Much of the concerns regarding engineered NP toxicity are based on knowledge from previous studies on ambient and environmental particles. E.g., the effects of exposure to silica dust particles have been studied intensively due to the carcinogenicity of crystalline silica. However, the increasing usage of engineered amorphous silica NPs has emphasized the need for further mechanistic insight to predict the consequences of exposure to the amorphous type of silica NPs. Recently, the parallelogram approach was proposed as a scheme to assess biological effects of nanomaterials (Krug and Wick, 2011). Accordingly, the present study focused on the cytotoxicity of amorphous silica NPs in six different cell lines selected to explore the significance of tissue type and species. The cells were selected as three pairs of human/mouse cell lines derived from lung epithelium (A549 and ASB-XIV), colon epithelium (CaCo-2 and Colon-26) and macrophages (THP-1 and J774A.1). Viability data demonstrated that macrophages were most sensitive to silica NP and interestingly, murine cell lines were generally found to be more sensitive than comparable human cell lines. Further studies were conducted in the human epithelial lung cell line, A549, to investigate the mechanism of action. A concentration-dependent increase of cellular reactive oxygen species was demonstrated in silica NP exposed A549 cells. However, induction of oxidative stress related pathways was not found after silica NP exposure in gene array studies on global gene expression. Instead, up-regulated genes primarily related to lipid metabolism and biosynthesis whereas down-regulated genes were enriched in several processes, including transcription, cell junction and extra cellular matrix (ECM)-receptor interaction. Accordingly, our data suggest that oxidative stress might not be the main mediator of amorphous silica NP toxicity. Furthermore, the differential response in different cell lines suggests careful consideration in the choice of cell models and cautions that interspecies extrapolation may have to consider higher sensitivity in mice towards NPs.