Our study is one of few to use a dry powder generation system for inhalation exposure to HARNs. Generation and exposure systems utilizing dry aerosols are the best options to recreate human exposure conditions, since there have been many challenges with characterization of nanomaterials in different milieu (e.g. water or culture media). These challenges are even more pronounced for HARNs mainly due to the higher aspect ratio of fibers in comparison to other nanomaterials as well as increased surface area that can lead to higher surface interactions than other nanomaterials .
The Al3+ concentration detected in the lungs of mice after 4 wk exposure to Al nanowhiskers was 0.505 μg Al/mouse. Considering that 30% of AO nanowhisker mass is Al, the mouse burden of AO nanowhiskers would be 1.68 μg AO nanowhiskers/mouse, which corresponds to the lung burden accumulated by a 70 kg person exposed to a concentration of 1 mg/m3 for 44 hrs (assuming breathing 15 breaths/min, 600 mL/breath, and a pulmonary deposition fraction for 150 nm particles of 0.2  and delivered AO nanowhisker dose to human 4712 μg). Extrapolating mass of AO nanowhiskers per alveolar epithelial surface area for mouse (0.05 m2) and the human (102 m2) , the mouse burden would be 34 μg/m2 and human burden would be 46 μg/m2.
The nanowhisker aerosol produced by the novel generation process applied during this study was characterized by an SMPS and TEM. Those results suggest that this dry-dust generation method was capable of producing a size distribution comparable to those produced in our previous studies [18, 29–32] in which a powder suspension was nebulized and the droplets dried prior to entry into the chamber. Our previous efforts to use sonic energy to disperse nanopowders were not as successful when applied to other nanopowder types and without the alterations applied during this study . The addition of the venturi aspirator and static-charge grounding strap to an aluminum (rather than plastic) elutriator were adaptations to the ADAGE that resulted in a consistent output of nanowhisker agglomerates with a GM of 150 nm.
We attempted to characterize the particles in the lungs after exposure using bright and dark field microscopy as well as TEM-EDS. However, we were unable to locate particles in lung tissue using TEM. This might be attributed to low retained doses of AO nanowhiskers in the lungs and low amount of Al in the AO nanowhiskers, thus it is possible that Al ion concentrations in the lung tissues were too low to produce a signal in TEM‐EDS.
As with our previously studied materials [18, 29, 31], characterization of the nanomaterials confirmed that characterization data provided by manufacturers needs to be independently determined due to batch to batch variability and the fact these materials are not always characterized using state-of-the-art instrumentation. Based on the XRD patterns, it appears that AO nanowhiskers were not a pure aluminum oxide phase such as Al2O3 but instead were a mixture of bayerite (Al(OH)3) and boehmite (AlOOH) . Thus, complete and independent characterization of material used for toxicity evaluation is essential .
Many recent studies have related bioavailability and subsequent toxicity of nanomaterials to the ability of the nanomaterial delivered to biological or environmental milieu to release soluble metal ions to proximate cells or organisms [34–36]. Our previous study found this association between nanoparticle toxicity and dissolution . However, with AO nanowhiskers, we found that 35% of nanowhiskers were dissolved in ALF buffer (pH 4.5) in 2 wks and 58% in 4 wks. About 15% of AO nanowhiskers dissolved in Gamble’s solution (pH 7.4) which represents extracellular fluid (Figure 4) in 4 wks. We found that it was difficult to dissolve AO nanowhiskers even in concentrated nitric acid. The observation of relatively high dissolution of AO nanowhiskers in ALF but low dissolution in vivo suggests that AO nanowhiskers undergo ligand-promoted dissolution in the simulated biological buffers due to the presence of citric acid .
In ligand–promoted dissolution, ligands form surface complexes that can polarize metal-oxygen bonds, thus facilitating the detachment of metal species from the surface. Ligands are capable of forming surface chelates, i.e., bi- or multi-dentate ligands such as oxalate or citrate, can facilitate these interactions and enhance the dissolution of aluminum oxide . It has been shown that citrate adsorbs to aluminum oxyhydroxide surface in a predominantly inner-sphere manner which facilitates the dissolution of this mineral . Citrate is a molecule of both environmental and biological importance since it is present in many naturally-occurring systems and in blood plasma at a concentration of 0.1 mM [40, 41]. It is used in ALF solution to represent protein binding to foreign objects. Here the interaction of citrate with AO nanowhiskers plays an important role in the nanomaterial dissolution. The solutions of artificial biological fluids are only facsimiles of actual lung fluids and some AO nanowhiskers were likely cleared from the pulmonary system after exposure. Thus, it is most likely that the nanomaterials were not present in lung fluids for as long as they were present in our simulated lung fluids for our dissolution studies (2 or 4 wks). Low solubility of AO nanowhiskers in epithelial lining fluid was also confirmed by very low levels of Al ions in supernatants of BAL fluid measured by ICP-MS.
There was significantly higher recruitment of alveolar macrophages in BAL fluid with longer exposure to AO nanowhiskers. However, the number of neutrophils and lymphocytes were not elevated meaningfully. Thus, it seemed that macrophages were able to control the aerosol load in the pulmonary system without upregulation of cytokine production or cytotoxicity. The experimental design with the total exposure period lasting up to 28 days would be sufficient for the mice to develop fibrosis in the lung tissues, however our histopathology evaluation of the lung tissues as well as pulmonary mechanics assessment found no pathological changes. This may be attributed to the low fibrogenicity of the material or low lung burden of AO nanowhisker in exposed mice. These findings are consistent with our measurements of total protein, LDH activity as well as cytokines/chemokines measured in BAL fluid. We also analyzed basic hematology parameters in the blood for possible systemic effects of inhaled aerosol of AO nanowhiskers, but did not find any deviations from our control animals.
A rat inhalation study of AlOOH with primary particle range of 10 – 40 nm (with mass median aerodynamic diameter (MMAD) of agglomerated particles in inhalation chambers 1.7 and 0.6 μm, respectively) for 4 wks (6 hrs/day, 5 days/wk), found significant pathology changes in BAL parameters and lung pathology only at the highest exposure concentration (28 mg/m3), but not at concentrations of 0.4 and 3 mg/m3
. Despite a number of differences between Pauluhn and our study (e.g. different animal models, morphology and size of nanomaterials, and length of exposures) their findings are noteworthy. Similarly, like our study, this study did not find any substantial extrapulmonary translocation of Al.
In our presented study, the estimated lung burden of AO nanowhiskers per mouse after 2-wk-exposure was 19.6 μg/mouse and after 4-wk-exposure it was 39.2 μg/mouse. We observed minimal pulmonary responses after this exposure similarly as after sub-acute (2-wk) exposure to silver (Ag) NPs with estimated lung burden of 29.4 μg/mouse . On the other hand, in our previous study of copper (Cu) NPs  we observed substantial neutrophilic response after sub-acute exposure to Cu NPs with estimated lung burden of 32 μg/mouse. Certainly, there are many physicochemical characteristics of nanomaterials that play role in their toxicity (e.g. bulk and surface chemical compositions, size, phase, propensity for dissolution, ability for clearance or translocation etc.) and they need to be taken into the consideration when these materials are evaluated.
Our inhalation exposure to AO nanowhiskers did not result in “frustrated phagocytosis” and subsequent inflammation, which might have been expected since shorter fibers are more likely to be phagocytosed by alveolar macrophages than longer fibers. Another reason might be that the AO nanowhiskers after generation to the exposure chamber appeared more like tangled bundles of materials rather than individual fibers. And thus it is possible that they behaved more like particles that could be cleared from respiratory system by the mucociliary escalator. It appears that alveolar macrophages in our study were not overloaded since we did not observe an increased production of pro-inflammatory cytokines/chemokines and subsequent higher recruitment of neutrophils that would otherwise lead to pathologic outcomes. Furthermore, AO nanowhiskers may have endured some other changes after entering the respiratory system, such as breakdown that might further increase their clearance from pulmonary system.
A number of studies report that surface area of particles is associated with the magnitude of airway inflammation [43–45]. The surface area of AO nanowhiskers (320 ± 4 m2/g) was the largest of all nanomaterials studied thus far by our group [18, 29–32]. Therefore and also considering current knowledge about HARNs concerning their toxicity, we hypothesized that there might be a potential for a higher toxicity of this nanomaterial. However, as in our previous study with two TiO2 nanoparticles with different surface areas  as well as other studies [46, 47], our results add the evidence that there are many physicochemical properties of nanomaterials that play a role in toxicity. Larger surface area of nanomaterials is only one consideration and is not always responsible for higher toxicity. Other studies have shown that the shape of fibrous material has a decisive role in their toxicity, however AO nanowhiskers that we investigated exhibited low toxicity in this study. Cell surface injury was observed in macrophages exposed to fibrous TiO2, but no significant changes in macrophages were found in case of particulate TiO2
. In vitro study of carbon nanotubes and nanofibers found that long, straight, well-dispersed nanofilaments caused significantly higher production of TNF-α and reactive oxygen species (ROS) in human peripheral blood mononuclear cells than highly curved or tangled materials .