Crystalline structures such as quartz particles have previously been reported to activate the NALP3 inflammasome [14, 25], a mechanism suspected to play an important role in the development of diseases such as silicosis . In the present study we show how different non-crystalline (amorphous) silica particles from nano- to submicro-sizes are potent inducers of inflammasome activation, thus acting via the NALP3 assembly, leading to a caspase-1-dependent pro-IL-1β maturation and IL-1β release. When related to surface area the order of potency for IL-1β release in the lung macrophages was MinUsil 5 > fumed, silica > Si500 > Si50. The non-crystalline and crystalline silica particles are suggested to operate via similar mechanisms, since inhibitors of uptake and phagosomal stabilization seem to affect the inflammasome activation by these different particles in a similar manner, as demonstrated in RAW-macrophages.
Our results support and elaborate previous findings suggesting that a crystalline structure is not required for silica particles to activate the inflammasome. Non-crystalline silica particles have been found to induce a boost of IL-1β release in LPS-primed murine macrophages , TPA-primed THP-1 monocytic cells  murine dendritic cells  and keratinocytes . In the dendritic cells the existence of an inflammasome mechanism was demonstrated by using cells from NALP3- and caspase-1-deficient mice . For the THP-1 cells the involvement of an inflammasome mechanism was supported by inhibition of IL-1β release by zYVAD, an inhibitor of caspase-1 . In the present study the inflammasome mechanism induced by different non-crystalline silica particles was demonstrated both by use of the caspase-1 inhibitor zYVAD and by siRNA against the NALP3 in RAW274.7 macrophages.
Some studies also point to other nanoparticles that may induce NALP3 inflammasome responses, such as TiO2-particles, amino-functionalized polystyrene particles and carbon black particles in macrophages [29, 31–33]. In contrast, nanoparticles, like diesel exhaust particles, ZnO-NPs and carboxyl-functionalized polystyrene NPs [25, 29, 31], do not seem to show such responses. A remaining question is which nanoparticle characteristics and types of functionalisation are required to elicit NALP3 inflammasome responses.
In previous studies the IL-1β responses to non-crystalline particles varied with size, and also with cell type used. Morishige and coworkers  observed that the non-crystalline silica particles of 1000 nm size induced an IL-1β release in TPA-primed THP-1 cells, whereas the 30 to 300 nm counterparts were without effect. In murine macrophages approximate similar IL-1β levels were reported in the presence of a high concentration of non-crystalline silica particles of nano (15 nm)- and micrometer (1.5 μm)-size, whereas only the nano-sized particles induced IL-1β release in human keratinocytes . Winter and coworkers , showed that 14 nm silica nanoparticles induced a marked IL-1β release in LPS-primed murine dendritic cells. In the RAW264.7 macrophages and rat primary lung macrophages we observed that both silica particles from the nano-size (Si50, fumed silica) and submicro-size (Si500 and fused silica), besides the crystalline MinUSil 5, induced marked IL-1β responses.
The potential to induce inflammatory responses at low concentrations is a critical question. In dendritic cells, the potency of the non-crystalline silica nanoparticles (14 nm) on a mass basis was approximately similar to quartz (DQ12) . In our study non-crystalline Si50 was less potent than quartz (MinUsil 5) on a mass basis, when examined in primary lung macrophages. The fumed silica with the smallest size (5–20 nm), but also a different composition, was however on mass basis the most potent of all the particles examined, giving IL-1β responses from rather low concentrations (1–5 μg/ml). This supports the potential relevance of these findings, suggesting that additional toxic studies on these particles could be of interest. In other studies the concentrations of non-crystalline silica particles used to induce IL-1β release were mostly in a higher range [27, 28], than those used in the present study.
In general, in many previous studies using cell cultures, nano-sized particles seemed more potent compared to larger particles of the same composition, when presented on a mass basis. However, these differences tend to disappear when relating the responses to particle surface area both upon in vitro and in vivo[34, 35] exposure. In accordance with this, we find that when assessing the IL-6 and IL-8 response to Si50 and Si500 in BEAS-2 B-cells, the particle surface area seems to be the critical determinant (unpublished results). However, with regard to the IL-1β responses in primary rat lung macrophages and in particular in mouse RAW-macrophages, the Si500 silica particles seemed markedly more potent than Si50 of the same composition, when presenting the responses in relation to particle surface area. These somewhat surprising finding could be due to a different role of particle uptake for various cytokine responses. Our previous study indicated that the increased IL-6 and IL-8 release following Si50 in BEAS-2 cells seemed to occur independently of particle uptake . In contrast, the IL-1β response in RAW264.7 macrophages, shown in the present study, seemed to be dependent on uptake of the silica particles (see below). Notably, it is known that particle uptake via phagocytosis is less effective with nano-sized particles than with larger particles around 0.5 μm . Thus, the present finding with Si500 as more effective inducers of IL-1β than Si50 in RAW264.7 macrophages and primary rat lung macrophages at comparable particle surface areas could be related to more effective particle uptake, although it cannot be excluded that other explanations may be involved. Previous inhibitor studies have shown that uptake of micro-sized crystalline silica particles in alveolar macrophages occurring via actin-dependent phagocytosis  is required for eliciting inflammasome response . In our study, uptake of all the non-crystalline silica particles of various sizes also seemed to be mediated via a mechanism that was actin-dependent, as cytochalasin D inhibited the IL-1β responses. However, both phagocytosis and macropinocytosis are actin-based pathways . For nano- and submicro-sized particles, that are relevant for the present study, it has been reported that the extent of uptake in macrophages differ, with several mechanisms operating for particles of different sizes and functionalisation [37, 39–43]. Furthermore, the mechanisms of uptake are dependent on the differentiation state of the macrophages and the proteins that the nanoparticles encounter in culture media. Thus, in primary macrophages opsonised by serum proteins nanoparticles (polystyrene) were taken up by phagocytosis, whereas other mechanisms operated in monocytic THP1 cells with and without differentiation with TPA . The similar patterns of responses to cytochalasin D and the other inhibitors to all the different particles observed in the present study were surprising, as it points to a common mechanism operating over the whole particle size range. One possible explanation could be that the secretion of IL-1β from the cells was actin-dependent. Previous studies in bone marrow dendritic cells have however shown that cytochalasin D does not interfere with the secretion process of IL-1β . More conceivably, the agglomeration state of the particles could be of importance for the similar responses of nano-and submicro-sized silica particles to cytochalasin D. Thus, the macrophages might sense all the different particles as sub-micrometer agglomerates, pointing to phagocytosis as a common mechanism for particle uptake. This was discussed by Winter et al. , but the agglomeration state of the non-crystalline particles was not assessed. In the present study the hydrodynamic size of all the particles in the culture medium with serum, was from 90 nm and larger. The optimal particle size for phagocytosis has been reported to be relative large, 250 nm to 3 μm, whereas nanoparticles less than 250 nm were less effectively phagocytosised [37, 41]. Further studies are, however, required to approach the importance of the agglomeration process for the inflammasome activation.
With further respect to mechanisms, our findings suggest that the IL-1β responses induced by the non-crystalline particles in the RAW264.7 macrophages involve activation of the vacuolar H+-ATPase, and induction of cathepsin B release, subsequently leading to phagosomal destabilization, as previously reported for quartz particles . In a recent study, Morishige  found that the IL-1β response induced by 1000 nm non-crystalline silica particles was markedly reduced by inhibitors of these processes. Here, we report that the IL-1β responses by all the different silica particles were nearly abolished upon pre-treating the RAW264.7 macrophages with relevant inhibitors such as bafilomycin A1 and CA-074-Me. Thus, our findings corroborate that non-crystalline silica particles of different sizes and composition and crystalline silica particles, act via similar mechanisms. Previous studies in endothelial cells have shown that non-crystalline silica particles may induce cytotoxic responses, with the particles in nano-size as most potent on a mass basis . Potentially, such cytotoxicity could influence the silica-induced IL-1β responses in our mechanistic studies performed at higher concentrations, in which different inhibitors were used. At the early time points (6 h) used to assess the IL-1β responses, however, the particles induced little toxicity as assessed by LDH-release. In comparison the IL-1β release was much larger, thus strongly arguing that the observed responses are not due to release of pro-IL-1β as a result of plasma membrane rupture. This is also contradicted by the fact that zYVAD and siRNA against NALP3 inhibit the SiNP-induced IL-1β responses.
The main goal in the present study was to investigate the ability of non-crystalline particles to act via inflammasome activation in macrophages. We have also compared the relative potency of the non-crystalline silicas to crystalline silica (MinUsil 5) with respect to IL-1β release in the primary rat lung macrophages. When related to surface area the crystalline silica was more potent than the non-crystalline silicas (in particular compared to Si50 and Si500), which could suggest a role of crystallinity. However, the findings showing that fumed (non-crystalline) silica particles induced almost the same response as crystalline silica, suggest that other factors/mechanisms also may contribute. Thus, the overall response pattern obtained may be influenced by differential uptake mechanisms for the different silica particles in the lung macrophages (as previously discussed), but also to other differences in particle surface reactivity than caused by crystallinity. Notably, studies with different quartz particles (nanoscale- and fine quartz) by Warheit and coworkers  indicate that the toxicity of different particles is more dependent upon particle surface activity effects than particle size and surface area.
Based on these studies it would be interesting to explore any possible acute inflammatory effects of non-crystalline silica particles, in particular for relevant concentrations of nano-sized particles in LPS-primed animals. Any possible implications for possible long-term effects are not likely as no or minor effects from long-term industrial use of non-crystalline (amorphous) silica particles have been reported. This is probably due to the reduced retention time of the non-crystalline silica particles in the lung tissue compared to quartz.