BACKGROUND: Exposure to respirable crystalline silica particles, as opposed to amorphous silica, is associated with lung inflammation, pulmonary fibrosis (silicosis), and potentially with lung cancer. We used Affymetrix/GeneSifter microarray analysis to determine whether gene expression profiles differed in a human bronchial epithelial cell line (BEAS 2B) exposed to cristobalite vs. amorphous silica particles at non-toxic and equal surface areas (75 and 150 × 106µm2/cm2). Bio-Plex analysis was also used to determine profiles of secreted cytokines and chemokines in response to both particles. Finally, primary human bronchial epithelial cells (NHBE) were used to comparatively assess silica particle-induced alterations in gene expression. RESULTS: Microarray analysis at 24 hours in BEAS 2B revealed 333 and 631 significant alterations in gene expression induced by cristobalite at low (75) and high (150 × 106µm2/cm2) amounts, respectively (p < 0.05/cut off = 2.0-fold change). Exposure to amorphous silica micro-particles at high amounts (150 × 106µm2/cm2) induced 108 significant gene changes. Bio-Plex analysis of 27 human cytokines and chemokines revealed 9 secreted mediators (p < 0.05) induced by crystalline silica, but none were induced by amorphous silica. QRT-PCR revealed that cristobalite selectively up-regulated stress-related genes and cytokines (FOS, ATF3, IL6 and IL8) early and over time (2, 4, 8, and 24 h). Patterns of gene expression in NHBE cells were similar overall to BEAS 2B cells. At 75 × 106µm2/cm2, there were 339 significant alterations in gene expression induced by cristobalite and 42 by amorphous silica. Comparison of genes in response to cristobalite (75 × 106µm2/cm2) revealed 60 common, significant gene alterations in NHBE and BEAS 2B cells. CONCLUSIONS: Cristobalite silica, as compared to synthetic amorphous silica particles at equal surface area concentrations, had comparable effects on the viability of human bronchial epithelial cells. However, effects on gene expression, as well as secretion of cytokines and chemokines, drastically differed, as the crystalline silica induced more intense responses. Our studies indicate that toxicological testing of particulates by surveying viability and/or metabolic activity is insufficient to predict their pathogenicity. Moreover, they show that acute responses of the lung epithelium, including up-regulation of genes linked to inflammation, oxidative stress, and proliferation, as well as secretion of inflammatory and proliferative mediators, can be indicative of pathologic potential using either immortalized lines (BEAS 2B) or primary cells (NHBE). Assessment of the degree and magnitude of these responses in vitro are suggested as predictive in determining the pathogenicity of potentially harmful particulates