Nanoparticle dispersions are being developed in an attempt to improve the utility of poorly soluble drug candidates and may have additional benefits such as improved oral bioavailability or cellular targeting. However, there is currently a lack of information regarding the interaction of nanodispersions with biological systems. The purpose of this study was to investigate the influence of nanoparticle properties on the cytotoxicty of nanodispersions using ritonavir as a paradigm. A library of 68 nanodispersions with varying combinations of 23 excipients (10 polymers and 13 surfactants) was generated using previously published methodology (Zhang et al., 2008). Particle size (z average) ranged from 212 – 3682 nm and surface charge (zeta potential) ranged from -92 – +36 mV. The cytotoxicity of each nanodispersion was assessed in hepatic (HepG2), intestinal (Caco-2), lymphocyte (CEM), monocyte (THP-1) and macrophage (A-THP-1) cell systems. Cells were separately seeded at a density of 2.5 x 104
/ 100 µl
into each well of a 96 well plate and incubated for 24 hours at 37o
C and 5% CO2
. Media was then aspirated and replaced with media containing 0.1, 1, 10, 100, 500 or 1000 µM of each dispersion and incubated for a further 24 hours. Subsequently, 100 µl of CellTiter-Glo® Reagent (Promega, USA) was added to each well and incubated at room temperature for 10 minutes. Luminescence was then measured using a Tecan Genios plate reader. Data were used to calculate IC50
values by non-linear regression in Graphpad prism. SPSS was then used to assess the impact of either particle properties (z average, zeta potential and polydispersity) or excipients on the resultant IC50
s in univariate and multivariate models. A wide variability in cytotoxicity was observed in all cell systems with IC50
values ranging from 0.3 - 734 mM in A-THP-1 cells (IC50
of an aqueous solution in this cell system was 8.1 mM). Particle size and surface charge but not polydispersity were related to the cytotoxcity of the dispersions. For example, a linear relationship between z average and logIC50
= 0.09) and a quadratic relationship between zeta potential and logIC50
= 0.09) was observed in A-THP1 cells. Furthermore, in a multiple regression model, z average and zeta potential accounted for 25% of the variability in A-THP1 cytotoxicity. A number of polymers (F127, hydrolysed gelatine) and surfactants (chremaphor, tween 80, brij58, hyamine, CTAB) were also associated with cytotoxicity and this was partially explained by their impact upon particle properties. Importantly, none of the excipients were cytotoxic at any of the concentrations used. In conclusion, these data indicate that nanoparticle properties directly influence cytoxicity of the resulting dispersions. Further work is now required to investigate the mechanisms for these observations and to establish the source of the unexplained variability in cytotoxicity.
Zhang H, Wang D, Butler R, Campbell NL, Long J, Tan B, Duncalf DJ, Foster AJ, Hopkinson A, Taylor D, Angus D, Cooper AI and Rannard SP (2008) Formation and enhanced biocidal activity of water-dispersable organic nanoparticles. Nat Nanotechnol 3:506-511.