction steps, excessive monomer loading, and lengthy chromatographic separations) and structural constraints of high-generation dendrimers (reduction of good monodispersity and ideal dendritic construction due to de Gennes dense-packing phenomenon). However, the biomedical uses of tecto-dendrimers remain unexplored. In this work, after synthesizing saturated shell core-shell tecto-dendrimers using amine-terminated polyamidoamine (PAMAM) generation 5 (G5) as core and carboxyl-terminated PAMAM G2.5 as shell (G5G2.5 tecto-dendrimers), we surveyed for the first time the main features of their interaction with epithelial cells. METHODS Structural characterization of G5G2.5 was performed by polyacrylamide gel electrophoresis, matrix-assisted laser desorption time-of-flight mass spectrometry, and microscopic techniques; their hydrodynamic size and Z-potential was also determined. Cellular uptake by human epidermal keratinocytes, colon adenocarcinoma, and epidermal melanoma (SK-Mel-28) cells was determined by flow cytometry. Cytotoxicity was determined by mitochondrial activity, lactate dehydrogenase release, glutathione depletion, and apoptosis/necrosis measurement. RESULTS The resultant 60%-67% saturated shell, 87,000-dalton G5G2.5 (mean molecular weight) interacted with cells in a significantly different fashion in comparison to their building blocks and to its closest counterpart, PAMAM G6.5. After being actively taken up by epithelial cells, G5G2.5 caused cytotoxicity only on SK-Mel-28 cells, including depletion of intracellular glutathione and fast necrosis that was manifested above 5 μM G5G2.5. It cannot be discounted that traces of LiCl within G5G2.5 were involved in such deleterious effects. CONCLUSION These preliminary results suggest that at concentrations that do not damage healthy keratinocytes, G5G2.5 could display antimelanoma activity.