Abstract
Clinical use of CuO nanoparticles (NPs) as antibacterials can be hampered by their toxicity to human cells. We hypothesized that certain surface functionalizations of CuO NPs may render NPs toxic to bacteria, but still be relatively harmless to human cells. To control this hypothesis, the toxicity of differently functionalized CuO NPs to bacteria Escherichia coli vs human cells (THP-1 macrophages and HACAT keratinocytes) was compared using similar conditions and end points. CuO NPs functionalized with polyethylene glycol (CuO-PEG), carboxyl (CuO-COOH, anionic), ammonium (CuO-NH4+, cationic) and unfunctionalized CuO NPs and CuSO4 (controls) were tested. In general, the toxicity of Cu compounds decreased in the following order: CuO-NH4+ > unfunctionalized CuO > CuSO4 > CuO-COOH > CuO-PEG. Positively charged unfunctionalized CuO and especially CuO-NH4+ proved most toxic (24-h EC50 = 21.7-47 mg/l) and had comparable toxicity to bacterial and mammalian cells. The multivariate analysis revealed that toxicity of these NPs was mostly attributed to their positive zeta potential, small hydrodynamic size, high Cu dissolution, and induction of reactive oxygen species (ROS) and TNF-α. In contrast, CuO-COOH and CuO-PEG NPs had lower toxicity to human cells compared to bacteria despite efficient uptake of these NPs by human cells. In addition, these NPs did not induce TNF-α and ROS. Thus, by varying the NP functionalization and Cu form (soluble salt vs NPs), it was possible to "target" the toxicity of Cu compounds, whereas carboxylation and PEGylation rendered CuO NPs that were more toxic to bacteria than to human cells envisaging their use in medical antibacterial products.