background: As the application of silica nanomaterials continues to expand, increasing chances of its exposure to the human body and potential harm are anticipated. Although the toxicity of silica nanomaterials is assumed to be affected by their physio-chemical properties, including size and surface functionalization, its molecular mechanisms remain unclear. We hypothesized that analysis of intracellular localization of the particles and subsequent intracellular signaling could reveal a novel determinant of inflammatory response against silica particles with different physico-chemical properties. results: We employed a murine intratracheal instillation model of amorphous silica nanoparticles (NPs) exposure to compare their in vivo toxicities in the respiratory system. Pristine silica-NPs of 50 nm diameters (50 nm-plain) induced airway-centered lung injury with marked neutrophilic infiltration. By contrast, instillation of pristine silica particles of a larger diameter (3 μm; 3 μm-plain) significantly reduced the severity of lung injury and neutrophilic infiltration, possibly through attenuated induction of neutrophil chemotactic chemokines including MIP2. Ex vivo analysis of alveolar macrophages as well as in vitro assessment using RAW264.7 cells revealed a remarkably lower cellular uptake of 3 μm-plain particles compared with 50 nm-plain, which is assumed to be the underlying mechanism of attenuated immune response. The severity of lung injury and neutrophilic infiltration was also significantly reduced after intratracheal instillation of silica NPs with an amine surface modification (50 nm-NH) when compared with 50 nm-plain. Despite unchanged efficacy in cellular uptake, treatment with 50 nm-NH induced a significantly attenuated immune response in RAW264.7 cells. Assessment of intracellular redox signaling revealed increased reactive oxygen species (ROS) in endosomal compartments of RAW264.7 cells treated with 50 nm-plain when compared with vehicle-treated control. In contrast, augmentation of endosomal ROS signals in cells treated with 50 nm-NH was significantly lower. Moreover, selective inhibition of NADPH oxidase 2 (NOX2) was sufficient to inhibit endosomal ROS bursts and induction of chemokine expressions in cells treated with silica NPs, suggesting the central role of endosomal ROS generated by NOX2 in the regulation of the inflammatory response in macrophages that endocytosed silica NPs. conclusions: Our murine model suggested that the pulmonary toxicity of silica NPs depended on their physico-chemical properties through distinct mechanisms. Cellular uptake of larger particles by macrophages decreased, while surface amine modification modulated endosomal ROS signaling via NOX2, both of which are assumed to be involved in mitigating immune response in macrophages and resulting lung injury.