Specific pathogen-free male Fischer 344 rats (Harlan; Frederick, MD; 175–200 g body weight) were housed in filter-top plastic cages and given free access to food (5001; Purina Mills, LLC, St. Louis, MO) and water in a humidity and temperature controlled room with a 12 hr light–dark cycle. Prior to use in experimental protocols, all animals were acclimated for at least 1 week in an Association for Assessment and Accreditation of Laboratory Animal Care International-accredited facility. Animals were treated humanely and with regard to the alleviation of suffering in accordance with a protocol that was approved by the University of Rochester’s Committee on Animal Resources. Animals were randomly distributed among the treatment groups, with 3–5 rats per group for dosimetry and separate groups of 5 rats for assessing acute inflammatory parameters.
Achieving similar deposited doses of TiO2
A high deposited dose of ~200 μg TiO2 was selected so that we could compare our findings with existing data sets generated in our own laboratory and other literature reports [34, 52, 72, 73]. The lower dose of ~45 μg was chosen after performing preliminary experiments that interrogated the dose response curve for neutrophil influx 24 hr after intratracheal instillation, in which we began observing significant inflammatory effects compared to saline controls. While it is straightforward to achieve the target doses for instillation exposures, lower RT deposition fractions derived from the Multiple Particle Path Dosimetry (MPPD) model  were used to calculate aerosol concentrations that would produce the desired doses upon inhalation exposure. In addition, the deposited doses (ILBs) following inhalation and intratracheal instillation exposures were verified through quantification of Ti content in lung tissue (see below).
Whole body inhalation exposures
Whole body inhalation was used as a low dose rate delivery method. Other exposure methods, such as nose-only inhalation, can minimize the deposition of NPs on the animals’ skin and fur and thus reduce oral uptake of NPs. However, a significant disadvantage to such methods is that the animals are subjected to higher levels of stress and would not be acclimated to exposure conditions following a single or even repeated exposure for only 4 days . Due to the uncertain contribution of this stress to response outcomes, we decided to compare our intratracheal instillation response outcomes to those following whole body inhalation NP exposure. Furthermore, much of the historical literature to which we wished to compare our results describes effects of TiO2 NPs that were delivered via whole body inhalation exposures.
Rats were randomly placed in a 60 L compartmentalized, polycarbonate (Lexan) chamber, which was under slight negative pressure with an internal horizontal flow of 35 L/min. TiO2 powder (AEROXIDE P25 powder, Evonik, Germany; primary particle size, 25 nm; 80% anatase, 20% rutile crystal phase; BET surface area 57 m2/g, ) was fed by a screw mechanism into a jet mill (JET-O-MIZER™, Fluid Energy Equipment Division, Telford, PA), producing aerosol concentrations of either 33 ± 4 mg/m3 or 13 ± 1 mg/m3 (for 4 hr). The higher aerosol concentration of 33 ± 4 mg/m3 was used to achieve the higher deposited dose for the single exposure. The 13 ± 1 mg/m3 aerosol concentration was used for the single, low deposited dose and also for the repeated exposure for 4 hr over 4 consecutive days. Control animals were exposed to filtered air. The TiO2 aerosol mass concentrations were determined gravimetrically via filter (PALLFLEX® Emfab; PALL Life Sciences, Port Washington, NY) samples drawn from the chamber that were collected every 15 min. The MMAD was determined by an eight-stage Nano-MOUDI impactor (MSP Corp., Shoreview, MN) and the CMD by a model 1000XP wide range particle spectrometer (MSP Corp.). The size characteristics of the TiO2 aerosols were similar for the single and repeated exposure concentrations (Table 1).
Intratracheal instillation exposures
TiO2 NPs were cup horn sonicated (Sonics VCX 750 Vibra Cell, Sonics and Materials, Inc., Newtown, CT) for 5 sec at 29% amplitude in 0.9% sterile saline and vortexed for ~30 sec immediately prior to intratracheal instillation. For the experiments in Additional file 1: Figure S1, the NPs were sonicated for 5 sec, 5 min or 30 min in either dispersion medium or saline. Rats were anesthetized with 4.5% isoflurane until their breathing was slow and shallow, after which the rats were placed in a supine position with the head elevated. A modified pediatric otoscope was used to visualize the vocal cords. A 20 gauge, 1.5 inch, Teflon® catheter sheath was inserted through the vocal cords until the tip was ~3-5 mm above the bronchial carina. In synchrony with the inspiratory phase of the breathing cycle, 250 μL of the TiO2 suspension was instilled into the lung. Controls were exposed to saline that was sonicated as described above. Mock intratracheal instillations directly into platinum crucibles showed that no significant losses of TiO2 occurred in the instillation needle or syringe when compared to the deposited doses in the lungs.
The hydrodynamic size of a 1:10 dilution of the 800 μg/mL suspension of TiO2 for intratracheal instillation was determined by DLS (Nano ZS Zetasizer, Malvern Instruments, Westborough, MA) and LDS (Partica LA-950 V2; Horiba Instruments, Inc., CA) as previously described . The measurements by DLS and LDS were taken after the material was sonicated for ~5 sec and then vortexed for ~30 sec, every 15 minutes for 1 hr, mimicking the preparation used prior to instilling the material and accounting for the time it takes to instill all of the rats in one group (~1 hr).
Dissolution of TiO2
The dissolution rate of TiO2 NPs was determined with a dynamic flow-through system as previously described . Briefly, NPs (~0.8 mg) were suspended in 1 mL of dissolution buffer before being injected into the upper chamber of a dialysis cell fitted with a 3,500 molecular weight cellulose ester asymmetric membrane (Spectra/Pore®, Gardena, CA; effective pore size ~3.5 nm). The Ti-free dissolution buffers simulated extracellular lung lining fluid (pH = 7.4) and intraphagolysosomal fluid (pH = 4.5), respectively. The buffers flowed into the dialysis cells at a rate of 60 μL/min, or ~3 mL/hr; the outlet ports of the dialysis cells were connected to a fraction collector. The dialysis cells were submerged in a 37°C water bath in a dark room. A fraction collector with metal-free, pre-weighed polypropylene tubes was used to collect the dialysates over the course of 7 days. The sample weight for each tube was recorded. The solubilized amount of Ti in the fractions were all below the instrument limit of detection (~10 ng/mL) by atomic emission spectroscopy (Beckman Spectraspan V, Fullerton, CA).
Quantification of TiO2 NPs in lung tissues
TiO2 NP-exposed rats were euthanized with an overdose of 2, 2, 2-tribromoethanol (Avertin; 25 mg/100 g body weight, i.p.); the pelts were removed to eliminate possible transfer of TiO2 to the lung tissue from the animals’ fur as previously described . Lung tissues were harvested immediately following exposure (ILB), 24 hr and 7 days post exposure by excising the lung above the bifurcation of the main bronchi. Tissue samples were dried at 85°C and then ashed at low temperature (50–100°C) in a solid state plasma asher (March Instruments Inc., Concord, CA), in which organic material is gently oxidized to CO2, leaving only TiO2 and inorganic ash. Samples were then fused with sodium carbonate/sodium borate (2:1; Sigma, St. Louis, MO) at 1500°C in platinum crucibles for 20 min or until a clear melt was formed. The melt was cooled and then dissolved in 2.5 N sulfuric acid and diluted 1:2 with ultra pure water. The concentration of Ti was quantified using atomic emission spectroscopy and the mass of TiO2 in each sample was then determined stoichiometrically. Control and naïve animals were found to have background levels of TiO2 in the lung below the instrument limit of detection for atomic emission spectroscopy (~10 ng/mL).
Cellular and biochemical parameters in bronchoalveolar lavage fluid
Separate groups of rats were euthanized at 4, 8, 24 hr and 7 days after instillation or after the beginning of the inhalation exposures with an overdose of Avertin followed by exsanguination. The lung/heart block was excised and excess tissue removed prior to the lungs being lavaged with sterile, 0.9% saline (5 × 5 mL), keeping the first two lavage supernatants separate from the remaining ones following centrifugation (10 min, 350 × g, 4°C). BAL cell viability (trypan blue exclusion), number, and the percentage of different cell types (Hema 3®; Fisher Scientific, Kalamazoo, MI) were determined. Total protein concentration was measured as an indicator of cytotoxicity and epithelial barrier permeability with the bicinchoninic acid (BCA) assay using reagents purchased from Thermo Scientific (Rockford, IL). Lactate dehydrogenase and β-glucuronidase activities, as indicators of cell membrane and lysosomal membrane integrity, respectively, were determined using reagents from Sigma.
Preparation of lung homogenates
Flash frozen, right lung tissues were homogenized on ice for 30 sec in 4.5 mL of radioimmunoprecipitation assay (RIPA) buffer, comprised of reagents from Sigma (50 mM Tris–HCl, 150 mM NaCl , 0.25% deoxycholic acid, 1 mM EDTA, 0.1 mM PMSF) and Roche (Indianapolis, IN; 1% nonindet P-40, 10 μg/mL aprotinin and 10 μg/mL leupeptin). Samples were centrifuged for 1 hr at 19,800 × g and 4°C in 50 mL, round PPCO tubes (Nalgene, Rochester, NY). The protein content of the supernatants was measured using the BCA assay.
Measurements of inflammatory mediators
The lung homogenate and the BAL fluid supernatants were used for measuring MCP-1, MIP-2 (CXCL2/CINC-3), TNF-α, and IL-10 levels by ELISA using antibodies and protocols from BD Biosciences (San Diego, CA), R&D Systems (Minneapolis, MN), eBioscience (San Diego, CA) and Invitrogen (Frederick, MD), respectively. HO-1 in BALF cell pellet lysates (lysis buffer: 1.25x protease inhibitor cocktail P2714 from Sigma, 1% Triton X and 0.2 mM PMSF) and supernatants were also assessed by ELISA (Enzo Life Sciences, Farmingdale, NY).
The dosimetry results were analyzed for time related changes from ILB by one-way ANOVAs for both exposure methods. Differences between the exposure methods at the same post-exposure time points were assessed by a Student’s t-test. Response endpoints for the high dose single exposures were analyzed by two-way ANOVA to detect differences over time and between exposure methods. Separate control groups for every post-exposure time point were not included due to ethical reasons, specifically to limit the total number of animals used in the study, as it was not considered likely that within-exposure-method responses in controls would change consistently over the time course of this study. Therefore, controls were evaluated only at 24 hr post exposure (the endpoint of highest possible acute inflammation) and 4 hr post (the earliest time point assessed for the onset of the inflammatory responses). Responses in controls were only found to have significant time related differences within inhalation exposure for cell viability, BALF MCP-1 and homogenate MIP-2 and within instillation exposure for BALF protein, homogenate MIP-2, IL-10 and BALF pellet HO-1. Because there was no consistency in terms of which time point had higher values and the small differences were not likely to be biologically significant, the controls were pooled. The low dose response and repeated exposure data were analyzed by two-way ANOVAs with exposure method and dose of TiO2 as the main factors. Data were appropriately transformed if analyses of residuals suggested deviations from the assumptions of normality and equal variance. Two outliers were identified based on analyses of residuals (β-glucuronidase and BALF protein) and were removed from figures, tables and corresponding ANOVA tests. All comparisons were considered statistically significant when p < 0.05.